TURNS OUT I WAS WRONG! See it work here: th-cam.com/video/II5M968YBCM/w-d-xo.html I just did a test where I replaced the internal drive with a different one from a PC and it was able to format a disk! There is something wrong with my drives and I suspect it is belts.
Maybe a long shot, but, what about disassembling the prepare tool? Z80 assembler is fairly simple, using the error message as an entry point of search the code that verifies the disk format could be traced, and as you already know the format of the disk what must fail is the validation of the initial content of the disk, that must be some kind of data and the program should have it as a table somewhere inside it, maybe also has a checksum or similar, but the program must contain what needs to be there to validate it...
That's just what I was going to say. Even just scanning through the readable strings on the boot disk might provide a clue regarding what it's looking for.
Ok, so I know absolutely nothing about Z80 at all and have no damn idea what I'm doing ....but I tried z80dsam and stuff starting at 0x6000 looks kinda like code. I can see millions of small functions seemingly doing... something? Most of the functions are too short - it does not look like a human written that kind of assembly code. So this is probably compiled from a high-level language. But JP addresses seem to be increasing more or less consistently (code tends to jump to nearby code) so I'm going to say it's not complete gibberish The issue I see is that this seems to be a FAT-like filesystem, and code segments are not always consecutive, or belong to different programs that occupy same memory at different time. We cannot really make much sense of the code if we don't know exactly where it is stored in memory - otherwise call addresses make no sense, and so a "smart" disassembler/decompiler gives up immediately. Manually pasting the code fragments to where they seem to belong in memory is not impossible, but it's a lot of work. The strings addresses should be referenced in the code somewhere. But because I cannot associate code with addresses properly I cannot find those. Code right before the strings seems to be around 0x9c0, but after it seems to be around 0x870 again, going up to ~0x9c0 before abruptly switching to ~0xe0d0. At what address are the strings loaded then? Who knows! By the way, strings are padded with long strings of spaces with no prefix or delimiter - this looks like something FORTRAN would do, although other languages/compilers may have used similar convention. So the first step seems to be - figure out the filesystem, figure out how the code is loaded into memory, then disassemble.
@jwhite5008 I've been checking the ROM and it seems the main processor is not the Z80, this is from shelby's uploaded ROM: "The machine has several CPUs in it, but this is most likely for the 8088" and decompiling the rom with z80dasm yields results that don't make sense for a Z80, the first instruction on it if it were a Z80 is "ret PE" which makes no sense, and the usual RST vectors, specially 038h don't make sense... Assuming that's true the boot address of the 8088 is 0xFFF0 (I do program assembler for the Z80 but not for the 8088 so this is info that I have found) so the ROM must be mapped to end at 0xFFFF, and that seems to match, at 0x7FF0 in the ROM there is 0xEA that seems to be a far JMP instruction what do makes sense... So from here I have no idea where to continue as I don't know 8088's assembler :(
@@drgusman Maybe the ROM is mapped into some weird address and the beginning of it isn't even code? We really need to figure out memory map of this thing...
@@jwhite5008 Well, if the ROM is really for the 8088 what I found makes sense, instead like the Z80 that its reset vector is 00h the 8088 boots from FFF0h, so, the ROM should be mapped at the end of the 64kb space, from 8000h to FFFFh. If that's correct what is there is a long jump (and casually there is something that looks like a table filled with "00 00 7F FF" that suddenly is "broken" at that address) that if I understood correctly how the 8088 maps the memory jumps to the physical address FF000h, the last 4Kb of the 1Mb memory space. That and the lack of any kind of string in the ROM makes me wonder if there may be more ROMs on the machine...
...did Xerox sell boxes of preformatted disks specifically for this thing back in the day? That's the only explanation I can think of, but then I would assume (if not hope) they would have said so in the manual...
@@TechTangents And another thing. The manual says if the instruction disk is bad, you need to, quote: "Make another copy from the original instruction disk." Can it do that? I.e. if you select disk copy and insert a random disk with incompatible format, can it actually copy over the instruction disk? And if that fails, try overwriting a duplicate of instruction disk with that same disk. There is a screen message "Copy over instruction disk?" which to me suggests that an instruction disk can be overwritten. If it can do that - this means it can in fact format disks. Something else is wrong then. If it tries but fails - that means there may still be a problem with the drive or drive-control circuitry. If it refuses - we are probably out of luck?
Another thought that came to mind since you know the drives are both able to write is if pins 22 and 24 (WRITE_DATA and WRITE_GATE) somewhere between the drives and he Memorywriter's disk controller are damaged. It could be the typewriter *CAN* format its own media but when it goes to lay down the new tracks it's not actually able to write the new data and the WRITE_PROTECT line isn't flagged, so as soon as it writes a track/cylinder/sector/whatever it reads back for integrity/checksum and fails because it didn't actually do anything. Xerox did some dumb things with their word processors but I can't see them pulling a Rainbow 100 and telling everyone (initially) they could only buy diskettes pre-formatted. Also I'm incredibly lucky your Wheelwriter has the tiny little CRT and the proper arm. Even I don't have that!
I tried that just yesterday actually in this stream: th-cam.com/video/PLfLXoX3GFE/w-d-xo.html the write signals seem fine. Write protect is also working correctly and it has a different error for that. There's potential that my replacement belts are causing the issue but I still need to try a different drive next I think. It's crazier that it also has a base plate the monitor arm connects to! I didn't even know that existed until I went to pick it up.
The floppy blank disk format is similar to the BBC Micro ADFS-S format 1 side 40 tracks 16 sectors 256 bytes. You might be able to use omniflop to format you a blank disk. Also reminds be of one of the many cpm disk formats that where out there
Oh boy....You don't want to know how many of these "daisy-wheel " Xerox (and many other brands) we had at home baxk in the 70/80's.....Worked on them a lot as a teenager...My dad imported, repaired and sold those machines by the loads...Same goes for the IBM "Golfball" typewriters and Olivetti's....
I don't give a flying pigs arse about any of the stuff he's fixing, but hell's horses do I love watching this bloke fix stuff. His drive, passion and knowledge are all so infectious! Makes me want to learn how to fix my dead old Voodoo Banshee 💔
This is a sanity-check and therefore probably won't work, but as you didn't say it, it's worth explicitly asking: Have you tried preparing a blank disk in the left-hand drive? edit: it was the drive itself, seems that the overtight belt was causing the drive-speed to be marginal.
I suspect that it needs structures (like a directory for example) pre-created, as that way it wouldn’t need a special case for blank disks. Given a copy of the boot you might be able to determine that structure.
Nice summary! I've seen you working for hours on this device, compliments on the outcome. Now, this video just needs to find its way to that one (retired?) Xerox employee who still has a box of floppy disks.
I am an IBM Selectric mechanic for Philly Typewriter in Philadelphia, PA. I have been branching out towards Wherlwriters, Brothers and Xerox machines. This video helped me out a lot thanks. I do have an original Wheelwriter monitor but it started to smoke. However I’m looking forward to more of these videos.
Those bearings usually do not go bad easily. They use very hard steel on these. I had similar sized bearings almost stuck on from a very weathered ac unit, applied plenty of wd40, after I got it lose, spinned it with a drill. After rust etc came out, I washed them with dish soap, dried and applied grease. They were not even lose, spinning like brand new, no noise or else. Still working perfectly. They need grease not machine oil or wd40. Edit: They worked without grease, squeaking for months before taken care of, then they got stuck.
A trick to fix old bearings is pack in some graphite grease then run them in a drill. The graphite grease is a bit like toothpaste and polishes the bearings and race so it runs even smoother.
I once owned an IBM Selectric printer, which was the eponymous typewritter without keyboard, but came with parallel port. Typed nice letters, but was pretty loud. My brother used it for a while as well. Could be used as a printer from DOS, Windows, OS/2, whatever, but obviously letter only device. Also came with IBM word processor package that booted straight into the word processor without need for a system disk on the PC it was attached to.
The start of the FATs (0x11000) on the system disks start with "SABRDSYSTEM", which is probably sort-of a volume label (and I bet is checked so you can't prepare the system disks; most likely only half for any malicious reason, and probably just so you can't accidentally wipe your system disks). Searching for parts of that seems to pickup a couple potential other candidate labels that might allow you to force your way into a prepared disk.
Hmm, but there is NO other place in the option disk where this string exists. Yes, it could check with the label of the disk itself, but why would the code then need to seek to a different sector and read it from there whereas they could have copied the string? I think this is not it.
SABRDSYSTEM and SABREDATA exists on system and option disks also there's a file called FDISK on the option disk, wonder if there's a way to directly invoke it
@@bleeedthebrakes hmm, invoking it might be possible by swapping filenames but without proper setup it will most likely just crash. I do see what looks like file names in the dump, but it's just a guess, or do you have an idea on how to decode the filesystem?
@@jwhite5008 Sadly no, I'm just throwing ideas around. The table seems to have a structure of 6 bytes of unknown data + 0x00 0x00 + one hex value tied to the file and persistent across disks + filename string + four 0x20
That ink ribbon isn't a traditional wet ink ribbon, but transfers something like toner from the clear tape of the ribbon to the paper, when you reuse sections and it has gaps that's because the character you just typed overlapped with the previously struck character the first time the ribbon was used. You can see the ghost of what's already been typed to the right of the ribbon as you go. Fax machines often worked the same way.
I had one of these at home in the early 90s as I remember doing my home work on it, I absolutely loved this! It was no longer needed at my dad's work so he was able to bring it home. I'm racking my brain to try and remember if I ever prepared a new floppy or not, but although I remember the main programme and option disks, I can't remember what I used as the data floppy disk! I wonder if you could go through the code and see what causes that error to be flagged - it might be some descriptor on the disk that matches some other old computer or something?
I used to service those things in the early late 80's and 90's, the touch panel often cracked and was not easy to find, so surprised yours is still good. Otherwise a huge machine with only basic word processing abilities (at least compared to a Pc of the time). I generally didn't like them over the equivalent IBM or Olivetti which were closer to being a full pc. These things were just on their own and the abstract panel I never got to know what it did either.
I'm so happy to see an update for this machine! I like to think this video inspired me to get my own word-processor machines and they've be super fun to work with. I still kind of want to pull the processor boards and stick in a raspberri pi, but there is just never enough time in the day.
Gotta say, $30 for ANY TTL monitor is a pretty sweet deal, let alone such a rare and proprietary one! As for dealing with Shugart drives, yeah, it's a pretty common experience with Tandy owners as well. Sure, normal PC drives work, but you either have to worry about fiddling with the jumpers, or building a custom ribbon to install a replacement PC drive. It's for that reason I always try to save 3.5" floppies that have jumpers/switches at the back, as it can make a direct replacement a lot simpler.
That reminds me of something they had at my elementary school. It was a typewriter with a row of 16 segment displays the same length as it would type on the page, it would apply to the paper the full line on the display when the return key was hit.
Need more typewriter/word processor videos! Can't wait to see the one on the Magavox Videowriter you teased in the last mail call video or really any of these machines as they're such an interesting concept and window of history
Since you were able to take archives of the two floppies, were you able to restore this images to new disks and have them work in the machine? If so, you might be able to use the formatting tool to whack one of those. And if it doesn't let you because it recognizes that they are "system disks", you might be able to just hex edit the directories to remove files so it works? Experimenters dream project, right? 🙂Good work so far!
My thoughts exactly. It is either looking for some flag in the system-reserved area or files present. I think the first thing to try is to overwrite file *names* in the hex editor with "AAAAAA" or something. If it's the flag that is a lot harder though. Did you try formatting a duplicate copy of EVERY compatible disk you have? Do the all give exactly the same error?
There is a way to figure this out. You decompile the os disc. You look for the place that returns the error message, then work backwards to see what checks it's performing. Then you can make a disc to fit those requirements. Might be as simple as a magic string. You might have scanned passed it, when you looked at the ASCII in the disc code. It won't be exactly easy, but not undoable either. The CPU in that thing is no core i7. Likely some 8 bit thing like a z80. Very simple instruction sets.
I have a Xerox 860 in my shed...you want it? I no longer have the printer, though...I have a bad back, and those things weigh a ton. I even have some blank 8" floppies to go with it. I really ought to find it a home. I only took it because I'd used one in the navy, in the late 80s.
ive never thought about how lucky i was when i bought an Olivetti CWP-1 back in the mid 2010s for super cheap with monitor with 0 issues until now, it didnt need a weirdly formatted disk to work (but it does have a custom format, it can format any disk into its format and I'm still unable to dump anything i put into one), i didn't get a manual nor the "initialization disk" or whatever it is tho, but i can at least use as much as i can figure out about it lol, i love 80s word processors so much too, cant wait for the next vid!
11:37 Which Epson sheet scanner is that? What happened to your CZUR? Just wondering about the effort of decombing/despiraling the manual. As always the whole journey is fascinating. This typewriter is so much more intricate than I would imagine would be necessary. Thank you for your efforts!
That is an Epson GT-S50, the CZUR is still here, it's just about using the right tool for the job. The feed scanner is faster and easier for something I can debind. And the manual uses a metal spine that just opens up so it wasn't difficult.
I wonder if a dive into the ROM or OS will show what flag or marker it's looking for on the disk. The fact it can detect a boot drive is linked to it looking for a prepped blank disk with a marker on the the disk
@@spookisghostly4619Be warned though, the Selectric typewriters are very complicated mechanically and apparently are hard to work on! My parents had one when I was a kid in the 80's and 90's and I remember they were always having problems with it! I love vintage machines and especially typewriters, but I have no interest in getting a Selectric!
@@ct6502-c7w oh I'm aware it's one of the most complicated things I've learned about honestly and I have a fascination with typewriters, film cameras, tape recorders, ect it just sort of feels like the next step id make eventually
Does it detect if the disc is read-only at all? If not, is it possible that what is used for detecting that has seized up or malfunctioned, keeping the drive from being able to write to the disk coupled with unhelpful errors?
@@guepar58 Also, isn't it completely up to the software to honor the Write Protection or not? If so, could it be that some of it just doesn't care and won't throw an error even if Write Protection is enabled or faulty? Edit: "After some research It seems like this was mostly controlled or gated by the drive and not software running on the host.
@@TechTangents Is there some sort of magic byte at some point on the disc that the OS is looking for so that a user who might not know what they're doing doesn't completely overwrite their system disks? Or, probably more likely just in order to sell special preformatted disks at a markup to regular floppy discs? If so, maybe someone can reverse engineer the OS in order to find out what it's looking for.
@@TechTangents It looks like Xerox sold branded diskettes for different things. One thing I could find was Gold Series followed by For Office Imaging Technology. Though, those are unformatted 2 sided 40 track regular discs with supposedly high quality.
nice work! i was thinking about his machine, a couple of months ago i saw one (in very very bad shape) in a flea market and i was thinking if you were going do an update on it
I think most of them are in bad shape unfortunately. This whole device category is a hard sell for collectors now and they are languishing. Hopefully we can find enough stuff to be able to keep them going!
I have a Xerox 6040 Memory Writer, it has 3.5 floppy disks. But I do not have the Daisy wheel so it does not pass the self-test and will not operate. much sad
This video has made me wonder how hard it would be to create a custom typewriter using a raspberry pi, some steppers and a resin-printed fontwheel 🤔Could be an interesting electro-mechanical experiment 🤔
@@RetroJack Will do, I'm starting a YT channel for documenting this sort of thing, beginning with a rebuild of an old 3D printer, followed by creating a home assistant light switch
@@clonkex : Assuming you're going for daisy-wheel style, remember that you also want a solenoid for the actual "print" step- this should probably be controlled by a simple relay, which can usually be found at either car part stores or any farm store that features some vehicle lights (they get used to control extra lights from a single switch); control the relay itself with a transistor, and a flyback diode.
@@clonkex : For light switches, assuming- 1) this is another Raspberry Pi project, and 2) you're looking at standard wall voltages: You can get "contactors" (high power relays) at "contractor supply stores", and control them with relays. Frankly, a contactor is massive overkill for household lighting, but the more normal relays that stores around me sell say nothing about their voltage capabilities, so safety first.
Yeah, I think you need to rub the camera again, it still seemed like it wasn't quite focused, double vision on the floppy drives and the screen there. 😉 I'll just assume it was fixed for the second channel.
@@gorak9000 It isn't. It's a Bessey, and while not a precision machining vise by any means it's OK for what it is. I do suggest applying some non-drying silicone grease to the leadscrew and slide pins when you first take it out of the box, however.
I have a Smith Corona word processor that won't format disks either. I tried single density, double density, 720K and 1.44MB, nothing works. The drive runs...it just won't format anything.
It looks to me that Xerox produced their own custom forma floppies for the machine, probably to keep customers returning to Xerox. Have you tried constructing a virtual image for a blank floppy that mimics the Xerox format and using a raw write to construct it on a blank 5.25" disk?
Have you tried covering the write protect notch? In some old systems the write protect the opposite way ibm machines did; tape over notch allows writing.
@@jwhite5008 Gotek / Flash Floppy supports everything. If not just ask the Flash Floppy dev and he will add support. I worked with him several times to add some unsupported features he's a great guy and very helpful and enthusiastic.
I think I know why you might not be able to format the discs, including the boot disc: it's write protected. 5.25 inch floppies have a write protect notch on one side: if the notch is visible the disc cannot be written to. To disable this, cover it with a small piece of opaque tape folded around the edge of the disc. The notch I'm talking about is visible at 18:15 when you hold the disc up in front of the machine. I can't tell just from watching if the other discs have the same notch. I grew up with these things. Yes, I'm old. :-D You young whippersnappers, etc. :-D
@@ownpj Yeah he's wrong. I have 1000's of 5 1/4" disks here. Write protect is with the notch visible. Back in the day most original software disks didn't have a notch so they couldn't accidentally be erased. Blank disks were expensive so we cut a notch on the other side and flipped the disk and used the other side heh!
interesting problem making blank media? you had ago at copying blank data area from a working disk, to your new experimental media disc? did not work? I am assuming, now the ins and outs working at least the diagnostic anyway? in your demo, the message the comes up on the bottom left of the screen? where the text coming from? I now why it putting the message on the screen but where are actual letters stored are in the machines bios, or what passes for bios? and the bit maybe clue, like PC floppy hard disks etc, the is is not system disk message stored on the media it self? if you even load really blanked disk, your don't get any message at all? so this is the hunch, could not just look for message on the working disk, and, find it bit for bit location? the most bit something bios look and reads that message from the disk? and the string data in the bios, more hunches is going the the same line of text, in machine code or hex or what ever? more hunches, will there be a good change in the boos memory, close by, a list of expected text messages, that could be read disk loading time? to check for? and just do the steps backwards, put text data on floppy disk? and see what happens?
Would it be funny to just contact xerox for some help formatting a disk i bet this thing is still used in government offices around the world still. I see type writers in businesses still too usually the ibm branded ones.
Hey man. I dont know if you will see this comment but a guy in nj is selling this exact xerox with the crt and everything on facebook marketplace. I tried to find your email so i could send you the link to the listing but couldnt find one. Lmk how to send it to you if you’re interested.
So basically Xerox got into the "ink cartridge" scam before ink cartridges were a thing. "You cannot format a blank disk, you must buy empty ones from us! Mwahahahahahah!" -Xerox
That ink tape is worse. It's basically toner on a clear tape. After typing on it the toner is removed and the tape is clear in that place. So they can't be re-inked. Once used they are scrap.
Please dump the roms this needs to be emulated, just because heh! Please also take readable pics of the board/chips so that the emulation will have the correct hardware and a chance of working correctly.
When you E5'ed the disk, maybe you did also E5'ed the bytes telling the machine which disk it is. Who knows Xerox did it on purpose to make sure users did buy THEIR blank disk and not a non-Xerox blank disk. Who knows how many people failed to format a non-Xerox disk like you did.
So maybe their custom disks were a way to make money from users. Gross tactic and makes their product useless on its own. Maybe someone on the internet would help you hack it.
TURNS OUT I WAS WRONG!
See it work here: th-cam.com/video/II5M968YBCM/w-d-xo.html
I just did a test where I replaced the internal drive with a different one from a PC and it was able to format a disk! There is something wrong with my drives and I suspect it is belts.
Second channel 👀👀 What's it called so I can subscribe
@@UpsetFather Tech Tangents II (also, Tech Tangents: Live)
Congratulations and Yayifications!
Maybe the drive has a broken readonly led or sensor?
Check that the write protect sensor(s) are working and that there are no jumpers enabled on the drives to set them permanently to R/O
Maybe a long shot, but, what about disassembling the prepare tool? Z80 assembler is fairly simple, using the error message as an entry point of search the code that verifies the disk format could be traced, and as you already know the format of the disk what must fail is the validation of the initial content of the disk, that must be some kind of data and the program should have it as a table somewhere inside it, maybe also has a checksum or similar, but the program must contain what needs to be there to validate it...
That's just what I was going to say. Even just scanning through the readable strings on the boot disk might provide a clue regarding what it's looking for.
Ok, so I know absolutely nothing about Z80 at all and have no damn idea what I'm doing
....but I tried z80dsam and stuff starting at 0x6000 looks kinda like code.
I can see millions of small functions seemingly doing... something? Most of the functions are too short - it does not look like a human written that kind of assembly code. So this is probably compiled from a high-level language.
But JP addresses seem to be increasing more or less consistently (code tends to jump to nearby code) so I'm going to say it's not complete gibberish
The issue I see is that this seems to be a FAT-like filesystem, and code segments are not always consecutive, or belong to different programs that occupy same memory at different time.
We cannot really make much sense of the code if we don't know exactly where it is stored in memory - otherwise call addresses make no sense, and so a "smart" disassembler/decompiler gives up immediately.
Manually pasting the code fragments to where they seem to belong in memory is not impossible, but it's a lot of work.
The strings addresses should be referenced in the code somewhere. But because I cannot associate code with addresses properly I cannot find those. Code right before the strings seems to be around 0x9c0, but after it seems to be around 0x870 again, going up to ~0x9c0 before abruptly switching to ~0xe0d0. At what address are the strings loaded then? Who knows!
By the way, strings are padded with long strings of spaces with no prefix or delimiter - this looks like something FORTRAN would do, although other languages/compilers may have used similar convention.
So the first step seems to be - figure out the filesystem, figure out how the code is loaded into memory, then disassemble.
@jwhite5008 I've been checking the ROM and it seems the main processor is not the Z80, this is from shelby's uploaded ROM: "The machine has several CPUs in it, but this is most likely for the 8088" and decompiling the rom with z80dasm yields results that don't make sense for a Z80, the first instruction on it if it were a Z80 is "ret PE" which makes no sense, and the usual RST vectors, specially 038h don't make sense... Assuming that's true the boot address of the 8088 is 0xFFF0 (I do program assembler for the Z80 but not for the 8088 so this is info that I have found) so the ROM must be mapped to end at 0xFFFF, and that seems to match, at 0x7FF0 in the ROM there is 0xEA that seems to be a far JMP instruction what do makes sense... So from here I have no idea where to continue as I don't know 8088's assembler :(
@@drgusman Maybe the ROM is mapped into some weird address and the beginning of it isn't even code? We really need to figure out memory map of this thing...
@@jwhite5008 Well, if the ROM is really for the 8088 what I found makes sense, instead like the Z80 that its reset vector is 00h the 8088 boots from FFF0h, so, the ROM should be mapped at the end of the 64kb space, from 8000h to FFFFh. If that's correct what is there is a long jump (and casually there is something that looks like a table filled with "00 00 7F FF" that suddenly is "broken" at that address) that if I understood correctly how the 8088 maps the memory jumps to the physical address FF000h, the last 4Kb of the 1Mb memory space. That and the lack of any kind of string in the ROM makes me wonder if there may be more ROMs on the machine...
...did Xerox sell boxes of preformatted disks specifically for this thing back in the day? That's the only explanation I can think of, but then I would assume (if not hope) they would have said so in the manual...
Yep, Xerox order number #9R80442 which I could only find a reference as being for this line of word processors.
@@TechTangents And another thing.
The manual says if the instruction disk is bad, you need to, quote:
"Make another copy from the original instruction disk."
Can it do that?
I.e. if you select disk copy and insert a random disk with incompatible format, can it actually copy over the instruction disk? And if that fails, try overwriting a duplicate of instruction disk with that same disk.
There is a screen message "Copy over instruction disk?" which to me suggests that an instruction disk can be overwritten.
If it can do that - this means it can in fact format disks. Something else is wrong then.
If it tries but fails - that means there may still be a problem with the drive or drive-control circuitry.
If it refuses - we are probably out of luck?
Another thought that came to mind since you know the drives are both able to write is if pins 22 and 24 (WRITE_DATA and WRITE_GATE) somewhere between the drives and he Memorywriter's disk controller are damaged. It could be the typewriter *CAN* format its own media but when it goes to lay down the new tracks it's not actually able to write the new data and the WRITE_PROTECT line isn't flagged, so as soon as it writes a track/cylinder/sector/whatever it reads back for integrity/checksum and fails because it didn't actually do anything.
Xerox did some dumb things with their word processors but I can't see them pulling a Rainbow 100 and telling everyone (initially) they could only buy diskettes pre-formatted.
Also I'm incredibly lucky your Wheelwriter has the tiny little CRT and the proper arm. Even I don't have that!
I tried that just yesterday actually in this stream: th-cam.com/video/PLfLXoX3GFE/w-d-xo.html the write signals seem fine. Write protect is also working correctly and it has a different error for that. There's potential that my replacement belts are causing the issue but I still need to try a different drive next I think.
It's crazier that it also has a base plate the monitor arm connects to! I didn't even know that existed until I went to pick it up.
@@TechTangents Will you please dump the ROMs of that machine and state the make and model of the CPU chip? I'd love to pass a disassembler over it.
The floppy blank disk format is similar to the BBC Micro ADFS-S format 1 side 40 tracks 16 sectors 256 bytes.
You might be able to use omniflop to format you a blank disk.
Also reminds be of one of the many cpm disk formats that where out there
Oh boy....You don't want to know how many of these "daisy-wheel " Xerox (and many other brands) we had at home baxk in the 70/80's.....Worked on them a lot as a teenager...My dad imported, repaired and sold those machines by the loads...Same goes for the IBM "Golfball" typewriters and Olivetti's....
I don't give a flying pigs arse about any of the stuff he's fixing, but hell's horses do I love watching this bloke fix stuff. His drive, passion and knowledge are all so infectious! Makes me want to learn how to fix my dead old Voodoo Banshee 💔
This is a sanity-check and therefore probably won't work, but as you didn't say it, it's worth explicitly asking: Have you tried preparing a blank disk in the left-hand drive?
edit: it was the drive itself, seems that the overtight belt was causing the drive-speed to be marginal.
I suspect that it needs structures (like a directory for example) pre-created, as that way it wouldn’t need a special case for blank disks.
Given a copy of the boot you might be able to determine that structure.
That's about what I was thinking, I bet it's identical to a blank OS disk with a couple different bytes in sector 0 or something like that.
If the roms can be dumped it should be possible to determine what it expects from the user disk.
Nice summary! I've seen you working for hours on this device, compliments on the outcome. Now, this video just needs to find its way to that one (retired?) Xerox employee who still has a box of floppy disks.
Absolutely.
This is a real saga, I seriously doubt I would put all this effort into such a machine. Kudos!
I am an IBM Selectric mechanic for Philly Typewriter in Philadelphia, PA. I have been branching out towards Wherlwriters, Brothers and Xerox machines. This video helped me out a lot thanks. I do have an original Wheelwriter monitor but it started to smoke. However I’m looking forward to more of these videos.
Those bearings usually do not go bad easily. They use very hard steel on these. I had similar sized bearings almost stuck on from a very weathered ac unit, applied plenty of wd40, after I got it lose, spinned it with a drill. After rust etc came out, I washed them with dish soap, dried and applied grease. They were not even lose, spinning like brand new, no noise or else. Still working perfectly. They need grease not machine oil or wd40.
Edit: They worked without grease, squeaking for months before taken care of, then they got stuck.
A trick to fix old bearings is pack in some graphite grease then run them in a drill. The graphite grease is a bit like toothpaste and polishes the bearings and race so it runs even smoother.
I once owned an IBM Selectric printer, which was the eponymous typewritter without keyboard, but came with parallel port. Typed nice letters, but was pretty loud. My brother used it for a while as well. Could be used as a printer from DOS, Windows, OS/2, whatever, but obviously letter only device. Also came with IBM word processor package that booted straight into the word processor without need for a system disk on the PC it was attached to.
I want one of these so badly now
Got those hybrid bearings in stock at work, about to replace a couple again soon.
They are common in electric motors.
The start of the FATs (0x11000) on the system disks start with "SABRDSYSTEM", which is probably sort-of a volume label (and I bet is checked so you can't prepare the system disks; most likely only half for any malicious reason, and probably just so you can't accidentally wipe your system disks). Searching for parts of that seems to pickup a couple potential other candidate labels that might allow you to force your way into a prepared disk.
Hmm, but there is NO other place in the option disk where this string exists.
Yes, it could check with the label of the disk itself, but why would the code then need to seek to a different sector and read it from there whereas they could have copied the string? I think this is not it.
SABRDSYSTEM and SABREDATA exists on system and option disks
also there's a file called FDISK on the option disk, wonder if there's a way to directly invoke it
@@bleeedthebrakes hmm, invoking it might be possible by swapping filenames but without proper setup it will most likely just crash.
I do see what looks like file names in the dump, but it's just a guess, or do you have an idea on how to decode the filesystem?
@@jwhite5008 Sadly no, I'm just throwing ideas around.
The table seems to have a structure of 6 bytes of unknown data + 0x00 0x00 + one hex value tied to the file and persistent across disks + filename string + four 0x20
As I don't have any more advanced background knowledge about filesystems, this is all I can see(which most likely doesn't mean anything)
That ink ribbon isn't a traditional wet ink ribbon, but transfers something like toner from the clear tape of the ribbon to the paper, when you reuse sections and it has gaps that's because the character you just typed overlapped with the previously struck character the first time the ribbon was used. You can see the ghost of what's already been typed to the right of the ribbon as you go. Fax machines often worked the same way.
Thermal printers also work that way when using color ribbons. For example the Okimate 20.
I had one of these at home in the early 90s as I remember doing my home work on it, I absolutely loved this! It was no longer needed at my dad's work so he was able to bring it home. I'm racking my brain to try and remember if I ever prepared a new floppy or not, but although I remember the main programme and option disks, I can't remember what I used as the data floppy disk!
I wonder if you could go through the code and see what causes that error to be flagged - it might be some descriptor on the disk that matches some other old computer or something?
This was recently listed on my local Craigslist in a complete and working condition. I was briefly tempted to acquire it, but someone beat me to it.
I used to service those things in the early late 80's and 90's, the touch panel often cracked and was not easy to find, so surprised yours is still good. Otherwise a huge machine with only basic word processing abilities (at least compared to a Pc of the time). I generally didn't like them over the equivalent IBM or Olivetti which were closer to being a full pc. These things were just on their own and the abstract panel I never got to know what it did either.
You're a maestro! Excellent job.
Thanks for sharing. Even if it doesn't work as a word processor, you may try to use it as a Linux terminal :)
I'm so happy to see an update for this machine! I like to think this video inspired me to get my own word-processor machines and they've be super fun to work with. I still kind of want to pull the processor boards and stick in a raspberri pi, but there is just never enough time in the day.
Gotta say, $30 for ANY TTL monitor is a pretty sweet deal, let alone such a rare and proprietary one!
As for dealing with Shugart drives, yeah, it's a pretty common experience with Tandy owners as well. Sure, normal PC drives work, but you either have to worry about fiddling with the jumpers, or building a custom ribbon to install a replacement PC drive. It's for that reason I always try to save 3.5" floppies that have jumpers/switches at the back, as it can make a direct replacement a lot simpler.
Yes! Typewritter action!
That reminds me of something they had at my elementary school. It was a typewriter with a row of 16 segment displays the same length as it would type on the page, it would apply to the paper the full line on the display when the return key was hit.
Need more typewriter/word processor videos! Can't wait to see the one on the Magavox Videowriter you teased in the last mail call video or really any of these machines as they're such an interesting concept and window of history
Wow, a literal Wipe Transition, impressive
You have to take the new floppy and use the program in the machine to format it to cpm for the data disk
Since you were able to take archives of the two floppies, were you able to restore this images to new disks and have them work in the machine? If so, you might be able to use the formatting tool to whack one of those. And if it doesn't let you because it recognizes that they are "system disks", you might be able to just hex edit the directories to remove files so it works? Experimenters dream project, right? 🙂Good work so far!
My thoughts exactly. It is either looking for some flag in the system-reserved area or files present.
I think the first thing to try is to overwrite file *names* in the hex editor with "AAAAAA" or something.
If it's the flag that is a lot harder though.
Did you try formatting a duplicate copy of EVERY compatible disk you have? Do the all give exactly the same error?
I freaking love this channel. That's it, that's all I have to say.
Some canadian listed site is selling one of these with what looks like a selection of floppy disks, including a box of xerox floppies
There is a way to figure this out. You decompile the os disc. You look for the place that returns the error message, then work backwards to see what checks it's performing. Then you can make a disc to fit those requirements. Might be as simple as a magic string. You might have scanned passed it, when you looked at the ASCII in the disc code.
It won't be exactly easy, but not undoable either. The CPU in that thing is no core i7. Likely some 8 bit thing like a z80. Very simple instruction sets.
I have a Xerox 860 in my shed...you want it? I no longer have the printer, though...I have a bad back, and those things weigh a ton. I even have some blank 8" floppies to go with it. I really ought to find it a home. I only took it because I'd used one in the navy, in the late 80s.
ive never thought about how lucky i was when i bought an Olivetti CWP-1 back in the mid 2010s for super cheap with monitor with 0 issues until now, it didnt need a weirdly formatted disk to work (but it does have a custom format, it can format any disk into its format and I'm still unable to dump anything i put into one), i didn't get a manual nor the "initialization disk" or whatever it is tho, but i can at least use as much as i can figure out about it lol, i love 80s word processors so much too, cant wait for the next vid!
7:59 🤯🤯as a skateboarding computer scientist subscriber
Thank God for you!
11:37 Which Epson sheet scanner is that? What happened to your CZUR? Just wondering about the effort of decombing/despiraling the manual.
As always the whole journey is fascinating. This typewriter is so much more intricate than I would imagine would be necessary. Thank you for your efforts!
That is an Epson GT-S50, the CZUR is still here, it's just about using the right tool for the job. The feed scanner is faster and easier for something I can debind. And the manual uses a metal spine that just opens up so it wasn't difficult.
I wonder if a dive into the ROM or OS will show what flag or marker it's looking for on the disk. The fact it can detect a boot drive is linked to it looking for a prepped blank disk with a marker on the the disk
Man, I love these videos. Maybe I'll get back into writing, make a Deus ex Machina style story featuring these classic machines.
I'm curious if a programming expert could find a way to disable the check for the necessary storage disk.
Can you make a close up of the leyboard and the funxtion keys?
4:44 KEYTRONIC foam and foil switches are always bad. ALPHAMERIC foam and foil switches are just fine. 🙂
Maybe this was stated in a previous video and I missed, but can the Memorywriter be used as a printer, the way that IBM Wheelwriter model does?
So near, yet so far. Awesome work!
This is really cool I absolutely love electromechanical stuff like this I really wanna get an IBM selectric one of these days
This is the most fallout esc thing especially with that monitor
@@spookisghostly4619Be warned though, the Selectric typewriters are very complicated mechanically and apparently are hard to work on! My parents had one when I was a kid in the 80's and 90's and I remember they were always having problems with it! I love vintage machines and especially typewriters, but I have no interest in getting a Selectric!
@@ct6502-c7w oh I'm aware it's one of the most complicated things I've learned about honestly and I have a fascination with typewriters, film cameras, tape recorders, ect it just sort of feels like the next step id make eventually
23:42
damn those equipment moving ninjas
8:11 That plumbing tool is what Luigi used to remove the rock from the meteorite in the 1993 movie!
Does it detect if the disc is read-only at all? If not, is it possible that what is used for detecting that has seized up or malfunctioned, keeping the drive from being able to write to the disk coupled with unhelpful errors?
Yes, it's my idea too maybe the write disk mechanism is blocked !!
@@guepar58 Also, isn't it completely up to the software to honor the Write Protection or not? If so, could it be that some of it just doesn't care and won't throw an error even if Write Protection is enabled or faulty?
Edit: "After some research It seems like this was mostly controlled or gated by the drive and not software running on the host.
I have tested it with the write protect space covered, it gives an error about that instead.
@@TechTangents Is there some sort of magic byte at some point on the disc that the OS is looking for so that a user who might not know what they're doing doesn't completely overwrite their system disks? Or, probably more likely just in order to sell special preformatted disks at a markup to regular floppy discs?
If so, maybe someone can reverse engineer the OS in order to find out what it's looking for.
@@TechTangents It looks like Xerox sold branded diskettes for different things. One thing I could find was Gold Series followed by For Office Imaging Technology. Though, those are unformatted 2 sided 40 track regular discs with supposedly high quality.
nice work! i was thinking about his machine, a couple of months ago i saw one (in very very bad shape) in a flea market and i was thinking if you were going do an update on it
I think most of them are in bad shape unfortunately. This whole device category is a hard sell for collectors now and they are languishing. Hopefully we can find enough stuff to be able to keep them going!
I have a Xerox 6040 Memory Writer, it has 3.5 floppy disks. But I do not have the Daisy wheel so it does not pass the self-test and will not operate. much sad
This video has made me wonder how hard it would be to create a custom typewriter using a raspberry pi, some steppers and a resin-printed fontwheel 🤔Could be an interesting electro-mechanical experiment 🤔
Let us know how you get on!
@@RetroJack Will do, I'm starting a YT channel for documenting this sort of thing, beginning with a rebuild of an old 3D printer, followed by creating a home assistant light switch
@@clonkex Nice - I look forward to seeing it!
@@clonkex : Assuming you're going for daisy-wheel style, remember that you also want a solenoid for the actual "print" step- this should probably be controlled by a simple relay, which can usually be found at either car part stores or any farm store that features some vehicle lights (they get used to control extra lights from a single switch); control the relay itself with a transistor, and a flyback diode.
@@clonkex : For light switches, assuming-
1) this is another Raspberry Pi project, and
2) you're looking at standard wall voltages:
You can get "contactors" (high power relays) at "contractor supply stores", and control them with relays. Frankly, a contactor is massive overkill for household lighting, but the more normal relays that stores around me sell say nothing about their voltage capabilities, so safety first.
Yeah, I think you need to rub the camera again, it still seemed like it wasn't quite focused, double vision on the floppy drives and the screen there. 😉 I'll just assume it was fixed for the second channel.
Have you checked the write-protect detector on the drives?
Have you tried decompiling the OS and reverse engineering the format based on what the OS is expecting for the blank media?
random thought, is the "write protect" sensor in the drive malfunctioning?
Ooooh! I have that exact bench vise! I'm leaning my elbow on it right now!
Looks rather cheap and nasty - like from the last rate chinese factory that hasn't quite figured out how to make stuff yet
@@gorak9000 It isn't. It's a Bessey, and while not a precision machining vise by any means it's OK for what it is. I do suggest applying some non-drying silicone grease to the leadscrew and slide pins when you first take it out of the box, however.
@@darkwinter7395 You can't see any markings on it in the video, and I have a very similar looking one from Harbor Freight that's pretty bad.
what a saga!!!!
I have a Smith Corona word processor that won't format disks either. I tried single density, double density, 720K and 1.44MB, nothing works. The drive runs...it just won't format anything.
It looks to me that Xerox produced their own custom forma floppies for the machine, probably to keep customers returning to Xerox.
Have you tried constructing a virtual image for a blank floppy that mimics the Xerox format and using a raw write to construct it on a blank 5.25" disk?
how is your fendar woot tshirt still intact?
Love the channel, man! Great job.
Wouldn't have attempted this I would've assumed I ddint have the proper tools experience or patience good job sir braver than I 😂
Tried hard sectored floppies for the memory?
It would be really cool getting access to the software 😅
Fun fact xerox reused that case. I had xerox that looked that but bthe 5.25 floppies were blocked off
Have you tried covering the write protect notch? In some old systems the write protect the opposite way ibm machines did; tape over notch allows writing.
He had
I will bet they didn't let it create the formatted disks so you had to go to them to get new ones.
Would a GoTek work in place of one of the real drives, I wonder?
I don't think it supports the non-standard disk format
@@jwhite5008 Gotek / Flash Floppy supports everything. If not just ask the Flash Floppy dev and he will add support. I worked with him several times to add some unsupported features he's a great guy and very helpful and enthusiastic.
I think I know why you might not be able to format the discs, including the boot disc: it's write protected. 5.25 inch floppies have a write protect notch on one side: if the notch is visible the disc cannot be written to. To disable this, cover it with a small piece of opaque tape folded around the edge of the disc. The notch I'm talking about is visible at 18:15 when you hold the disc up in front of the machine.
I can't tell just from watching if the other discs have the same notch.
I grew up with these things. Yes, I'm old. :-D You young whippersnappers, etc. :-D
Are you sure? 5.25s are write protected if the notch is covered.
@@ownpj Yeah he's wrong. I have 1000's of 5 1/4" disks here. Write protect is with the notch visible. Back in the day most original software disks didn't have a notch so they couldn't accidentally be erased. Blank disks were expensive so we cut a notch on the other side and flipped the disk and used the other side heh!
Please send that letter to a real Xerox office, just to see what they send in reply!
interesting problem making blank media? you had ago at copying blank data area from a working disk, to your new experimental media disc? did not work? I am assuming, now the ins and outs working at least the diagnostic anyway? in your demo, the message the comes up on the bottom left of the screen? where the text coming from? I now why it putting the message on the screen but where are actual letters stored are in the machines bios, or what passes for bios? and the bit maybe clue, like PC floppy hard disks etc, the is is not system disk message stored on the media it self? if you even load really blanked disk, your don't get any message at all? so this is the hunch, could not just look for message on the working disk, and, find it bit for bit location? the most bit something bios look and reads that message from the disk? and the string data in the bios, more hunches is going the the same line of text, in machine code or hex or what ever? more hunches, will there be a good change in the boos memory, close by, a list of expected text messages, that could be read disk loading time? to check for? and just do the steps backwards, put text data on floppy disk? and see what happens?
Would it be funny to just contact xerox for some help formatting a disk i bet this thing is still used in government offices around the world still. I see type writers in businesses still too usually the ibm branded ones.
Can you edit the boot software?
I wonder what Xerox was thinking when they designed that thing? 🤨
In the Xerox letter - loyal user, not loyal users. ;-)
And now that the video is out, someone emails you that he has the discs or someone puts them on ebay 🤣
Hey man. I dont know if you will see this comment but a guy in nj is selling this exact xerox with the crt and everything on facebook marketplace. I tried to find your email so i could send you the link to the listing but couldnt find one. Lmk how to send it to you if you’re interested.
Did I not see the data general come back to life yet? Also great video on word processors.
Based KDE Plasma user
Nice video.
Is its as stupid as needing hard sector floppy disks?
I don't think it would read anything then.
And the padding at the end of each track suggests it's not hard-sectored
I swear to you, you are the luckiest !@£$%^& in the retro-tech World.
So basically Xerox got into the "ink cartridge" scam before ink cartridges were a thing.
"You cannot format a blank disk, you must buy empty ones from us! Mwahahahahahah!"
-Xerox
That ink tape is worse. It's basically toner on a clear tape. After typing on it the toner is removed and the tape is clear in that place. So they can't be re-inked. Once used they are scrap.
Vibration detector: microphone with a stick.
It wants you to put in a blank disk, lol.
get onkeys to rewrite nthe the boot disk
What if you created another floppy disk from a system disk but corrupt the files to where the system doesn't recognize it as a system disk
Please dump the roms this needs to be emulated, just because heh! Please also take readable pics of the board/chips so that the emulation will have the correct hardware and a chance of working correctly.
5:41 Ow, flashing lights hazard.
Can it run doom
16:33 Not inane... *VENDOR LOCK-IN.*
When you E5'ed the disk, maybe you did also E5'ed the bytes telling the machine which disk it is. Who knows Xerox did it on purpose to make sure users did buy THEIR blank disk and not a non-Xerox blank disk. Who knows how many people failed to format a non-Xerox disk like you did.
So maybe their custom disks were a way to make money from users. Gross tactic and makes their product useless on its own. Maybe someone on the internet would help you hack it.