Help please: Cromemco Z-1 with PFD

[Martin Eberhard]

Well, time to stop lurking...

I am bringing up a Cromemco Z-1 with a Cromemco PFD disk drive, and I have reached the point where I could use a few hints from Cromemco experts.

I have debugged and repaired the Persci 277 disk drive using a Processor Technology Sol 20/Helios II, and the Processor Technology SIMU software. The drive now seems to work correctly on the Sol 20. There a couple of dozen jumpers that must change when moving a Persci 277 from a Helios II environment to a Cromemco environment. I have moved them over for the Cromemco. (If anyone's interested in the Persci 277 jumper changes, let me know.)

Again, after a fair bit of cleanup and debugging, the Z-1 seems to work correctly at 2 MHz and 4MHz, though I have not tested it extensively. The following boards are installed:
(1) Cromemco ZPU
(4) Cromemco 16Kz
(1) Cromemco 16FDC

When connected to a Wyse WY-30 terminal (which I designed :-) ), RDOS 2 comes up and seems to work as advertised.

Lots of Cromemco software seems to be available in Image Disk format (.IMD), so I got a copy of Dave Dunfield's Image Disk software. I hacked up an old Dell PC to be able to read/write both a Shugart SA850 and an old 360K 5.25" drive. (John Wilson's FDADAP board works great for interfacing to 8" drives.) Image Disk seems to work with this setup - I can write files to a floppy and read them back correctly. Further, I can write 8" disk images for CP/M and read them correctly on one of my CP/M machines.

I pulled several images of Cromemco CDOS disks from the web, in Image Disk (.IMD) format. Rumaging through, I found several CDOS images that were for 8" disks, and were single-sided, single density. I also found a few single-sided 5.25" images.

When I attempt to boot CDOS using the Persci drives, I get poor results. CDOS comes up (sometimes), but I can't always read the directory, and I often get "Read-after-write" errors. Executing STAT seems to trash the disk. I note that CDOS complains if I remove the write-enable tab from the disk.

I suspect that the problem is the disk controller, not the drive. I switched to a known-good SA400L drive, and attempted to boot a 5.25" CDOS image - no luck. with this drive, I never got to the CDOS prompt - after trying several CDOS images. However, I can successfully and reliably read individual sectors and sequences of sectors using RDOS commands. This is true for both single-density and double-density tracks.

I put a frequency counter on the 16FDC data separator and tuned it to 500.46 KHz. (it was a little out of spec.) No luck.

So here are a few questions:

1. Does CDOS write to a floppy during normal operations like DIR and STAT? Or is the read-after-write error a symptom of a larger problem, like a bad CDOS load that is executing garbage?

2. Can someone point me to a known-good single-sided 8" CDOS image? Also a single-sided 5.25" image would be nice.

3. Even better, would someone snail-mail me a single-sided CDOS disk?

4. Any advice on debugging the 16FDC? What should I be looking for?

This Z-1 will be a sweet machine when it's done. It's now in great shape, and the walnut cabinet of the PFD is gorgeous. Your help restoring this machine is appreciated.

Thanks!

Martin

[Michael Leon Peterson]

Martin,

 

As soon as I get my 5.25 CDOS disk from Marcus I will be able to create additional 5.25 and 8 in for you and provide that.  I am expecting them to arrive next week, Mon or Tue. 

 

I would expect the issues to be more with the Persci 277, rather than the 4 FDC, the heads on those will wear and the response goes away.  I used to repair those by replacing the heads even on 299.  What version of CDOS are you using?

 

Michael

--
You received this message because you are subscribed to the Google Groups "Cromemco" group.
To view this discussion on the web visit https://groups.google.com/d/msg/cromemco/-/uNdCbbOEu8kJ.
To post to this group, send email to crom...@googlegroups.com.
To unsubscribe from this group, send email to cromemco+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/cromemco?hl=en.

[Martin Eberhard]

Hi Michael,

Thanks! send me a note when you have CDOS. Not sure what sort of drives you have, but the ones on my Z-1 are both single-sided - the Persci 277 and the Shugart 400L.

Yes, ordinarily, I would suspect the Persci 277 as well. But... the Persci works fine on the Helios II (after quite a bit of rebuilding). Granted, on the Helios things are set up a bit differently, but the heads and servo must be at least reasonably ok. Also, the 16FDC does not seem to work with a known-good Shugart SA400L (5.25" drive). This makes me suspect either the 16FDC, or the various copies of CDOS I have tried.

I've found a few CDOS versions out there - all with more or less the same result. I have tried IMD files that were labeled CDOS 2.17, 2.36, 2.53, and a few others.

Note that my controller is the 16FDC, not the 4FDC. The 4FDC is built around the Western Digital WD1771. The 16FDC is built around the Western Digital WD1793. There are some basic incompatibilities between these two controller chips - I have seen that disks written using a WD1771 won't necessarily read correctly with a WD1793. The most obvious difference is that the 1771 does not write an Index Address Mark at the beginning of each track, while the 1793 does. Additionally, the 1773 writes just 10 bytes of FF after each sector, while the 1793 writes 27 bytes of FF after each sector.

I kind of suspect that the .IMD files that I have found were written originally using a 4FDC, and this somehow made the IMD file incompatible with a 16FDC. Not sure if this makes any sense...

Martin Eberhard

[Bill Sudbrink]

I had the same problems with my 16FDC.  After eliminating everything else, it turned out to

be the 1793 FDC chip that had gone “flakey”.  My 16FDC had one of the early 1793s, ceramic

with the “painted over” quartz window on top.  I replaced it with a later (only a year later if I

recall correctly) plastic encased chip and everything has been rock solid ever since. 1793s are

getting a little thin on the ground but I think Unicorn still has them.

 

I guess (but don’t know for sure) that the early 1793’s with the windows had some

erasable/programmable element under the window.  If that is the case, then I figure that

programming is fading.  Continuing with that speculation, the plastic parts have whatever

element that is hard programmed (not erasable) and are therefore more reliable.

 

Bill Sudbrink

 

From: crom...@googlegroups.com [mailto:crom...@googlegroups.com] On Behalf Of Martin Eberhard
Sent: Thursday, August 11, 2011 8:11 PM
To: crom...@googlegroups.com
Subject: Help please: Cromemco Z-1 with PFD

 

Well, time to stop lurking...

--

You received this message because you are subscribed to the Google Groups "Cromemco" group.
To view this discussion on the web visit https://groups.google.com/d/msg/cromemco/-/uNdCbbOEu8kJ.
To post to this group, send email to crom...@googlegroups.com.
To unsubscribe from this group, send email to cromemco+u...@googlegroups.com.
For more options, visit this group at http://groups.google.com/group/cromemco?hl=en.

---

No virus found in this message.
Checked by AVG - www.avg.com
Version: 10.0.1392 / Virus Database: 1520/3828 - Release Date: 08/11/11

[bil...@aol.com]

 

 

As soon as I get my 5.25 CDOS disk from Marcus I will be able to create additional 5.25 and 8 in for you and provide that.  I am expecting them to arrive next week, Mon or Tue. 

 

I would expect the issues to be more with the Persci 277, rather than the 4 FDC, the heads on those will wear and the response goes away.  I used to repair those by replacing the heads even on 299.  What version of CDOS are you using?

 

I have given up trying to get my PerSci 299 to work correctly.  It seems like I fix on cap and another one goes. After a while you're looking for something more reliable.

 

Bill

[Martin Eberhard]

Want to sell me your 299? (I am serious.)

After fixing the first 6 or 7 caps, I realized that all of the failed caps were tantalum caps that were connected to the +24V rail, either directly or through a limiting resistor. Clearly, these caps suffered long-term damage at this voltage, despite being rated for 35V. So I simply replaced every one of the tantulums an the 24V rail, on every single board. Tedious, but effective. (Replacement caps were rated at 50V...)

The other nightmare of the Persci drives is the incandescent bulb-based head servo system. Took me a while to get the hang of debugging this. I have not been able to find exact replacement bulbs. But I have found bulbs that are physically the same size, with somewhat different resistance, and therefore different light output at the same voltage. I figured out how to use these, adjusting the lamp voltage such that the *power* consumption of the bulb is the same as the original. This results in close to the same light output.

The reason I want to keep the Persci drive (instead of switching out for a much more reliable and simple Shugart drive) is because I want to restore this old machine to something close to its original condition. The Z-1 is Cromemco's first computer, and the PFD is their first disk drive subsystem. This machine shows the early thinking of the folks who started Cromemco, and shows the general state of the computer field at the time. Remember computers with walnut cabinets? The Sol-20, the Northstar Horizon, the Micronics disk subsystem. IMHO, the Cromemco PFD is one of the nicest and is somewhat rare.

I hope to run CDOS, and also the Z80-based version of Cromix. Someone else on this forum made a very succinct comment about Cromix running on a Z80. What an achievement!

If I had a 4FDC, I would be tempted to use it instead of the 16FDC, even though this means incompatibility with the "industry standard" IBM 3740 disk format. This is why I keep 4 16KZ memory boards, rather than a lower-power and more reliable 64KZ.

I plan to add several more older Cromemco boards to the system: aBytesaver, a TUART, and a D+7A. I am still looking for a TV Dazzler and a Cromemco joystick. Anybody have either of these they'd like to sell?

Martin

[Martin Eberhard]

Very interesting suggestion. My 1793 is also one with a window that has been painted over. You are right - it has some EPROM-based microcode, and it may have "dried up." I will try to track down a replacement 1793.

Thanks for the hint!

Martin

[MikeS]

Hello, Martin, and welcome!

 

Of course it could just be the disks, especially if they're 5.25" versions; most PCs using a 765 derivative have trouble with the narrow gap between sectors, if they can even do single density at all (see Dave Dunfield's TESTFDC program and registry of motherboard/controller capabilities).

 

There were also issues with some revisions of the 16FDC not working nicely with some drive types; see:

http://retrotechnology.net/herbs_stuff/16fdc.html

 

Maybe not everyone is aware that Dave also has a package called CROMRT on his site that lets you create a bootable 5.25" CDOS disk from scratch using RDOS and the console port. If you have trouble booting from a PC-created disk or even one from a different Cromemco system/drive like Michael P's then this might be the answer, since you're actually creating the disk on the same drive you're going to boot from.

 

Dave and I actually had a little friendly competition both working on that project some years ago; needless to say Dave finished his first, it was less of a kludge, and a much neater package than mine would have been (if I'd actually finished it ;-) )

 

FWIW, in my experience using 1.2M HD 5.25" drives (emulating 8") to transfer images and data between Cromemco and PC has been the most reliable, and usually even faster than using serial ports.

 

Good luck; sounds like a sweet machine indeed!

 

mike

 

----- Original Message -----

From: Bill Sudbrink

To: crom...@googlegroups.com

Sent: Friday, August 12, 2011 9:10 AM

Subject: RE: Help please: Cromemco Z-1 with PFD

I had the same problems with my 16FDC.  After eliminating everything else, it turned out to

be the 1793 FDC chip that had gone “flakey”.  <snip>

Bill Sudbrink

 

From: crom...@googlegroups.com [mailto:crom...@googlegroups.com] On Behalf Of Martin Eberhard

<snip>

[Martin Eberhard]

Thanks, Mike.

I have read most of Herb's stuff on the Persci drives, and per your suggestion, I just went and read the related discussion about the 16FDC. It seems the 16FDC was originally intended for the Persci 277 and 299 drives - so this setup should work.

Also, this particular 16FDC was originally connected to this particular 277 - though they both spent a few decades in a rat-infested garage since they last ran :-( They (and the Z-1) were pretty nasty looking when I got them.

In the next few weeks, I will attempt to create a CDOS disk via RDOS, as you suggest. This should tell me if the problem is with the IMD-created disk or with my hardware...

Martin

[Martin Eberhard]

Bill,

You are a genius!

I got a hold of a (Siemens) 1793 controller chip to replace the
WD1793, which like yours had a painted-over EPROM window.

I used Dave Dunfield's Image Disk utility on my hacked Dell PC to
write a CDOS 2.54 8", Single-sided, single density bootable disk.
And... It works great! I am able to:
* Boot CDOS 2.54 reliably
* Run STAT (works fine)
* run INIT to format a disk in B: (works perfectly)
* run WRTSYS to write the system tracks from A: to b: (works
perfectly)
* use XFER to copy (with verify) all the utility files from A: to B:

The good news is that my lovely PFD with its Persci 277 drives works
well - my rebuild of the Persci drive was a success. INIT complains
that my disk rotational speed is a tad low - I will adjust it upward a
few hertz soon.

I am well on my way now to having a pretty neat old-school pure-
Cromemco Z-1 system. The two pieces I still seek are a Dazzler and a
Cromemco joystick.

Thanks so much for the hint, Bill!

Martin



<SNIP> On Aug 12, 6:10 am, "Bill Sudbrink" <wh.sudbr...@verizon.net>
wrote:

> I had the same problems with my 16FDC.  After eliminating everything else,
> it turned out to
>

> be the 1793 FDC chip that had gone "flakey".... </SNIP>

[Bill Sudbrink]

Congrats! It's a fun box.

Everything I run on my box started out from Dave's images.
It is really a great service he has done for us.

I'm working on porting some other games to the Dazzler...
The one I really want to do is "Crush, Crumble and Chomp".
It was a really fun little game, all in basic. I know I
have a listing for the IBM PC with CGA somewhere... If I
can find it, porting to the dazzler under cromemco basic
shouldn't be hard.

By the way, you know that to use a cromemco JS-1 you need
a D+7A, right?

Bill

> -----Original Message-----
> From: crom...@googlegroups.com [mailto:crom...@googlegroups.com] On
> Behalf Of Martin Eberhard

> --
> You received this message because you are subscribed to the Google
> Groups "Cromemco" group.

> To post to this group, send email to crom...@googlegroups.com.
> To unsubscribe from this group, send email to
> cromemco+u...@googlegroups.com.
> For more options, visit this group at
> http://groups.google.com/group/cromemco?hl=en.
>

> -----

> No virus found in this message.
> Checked by AVG - www.avg.com

> Version: 10.0.1392 / Virus Database: 1520/3842 - Release Date: 08/18/11

[Martin Eberhard]

Agree completely about Dave's program and images. I think John
Wilson's FDADAP board (http://www.dbit.com/fdadap.html) is another
such tool - it works nicely with Image Disk to connect 8" drives,
eliminating the need for generating the TG43 signal (that the PC does
not generate, but 8" drives require). John seems to be doing the same
sort of restoration thing with DEC machines.

Yes, I've owned a D+7A for a while, looking for a joystick. I'll
probably rig something up homebrew until I get a real Cromemco one.
I'm also planning to install my Bytesaver and a few other vintage
Cromemco boards.