PDA

View Full Version : Kaypro 10 CP/M 2.2H Reload Issues



TangentDelta
December 25th, 2017, 04:25 PM
Hello! I recently acquired a Kaypro 10 "83". I'd like to get put a KayPlus BIOS in it, but in order to do that, I need to get a working load of CP/M going first.

The machine was shipped to me with the drive safely parked with SAFETY.COM. On the load of CP/M 2.2H it had when I got it, FINDBAD was not reporting any bad blocks on the A partition, but the B partition would always give me bad block errors. In my experimentation, I stupidly formatted the hard drive and lost the working load of CP/M 2.2H that it had.

I have the full set of 8 CP/M 2.2H floppies, and I found images of them online. I'm able to boot to the first floppy, and it immediately starts the EX script. Once it does its cursory drive check, it attempts to run PUTSYS.COM and immediately gives me a "DRIVE WRITE FAULT" message and locks up. I ran the FORMAT.COM utility on user 13 with the correct parameters (1,1; 0,3; 0,305) and got about 40 bad blocks during the format. I rebooted the system and tried to load CP/M again, and got the same "DRIVE WRITE FAULT" message.

So, at this point, am I looking at a bad drive? Is there a low-level-format routine on the hard drive controller that I can jump to with a monitor?

https://i.imgur.com/0qN7qGZ.jpg

ldkraemer
December 27th, 2017, 06:54 AM
Maybe this posting or perhaps another one on formatting will help you get started.

http://www.vcfed.org/forum/showthread.php?36747-Format-Hard-Disk-on-Kaypro-10
The second from last posting will likely work for you.

Larry

durgadas311
December 27th, 2017, 08:51 AM
I think there are additional steps required to map-out bad blocks, but if you have info on the bad blocks we should be able to tell whether they occur in the area used by PUTSYS. I'm not sure where the WRITE FAULT message comes from, but at this point I am thinking PUTSYS got an error result from the ROM write routine. Not sure if that is from the controller or due to media/verification errors.

TangentDelta
December 27th, 2017, 09:30 AM
Maybe this posting or perhaps another one on formatting will help you get started.

http://www.vcfed.org/forum/showthread.php?36747-Format-Hard-Disk-on-Kaypro-10
The second from last posting will likely work for you.


Those are the steps I followed exactly. It finds about 20 bad blocks and then asks if I want a second pass, which I decline.


I think there are additional steps required to map-out bad blocks, but if you have info on the bad blocks we should be able to tell whether they occur in the area used by PUTSYS. I'm not sure where the WRITE FAULT message comes from, but at this point I am thinking PUTSYS got an error result from the ROM write routine. Not sure if that is from the controller or due to media/verification errors.

I'll try to document the bad blocks when I get a chance.

TangentDelta
January 9th, 2018, 03:48 PM
Alright, so I finally got around to working on the Kaypro today. I ran FORMAT.COM on user 13 and took a picture of the screen.

https://i.imgur.com/jDvZbRo.jpg

Only one error. I was getting ~30 errors the last time I ran the utility.

After formatting the drive I ran SAFETY.COM and power-cycled the machine. It booted to the floppy, and I pressed return at the prompt.

https://i.imgur.com/wjVBY3U.jpg

...and I'm getting the same error message.

This issue really has me scratching my head.

TangentDelta
January 9th, 2018, 05:55 PM
Quick update. I found a disassembly listing for PUTSYS.COM (http://www.thebattles.net/oddments/kaypro/putovl.mac). I used DDT to insert code that loads the error register at port 81h into the A register when it displays "DISK WRITE FAULT".

When it fails, the error register has the value 10h. Looking at the WD1002 datasheet (https://computerarchive.org/files/comp/hardware/WD1002-05%20Winchester%20Controller%20(19xx)(Western%20Di gital)(pdf).pdf), the error indicates an "ID NOT FOUND". I get the same error when I try to run CHECK.COM as the service manual suggests.

So, why would CHECK.COM be unable to access the disk yet FORMAT.COM can? They seem to be identical.

durgadas311
January 11th, 2018, 04:30 AM
I believe "ID NOT FOUND" refers to a failure to read the sector header, which indicates a bad track (although there could be other causes). I've not scrolled back to see what FORMAT was able to do, but format will rewrite the entire formatting and unless you do a verify you won't know if the sector header is readable. If the sector header was good immediately after the format, but was found to be bad shortly after that, it could indicate degraded media (or e.g. weak write current or other electronic failures).

TangentDelta
January 11th, 2018, 09:33 AM
I believe "ID NOT FOUND" refers to a failure to read the sector header, which indicates a bad track (although there could be other causes). I've not scrolled back to see what FORMAT was able to do, but format will rewrite the entire formatting and unless you do a verify you won't know if the sector header is readable. If the sector header was good immediately after the format, but was found to be bad shortly after that, it could indicate degraded media (or e.g. weak write current or other electronic failures).

How would I determine if the sector header is unreadable? Is there a more in-depth diagnostics tool available? If the drive is toast, what would be the chances that I could repair it? Could I replace the entire thing with a solid-state solution of some sort?

durgadas311
January 11th, 2018, 09:42 AM
I'm not familiar with the tools Kaypro provided, but there should be some method to test the drive. I believe there are SSD replacements, but do not have experience with that, either. Perhaps others can help.

Alphasite
January 11th, 2018, 01:14 PM
I'm not familiar with the tools Kaypro provided, but there should be some method to test the drive. I believe there are SSD replacements, but do not have experience with that, either. Perhaps others can help.

There are three that I've bought. The first is David Gesswein's reader/emulator:

https://www.pdp8.net/mfm/mfm.shtml

I've used it to replace hard drives in my TRS-80 external drives but haven't tried on my Kaypros. The tested configurations page lists the Kaypro 10 as working.

The second is the DREM:

https://www.drem.info/

I just received mine a week or so ago and haven't had a chance to test it.

The third is the FreHD:

http://ianmav.customer.netspace.net.au/trs80/emulator

It was designed for the TRS-80 and requires an adapterboard between the HD connector on the Kaypro motherboard and the FreHD to convert the port addresses. I've used this with my Kaypro 2X that had a fully-populated motherboard with the Kaypro 10 ROMs and one of the 3rd party ROMs. There's a thread here about it that someone started.

TangentDelta
January 11th, 2018, 03:20 PM
I believe "ID NOT FOUND" refers to a failure to read the sector header, which indicates a bad track (although there could be other causes). I've not scrolled back to see what FORMAT was able to do, but format will rewrite the entire formatting and unless you do a verify you won't know if the sector header is readable. If the sector header was good immediately after the format, but was found to be bad shortly after that, it could indicate degraded media (or e.g. weak write current or other electronic failures).


I tried formatting the first few sectors and then immediately running CHECK.COM, but it still reports a NID error.


There are three that I've bought. The first is David Gesswein's reader/emulator:

https://www.pdp8.net/mfm/mfm.shtml

I've used it to replace hard drives in my TRS-80 external drives but haven't tried on my Kaypros. The tested configurations page lists the Kaypro 10 as working.

The second is the DREM:

https://www.drem.info/

I just received mine a week or so ago and haven't had a chance to test it.

The third is the FreHD:

http://ianmav.customer.netspace.net.au/trs80/emulator

It was designed for the TRS-80 and requires an adapterboard between the HD connector on the Kaypro motherboard and the FreHD to convert the port addresses. I've used this with my Kaypro 2X that had a fully-populated motherboard with the Kaypro 10 ROMs and one of the 3rd party ROMs. There's a thread here about it that someone started.

I sent David an email earlier today with an inquiry on the availability of a kit, or pre-built unit.

The WD1002 seems like an fairly simple controller. I'll do some research to see what it would take for me to build my own emulator using an Arduino Mega.

bcc
January 13th, 2018, 11:14 PM
I believe your drive is suspect, and power levels need to be measured on the drive itself.
You may want to rerun the format on the first 5 tracks or so, several times. (So start at 0 end at 5 instead of 305).
Run the 0-5 cylinder format a number of time. You could probably do this with a spare known good drive also (just Cylinder 0-5)
The bad spots you found are far away from the area of the drive that putsys is dealing with.

I have a utility from Balcones Computer Corp called SC (for Structured Copy) which has an option to verify the drive. It's pretty quick and reads and reports errors by Group ID. There is another utility called FLAW that allows me to then flaw out the bad spot by Group ID. I believe if Group ID 0 is flawed, Putsys will fail as you described. Likewise if cylinder 0 reports errors the HD may be toast unless you can do some serious adjusting of the location of track 0...some drives have an external means of accomplishing this feat, although I have yet to attempt it.

Good luck.

TangentDelta
January 18th, 2018, 03:22 PM
I tore down and de-oxed the old ST-212 drive (without opening it, of course) that the K10 came with. Initially, the drive was working perfectly outside of the system. I was able to format the disk, and CHECK.COM did not error out. PUTSYS.COM was able to do the system installation and make the drive bootable. I put it all back together, and it was right back to doing the same thing.

I was able to snag an ST-251 out of an old 386 DOS machine, and got it fitted into the K10. I was able to format it with no errors (other than the flaws from the factory), and do a complete CP/M 2.2H install!

I'm going to put the ST-212 into the 386 machine and see what I can do with it. Maybe I'll be able to run better diagnostics on it. It feels to me like a loose connector somewhere, or a cold solder joint.

Thanks for all of the suggestions and help!

TangentDelta
January 18th, 2018, 05:59 PM
So, new problem. I created the media as described in the KayPlus manual, but the KayPlus ROM does not recognize it. KPCNFG83.COM sees the disk I created as a valid KayPlus system disk, and lets me change the settings on it. When I swap the stock ROM out for the KayPlus ROM (with the correct adapter socket and wire soldered to pin 2) the system never tries to boot. It sits and waits for me to press a key, which takes me to the monitor. From there, if I try to boot to the floppy, I just get a SYSTEM? error message.

bcc
January 21st, 2018, 06:08 PM
So, new problem. I created the media as described in the KayPlus manual, but the KayPlus ROM does not recognize it. KPCNFG83.COM sees the disk I created as a valid KayPlus system disk, and lets me change the settings on it. When I swap the stock ROM out for the KayPlus ROM (with the correct adapter socket and wire soldered to pin 2) the system never tries to boot. It sits and waits for me to press a key, which takes me to the monitor. From there, if I try to boot to the floppy, I just get a SYSTEM? error message.

I don't want to come across as heckling you, but I often find there are simple solutions to a bunch of problems...

Still thinking about the power supply and/or system memory. If you have access to 8 memory chips and they are socketed, it may be worth a try swapping them out and going with a faster version (such as a "4164-2" instead of a "-3").

I have not used the KayPlus Rom...I suppose you have already check things like checksums and verified voltages are not anywhere lower than 4.8VDC anywhere in the system (HD is probably as far away as you can get...check there). Is the KayPlus ROm type identical? (for example 24 pins and same voltage requirements on the same pins). Also make sure that power connector intended to power the WD-1002 board which has only 2 wires isn't trying to power up any drives...floppy or hard. Are any ribbon cables or power cables different between starting and ending with the problem?

The +5VDC output of the power supply can be adjusted (small adjustments) in my experience, but probably not the +12VDC output.

FYI,
On my K-10 I dumped the Aztec supply and replaced it with a commercial Power-One (now BEL) supply. the +12VDC supply could not supply enough current to spin up the big drives.
(It's not a cheap solution, but I know for certain how reliable the supply is as I used to work for Power-One in my first electronics job. The one I used is an exact replacement on the mounting holes as well and is speced to supply a higher peak current on the +12 supply when demanded...I also target using a 3.5" drive with lower power loading when possible.

Keep at it!

TangentDelta
February 2nd, 2018, 05:27 AM
I don't want to come across as heckling you, but I often find there are simple solutions to a bunch of problems...

Still thinking about the power supply and/or system memory. If you have access to 8 memory chips and they are socketed, it may be worth a try swapping them out and going with a faster version (such as a "4164-2" instead of a "-3").

I have not used the KayPlus Rom...I suppose you have already check things like checksums and verified voltages are not anywhere lower than 4.8VDC anywhere in the system (HD is probably as far away as you can get...check there). Is the KayPlus ROm type identical? (for example 24 pins and same voltage requirements on the same pins). Also make sure that power connector intended to power the WD-1002 board which has only 2 wires isn't trying to power up any drives...floppy or hard. Are any ribbon cables or power cables different between starting and ending with the problem?

The +5VDC output of the power supply can be adjusted (small adjustments) in my experience, but probably not the +12VDC output.

FYI,
On my K-10 I dumped the Aztec supply and replaced it with a commercial Power-One (now BEL) supply. the +12VDC supply could not supply enough current to spin up the big drives.
(It's not a cheap solution, but I know for certain how reliable the supply is as I used to work for Power-One in my first electronics job. The one I used is an exact replacement on the mounting holes as well and is speced to supply a higher peak current on the +12 supply when demanded...I also target using a 3.5" drive with lower power loading when possible.

Keep at it!

All of the voltages on the system board looked good. I only ever saw the 5V rail dip down to 4.9V during my testing, though to be fair that was before I put the new ST-251 in the system. I'll see if I can get another reading off of it this weekend when I have some time.

Unfortunately my memory chips are not socketed. According to the memory tests I ran from the KayPlus monitor, the memory is good. If you know of a more accurate memory test that doesn't involve desoldering 8 chips, let me know.

One thing that I read online was that the KayPlus ROM does not work correctly on the "new" style '83 Kaypro 10s. It seems to be due to the ~RESET line for the WD-1002 being inverted on the processor board for some reason. If this was causing my problems, wouldn't you think unplugging the WD-1002 would still let me boot to the floppy disk?

I've been using the K10 just fine with the ST-251 for the past few weeks. I have Kermit up and running and can transfer files to-and-from my 386 tweener machine. I haven't had any weird hiccups so far. If it's a voltage problem, wouldn't I be running into issues just using it as it is? Could the 5V rail be just on the edge of failing, and the beefier ROM I'm using for KayPlus is putting it over the edge?

bcc
February 3rd, 2018, 02:12 PM
Actually the Aztec power supply does not have an output voltage adjustment potentiometer...the California DC power supply does.

Since you say you've been using the system without incident, your ram is probably OK.

The 5V line is ideally set above 5V..I typically set it at just below 5.25V knowing there will be some voltage drops along the way to the devices. If you know the supply isn't going below 4.8 volts you are probably OK...but as you said, the 4.9Vdc was observed before adding the 5.25" ST-251.

I seriously doubt the new ROM power draw would be a viable factor.

I am more inclined to think the WD1002 board data separator circuit, adjustable C16 may play a part in the "bad sector" issues.
Yes, just removing the interface to the HD will allow floppy boot.

The inverted reset issue (WD1002) may be the intermediate board version that was being discussed...the "A" version probably addressed that issue...see my post that shows the intermediate board schematic. My K10 uses the "A" version with the 81-302C Rom. (have not tried the KayPlus Rom yet)


(I sent a message earlier today about that C16 possibility, but my messages may take 7-10 days to hit this forum given past experience...today is 2/3/18 so my guess is that you will read this message around 2/10/18).

bcc
February 6th, 2018, 10:33 PM
Actually the Aztec power supply does not have an output voltage adjustment potentiometer...the California DC power supply does.

Since you say you've been using the system without incident, your ram is probably OK.

The 5V line is ideally set above 5V..I typically set it at just below 5.25V knowing there will be some voltage drops along the way to the devices. If you know the supply isn't going below 4.8 volts you are probably OK...but as you said, the 4.9Vdc was observed before adding the 5.25" ST-251.

I seriously doubt the new ROM power draw would be a viable factor.

I am more inclined to think the WD1002 board data separator circuit, adjustable C16 may play a part in the "bad sector" issues.
Yes, just removing the interface to the HD will allow floppy boot.

The inverted reset issue (WD1002) may be the intermediate board version that was being discussed...the "A" version probably addressed that issue...see my post that shows the intermediate board schematic. My K10 uses the "A" version with the 81-302C Rom. (have not tried the KayPlus Rom yet)


(I sent a message earlier today about that C16 possibility, but my messages may take 7-10 days to hit this forum given past experience...today is 2/3/18 so my guess is that you will read this message around 2/10/18).

=========================
ERRORS WITH PUTSYS:

I responded to the OP as indicated in the previous post, but during that initial moderator review period that POST never made it here. ..

The following was sent as a PM to the OP, but with some edits I believe it could come in handy for others that experience HD read errors, so including the info here.

===========================
Just try to circumvent running Putsys right off the bat if it is failing.

(You will probably find the data separator C16 is at fault. You are very close to the right setting since it has worked before.)

Format those first few tracks with the Kaypro FORMAT program. (PUTSYS only uses 1 or 2 tracks starting at CYL 0)...don't bother to flaw bad spots yet. Try CYL 0,5 If errors then try the C16 adjust below. If no errors, try PUTSYS, but read the rest of this message. Finding the midpoint is crucial.

Then try to just "dir" the B: drive or startup a format (format won't stop on errors which is what you want...but DIR's won't erase the existing data on the HD...if you know the HD history you could target formatting an area way beyond your data usage).

Note the location of the C16 cap slot.
If you get a read error, give a slight turn clockwise to C16 and try again... repeat again... and again (enter <CR> to the error message meaning "retry" fir DIR's), but don't bust the cap by turning too far!

If no change return the slot to about where it was and repeat, but turn in the counter clockwise direction...slight adjustments till errors subside

HEAT will affect the setting. It might work hot, but let it cool and maybe wont work till it is hot again...so readjust to find that median point where it works both hot and cold. (In fact I think Kaypro docs indicate warming up the system for about 45 minutes before formatting...the C16 adjust is the WD1002 achilles heel.). I would not suggest freeze spray, but you'll end up needing to check things both hot and cold...thus the midpoint target.

Try to find the setting with no errors...a midpoint between clockwise errors and counter clockwise errors...then try PUTSYS to see if it passes. Reboot with no floppy if PUTSYS passed. If it booted and you can get a directory, start over and format the whole drive (on a new setup only), flaw bad spots, see how it works both hot and cold, load CP/M files, pip stuff around

If anyone has turned the C16 cap too far the cap may no longer change value...you need a 0-60pF variable cap (or there abouts). You will find them on ebay cheap...they are a brown color and probably better quality than the original.

I have a FUJI drive that genuinely has about 25 bad spots, but a good drive will have just a couple or none at all.

I have used this method many times maintaining the WD1002-05 in the field.
WD includes a Maintenance method using a frequency counter and dual trace oscilloscope which is fine if you want to go there. WD also mentions the need to repeat the adjustment over time.