1984 Quantum Q540 hard drive sounds - This drive had NO platters, but I got it working!
ฝัง
- เผยแพร่เมื่อ 22 ม.ค. 2024
- Here's yet another crazy story. This drive was on French website leboncoin for a few months. The listing said it had no platters, that they were removed in the 90s, and pictures showed that it indeed had no platters but the heads looked like they might be intact. But I saw online that the Q500 series, even though it uses an optical encoder for positionning, also had wedge servo, thus I assumed it would probably never do anything so I wasn't buying it. Until last week when I talked to @TheDiskMaster on discord and he convinced me I should try. The drive was 8€ (16€ after shipping and site fee)
When I got the drive, sure enough it seemed like it couldn't do anything. I put platters just in case and still nothing. Until I found out jumper E3 disables using the wedge servo! I later found a Quantum documentation stating about E3 that "When jumpered, disables servos and causes the drive to use only the glass scale for positioning; for diagnostic use only." It's incredible that Quantum left us this option, and gave a chance to this platter-less drive to live again!
After messing with it some more, including swapping another set of platters in, I got it working good enough to boot DOS and Windows 3.1 using head 0-3. Sadly, head 4 does not work properly at all, and I have to wonder if maybe head 4 failed back in the day and that's why they took the drive apart and removed its platters in the 90s.
This drive sounds great, its seeking sounds are very unusual, but it is a very early voice coil drive from 1984 so that's not too surprising. The way it's constructed is quite interesting as well, with the magnets and the coil being hidden under a metal plate that has other things monted on it. Sadly the head locking part is broken in multiple peices so it isn't in there, but it used the same principle that Quantum still used on their drives even in the early 2000s : Using air from the spinning disks to move a plastic part out of the way. Because the lock isn't there, I need to be careful as the heads can very very easily get out of the landing zone.
The Q500 series of drives might very well be the only voice coil drive where a complete platter swap like this is possible thanks to using an optical encoder (which was a Quantum thing) and having that diagnostic jumper to disable using the wedge servo. Other voice coil drives would've used a servo surface and definitly wouldn't be able to seek at all if it wasn't there.
I'm glad that this worked out like it did, as even if it is far from perfect, it is a very nice addition to the collection that I did not have to pay an outrageous amount for. - วิทยาศาสตร์และเทคโนโลยี
Update about the Q540 : After making this video some people on discord suggested swapping the heads so I could use 7 heads instead of only 4. I modified the PCB (cut a trace and a leg of an IC, and added two wires) to swap the signals selecting head 4 and 7 around, and it totally worked! I can now use the drive with 7 heads only loosing 1/8th of its capacity due to the bad head rather than half!
Nice! Now you can fit Windows 95 onto the drive!
Nice! The drive is almost working perfect!
@@JankPods0201That will not work, Windows 95 needs at least 40mb to theoretically work. The drive's capacity after all those manipulations is 25-30MB.
I'm glad I talked you into getting this even without platters, I knew it was a good decision
Yeah I’m glad you did. The jumper to disable the servo was very unexpected but it is amazing that it exists! I know you are interested by one, and now you know even if the servo is missing (head 3 broken or having a massive head crash etc…) it can still seek at least thanks to the jumper.
Started with no platters. Finished up with platters and booting an OS. That's something of a miracle!
good work fixing it up^^
love seeing old tech made to do what it was intended for
Thanks, I'm glad I could make it work, it is a really nice early drive!
You brought it back from the dead!
Yes and I'm glad I could do that as they aren't the most common drive you can find, and it sounds great.
It sounds almost musical! Initially, just hearing the opening sounds, I genuinely thought for a moment that you had written a piece of music for it and had customised it in order to perform it on the drive. Though I'm not quite sure if I'll think the same tomorrow - when I've been composing for long hours through the night, I experience very extreme pareidolia and hear musical patterns in almost anything, to the extent that heating up a snack before I sleep can mean I get to enjoy a musical performance by my oven and the various objects and appliances that inhabit the kitchen.
This is awesome. Hard drive necromancy with a Frankenstein twist.
Actually kinda true, I didn’t even expect it would be possible but it was. These aren’t the most common drives so I’m glad I could make it work even if it wasn’t in a conventional way.
An MFM drive that's driven by a voice coil? Now that's unusual! I've seen MANY MFM Drives driven by a stepper, but Never a Voice coil! Glad you got it working with what you have on hand!
While stepper drives were more common during that time, voice coil drives did exist, but were more expensive. They were not all rotary voice coil either, Seagate drives like the ST-4096 for exemple used a linear voice coil, which is very interesting and I hope some day I'll find a linear voice coil drive at a low enough price for me to get.
I just saw the inside of an ST-4096 And it is EXACTLY What i thought it would be when you said "linear voice coil" the heads just go Back and forth like a MiniScribe 3425! I hope you do get your hands on a Linear voice coil drive!
I was reading the q540 manual the other day and saw the note about the servo jumper. My Xerox Daybreak has one of these, and I wondered how well it would work with that jumper enabled. Good to know it's possible to recover one of these disks!
It seems like the drive works decently when hot, but when cold it struggles a lot. Now I don’t know if that’s because of the replacement platters or what, but I know the servo is supposed to help with thermal expansion. If the servo is working fine, it’s probably better to let the drive use it, but in a case like mine, it’s nice that it exists and the drive can function like that.
Incredible you got this drive working, not only that, but able to store data and boot! That is really an amazing accomplishment, I am surprised to the extent you went to try and make it usable! This also may even help save other drives like this in the future realizing it can be made to work with replacement platters, neat!
I had no idea that Quantum had a jumper for disabling the wedge servo, and it is stored on surface three according to documentation? Also it is really interesting how the drive struggled to LLF with the ESDI platters put in at first, almost like it was maybe trying to read the information on the platters before wiping them? I am curious as to how swiping the magnet over the platters after putting them in made it work fine, because I thought that the drive wouldn't try to be reading from the platters when doing the LLF....
I don’t really know what happens during an LLF, but remember the corroded WD and its seemingly dead head 3? When trying to LLF it was also getting stuck for very long on each cylinders, so I guess the controller can find out when something is very wrong during the LLF. I’m not sure why the magnet helped, my theory is that the heads were picking up some magnetic fields from the NEC tracks that were narrower and possibly ended up in between where the Quantum was laying its tracks and that confused it when trying to read because the heads picked up two different signals?
I wonder if that was the thing with the Maxtor platters too or if it truly was the different material that made those not work…
Yes, it was awesome when I found that jumpers and saw it had a chance at being put back to life. I knew platter swaps can be done on some drives, the most primitive ones, but I really didn’t expect it would work on this considering I knew it had wedge servo. For some reason Quantum added a jumper to disable using the servo, which is the only reason it could work at all.
This drive sounds great and I’m glad I could make it work. Obviously if head 4 wasn’t messed up it would’ve been even better but considering how it started… Maybe head 4 having an issue could be why they took the platters off in the 90s.
@@arnlol Hmmm I am thinking that you are right on that, I guess MFM drives aren't as dumb as I thought. The heads picking up signals from the NEC tracks causing that would make sense, but odd how the drive would care at all since it is strictly set to use the optical sensor for positioning, I would assume it wouldn't care about the signals from the heads when writing to the disks during a LLF unless it was going to verify what was written, which SpeedStor doesn't do until after the LLF. Unless the drive electronics are indeed doing more than that.
Maybe I'm a bit confused here, but didn't the ESDI NEC only have one or two platters and this Q540 needed four? I see you used a combination of different platters, but I'm confused which platters came from what, in total you mentioned I believe about three different drives you took platters from, but what remained in the drive in the end?
Quantum probably added the option to disable the wedge servo because they were feeling generous and decided to be nice that day in case the platter with the wedge servo got damaged or in this case missing ; ))
Yeah shame about head 4 on the drive, unless it got misaligned when removing the platters originally? I think it is more likely that you are correct, head 4 decided to fail back in the day, the drive had issues, and in turn was why they did take the platters out of the drive. At least the drive still is around, who knows about the platters that came from it!
I don’t know if the drive truly can detect that, I think it’s more likely the signals got picked up by the heads and got on the interface, and it’s the MFM controllers that’s like "what’s going on it isn’t what I’m expecting".
I only grabbed platters from two drives, the Maxtor which did not work at all, then the NEC. The NEC had 4 platters, which is why I even considered it. The center cap that holds the platters in place is from a seagate, it was in one of those bad ST-225N I got a while ago, with a random screw I found that held in the center hub but was too long to use the original one from Quantum. I just had that laying around for no reason, I just kept that when I threw those bad ST-225N away. So yeah it involved 3 drives but only 2 as far as the platters go.
As for the servo, I guess that could’ve been used to attempt some data recovery if surface 3 got damaged? Also from the docs, another jumper disables servo protection, and I wonder if with a special controller or program you could write the servo back… Maybe they used the glass positioning and the actual heads to write the original servo? I don’t think we will ever know, but I’m glad that this jumper existed to allow the drive a chance at living again.
I didn’t notice anything special on head 4, and I assume it being miss aligned probably wouldn’t matter much since LLF lays new tracks anyways. I guess it’s possible that the issue with head 4 is an electronic problem. Well it’s the first drive with more than 4 heads I’ve dealt with so who knows, maybe it’s my MFM controller that has an issue (I don’t think so, but it is possible I guess)
17.5mb isn’t too bad, similar capacity to most full heights from that era.
Wow!!
Wow, that's great! You did something similar to what I'm planning to do to a dead IBM WDI325, whose platters were contaminated (that's a quite common fate for those drives). I cleaned the drive chamber and the heads, I already tried putting in it some platters scavenged from newer drives but the old IBM wasn't happy with them, additionally everytime I turned the drive off the heads stuck to them, probably because their surface is way too smooth for those old (comparatively) big heads. I guess only platters coated with iron oxide will work on that drive. I'm waiting to put my hands on a couple of suitable platters and when I'll do, well... fingers crossed.
I really do not know these drives, but I have heard that some IBM drives like that are often having major issues. If the drive does not have servo and can be low level formatted, then it might work. I think if you do that and it does work, there will probably be some people dealing with IBM machines that might be interested to know. Good luck!
@@arnlol It's a 20 MB stepper motor drive, so no servo tracks here. It was made for the ISA-based IBM PS/2 30 line of computers and has a proprietary interface (well, the interface to the mainboard I think it's actually just the 8-bit ISA bus brought on a ribbon cable toghether with the power supply lines). On the PS/2 diagnostic disk there's a utility to low-level format those drives and when I tried that on my drive with the replacement platters inside, it didn't work. The drive in fact didn't even did the initial seek test. We'll see with the appropriate platters how the drive will behave.
Great work, I'm amazed it's even close to balanced after getting the platters replaced. Did it require some fiddling to sort out any vibrations, or are the tolerances loose enough to where it doesn't really matter?
No I didn’t have to balance it, and I’ve messed with other drives before and it wasn’t an issue either. I think the tolerance between the spindle motor and the platters is good enough that you can’t really make it unbalanced enough for it to matter, at least at this kind of speed. I have some 7200 rpm drives that haven’t been opened and yet they vibrate a lot, even though they work properly, so I assume at higher speeds it probably matters much more.
Is there like any chance, by any i mean even 0.00001% chance that you would bring back the garry's mod apedobear server?
Well the thing is, why would I do that? The server closed then reopened once already in like 2021-2022 and it did not attract players, like a few people played for like a few days then back to 0. I kept the server open for a while in 2022 when it had no players. When the server closed again at the end of 2022, seekers also deleted the discord server as well. If the server re-opened now it would be even worse, it wouldn’t have any players not even ones from the old days. I think the game mode might have completely died as well, I don’t think Waurum even exists anymore (at least I couldn’t find it on gametracker). If I really wanted, I could, I still have a backup from like the middle of 2022 I’m pretty sure, so I could, I just don’t really see the point in doing it when it isn’t going to have any players.
I undestand you completely, but for the love of god please keep that back up safe. Last question, whats the average cost for the server to stay online?@@arnlol
Ohh i just wanna play the og pedobear it was so much better have u considered by any chance selling the server or something:(?
Without the platters, it's a perfect hard drive for a mime.
A ha, That's actually funny!
Thats also very cool that you got it to run speedsys by formatting it! Was that a low level format? I imagine you couldn't do this with a modern drive as I think thats done when its being assembled by machines?
Yes, MFM drives can be low level formatted (and often need to be, as the format laid by a controller might not be compatible with another one). The low level formatting ability is what allows a platter swap to be possible on MFM drives (but not even all of them, as some have servo etc, and having a jumper to disable the servo like this drive has isn’t an usual thing), and yeah it isn’t possible to do that on more modern drives.
Very nice vintage drive! It's a shame it came with no platters, but at least you got it to seek!. That is quite the drive though. I assume its a 51/4" full height ?
Yes, it is indeed a 5.25 full height drive. It is pretty heavy, when I picked the package up, knowing it didn’t even have platters inside I was surprised by the weight.
still more functional than my Seagate barracuda
Considering I got it to work decently well, I guess if your barracuda is dead then that's technically true
Does head 4 have any physical damage? Maybe a bond wire has broken off of the head? Would be super tricky to solder it back on, but it’s worth a shot imo.
Incredible that the drive works anyway, and the drive so happens to have a diagnostic jumper to disable using the disk surface for positioning.
I don’t think so, and it isn’t completely dead either as it took the LLF and the sectors "existed" (it wasn’t all IDNF), but basically every track on it was reporting as bad on a surface analysis. When I saw that I even tried swapping the platters around in case it was one of the surfaces that wouldn’t work for some reason but nope head 4 was still making every track as bad sectors. I dont know why, I guess either the head is damaged in some way even if I didn’t notice anything or maybe there’s an electronic issue that prevents it from working.
Hello arnold0, nice video. Where do you install Windows to test the noise from the hard drives? Do you install it on a computer or on an emulator, etc.????
I use old machines, either a Pentium, a Pentium 3 or an Athlon XP depending on which drive I want to test/record and what OS I want to run on it.
@@arnlol Very cool Arnold. I started to like and collect hard drives after watching your videos and now I have 7 hard drives, two aren't working but now I'm looking for a Seagate 157-a which is very difficult to find... Here in Brazil it costs more or less 500 reais which is equivalent to 100-200 dollars. I'm 13 years old and I really like HDs.
I had no idea this was even possible.
It is only on MFM drives that you have any chance at trying something like thins thanks to the ability of low level formatting the drives. That's assumig the drive doesn't have servo on the disks, which a lot of stepper drives don't. It's really lucky that it worked on that one though, thanks to Quantum for having that jumper that disables the on-disk servo. It's probably the only voice coil drive where this could be done
Do you have a clean room lab? Or work ar data recovery? That is some serious work you did there
No I do not, and it may or may not be part of the reason why it has gotten a lot of bad tracks.
Early drives like this do not seem to care too much, trying this kind of thing with a modern drive it definitely wouldn’t work and the heads would probably almost instantly get destroyed.
@@arnlol probably because sectors are bigger so in modern drives a dust particle destroys hundreds of sectors instead of one or two in the older ones
@@kernel_data_inpage_error Yes, and modern read/write heads are also way smaller, fly closer to the platters and are extremely fragile.
this is impressive! , is it some sort of early coil drive?
Yes it is a very early voice coil drive from 1984, the magnets and coils are under a metal plate so they cannot easily be seen. The optical sensor is used for positioning, normally helped by wedge servo on the platters, but luckily a jumper can disable using the wedge servo which allowed non original platters to be used.
what is a wedge servo?@@arnlol
Take a look at this, This link will give you the answer. en.m.wikipedia.org/wiki/Embedded_servo
It's crazy how you can take platters from a totally different drive and have them work! Are the platters the same capacity as the one originally in this drive?
No they were not the same capacity, the NEC was a 157MB drive. But since you can low level format MFM drives, which lays new tracks and sectors, it doesn't really matter. But in this case I was very lucky that Quantum included that jumper to disable wedge servo, as servo data will definitly not be there on non original platters. I have heard other people have been able to pull off a platter swap on MFM drives, though those were on stepper motor based ones.
This hard drive still working?
Yes it works, but MFM drives like this are basically the only ones that could work after doing a crazy thing like that, and not all of them even. This one might be the only example of a drive with servo where this can be done thanks to a jumper that disables the servo.
No IDE drives or anything newer could work after doing something like this as they cannot be low level formatted and almost all have embedded servo.
Frankenstein hard drive.
I guess some MFM drives are reliable : ) Good old Quantum...
Well I wouldn’t call it reliable, it’s got some issues, but considering what it went through… It’s nice that I got it to work even if it’s far from perfect
The rotating center is unbalanced, that's a problema nd will become worse, you need to fix it and provide a cover and a clean environment
Yeah I think the center is a bit unbalanced but there's not much I can do, it's not like I'm going to run it all the time. I did put its cover back on it (it didn't have platters, but it did have the top cover)
great, none of the ssd will work after 40 years.
I'm curious how true this will be, I bet when those drives were made people also said none of them would still work 40 years later, and it seems like mechanical parts would be more likely to fail over time compared to ICs. Now there's the thing about flash slowing losing charge over time etc so yeah, will that cause them to fail or not remains to be seen I guess.
@@arnlol mechanical might fail but can be serviced or data recovered, but you can't service semiconductor, data is as good as gone if chip fails.
Nice job, you revived it!
Yep! It’s far from perfect but considering how it was to begin with… It’ll be a nice part of the collection :)
@@arnlol i have a question, is that a voice coil drive?
@@engineer359 Yes it is, and it very well might be the only one where a complete platter swap like this is even possible thanks to the light positioning system that Quantum used + the jumper to disable the servo data.
@@arnlol that's nice thing, I'm glad that you make it work pretty well without much problem