This year I have managed to repair 22 5.25" floppy drives and I must say it was quite interesting and challenging to do. I only had to cannibalize one drive to do it. I couldn't have done it without IMD, which I had used before in my attempts to get 8" floppy drives formatted on a PC. There is one drive left for which I cannot get any data from the heads at all, no matter how much realignment I try to do. All these drives were sold to me untested, for parts or repair, but actually I was surprised to find out that about a quarter of them were fully functioning and that I could not find any fault on them. I think it helps if a Pentium I or 486 motherboard is used to test them instead of Pentium II/III/4 motherboards, as those might have trouble dealing with these drives that were already considered obsolete in that era. Other common faults I found and fixed were: 1. The rail on which the head assembly moves is caked in dust, causing the stepper motor to have trouble and also causing the head assembly to skip tracks --> Clean and lubricate the rails 2. Both heads don't give any data at all in IMD --> Clean thoroughly with a cleaning disk with isopropyl alcohol applied to it. Allow IMD to make around 10 passes. (Possibly someone used a chemically decomposing disk that thoroughly fouled up the heads, or perhaps used the wrong cleaning chemical) 3. IMD shows that not all 18 or 30 sectors can be read during alignment test --> Realign head assembly using IMD, sometimes the whole assembly, sometimes just head 1 3. The drive keeps spinning the disk permanently after inserting a disk and the disk cannot be accessed. --> Check if the disk locking mechanism is fully engaged, or the track hole sensor or disk insertion sensor are fouled up or not properly aligned. Repair the locking mechanism block (sometimes it is cracked) by glueing or melting it back together (soldering iron). Possibly place some duct tape underneath the pusher block to have it exert more force. Realign the sensors while the disk is spinning and while running IMD to find out the correct position 4. The drive does not pre-spin upon disk insertion, leading to a badly aligned and wobbly disk --> check if the detection mechanism still has its spring and replace it if it is missing or broken. A containment for the problem is also to manually 'pump' the locking lever after disk insertion to make sure it is aligned to the spindle properly 5. The drive kind of works with reading, but has trouble writing. At IMD alignment it does not show the full 18 or 30 sectors read at each track. Perhaps at some tracks there is 18 or 30, but not at all of them --> Check the head 1 spring and put duct tape or a small foam block underneath to change the force the spring exerts. Realignment in stepper direction is not always the solution 6. Jumper settings are not compatible with the PC, leading to strange behavior that has no mechanical reason --> Find out what the jumpers are for (via documentation, or other sources like the above video) and correct their placement Then there were some other less common faults that were more difficult to solve, like: 7. Obvious damage like a circuit board with a corner snapped off, or missing mechanical components, like the spindle motor --> try to salvage what is possible. Cannibalize a similar drive. Figure out the severed connections on the circuit board and solder on wires as needed. What is needed can be inferred by observing the drive behavior. Documentation about the circuit board is nearly impossible to find. 8. No visible damage at all, but drive behavior is still out of specification, for example it doesn't detect track 00 --> find out where the problem is in the circuit board by comparing with a functioning drive of similar model if possible. I found out that voltage on the track 00 sensor was too low on several drives and I applied an additional resistor between it and the 12V, causing it to be in spec again. What I used to fix those 22 drives: - Pentium I test PC - good working drives and secondary PC for reference - new-old-stock unused disks - the IMD program - duct tape - WD-40 - cleaning disks - isopropyl alcohol - multimeter - soldering iron - misc parts like replacement springs, bolts, resistors, wires - common sense - deductive logic - patience
@@PCRetroTech You're welcome. I am simply hoping we can save as many of these drives as possible. Your video, especially the part on the jumpers was helpful to me to try things I hadn't thought about before.
@@MacmorMach In my experience with the current era, you either buy a single good drive that is expensive or a single cheap drive that probably isn't in a good condition. Buying a single good cheap drive nowadays is only possible if you get it from someone who considers it junk, but there aren't many such opportunities left. Or you can buy a large batch of non-functioning ones, like I did, and repair them.
WD 40 is Not a lubrication liquid. WD literally stands for Water Displacement. It is a degreaser. It will leave residue and it will migrate. Everyone should just throw away WD-40. When you use it and it makes the device stop squeaking, it is only a temporary fix. If you use proper lubricant the device won't squeak either. For isopropyl you want to use the lowest water content that you can find, at least 85% and you don't want to use it on anything rubber because the rubber will degrade. In most cases the rails in a drive should not have any lubrication on them. They are highly polished and they slide in copper sleeves. Lubrication will attract dirt, and could cause the polished rails to have greater drag(its counter intuitive) since there is now some substance between the two surfaces. If you are using a drive which drives the head by a worm gear, you want to put a small amount of white lithium grease on it. There is also electrical contact cleaner but you probably wouldn't need it inside of a drive but it will clean switches and not leave volatile residue on the contact surface.
I frequently buy old PC's from thrift stores, garage sales, and estate sales. Not for the PC itself but for the floppy and hard drives and expansion cards that are inside!!! I recently purchased an old PC that was loaded with expansion cards. Including a Trackstar, an Apple emulator card.
Normal density (Mitsubishi MF504B,C 5.25" HD Drive) Logical ‘1’ selects high density read/write operation and logical ‘0’ selects normal density. With option jumpers SS shorted and SB open, this line changes the motor rpm and read/write 'operations are performed 400mS after density change, to allow the motor speed to stabilize. When motor rpm is switched the head must be moved to track 00 before reading or writing. With SB shorted and SS open the motor speed is always 360rpm so no waiting or track 00 seek is necessary.
that Chinon Drive where you talked about the last jumper and old 8" drives, whould make this suitable out-of-the-box to be configure for any (retro) machine that requires Shugart compatible drive interface, meaning it could be configured to work properly with an Amiga, MSX etc without doing any bodges
I managed to get the remaining two drives working guys, plus another one of the drives that had some issues towards the final track. This required more advanced techniques. I'll describe those here, but please don't blame me if you screw up your drives attempting this. In fact, I got one of the three working before accidentally slipping with my screwdriver and completely destroying one of the heads! In all cases, the issue was head alignment. In the case of the drive that sounded stuck, I had to move the stepper motor assembly so that it could move the head back to track zero. You can use the head align tool in IMD and move it back and forth until it can read track zero on head zero. Press D to check it can actually read the data. From here, all three drives had the same problem: everything was fine on head 0 (the one towards the bottom of the drive), but something would go wrong on head 1. In all three cases, the solution was the same: unscrew the two screws that hold the head in place and move it about until the head alignment program showed 18 matches. This was particularly hard on one drive because you have to remove a tensioning spring that is required to actually read data, so you have to apply pressure with your finger whilst you move the head about while watching the data in IMD. It didn't help that the sled was cracked and I had to superglue that back together before I could get anywhere. So I've now successfully rescued all my drives, except the one I accidentally broke beyond all possibility of repair (I literally sheared the head right off). I tested that the drives can now read 20 disks formatted on a known working drive and filled with data up to the last track!
Fantastic - I didn't know about IMD. I feel like head calibration is going to be increasingly important to save the drives that are still functional today, but will slowly fall out of alignment over time. And, of course, the expert information out there isn't well documented online (having been of most use before the Internet was popular) and literally dying out with the 80s long-beards. I feel you on the damaged head. I started disassembling floppy and CD drives as I acquired them, and giving them all a thorough cleaning and re-lubrication. There was one 3.5" drive that I was having trouble finding out how to remove the cover, and pulled a couple screws out of the bottom. Those were the ones holding the Track 0 sensor in place. So that drive is now wildly out of calibration. I may try to recover it now that I know about IMD, or pass it on to someone who has the motivation to do so. Glad I didn't just throw it away, though!
0:00 before watching the video I'm going to guess. From experience old lubricant can cause motors to be completely stuck or sticky enough to cause the drive to give up...next is the jumper settings, some drives need them set correctly...then I'll say stuck or sticky head mechanisms/dirty heads...and lastly incompatible disks, wrong formatting etc...this may be kinda cheating I've just listed all the issues I know lol. But if you read all of my nonsense I Appreciate It! Thank You For Your Time And Effort! I Hope You Are Having A Great Day Or Night!
FYI that head alignment test is called "an azimuth test" and you can hook up a scope to to see the the _analog_ signal coming off the head, IMD is just doing a _digital_ test using "data integrity checks" (data does not change between passes, and track overlap doesn't distort sector reading). Azimuth alignment with a scope is how you "fix" a Wren / Falcon / Phoenix Drive platters when its not reading, as Usagi Electrical showed when restoring the CDC mini computer hard drives.
May I recommend the cable transfer program FastLynx? It can transfer files over RS-232 null modem cables or parallel port cables including those for Laplink. One of its fanciest features is the ability to transfer itself to a DOS PC over null modem, so you don't even need a working floppy drive.
I feel the same way about you regarding retro programming. It is just not the same using a modern day computer to learn programming as was done in the earlier days.
The FG jumper usually stands for Frame Ground, it connects the ground signal from the data connector to the metal frame of the drive, it's do deal with ground loop problems. Most of the time that jumper won't matter
Thanks for this video. Lots of good helpful info. On the one drive, you mention an "IM" and "NM" jumper setting; that could possibly be for maskable or non-maskable interrupts.
This video reminded me of the days when you had to specify cylinders and sectors in the bios of hard drives. Also, I have a memory that you had to put that in COBOL programs as well. Thank you.
Thank you for your video. I purchased this floppy drive unit but wrongly purchased the data cable and your video has solved the problem for me. Otherwise I won't be able to use this device forever.
It's amazing how many things I've learned from this video. I have a 5.25 floppy drive that works. When I can I send you the pin configuration. I am very grateful to you. All the best.
You are exactly correct. If I format my 360k disks in my original IBM PC and write the images on my Pentium with a 1.2mb drive I can then read them back on my IBM PC. If I format the 360k disk on my 1.2mb drive it will not read it back on the IBM PC or at least not 100% properly. Good Video.
Yeb cause the heads of a 1,2M are smaller. See as a 2 lane driveway. The 2 lanes together are 360K Wide, a 1,2M uses just 1 lane at 360K. So if you write a 360k formated floppy with a 1,2M just 1 lane is read and other lane is un toughed. If there were data on that 360k, at 2nd lane there is still some old data left. The 1,2M does not bother as it just read 1 lane, BUT the 360k drive reads both lanes at the same time. So now there is a mix-up of signals.
The more I watch videos about retro hardware, the more I remember what a pain in the ass was setting PCs up in the late eighties PnP was a blessing, although it had many issues at the beginning.
@@BlackEpyon indeed. Although I don't miss fiddling with the software side of things, like config.sys and qemm to squeeze as much ram as possible to run DOS games.I was very good at remembering all the various bits and pieces, nowadays I'd have to research them. 😭
@@bufordmaddogtannen Yeah, I've got a Tandy 1000 RSX that I'm running SCSI on because it's a slim-profile case that doesn't have room for extra drives. I could revert back to an IDE hard drive (limited to 504MB) if I had to, but the external CD-ROM and Zip drive are both SCSI, and some games do need the CD to install or run. That means moving everything possible into highRAM, because THAT particular game (StarTrek 25th Anniversary CD) needs 560K of conventional memory, and won't take 'no' for an answer!
@@BlackEpyon SCSI FTW. Been there, done that. Qemm has always been a savior, although I am not sure if the scsci drivers suck too much ram. Have you tried copying the content of the cdrom on the hard drive on a temp directory and adding a drive letter pointing to it using the subst command? It could save the trouble of loading the cdrom driver and free some memory. Sometimes it is enough to fool games.
@@bufordmaddogtannen Na, I tried. Some games won't even let you install from a directory, or change the directory it installs to. Having a 2GB SCSI hard drive in a 386 is TECHNICALLY within the bounds of the time period (if you were building a server), but boy howdy would that have been an expensive build for the day! I'm kinda surprised the BIOS even allowed me to choose to boot from SCSI.
Gday Gday I am totally in your world mate I have just started restoring some XT 8088 system i recently found at the local refuse site. Could not bare to let them rot. I totally want to get started on 8088 Assembly. Glad to see this video. Now im gonna watch the rest
That's interesting: that when the HD version of these drives came out, there was a generation of them that were upwards compatible between 360K and 1.2M, but even newer ones aren't. I'm used to the times when once a line of devices is made upwards compatible, that's the way they stay. Why would these have regressed somewhat to not being upwards/backwards compatible anymore?
5:14 - "Just buy any old computer case" - Man, even cases are getting stupid. $100-200 these days for an empty case!!! It's normally more cost effective to just buy built systems. It also seems to be cheaper than to build from parts. I've been piecing together a 486 VLB build and was in over $300 before I knew it!
There's a certain driver floating around, PalmZip I believe, that purports to get parallel Zip drives working on 8088 hardware. It's going to be horrifically slow, but it should at least work.
@@PCRetroTech Heh, I ran into that problem myself actually. Mostly I just have the drive now so I can recover old disks people bring me, hopefully. I actually had to repair my ATAPI drive, I'd managed to be enough of a bonehead that I nearly sheared one of the pins of the IDE connector off. I don't usually do SMD soldering, but the pads there were just big enough that while I was unable to get the original pin back in place how it was meant to go, I bodged a wire between it and the pin and got it back in the hole. It makes just enough contact now that the drive works, at least.
Why mix-up 360k Floppy's with use of 1,2M drive fails often.Cause the heads of a 1,2M are smaller. See as a 2 lane driveway. The 2 lanes together are 360K Wide, a 1,2M uses just 1 lane at 360K.And step over 2nd lane for next track, leaving that 2nd lane unused. So if you write a 360k formated floppy with a 1,2M just 1 lane is read and other lane is un toughed. If there were data on that 360k, at 2nd lane there is still some old data left. The 1,2M does not bother as it just read 1 lane, BUT the 360k drive reads both lanes at the same time. So now there is a mix-up of signals. A format DOES not whipe clean the disk, completely. So there is still some old data left at those lanes. The reverse is the same kind of mix up data. --> Then from WIN95 on, 360k drives Does not work at all any more! Even if Bios set up is correct. 360k only works in the DOS mode. For many old upscure Floppy's from old non DOS machines, take very care of your PC Floppy controller if it can handles that format. Not all floppy controllers can handle all kinds of formats. Indeed use TESTFDC for that to check out.
@@PCRetroTech Thats indeed next problem with use High dence disks in 360k drive. Often lead to bad data. Its a struggle. But in the end you get the hand of it.
I have a 5.25" floppy drive, works fine on a user unfriendly and heavy machine i don't want to use. On a computer i want to use, i can select the drive, but, for some reason the bios jumps back to a 3.5" floppy drive. The bios battery is good and the floppy drive setting is the only "stubborn" setting. (if no floppy drive is connected, the bios says correct "none" )
I was tidying and reorganising my 'nerdy clutter' as my wife calls it (even though she doesn't mind me keeping it all), and I swore I had a 5.25" 1.2Mb FDD in storage, and to my dismay I can't find it. What I can find is a load of 5.25" floppy disks I would like to make digital copies of before they all succumb to mould, or being wiped somehow. Anyone in the Hampshire area of the UK have a spare drive they're willing to part with along with cable? I'd be so happy to find one for as cheap as possible as I'm on benefits and I'm my wife's full-time carer.
It's weird, my experience has been that most 3.5" floppy drives seem broken or questionable while all of the 5.25" drives I've used have been extremely reliable. I wonder why that is.
No, i can second that. I still have my 5.25" disk drive in a working condition. It only had to go to a repair shop once. But the 3.5" disk drives did fail very soon. In the nineties, when floppy disk were still common i had to buy several 3,5" disk drives. The situation got worse, when the CPUs started to require active cooling fans and when the floppy disk drive as a storage medium become obsolete. I assume, the cooling fans do drag some dust into the floppy drives. And the situation is much worse when we are talking about the floppy disks. I threw away over 90 x 1,44 MiB 3,5" floppy disks in the nineties because of read errors, but only about 10 x 5,25" HD floppy disks.
Great video overall. Nevertheless, some information is misleading or incomplete. In IBM PC compatible computers, you generally want drives to run at a fixed speed. DD (360KB) drives should operate at 300RPM, and HD (1.2MB) drives should operate at 360RPM, no matter what the level on pin 2 is. The BIOS deals with DD diskettes running "too fast" in HD drives by choosing 300kbit/s instead of 250kbit/s data rate in the floppy controller. On the drive side, there is only one kind of stepping, which steps the drive one drive track. "double stepping" means that the disk in the drive is formatted in a way that the disk tracks are twice far apart than the drive tracks. This happens if you try to access a 40-track disk in an 80-track drive. The drive you used with IMD is thus obviously an 80-track drive. In IBM PCs (don't know about Amstrad pecularities), all 80-track drives are HD drives. So this drive is likely a 1.2MB drive (at which operation at 360RPM is normal), but we can't exclude the drive being a non-PC 720KB drive (80 tracks, but no high-density support). This also explains why testfdc failed: The drive spins so fast that the formatting or writing the test track takes more than one revolution of the drive, so formatting the last sector will overwrite the beginning of the track, or formatting will be aborted because the index hole is detected during formatting. testdfdc with "1.2" might have passed at that point. Using "DS0" to use a single drive before the twist as A: drive will likely not work as intended. On most drives, the DS0 to DS3 jumpers are just choosing which DS wire on the floppy cable will be used as DS for the drive. There is only *one* standard motor-on signal the floppy connector. The pin used by IBM for "A: motor on" before the twist is has been repurposed from one of the higher "DS" lines (DS2 or DS3). A drive before the twist jumpered to DS0 thus will respond to "B: motor on" and "A: drive select", which is not fine, unless you jumper the drive to use "drive select" instead of motor on to turn on the motor. In that case, you might get issues if the BIOS tries to keep the motor of drive A running while drive B is selected, and then re-selects drive A without waiting for an appropriate spin-up time (although I don't know whether the BIOS of IBM-compatible PCs generally does that). The advice to jumper Drive Ready / Disk Changed generally to "Drive Ready" is bad. The PC BIOS is able to handle a "disk change" signal, and will return "error code 6" on the first floppy access after a disk change. DOS uses this error code to flush the FAT / Directory buffers. Not having a working disk change signal may inadvertantly use the old directory/FAT on a new disk, which will inevitably lead to data corruption. But there is one extra piece of information, which makes your advice partially correct: Classic 360KB PC drives used pin 34 as "drive ready", and the BIOS/DOS don't expect disk change signalling to work if the CMOS type is "360KB". For 1.2MB drives, you better use "disk changed", though.
Thanks for the in-depth analysis and sharing your (considerable expertise). Other commenters have noted how useful the comments on the video is (due to the incomplete nature of what I presented). This greatly adds to the usefulness of the resource!
I have an old 386 based IBM compatible with a 3.5 drive that fits the 5.25 bay they exist and use the same ribbon cable style that slips on a pcb instead of the ide pin style
A trick I have used on floppy disks when all else fails is nice strong magnet fully erase all the formatting completely then reformat the disk in what ever computer your planning to use it in I used a magnet from a blown 12inch woofer and still had to make sure passed it all over the disk for it to be completely wiped I would only use this as a last resort but have had good success doing this.
@zaper, this way you can only magnetize the disk beyond the drive's ability to re-magnetize it during format. Use the demagnetizer with slowly decaying AC electromagnet
ENIAC is a dinosaur, but definitely not an 8088 PC. When ENIAC is the dinosaur and IBM System/360 the Mastodon, then the 8088 is an Egypt pyramids. So very new.
you deserve way more views than those pseudo retro guys who blantly slap an ssd or sdcard instead of using floppies because they lazy to do it the proper way. thats the proper way.
These drives are beautifull beasts. To me they are still usefull for anyone that work with pure text, manuscripts, scripts... caracters don't thake more space than back in the day. Also, 5 1/4 is more reliable than the 3 inch diskettes.
Yeah, if you have the drive out of a computer or external case then you might as well use the cotton-swab method to clean it; that's usually more thorough, based on what I've heard.
Ugh, I got my first 5 1/4" drive ever and it looks absolutely perfect, but doesn't do anything but spin the disk briefly when inserted. Not sure if it's my controller, cable, or drive... no spares to deduce what's going on! Sensor LEDs still work, no dirt or corrosion at all. Everything moves freely on track. Just nothing initiates.
Check all the obvious things first. Is it set up correctly in the BIOS settings. Is the cable plugged in the right way around and in the correct header on the cable. Are the heads clean. Are the heads present (sometimes they can get snapped off). Try another floppy disk (somehow you have to check if the disk is ok). Check to see if any parts are getting really hot. Good luck figuring it out. It can be hard without other parts to compare with. At least floppy cables are cheap!
@@PCRetroTech Did all that... cleaned heads... thing looks like it's out of the box. I set my jumpers to factory default per manual except setting it to single drive only. Last thing to try is a different cable/motherboard, and check for hot components. I do have a thermal camera.
@@beefstickswellington1203 @beefstickswellington1203 Sounds like a difficult case. I can't think of anything else obvious. Cable is the most obvious problem, and motherboard as you say. Unfortunately the latter is a bit expensive unless you know someone who has one you can borrow!
@@PCRetroTech i have a second board from 2004 but not even sure it supports 5.25. My first one was an intel dg865 from 2002 but it seems blind to the drive
@@beefstickswellington1203 Intel boards were good for being compatible with all standards. You might be ok with the 2004 board if it had a floppy connector.
YD380bB drive that does the clacking noise has a bad track 0 sensor. If you move the head to the inside of any drive, and apply power it will home to track 0 (either initially or after loading a disk) (Also, not the old old drives without microprocessors like the tm400) Forgive the rambling. There is a switch or sensor that tells the electronics when the heads are at track 0, this is your culprit
Beware that this is told very much from the PC standard, which is in a way lame they just changed a few things for the heck of it (Big Blue arrogance?), as Shugart was perfectly fine, and maybe even better, as it did not need weird twisting cables and allowed 4 drives instead of 2. Shugart was not only for 8 inch floppies, it worked on 3,5 and 5,25 every bit as well. My 1981 home computer runs 4 3,5 floppy drives of Shugart. Most home computers used shugart (MSX, CPC, Tandy, ZX, Beeb) with the 2 weird ones, off course, being Apple and Commodore.
@@PCRetroTech I'm almost exclusively using 8 bit platforms (C64, C128), and that means all drives are early to mid 1980s technology, and that in turn means few if any 'early' smd caps, and lots of evidently very reliable 1980s 'standard' components. That probably makes a difference. But for me, it is especially the disks themselves. Of the thousands of 1980s era 3.5" and 5.25" floppies I have (all dsdd), typically 1 out of 10 3.5" disks no longer works. For 5.25" disks it is more in the order of 1 out of 250 approx.
@@PCRetroTech Yeah, there might be an unused header for it on the motherboard, or there are cards to be bought for cheap. I've also heard of doing it with drivers. Not sure about adapting DB-25/DC-37 connection to DE-9 though. Amazing video, by the way. I'll need to keep it in mind. I have a Pentium I machine, so I'm able to do it if I ever encounter a drive that needs maintenance.
"If you put alcohol on it, it'll go straight through." No, it gets absorbed by the disk and the disk is dampened by it. That's the point: to use the mildly abrasive disk with the alcohol. If the liquid went straight through the disk then you wouldn't be able to use it that way.
I bought an AT (5170) that had two five and a quarter floppies. I tried them both in a 386 machine and neither seem to work, so I hope to use the IMD program to figure out what's wrong. This video is a godsend and thank you so much for making it!.
I wish you a lot of success. I have been working on a rare 26 pin floppy for the channel, but ran into a number of roadblocks. I haven't fully given up hope yet though. I just don't have a way to run IMD on the machine just yet.
@PCRetroTech One question about the ImageDisk tool you use at 7:21. If a floppy disk drive is misaligned, how does the ImageDisk tool find the errors? What i want to know is, do you need to format a floppy disk on a working correctly aligned drive to use that tool on a not working drive?
It is not that hard to put a 3 1/2 drive in a XT. Just need to change the floppy controller as it is not supported by the system bios, and the other way would have been to use a program called drive.sys
Better yet, there are new BIOS that can be put in an 8-bit ISA slot, "universal BIOS" that supports all size formats including 2.88MB, and is useful for booting virtual floppy drives. The main issue is PC BIOS, not any controller. MFM drives and IDEs are a completely different matter.
I put one of those unified IDE/Floppy/serial/parallell port cards in my PC XT. The card is 16-bit, in an 8-bit slot, so not everything works, but the floppy and the serial port seems to work fine.
@PCRetroTEch I just want to inform you, that information box below the video is there to put some urls on there, if you refereing to some sort of 3rd party software, like you did in the video.
IMD is no longer available from the company who produced it. It is available online, but the links are moved regularly to prevent deep linking. I suggest you use Google to find it.
My question on TH-cam: I know you have an Amstrad PC1512 SD and a 1640. How can you say the Amstrad only has one drive since the SD model I have as two connections on the MB cable? No twists as you say but...? The PC1512 (without the SD on the end) came with two floppy drives on that cable. The Tandon TM65-2L, which spins at 300RPM, that I have, came with DS0 selected like yours. Mine doesn't have the TM jumper but has a resistor pack DIP 16 rated at 3.6K ohms each or there abouts as I remember. I believe that has to be removed if it is not the last drive on the cable. Are you saying my Tandon will work even if it a newer ASUS MB that only allows for one floppy on the cable? Must it be after the twist or before? I fear, since I've tried this, and the twist reverses pin Drive Select 0.4v with Motor Select 5v. My Floppy controller chip looks a bit heat cracked at the moment. Not sure if Tandon drives from the mid to early '80's can handle the voltage change. JP3 is labelled BC or AC which I understand is the spindle motor on or controlled by motor control or drive select according to the owners manual so I use AC. JP2 DRIVE SELECT CONTROL = Drive Selected at all times (on) Drive Select controlled by DSO through DS3 (off) so no jumper there. JP1 is not mentioned. It has F-G or G-H. It came with G-H. I don’t know what that is. Probably the same as that TEAC FD-550 unknown @31:15 Any ideas? Thanks
Unfortunately I am not familiar with the Tandon drive you have, so I don't know what the jumpers mean. Certainly in the DD model the Amstrad had two floppies on the same cable. That is what I see in my machines too. For the rest I don't think I have clear answers. I'm not aware of any incompatibilities such as the ones you allude to. But that doesn't mean there weren't some Amstrad models with some unusual features that I don't know about.
@@PCRetroTech I saw on the Russian web site he said FG was changing how the DC grounds to the housing or goes through caps and resistor. On the Tandon TM65-2L I tried to change H-G (how I found the drive) to F-G [OFF too] bench testing with only a digital ohm meter from 5v or 12v input and the input grounds and resistor levels didn't change. I guess nobody knows what the F-G jumper is for. Only time will tell I guess. In the Chinon FZ-506 UK manual (also known as Toshiba ND-0801AG... at least they look identical to me) there is a jumper Pin07 that the Russian guy says is FG. My Toshiba does label it on the board as FG. The owners manual never refers to it or suggests it should have a "short jumper" on it. Then I decided to try the resistor measurement test with the JP07 ON/OFF with the Toshiba and guess what? The resistance levels did go up on all input lines. The input grounds were then zero to the case when jumpered where before they had 100K on them when OFF. 12v and 5v had similar results but opposite. eg: 12v had 101k OFF and 0.7 ON. What that does I'm not too sure. Maybe just if the housing is not screwed to the case it is needed? Probably more than that.
@@mrwebber35 Yeah I think someone mentioned in the comments that F-G is frame ground. But it's almost certainly not needed except perhaps on some really ancient drives. I don't remember what H-G stands for though. There are drive manuals that mention frame ground in the form of a lug which is supposed to be wired to the chassis/equipment ground.
@@PCRetroTech The Fujitsu Limited M2611T 45MB HD has what I think they are referring to. Has what can be described as a spade clip screwed onto the back that takes the regular YC-180terminal. Can never get that drive to be recognized on anything yet but haven't put much effort into it. If I finish the Amstrad I'll try this old drive on it.
But if I already had 5.25" PC drives to get out, wouldn't they already be on the right settings? I can see cleaning and tweaking though. However, yes, if I try to build a PC that I can put a floppy controller into a slot of or that has one on the motherboard, or that I can adapt to USB in, etc., I will watch out for all of these things when I buy a drive, and I might have been fairly unsuccessful if I hadn't seen this video, because by the time I learned about how to build a PC, it was already 1998 and 5.25"ers were already pretty well phased out of regular use. So now to bring one of these back into use via retro., this video might help me. But I don't have any 5.25" disks from the PC world that I need to worry about. All of mine are Commodore 8-bit; either 1541- or 1571-based. Anyway, other than the wording problems, this was still an interesting, good video!
Yeah, Windos 8+ won't run well on computers that came out in the Windows 7 era or before. When I was still using my 2007 PC, I tried 8 on it and it was SOOO slow and choppy, OMGosh! But Windows 7 would work on it just fine! Now I'm using my current computer that I just built a few months ago from 2020 and '21 parts (call me a procrastinator), after using a 2016 W10 AiO for a while, and this one runs W10 just great and will likely run W12 fine when that comes out too. (Not gonna switch to 11 unless/until it gets a lot better; it's the new "bad Windows version.")
"There's no 3.5" bay in this machine." That doesn't stop someone; that's why companies make adaptors! But then again, if anachrony is against your rule and you think these systems weren't meant to be upgraded with one, then just that alone would get in your way.
Why would HD floppy drives be less expensive than the ones with _less_ technology in them (normally higher technology = harder or takes longer to make = costs more)?
Apparently you can. I think it at minimum has to boot to DOS, you have to have mode.com on the machine (and obviously a working keyboard to run it). But that's it, apparently.
@@PCRetroTechI would use Norton Commander with NC Link - just null modem serial and 4 files necessary. If there is a dos 6.22 it is possible to use interlink for that - pretty handy and I use it often to do the setup of old machines. I have a boot floppy with fdisk, format and interlink client and other machine boots as interlink server and I just copy what I need on a fresh setup without messing with MFM controllers or just moving things around
i have an old TEAC 5.25 inch floppy drive. it does spin, however it stops after a while. the little light isnt turning on at all. It is in fact NOT connected to the motherboard, only to the PSU. Im using it with a socket 462 motherboard, slightly too modern but... what am i doing wrong that the light isnt turning on? any help is welcome
The light should indicate when there's activity on the drive, which wouldn't be the case if it isn't connected to the motherboard (unless you have the data cable around the wrong way). It should be possible to see if your board supports the drive by looking in the BIOS. Of course there also needs to be a floppy connector on the board, but I assume you have that.
@@mrkevinkillercrafterminecr895 No, that should be the right name. Sometimes you find them in boxes of assorted cables, though make sure the connector is the right one for your drive.
Haha, why did you think that setting up an HD drive as an SD one in the BIOS would _enable_ both instead of _limiting_ it to SD (or making it not work at all, because of misidentifying it)?
"2 minutes to copy a file onto a disk..." ...From what: the cloud? A USB stick? And onto what kind of disk: the hard disk, a floppy disk, or...? And it wouldn't just be 2 minutes; it depends on the file size.
"There's a reason for that." Yeah, but there's a reason for _everything,_ even when it's just accidental or otherwise insignificant. So... of actual importance would be that there's a _specific_ reason for it. Right?
I haven't encountered any problems with Y2K on the machines I have. I didn't do anything to the BIOS myself, but perhaps the previous owner did during the Y2K "crisis".
If you need the answer to "Why most 5.25" Floppy Drives seem broken, but aren't!" skip all the way to 33:00 minutes where the answer is finally tucked into this movie about drives. The first 33 minutes is a video called "All you want to know and a whole lot more about disk drives"
I wouldn't recommend skipping it. All that detail explains WHY those changes work, and will help to decipher what to do when your drive's jumpers don't match 100% with the ones in his collection. If you're the TLDR type, might I suggest using a computer with double or triple digit MHz clocks instead, and bypassing the whole 5.25" drive scene entirely.
@@nickwallette6201 Absolutely. If you’re just impatient and want instant gratification, vintage computing isn’t for you. These machines require patience, persistence and knowledge to repair. It’s in your best interest to know as much about floppy drives as possible if you are going to repair them.
Hey I'm struggling to get a Mitsubishi MF504B-386U setup, have tried to follow the video but I just cant seem to get the thing working correctly, any help would be greatly appreciated!
Wow thanks very much ! one of my drives is exact the same CHINON and my jumpers are exactly configuration like yours , but in a modern systems after year 2005 the floppy controllers are shit and only supports one floppy A: , but the other problem is that usually in my experience 5.25 floppy's did not refresh the contents of a previous floppy disk like you explained her e 24:00 so I have to remove this D-R jumper to fix the issue :)
I tested did not work :( without this jumper floppy trow "Not ready" :( I changed the floppy with other model and it works ,no problem for refreshing the disk contents , strange
Wait "refresh previous disk?!?!" You realize there is absolutely no memory in floppy disk controllers, at all, right? Nothing ever gets "refreshed" in floppy drives, ever. It reads, writes and is subject to management by Disk Operating Systems. Are you talking about a virtual drive?
The floppy drive controllers are extremely simple, mostly everything is done in physical circuitry. Complaining about them not being as good is like saying "wheels aren't as round as they used to be. Remember when wheels were circles? Why are new ones less circular?" Is there are serious translation problem here?
@@squirlmy yes I am talking about real hardware in real mode DOS. When I read one 5.25" diskette, then replace it with another ine, the system shows the contend of the first disk and it stays like that until I reboot the PC
This year I have managed to repair 22 5.25" floppy drives and I must say it was quite interesting and challenging to do. I only had to cannibalize one drive to do it. I couldn't have done it without IMD, which I had used before in my attempts to get 8" floppy drives formatted on a PC. There is one drive left for which I cannot get any data from the heads at all, no matter how much realignment I try to do.
All these drives were sold to me untested, for parts or repair, but actually I was surprised to find out that about a quarter of them were fully functioning and that I could not find any fault on them. I think it helps if a Pentium I or 486 motherboard is used to test them instead of Pentium II/III/4 motherboards, as those might have trouble dealing with these drives that were already considered obsolete in that era.
Other common faults I found and fixed were:
1. The rail on which the head assembly moves is caked in dust, causing the stepper motor to have trouble and also causing the head assembly to skip tracks --> Clean and lubricate the rails
2. Both heads don't give any data at all in IMD --> Clean thoroughly with a cleaning disk with isopropyl alcohol applied to it. Allow IMD to make around 10 passes. (Possibly someone used a chemically decomposing disk that thoroughly fouled up the heads, or perhaps used the wrong cleaning chemical)
3. IMD shows that not all 18 or 30 sectors can be read during alignment test --> Realign head assembly using IMD, sometimes the whole assembly, sometimes just head 1
3. The drive keeps spinning the disk permanently after inserting a disk and the disk cannot be accessed. --> Check if the disk locking mechanism is fully engaged, or the track hole sensor or disk insertion sensor are fouled up or not properly aligned. Repair the locking mechanism block (sometimes it is cracked) by glueing or melting it back together (soldering iron). Possibly place some duct tape underneath the pusher block to have it exert more force. Realign the sensors while the disk is spinning and while running IMD to find out the correct position
4. The drive does not pre-spin upon disk insertion, leading to a badly aligned and wobbly disk --> check if the detection mechanism still has its spring and replace it if it is missing or broken. A containment for the problem is also to manually 'pump' the locking lever after disk insertion to make sure it is aligned to the spindle properly
5. The drive kind of works with reading, but has trouble writing. At IMD alignment it does not show the full 18 or 30 sectors read at each track. Perhaps at some tracks there is 18 or 30, but not at all of them --> Check the head 1 spring and put duct tape or a small foam block underneath to change the force the spring exerts. Realignment in stepper direction is not always the solution
6. Jumper settings are not compatible with the PC, leading to strange behavior that has no mechanical reason --> Find out what the jumpers are for (via documentation, or other sources like the above video) and correct their placement
Then there were some other less common faults that were more difficult to solve, like:
7. Obvious damage like a circuit board with a corner snapped off, or missing mechanical components, like the spindle motor --> try to salvage what is possible. Cannibalize a similar drive. Figure out the severed connections on the circuit board and solder on wires as needed. What is needed can be inferred by observing the drive behavior. Documentation about the circuit board is nearly impossible to find.
8. No visible damage at all, but drive behavior is still out of specification, for example it doesn't detect track 00 --> find out where the problem is in the circuit board by comparing with a functioning drive of similar model if possible. I found out that voltage on the track 00 sensor was too low on several drives and I applied an additional resistor between it and the 12V, causing it to be in spec again.
What I used to fix those 22 drives:
- Pentium I test PC
- good working drives and secondary PC for reference
- new-old-stock unused disks
- the IMD program
- duct tape
- WD-40
- cleaning disks
- isopropyl alcohol
- multimeter
- soldering iron
- misc parts like replacement springs, bolts, resistors, wires
- common sense
- deductive logic
- patience
Thanks for the really comprehensive repair notes. I'm sure this is going to help quite a few people.
@@PCRetroTech You're welcome. I am simply hoping we can save as many of these drives as possible. Your video, especially the part on the jumpers was helpful to me to try things I hadn't thought about before.
Hi, where can i get a good cheap 5 1/4 1.2 mb drive?
@@MacmorMach In my experience with the current era, you either buy a single good drive that is expensive or a single cheap drive that probably isn't in a good condition. Buying a single good cheap drive nowadays is only possible if you get it from someone who considers it junk, but there aren't many such opportunities left. Or you can buy a large batch of non-functioning ones, like I did, and repair them.
WD 40 is Not a lubrication liquid. WD literally stands for Water Displacement. It is a degreaser. It will leave residue and it will migrate.
Everyone should just throw away WD-40. When you use it and it makes the device stop squeaking, it is only a temporary fix. If you use proper lubricant the device won't squeak either.
For isopropyl you want to use the lowest water content that you can find, at least 85% and you don't want to use it on anything rubber because the rubber will degrade. In most cases the rails in a drive should not have any lubrication on them. They are highly polished and they slide in copper sleeves. Lubrication will attract dirt, and could cause the polished rails to have greater drag(its counter intuitive) since there is now some substance between the two surfaces. If you are using a drive which drives the head by a worm gear, you want to put a small amount of white lithium grease on it. There is also electrical contact cleaner but you probably wouldn't need it inside of a drive but it will clean switches and not leave volatile residue on the contact surface.
I frequently buy old PC's from thrift stores, garage sales, and estate sales. Not for the PC itself but for the floppy and hard drives and expansion cards that are inside!!!
I recently purchased an old PC that was loaded with expansion cards. Including a Trackstar, an Apple emulator card.
Normal density (Mitsubishi MF504B,C 5.25" HD Drive)
Logical ‘1’ selects high density read/write operation and logical ‘0’ selects
normal density. With option jumpers SS shorted and SB open, this line changes
the motor rpm and read/write 'operations are performed 400mS after density
change, to allow the motor speed to stabilize. When motor rpm is switched the
head must be moved to track 00 before reading or writing. With SB shorted and
SS open the motor speed is always 360rpm so no waiting or track 00 seek is
necessary.
Fantastic! Thanks for the information!
@18:10 So I did use the wrong cable for a MFM hard disk drive! I never thought they are different. Thanks for your information!
This video is excellent,
Well covered, very informative and comprehensive.
We need to document this online
Would the Rossmann Repair Group wiki be appropriate for something as retro as this?
This was quite a good disc-ussion of the subject.
It's a humble contribution to this sector of computing. I'm glad it was useful to some of the retro-heads out there.
@@PCRetroTech It does help keep us on track.
"Disk-cleaning disks"?
What: you're gonna clean another disk with that?
that Chinon Drive where you talked about the last jumper and old 8" drives, whould make this suitable out-of-the-box to be configure for any (retro) machine that requires Shugart compatible drive interface, meaning it could be configured to work properly with an Amiga, MSX etc without doing any bodges
I managed to get the remaining two drives working guys, plus another one of the drives that had some issues towards the final track. This required more advanced techniques. I'll describe those here, but please don't blame me if you screw up your drives attempting this. In fact, I got one of the three working before accidentally slipping with my screwdriver and completely destroying one of the heads!
In all cases, the issue was head alignment. In the case of the drive that sounded stuck, I had to move the stepper motor assembly so that it could move the head back to track zero. You can use the head align tool in IMD and move it back and forth until it can read track zero on head zero. Press D to check it can actually read the data.
From here, all three drives had the same problem: everything was fine on head 0 (the one towards the bottom of the drive), but something would go wrong on head 1. In all three cases, the solution was the same: unscrew the two screws that hold the head in place and move it about until the head alignment program showed 18 matches. This was particularly hard on one drive because you have to remove a tensioning spring that is required to actually read data, so you have to apply pressure with your finger whilst you move the head about while watching the data in IMD. It didn't help that the sled was cracked and I had to superglue that back together before I could get anywhere.
So I've now successfully rescued all my drives, except the one I accidentally broke beyond all possibility of repair (I literally sheared the head right off).
I tested that the drives can now read 20 disks formatted on a known working drive and filled with data up to the last track!
Splendid. I have a broken 5 1/4 drive -- I'll give all this a go. Cheers!
Fantastic - I didn't know about IMD. I feel like head calibration is going to be increasingly important to save the drives that are still functional today, but will slowly fall out of alignment over time. And, of course, the expert information out there isn't well documented online (having been of most use before the Internet was popular) and literally dying out with the 80s long-beards.
I feel you on the damaged head. I started disassembling floppy and CD drives as I acquired them, and giving them all a thorough cleaning and re-lubrication. There was one 3.5" drive that I was having trouble finding out how to remove the cover, and pulled a couple screws out of the bottom. Those were the ones holding the Track 0 sensor in place. So that drive is now wildly out of calibration. I may try to recover it now that I know about IMD, or pass it on to someone who has the motivation to do so. Glad I didn't just throw it away, though!
Now i am wondering if the many 3.5 floppy disks i threw away 25 years ago really had read errors or were just a cause of a misaligned disk drive?
0:00 before watching the video I'm going to guess. From experience old lubricant can cause motors to be completely stuck or sticky enough to cause the drive to give up...next is the jumper settings, some drives need them set correctly...then I'll say stuck or sticky head mechanisms/dirty heads...and lastly incompatible disks, wrong formatting etc...this may be kinda cheating I've just listed all the issues I know lol. But if you read all of my nonsense I Appreciate It! Thank You For Your Time And Effort! I Hope You Are Having A Great Day Or Night!
FYI that head alignment test is called "an azimuth test" and you can hook up a scope to to see the the _analog_ signal coming off the head, IMD is just doing a _digital_ test using "data integrity checks" (data does not change between passes, and track overlap doesn't distort sector reading). Azimuth alignment with a scope is how you "fix" a Wren / Falcon / Phoenix Drive platters when its not reading, as Usagi Electrical showed when restoring the CDC mini computer hard drives.
Great coverage. It can definitely be difficult dealing with 5-1/4" drives, particularly if they have come from different types of machines. Thanks!
May I recommend the cable transfer program FastLynx? It can transfer files over RS-232 null modem cables or parallel port cables including those for Laplink. One of its fanciest features is the ability to transfer itself to a DOS PC over null modem, so you don't even need a working floppy drive.
I feel the same way about you regarding retro programming. It is just not the same using a modern day computer to learn programming as was done in the earlier days.
The FG jumper usually stands for Frame Ground, it connects the ground signal from the data connector to the metal frame of the drive, it's do deal with ground loop problems. Most of the time that jumper won't matter
Cool, thanks for the info. One mystery solved!
Thanks for this video. Lots of good helpful info. On the one drive, you mention an "IM" and "NM" jumper setting; that could possibly be for maskable or non-maskable interrupts.
Sadly, floppy drives have no ability to raise interrupts, maskable or otherwise. That is entirely the domain of the controller.
This video reminded me of the days when you had to specify cylinders and sectors in the bios of hard drives. Also, I have a memory that you had to put that in COBOL programs as well. Thank you.
I've heard it called drive geometry. I recall setting that stuff up with the hd on my first computer. IDE made it all go away. : )
Thank you for your video. I purchased this floppy drive unit but wrongly purchased the data cable and your video has solved the problem for me. Otherwise I won't be able to use this device forever.
It's amazing how many things I've learned from this video. I have a 5.25 floppy drive that works. When I can I send you the pin configuration. I am very grateful to you. All the best.
You are exactly correct. If I format my 360k disks in my original IBM PC and write the images on my Pentium with a 1.2mb drive I can then read them back on my IBM PC. If I format the 360k disk on my 1.2mb drive it will not read it back on the IBM PC or at least not 100% properly. Good Video.
Yeb cause the heads of a 1,2M are smaller. See as a 2 lane driveway. The 2 lanes together are 360K Wide, a 1,2M uses just 1 lane at 360K. So if you write a 360k formated floppy with a 1,2M just 1 lane is read and other lane is un toughed. If there were data on that 360k, at 2nd lane there is still some old data left. The 1,2M does not bother as it just read 1 lane, BUT the 360k drive reads both lanes at the same time. So now there is a mix-up of signals.
Floppy drives are the coolest thing
Good video and good comments, thank everyone
The more I watch videos about retro hardware, the more I remember what a pain in the ass was setting PCs up in the late eighties
PnP was a blessing, although it had many issues at the beginning.
The satisfaction of getting it all working is part of the nostalgia for me.
@@BlackEpyon indeed. Although I don't miss fiddling with the software side of things, like config.sys and qemm to squeeze as much ram as possible to run DOS games.I was very good at remembering all the various bits and pieces, nowadays I'd have to research them. 😭
@@bufordmaddogtannen Yeah, I've got a Tandy 1000 RSX that I'm running SCSI on because it's a slim-profile case that doesn't have room for extra drives. I could revert back to an IDE hard drive (limited to 504MB) if I had to, but the external CD-ROM and Zip drive are both SCSI, and some games do need the CD to install or run. That means moving everything possible into highRAM, because THAT particular game (StarTrek 25th Anniversary CD) needs 560K of conventional memory, and won't take 'no' for an answer!
@@BlackEpyon SCSI FTW. Been there, done that.
Qemm has always been a savior, although I am not sure if the scsci drivers suck too much ram.
Have you tried copying the content of the cdrom on the hard drive on a temp directory and adding a drive letter pointing to it using the subst command? It could save the trouble of loading the cdrom driver and free some memory.
Sometimes it is enough to fool games.
@@bufordmaddogtannen Na, I tried. Some games won't even let you install from a directory, or change the directory it installs to.
Having a 2GB SCSI hard drive in a 386 is TECHNICALLY within the bounds of the time period (if you were building a server), but boy howdy would that have been an expensive build for the day! I'm kinda surprised the BIOS even allowed me to choose to boot from SCSI.
I have network cards in all my retro machines and use mtcp - having a nas with ftp server active makes file transfers a breeze to all of my machines.
Excellent presentation, I learned a lot of things I didn't know back then...
Gday Gday
I am totally in your world mate I have just started restoring some XT 8088 system i recently found at the local refuse site.
Could not bare to let them rot.
I totally want to get started on 8088 Assembly.
Glad to see this video.
Now im gonna watch the rest
My condolences. Two words: memory segmentation. 😱
That's interesting: that when the HD version of these drives came out, there was a generation of them that were upwards compatible between 360K and 1.2M, but even newer ones aren't. I'm used to the times when once a line of devices is made upwards compatible, that's the way they stay. Why would these have regressed somewhat to not being upwards/backwards compatible anymore?
Thank you for a very informative video!
5:14 - "Just buy any old computer case" - Man, even cases are getting stupid. $100-200 these days for an empty case!!! It's normally more cost effective to just buy built systems. It also seems to be cheaper than to build from parts. I've been piecing together a 486 VLB build and was in over $300 before I knew it!
I hear you. In some places people don't even seem to be able to get cases. I was lucky that I got machines and cases when I did.
This was great 👍 my 5.25 drives all have CPUs in them, cough *commodore* cough.
I'm sure someone answered that FG is Field Ground, grounds the drive electronics to the chassis of the drive if I recall right.
You're correct sir
Frame Ground.
There's a certain driver floating around, PalmZip I believe, that purports to get parallel Zip drives working on 8088 hardware. It's going to be horrifically slow, but it should at least work.
Yes, I am using that, though Zip disks aren't so reliable unfortunately.
@@PCRetroTech Heh, I ran into that problem myself actually. Mostly I just have the drive now so I can recover old disks people bring me, hopefully.
I actually had to repair my ATAPI drive, I'd managed to be enough of a bonehead that I nearly sheared one of the pins of the IDE connector off. I don't usually do SMD soldering, but the pads there were just big enough that while I was unable to get the original pin back in place how it was meant to go, I bodged a wire between it and the pin and got it back in the hole. It makes just enough contact now that the drive works, at least.
Why mix-up 360k Floppy's with use of 1,2M drive fails often.Cause the heads of a 1,2M are smaller. See as a 2 lane driveway. The 2 lanes together are 360K Wide, a 1,2M uses just 1 lane at 360K.And step over 2nd lane for next track, leaving that 2nd lane unused. So if you write a 360k formated floppy with a 1,2M just 1 lane is read and other lane is un toughed. If there were data on that 360k, at 2nd lane there is still some old data left. The 1,2M does not bother as it just read 1 lane, BUT the 360k drive reads both lanes at the same time. So now there is a mix-up of signals. A format DOES not whipe clean the disk, completely. So there is still some old data left at those lanes. The reverse is the same kind of mix up data.
--> Then from WIN95 on, 360k drives Does not work at all any more! Even if Bios set up is correct. 360k only works in the DOS mode.
For many old upscure Floppy's from old non DOS machines, take very care of your PC Floppy controller if it can handles that format. Not all floppy controllers can handle all kinds of formats. Indeed use TESTFDC for that to check out.
The coercivity of the magnetic media is also different.
@@PCRetroTech Thats indeed next problem with use High dence disks in 360k drive. Often lead to bad data. Its a struggle. But in the end you get the hand of it.
I have a 5.25" floppy drive, works fine on a user unfriendly and heavy machine i don't want to use.
On a computer i want to use, i can select the drive, but, for some reason the bios jumps back to a 3.5" floppy drive. The bios battery is good and the floppy drive setting is the only "stubborn" setting. (if no floppy drive is connected, the bios says correct "none" )
That sounds very annoying. To me it sounds like a BIOS issue rather than a floppy drive problem, but who knows.
I was tidying and reorganising my 'nerdy clutter' as my wife calls it (even though she doesn't mind me keeping it all), and I swore I had a 5.25" 1.2Mb FDD in storage, and to my dismay I can't find it.
What I can find is a load of 5.25" floppy disks I would like to make digital copies of before they all succumb to mould, or being wiped somehow.
Anyone in the Hampshire area of the UK have a spare drive they're willing to part with along with cable? I'd be so happy to find one for as cheap as possible as I'm on benefits and I'm my wife's full-time carer.
That's a shame you can't find your drive. I'm not in the area, but hopefully you can find someone who can help you rescue your disks!
You need laplink... (parallel port link between machines) or even just a serial port connection.
some redundancy is enough and retro fitting some modern tech is not a bad thing you still get the feel of working with the old nostlgia driven fun
It's weird, my experience has been that most 3.5" floppy drives seem broken or questionable while all of the 5.25" drives I've used have been extremely reliable. I wonder why that is.
That must be a regional thing. Perhaps there's a very good 5.25" repairer or something.
No, i can second that.
I still have my 5.25" disk drive in a working condition. It only had to go to a repair shop once.
But the 3.5" disk drives did fail very soon. In the nineties, when floppy disk were still common i had to buy several 3,5" disk drives. The situation got worse, when the CPUs started to require active cooling fans and when the floppy disk drive as a storage medium become obsolete. I assume, the cooling fans do drag some dust into the floppy drives.
And the situation is much worse when we are talking about the floppy disks.
I threw away over 90 x 1,44 MiB 3,5" floppy disks in the nineties because of read errors, but only about 10 x 5,25" HD floppy disks.
Great video overall. Nevertheless, some information is misleading or incomplete. In IBM PC compatible computers, you generally want drives to run at a fixed speed. DD (360KB) drives should operate at 300RPM, and HD (1.2MB) drives should operate at 360RPM, no matter what the level on pin 2 is. The BIOS deals with DD diskettes running "too fast" in HD drives by choosing 300kbit/s instead of 250kbit/s data rate in the floppy controller.
On the drive side, there is only one kind of stepping, which steps the drive one drive track. "double stepping" means that the disk in the drive is formatted in a way that the disk tracks are twice far apart than the drive tracks. This happens if you try to access a 40-track disk in an 80-track drive. The drive you used with IMD is thus obviously an 80-track drive. In IBM PCs (don't know about Amstrad pecularities), all 80-track drives are HD drives. So this drive is likely a 1.2MB drive (at which operation at 360RPM is normal), but we can't exclude the drive being a non-PC 720KB drive (80 tracks, but no high-density support). This also explains why testfdc failed: The drive spins so fast that the formatting or writing the test track takes more than one revolution of the drive, so formatting the last sector will overwrite the beginning of the track, or formatting will be aborted because the index hole is detected during formatting. testdfdc with "1.2" might have passed at that point.
Using "DS0" to use a single drive before the twist as A: drive will likely not work as intended. On most drives, the DS0 to DS3 jumpers are just choosing which DS wire on the floppy cable will be used as DS for the drive. There is only *one* standard motor-on signal the floppy connector. The pin used by IBM for "A: motor on" before the twist is has been repurposed from one of the higher "DS" lines (DS2 or DS3). A drive before the twist jumpered to DS0 thus will respond to "B: motor on" and "A: drive select", which is not fine, unless you jumper the drive to use "drive select" instead of motor on to turn on the motor. In that case, you might get issues if the BIOS tries to keep the motor of drive A running while drive B is selected, and then re-selects drive A without waiting for an appropriate spin-up time (although I don't know whether the BIOS of IBM-compatible PCs generally does that).
The advice to jumper Drive Ready / Disk Changed generally to "Drive Ready" is bad. The PC BIOS is able to handle a "disk change" signal, and will return "error code 6" on the first floppy access after a disk change. DOS uses this error code to flush the FAT / Directory buffers. Not having a working disk change signal may inadvertantly use the old directory/FAT on a new disk, which will inevitably lead to data corruption. But there is one extra piece of information, which makes your advice partially correct: Classic 360KB PC drives used pin 34 as "drive ready", and the BIOS/DOS don't expect disk change signalling to work if the CMOS type is "360KB". For 1.2MB drives, you better use "disk changed", though.
Thanks for the in-depth analysis and sharing your (considerable expertise). Other commenters have noted how useful the comments on the video is (due to the incomplete nature of what I presented). This greatly adds to the usefulness of the resource!
I have an old 386 based IBM compatible with a 3.5 drive that fits the 5.25 bay they exist and use the same ribbon cable style that slips on a pcb instead of the ide pin style
Didn't know that the compatibility was this bad between those drives
Prob because of the age of the drives, losing their calibration.
A trick I have used on floppy disks when all else fails is nice strong magnet fully erase all the formatting completely then reformat the disk in what ever computer your planning to use it in I used a magnet from a blown 12inch woofer and still had to make sure passed it all over the disk for it to be completely wiped I would only use this as a last resort but have had good success doing this.
I use a bulk tape eraser for a reel to reel tape recorder on my floppy disks to erase them and it works great .
@zaper, this way you can only magnetize the disk beyond the drive's ability to re-magnetize it during format. Use the demagnetizer with slowly decaying AC electromagnet
I love watching a video about an item from the age of the dinosaurs.
ENIAC is a dinosaur, but definitely not an 8088 PC. When ENIAC is the dinosaur and IBM System/360 the Mastodon, then the 8088 is an Egypt pyramids. So very new.
oh my god same all the one i have here are not working and judging by this video i am gonna learn some cool stuff.
This is awesome fella.
Im excited
Keep up the great work mate! I have subscribed - Love your content :)
you deserve way more views than those pseudo retro guys who blantly slap an ssd or sdcard instead of using floppies because they lazy to do it the proper way. thats the proper way.
Thanks.
I'm #663 like... nobody has any objections I see. Thanks for telling us your adventures.
Would genuinely like to thank you for saving me ~200$
No problems. I am glad this helped you out!
These drives are beautifull beasts. To me they are still usefull for anyone that work with pure text, manuscripts, scripts... caracters don't thake more space than back in the day. Also, 5 1/4 is more reliable than the 3 inch diskettes.
No, this is not boring. This is memories... mmmm , i will try to power up my 386 and check my drive :)
I have a ton of floppys that I need to go through. Any ideas where I can get a 5.25" external (USB) floppy drive?
No, I have not seen such a thing. There was basically no market for them.
Yeah, if you have the drive out of a computer or external case then you might as well use the cotton-swab method to clean it; that's usually more thorough, based on what I've heard.
Ugh, I got my first 5 1/4" drive ever and it looks absolutely perfect, but doesn't do anything but spin the disk briefly when inserted. Not sure if it's my controller, cable, or drive... no spares to deduce what's going on! Sensor LEDs still work, no dirt or corrosion at all. Everything moves freely on track. Just nothing initiates.
Check all the obvious things first. Is it set up correctly in the BIOS settings. Is the cable plugged in the right way around and in the correct header on the cable. Are the heads clean. Are the heads present (sometimes they can get snapped off). Try another floppy disk (somehow you have to check if the disk is ok). Check to see if any parts are getting really hot. Good luck figuring it out. It can be hard without other parts to compare with. At least floppy cables are cheap!
@@PCRetroTech Did all that... cleaned heads... thing looks like it's out of the box. I set my jumpers to factory default per manual except setting it to single drive only. Last thing to try is a different cable/motherboard, and check for hot components. I do have a thermal camera.
@@beefstickswellington1203 @beefstickswellington1203 Sounds like a difficult case. I can't think of anything else obvious. Cable is the most obvious problem, and motherboard as you say. Unfortunately the latter is a bit expensive unless you know someone who has one you can borrow!
@@PCRetroTech i have a second board from 2004 but not even sure it supports 5.25. My first one was an intel dg865 from 2002 but it seems blind to the drive
@@beefstickswellington1203 Intel boards were good for being compatible with all standards. You might be ok with the 2004 board if it had a floppy connector.
This is quite some quality content. Thank you!
YD380bB drive that does the clacking noise has a bad track 0 sensor. If you move the head to the inside of any drive, and apply power it will home to track 0 (either initially or after loading a disk) (Also, not the old old drives without microprocessors like the tm400)
Forgive the rambling. There is a switch or sensor that tells the electronics when the heads are at track 0, this is your culprit
Beware that this is told very much from the PC standard, which is in a way lame they just changed a few things for the heck of it (Big Blue arrogance?), as Shugart was perfectly fine, and maybe even better, as it did not need weird twisting cables and allowed 4 drives instead of 2. Shugart was not only for 8 inch floppies, it worked on 3,5 and 5,25 every bit as well. My 1981 home computer runs 4 3,5 floppy drives of Shugart. Most home computers used shugart (MSX, CPC, Tandy, ZX, Beeb) with the 2 weird ones, off course, being Apple and Commodore.
Annecdotal 'evidence', but I find 5.25" drives and floppies much much more reliable than 3.5" ones, and easier to repair when they are malfunctioning.
For reliability I've found the opposite (or just got lucky with 3.5" ones), but I can definitely imagine repairing the earlier drives is easier.
@@PCRetroTech I'm almost exclusively using 8 bit platforms (C64, C128), and that means all drives are early to mid 1980s technology, and that in turn means few if any 'early' smd caps, and lots of evidently very reliable 1980s 'standard' components. That probably makes a difference.
But for me, it is especially the disks themselves. Of the thousands of 1980s era 3.5" and 5.25" floppies I have (all dsdd), typically 1 out of 10 3.5" disks no longer works. For 5.25" disks it is more in the order of 1 out of 250 approx.
@@c128stuff Oh wow! That is an interesting observation. I have had a LOT of 5.25" diskettes die.
"Rubber band..."
* belt (Rubber bands are those more flimsy office supplies.)
For file transfer, could you not use a serial connection?
Yes, I guess you could. Most later PCs have USB rather than RS-232, but I imagine there are probably ways around it.
@@PCRetroTech Yeah, there might be an unused header for it on the motherboard, or there are cards to be bought for cheap. I've also heard of doing it with drivers. Not sure about adapting DB-25/DC-37 connection to DE-9 though.
Amazing video, by the way. I'll need to keep it in mind. I have a Pentium I machine, so I'm able to do it if I ever encounter a drive that needs maintenance.
@@starchief93 Thanks. I'm glad you found it interesting!
Why would 8" floppy drives spin all the time (while hopefully having a head mount/unmount feature)?
"If you put alcohol on it, it'll go straight through."
No, it gets absorbed by the disk and the disk is dampened by it. That's the point: to use the mildly abrasive disk with the alcohol. If the liquid went straight through the disk then you wouldn't be able to use it that way.
I bought an AT (5170) that had two five and a quarter floppies. I tried them both in a 386 machine and neither seem to work, so I hope to use the IMD program to figure out what's wrong. This video is a godsend and thank you so much for making it!.
I wish you a lot of success. I have been working on a rare 26 pin floppy for the channel, but ran into a number of roadblocks. I haven't fully given up hope yet though. I just don't have a way to run IMD on the machine just yet.
"Screw adjustment or potentiometer..."?
The screw adjustment _is_ the potentiometer (one of those micro ones).
"So if you don't want your disk destroyed..."
Heh, it doesn't destroy the disk.
Very good video,I tried everything you meticulously mentioned but still my mitsumi won’t work just can’t read the disks but thanks for your help 👍🏽
That's a shame. Perhaps there is some electrical problem. Capacitors are the usual suspect.
@PCRetroTech
One question about the ImageDisk tool you use at 7:21.
If a floppy disk drive is misaligned, how does the ImageDisk tool find the errors?
What i want to know is, do you need to format a floppy disk on a working correctly aligned drive to use that tool on a not working drive?
Yes you need a known good disk. This is covered in the manual that comes with the program.
@@PCRetroTech Thank you a lot for your answer.
It is not that hard to put a 3 1/2 drive in a XT. Just need to change the floppy controller as it is not supported by the system bios, and the other way would have been to use a program called drive.sys
Sure, I imagine 8 bit controllers exist for that. I just don't happen to have one.
Better yet, there are new BIOS that can be put in an 8-bit ISA slot, "universal BIOS" that supports all size formats including 2.88MB, and is useful for booting virtual floppy drives. The main issue is PC BIOS, not any controller. MFM drives and IDEs are a completely different matter.
I put one of those unified IDE/Floppy/serial/parallell port cards in my PC XT. The card is 16-bit, in an 8-bit slot, so not everything works, but the floppy and the serial port seems to work fine.
Why is there even an option for having a floppy drive spin all the time like hard disks do when they're not asleep while the computer is on?
@PCRetroTEch
I just want to inform you, that information box below the video is there to put some urls on there, if you refereing to some sort of 3rd party software, like you did in the video.
IMD is no longer available from the company who produced it. It is available online, but the links are moved regularly to prevent deep linking. I suggest you use Google to find it.
@@PCRetroTech I already found it. I searched for Daves Old Computers.
The software is from Dave Dunfield.
So what are those other jumper headers for on that drive that you said was hardwired to DS?
My question on TH-cam:
I know you have an Amstrad PC1512 SD and a 1640. How can you say the Amstrad only has one drive since the SD model I have as two connections on the MB cable? No twists as you say but...? The PC1512 (without the SD on the end) came with two floppy drives on that cable. The Tandon TM65-2L, which spins at 300RPM, that I have, came with DS0 selected like yours. Mine doesn't have the TM jumper but has a resistor pack DIP 16 rated at 3.6K ohms each or there abouts as I remember. I believe that has to be removed if it is not the last drive on the cable. Are you saying my Tandon will work even if it a newer ASUS MB that only allows for one floppy on the cable? Must it be after the twist or before? I fear, since I've tried this, and the twist reverses pin Drive Select 0.4v with Motor Select 5v. My Floppy controller chip looks a bit heat cracked at the moment. Not sure if Tandon drives from the mid to early '80's can handle the voltage change. JP3 is labelled BC or AC which I understand is the spindle motor on or controlled by motor control or drive select according to the owners manual so I use AC. JP2 DRIVE SELECT CONTROL = Drive Selected at all times (on) Drive Select controlled by DSO through DS3 (off) so no jumper there. JP1 is not mentioned. It has F-G or G-H. It came with G-H. I don’t know what that is. Probably the same as that TEAC FD-550 unknown @31:15 Any ideas? Thanks
Unfortunately I am not familiar with the Tandon drive you have, so I don't know what the jumpers mean. Certainly in the DD model the Amstrad had two floppies on the same cable. That is what I see in my machines too. For the rest I don't think I have clear answers. I'm not aware of any incompatibilities such as the ones you allude to. But that doesn't mean there weren't some Amstrad models with some unusual features that I don't know about.
@@PCRetroTech I saw on the Russian web site he said FG was changing how the DC grounds to the housing or goes through caps and resistor. On the Tandon TM65-2L I tried to change H-G (how I found the drive) to F-G [OFF too] bench testing with only a digital ohm meter from 5v or 12v input and the input grounds and resistor levels didn't change. I guess nobody knows what the F-G jumper is for. Only time will tell I guess.
In the Chinon FZ-506 UK manual (also known as Toshiba ND-0801AG... at least they look identical to me) there is a jumper Pin07 that the Russian guy says is FG. My Toshiba does label it on the board as FG. The owners manual never refers to it or suggests it should have a "short jumper" on it. Then I decided to try the resistor measurement test with the JP07 ON/OFF with the Toshiba and guess what? The resistance levels did go up on all input lines. The input grounds were then zero to the case when jumpered where before they had 100K on them when OFF. 12v and 5v had similar results but opposite. eg: 12v had 101k OFF and 0.7 ON.
What that does I'm not too sure. Maybe just if the housing is not screwed to the case it is needed? Probably more than that.
@@mrwebber35 Yeah I think someone mentioned in the comments that F-G is frame ground. But it's almost certainly not needed except perhaps on some really ancient drives. I don't remember what H-G stands for though.
There are drive manuals that mention frame ground in the form of a lug which is supposed to be wired to the chassis/equipment ground.
@@PCRetroTech The Fujitsu Limited M2611T 45MB HD has what I think they are referring to. Has what can be described as a spade clip screwed onto the back that takes the regular YC-180terminal. Can never get that drive to be recognized on anything yet but haven't put much effort into it. If I finish the Amstrad I'll try this old drive on it.
@@mrwebber35 Will be interesting to see what you find.
But if I already had 5.25" PC drives to get out, wouldn't they already be on the right settings? I can see cleaning and tweaking though. However, yes, if I try to build a PC that I can put a floppy controller into a slot of or that has one on the motherboard, or that I can adapt to USB in, etc., I will watch out for all of these things when I buy a drive, and I might have been fairly unsuccessful if I hadn't seen this video, because by the time I learned about how to build a PC, it was already 1998 and 5.25"ers were already pretty well phased out of regular use. So now to bring one of these back into use via retro., this video might help me. But I don't have any 5.25" disks from the PC world that I need to worry about. All of mine are Commodore 8-bit; either 1541- or 1571-based. Anyway, other than the wording problems, this was still an interesting, good video!
Yeah, Windos 8+ won't run well on computers that came out in the Windows 7 era or before. When I was still using my 2007 PC, I tried 8 on it and it was SOOO slow and choppy, OMGosh! But Windows 7 would work on it just fine! Now I'm using my current computer that I just built a few months ago from 2020 and '21 parts (call me a procrastinator), after using a 2016 W10 AiO for a while, and this one runs W10 just great and will likely run W12 fine when that comes out too. (Not gonna switch to 11 unless/until it gets a lot better; it's the new "bad Windows version.")
"There's no 3.5" bay in this machine."
That doesn't stop someone; that's why companies make adaptors! But then again, if anachrony is against your rule and you think these systems weren't meant to be upgraded with one, then just that alone would get in your way.
Why would there be an option for disabling the drive from being able to tell when a disk was changed?
Hmm, it's weird that they have a way to force the wrong spin speed for the disk type.
Why would HD floppy drives be less expensive than the ones with _less_ technology in them (normally higher technology = harder or takes longer to make = costs more)?
you should check bitsavers for drive user and service manuals.
I wonder if you could send files with RS-232 to a newer machine for backup.
Apparently you can. I think it at minimum has to boot to DOS, you have to have mode.com on the machine (and obviously a working keyboard to run it). But that's it, apparently.
@@PCRetroTechI would use Norton Commander with NC Link - just null modem serial and 4 files necessary. If there is a dos 6.22 it is possible to use interlink for that - pretty handy and I use it often to do the setup of old machines. I have a boot floppy with fdisk, format and interlink client and other machine boots as interlink server and I just copy what I need on a fresh setup without messing with MFM controllers or just moving things around
i have an old TEAC 5.25 inch floppy drive. it does spin, however it stops after a while. the little light isnt turning on at all. It is in fact NOT connected to the motherboard, only to the PSU.
Im using it with a socket 462 motherboard, slightly too modern but... what am i doing wrong that the light isnt turning on? any help is welcome
The light should indicate when there's activity on the drive, which wouldn't be the case if it isn't connected to the motherboard (unless you have the data cable around the wrong way). It should be possible to see if your board supports the drive by looking in the BIOS. Of course there also needs to be a floppy connector on the board, but I assume you have that.
@@PCRetroTech yes, but not connected, as im lacking a cable for it
@@PCRetroTech im struggling to find 5.25 inch floppy cables. do they have a specific name?
@@mrkevinkillercrafterminecr895 No, that should be the right name. Sometimes you find them in boxes of assorted cables, though make sure the connector is the right one for your drive.
@@PCRetroTech Sadly, its been hours and im yet to find a good offering in germany, which is where i happen to live
"Any old computer case..."
* computer (The case is just the shell, not counting major computer components.)
Thx.
What would be the point of formatting one in double-step mode?
Haha, why did you think that setting up an HD drive as an SD one in the BIOS would _enable_ both instead of _limiting_ it to SD (or making it not work at all, because of misidentifying it)?
"Removed the cover off..."?
Doesn't "remove" already include the taking of something _off?_
"Q-tip" or "cotton bud" doesn't depend on where we're from; it's one brand vs. a generic term.
"2 minutes to copy a file onto a disk..."
...From what: the cloud? A USB stick? And onto what kind of disk: the hard disk, a floppy disk, or...?
And it wouldn't just be 2 minutes; it depends on the file size.
"There's a reason for that."
Yeah, but there's a reason for _everything,_ even when it's just accidental or otherwise insignificant. So... of actual importance would be that there's a _specific_ reason for it. Right?
Zip drives ARE floppy drives; just the most modern kind. Already anachronistic for a system like this.
Did you did something to the bios? Most old bios doesn't go pass year 1999?
I haven't encountered any problems with Y2K on the machines I have. I didn't do anything to the BIOS myself, but perhaps the previous owner did during the Y2K "crisis".
or a serial cable
"Plastic notch..."
* plastic _key_ (A notch is a space; not a thing that goes into something else.)
Hello I typed in a dragon 32 program can you help out are you a programmer thanks
I don't know how to program the Dragon 32. I suggest you ask on Stack Exchange.
Rubber band!, never heard anyone call them that.
Probably an Australian word. I'm from Australia.
@@PCRetroTech Same here, the closest I've heard 'rubber band' being used was for describing car automatic transmissions
Well just to show I'm not making it up:
www.reddit.com/r/australia/comments/4rsa8k/what_do_you_call_a_thin_band_made_from_rubber_in/
"And that's if Windows Update is not running, which it always is!"
Wait. How would there be an "if Windows Update is not running" if it "always IS"?
"This is why it can't read the drive"?
How would a drive read itself?
3.5 to 5.25 adapter tray
If you need the answer to "Why most 5.25" Floppy Drives seem broken, but aren't!" skip all the way to 33:00 minutes where the answer is finally tucked into this movie about drives. The first 33 minutes is a video called "All you want to know and a whole lot more about disk drives"
I wouldn't recommend skipping it. All that detail explains WHY those changes work, and will help to decipher what to do when your drive's jumpers don't match 100% with the ones in his collection.
If you're the TLDR type, might I suggest using a computer with double or triple digit MHz clocks instead, and bypassing the whole 5.25" drive scene entirely.
@@nickwallette6201 Absolutely. If you’re just impatient and want instant gratification, vintage computing isn’t for you. These machines require patience, persistence and knowledge to repair. It’s in your best interest to know as much about floppy drives as possible if you are going to repair them.
Hey I'm struggling to get a Mitsubishi MF504B-386U setup, have tried to follow the video but I just cant seem to get the thing working correctly, any help would be greatly appreciated!
... but does it run Doom?
Wow thanks very much ! one of my drives is exact the same CHINON and my jumpers are exactly configuration like yours , but in a modern systems after year 2005 the floppy controllers are shit and only supports one floppy A: , but the other problem is that usually in my experience 5.25 floppy's did not refresh the contents of a previous floppy disk like you explained her e 24:00 so I have to remove this D-R jumper to fix the issue :)
I tested did not work :( without this jumper floppy trow "Not ready" :( I changed the floppy with other model and it works ,no problem for refreshing the disk contents , strange
Wait "refresh previous disk?!?!" You realize there is absolutely no memory in floppy disk controllers, at all, right? Nothing ever gets "refreshed" in floppy drives, ever. It reads, writes and is subject to management by Disk Operating Systems. Are you talking about a virtual drive?
The floppy drive controllers are extremely simple, mostly everything is done in physical circuitry. Complaining about them not being as good is like saying "wheels aren't as round as they used to be. Remember when wheels were circles? Why are new ones less circular?" Is there are serious translation problem here?
@@squirlmy yes I am talking about real hardware in real mode DOS. When I read one 5.25" diskette, then replace it with another ine, the system shows the contend of the first disk and it stays like that until I reboot the PC
@@intel386DX Yes, you are right. This can happen. Not sure how, but it does.
Why not just install a network card in the machine?