CompuPro Disk 1 and Shugart SA-850/851 drives

[RichCini]

All --

I spent some time today away from the CompuPro Spectrum to play around with my IMSAI. In the past, I used 5.25" HD drives in place of 8" drives, and they work well. Today, I tried to connect my dual Shugart SA-851 drives to the CompuPro Disk 1 controller, making all of the jumper adjustments on both the Disk 1 and the SA-851 per the CompuPro manual. The disks format properly but when putting CP/M on the disk, it won't boot (I continue to get BDOS sector not found errors).

The heads are clean, the disks are brand new 3M DSDD disks, and the drive pair work perfectly with my Gazelle (so it's not a drive problem). Does anyone have a working combination of the Disk 1/1A and SA-850/851 drives and can maybe confirm if any of the jumpers deviate from the recommended ones in the manuals?

Thanks!
Rich

 

[hmb]

If I understand correctly, you can:
- boot the system, presumably with some version of CP/M from CompuPro
- run the CompuPro FORMAT utility, and every track verifies without error
... but if you then try to install CP/M on a newly-formatted disk, any attempt to boot from that new disk results in the BDOS "sector not found" error.

To get as far as you did, the jumper settings on your drive are probably correct.

Are you able to read and write files to these newly formatted disks correctly, even if you are unable to create a working boot disk using these disks?

It has previously been noted that something as simple as the floppy drive step rate (as set in the BIOS) being just a bit too fast for the drive to keep up will cause the error you describe. However, this explanation only makes sense if you happen to be booting from a disk with a version of CP/M that has had its BIOS (as well as the LOADER in later versions of CompuPro CP/M) patched with a different step rate than the version of CP/M you are now installing on the new floppy disks.

 

[RichCini]

hmb --

Thanks very much. It's not a CompuPro software configuration at all, but rather highly customized (I never ran the CompuPro formatter; I have one "in ROM") CP/M system using 5.25" HD drives instead of 8". But, something you said made me think of a change I made which I will undo as a test. In the original ROM build, I had the step rate @ 5ms but I dropped it to 3ms because that's what was specified in the SA850 manual. I'm going to return it to 5ms and see if it makes a difference.

Thanks!

 

[hmb]

If you are not using the CompuPro FORMAT utility and you are trying to create a boot disk, you also need to take into account several issues specific to how CompuPro formatted its boot disks which varied depending on the CP/M version. The following has been copied from a previous posting on this topic in the Google Groups S100 forum:

Early Compupro CPM-80 Master Disks were on SSSD (single-sided single density) 8" floppies. The Boot Tracks on the early versions (up to 2.2N) of CPM-80 were recorded SSSD, and the console ports set at 9600. Step rate and Head Settle times were set for single-sided drives (which have lower/slower performance speeds). Later CPM-80 Master Disks were recorded with the Boot Tracks set for 1024 bytes sectors, even if the rest of the disk was formatted 128, 256, 512, or 1024 byte sections.

The later CPM Master Disks (Versions Q, R, S) were recorded on SSDD (single-sided, double-density) 8" floppies. Boot Tracks are recorded with 1024 byte sections, regardless of how the rest of the disk is formatted, because with time, the Boot or Boot Loader had grown larger than could be fit on single-density Boot Tracks. Console was set for 19.2K baud rate. These Operating Systems assume that you are using DS 8" drives with the faster/higher step rate, and do not work with single-side drives that do not meet double-sided performance specs (step rate, head settle rate).

 

[RichCini]

Thanks. This is a fully custom implementation (actually a bootstrapping process when no boot disk of any type is available) which relies on a custom monitor and ROM-based formatter (using the multi-sector formatting capability of the 8272 with a fixed interleave) and building the CP/M image in memory which is then multi-sector written to the disk. It's not designed to work with actual original disks and doesn't use the cold start loaders in the Disk 1 ROM.

I'm more specifically interested in comparing jumper settings for the drive in a working environment as compared to what is shown in the Disk 1 manual for the 850/851 drives. I'm trying to eliminate as many factors as possible. Diskettes are brand new 3M DSDD disks.

I guess I could find an image of a CompuPro disk and make a physical copy of it (I have the ability to image disks) and see if that can work.

Rich

 

[hmb]

I'm more specifically interested in comparing jumper settings for the drive in a working environment as compared to what is shown in the Disk 1 manual for the 850/851 drives. I'm trying to eliminate as many factors as possible. Diskettes are brand new 3M DSDD disks. Rich

Getting back to your original question then... as pointed out in CompuPro's manual for the Disk1, the primary requirement for jumper settings is that the spindle motor needs to be running all the time. As long as that is happening, and the proper drive select jumper is placed, the drive should work.

 

[RichCini]

Yes that makes sense.Both of those conditions exist. Head load seems to work; head carriage steps properly. So, it must be something in the software. Arrrgh.

It so happens I do have enough spare CompuPro parts that I could install and configure an Interfacer I, reconfigure the Disk 1 for using the internal boot routines and make a CompuPro 8" disk (like CP/M 2.2K or 2.2Q) from images in the Maslin Archive. Might be a better approach at this point since the original setup really was a no-disk-available bootstrapping.

Will report back on that.

Thanks again!
Rich

 

[hmb]

It so happens I do have enough spare CompuPro parts that I could install and configure an Interfacer I, reconfigure the Disk 1 for using the internal boot routines and make a CompuPro 8" disk (like CP/M 2.2K or 2.2Q) from images in the Maslin Archive. Might be a better approach at this point since the original setup really was a no-disk-available bootstrapping.

Ah - I had not appreciated the rationale for what you were doing, which was to boot CompuPro CP/M without having an actual boot disk.

At this point I'd also do as you outlined, which is to assemble a basic CompuPro hardware setup and to try booting from a real boot disk. At this point it isn't clear the software you are using as a workaround to boot the system is the problem, since the Disk 1 itself may be at fault. The Disk 1 has an analog data separator, which can fall out of adjustment resulting in read errors. The Disk 1A was a significant improvement on the Disk 1 for many reasons, one of which is that it had a digital data separator. Although I'm sure many people have used the popular Disk 1 without problems, the Disk 1 that I purchased brand new gave me constant grief (ie frequent read errors) from the first day I used it. It was particularly bad at reading/writing DD disks that had been formatted on another system. I was happy when I finally replaced it with a Disk 1A years later.

 

[RichCini]

No problem. I actually gave a presentation on this process at a VCF meeting about 5 years ago. I will test the card again in the exact same setup -- I tested it a few months ago but as you point out, it could be peculiar to that card too.

Rich