PDA

View Full Version : Shugart SA800's, Behavior?



atod
January 6th, 2011, 07:12 AM
Hi Folks,

I just powered on my Shugart SA800-1's today and connected them up
to my CCS 2422 FDC. This is the first time I have worked with 8" drives
so had some questions.

I notice the motors are constantly spinning when power to the drive is on.
This is the case even when the FDC cable is disconnected. Is this normal?

The manuals has a bunch of mods Select Drive without Loading Head or Enabling Stepper, Select Drive and Enable Stepper Without Loading Head, Load Head Without Selective Drive Or Enabling Stepper. I'm assuming I don't want any of these?

Thanks

glitch
January 6th, 2011, 07:41 AM
Yes, it's normal for the motors to always spin -- the AC connection is straight-through in the SA800.

Your needs for the jumperable select modes depends on the system you're using. I don't know about the CCS 2422 in particular, but I'd imagine you want head load and stepper enabled when the drive is selected.

atod
January 6th, 2011, 10:11 AM
Wow, I just looked again in the manual. AC and DC power to the drive! Thanks for clarifying this. Do you know when the 8" models were released to run off only DC? Also with a selectable motor enable? Maybe that wasn't until 5.25"....

Thanks

MikeS
January 6th, 2011, 10:22 AM
Wow, I just looked again in the manual. AC and DC power to the drive! Thanks for clarifying this. Do you know when the 8" models were released to run off only DC? Also with a selectable motor enable? Maybe that wasn't until 5.25"....

ThanksCan't tell you when, but later drives like the Tandon TM-848 only required 5V and 24V and did have motor control signals, although in practice they were often configured to run constantly anyway.

Conversely, some 5.25 drives were used the same way, motor running constantly and a head load solenoid activated on drive select.

Chuck(G)
January 6th, 2011, 10:26 AM
Yes, there are a few late-model drives, such as the Tandon 848 that use only DC. But the SA-800 is a venerable workhorse.

But why fret about motor enable? it delays the startup time of the drive. Note that 8" drives also have head-load solenoids, so while the disk is not being accessed, it's not causing a significant wear to the drive. The drive spindle motors themselves are synchronous, brushless designs and will probably last longer than you will.

glitch
January 6th, 2011, 11:47 AM
You could add a relay to switch the motor AC on and off, but as mentioned it's not necessary. I definitely prefer the AC drive 8" drives (less power supply required), and having the disk always spun up does make the drives very fast. I've been meaning to look into wiring the 5.25" drives that come stock with the Leading Edge Model D for motor always on, as they have head load solenoids.

Incidentally, if you'd rather have a DC-only drive, I have a working Tandon TM-848 I'd be happy to trade you. It needs the Track 0 sensor realigned, as I had to remove it to grease the head carriage drive screw, but other than that it's in great condition.

atod
January 6th, 2011, 04:23 PM
Thanks for the replies.

I'm sitting here next to the drives with CCS2422 monitor trying to read bytes from a sector. I see the correct drive select light come on and hear a click of the head, however I'm unable to read. It looks like (according to the monitor), the FD 179X-02 is returning back a "seek error" in the status register. I used isopropyl alc on the heads and that didn't help. I'm going to go through the manual and the CCS assembly code to see how this ties together. Hopefully, I can find the error.

Does anyone have any pointers?

Thanks

MikeS
January 6th, 2011, 04:42 PM
Sorry if it's obvious, but are you sure the disk is OK?

atod
January 6th, 2011, 04:52 PM
Could my problem be because the disk isn't formatted? I would suspect the FD ASIC handles detecting that during a read. As I'm loading everything into this system via serial hex, I don't have the ccsinit format tool. Looks like I'll have to write that.

Nick

glitch
January 6th, 2011, 04:59 PM
Yep, that would be your problem. I ran into that with my Cromemco 4FDC, which won't format floppies on its own.

Lou - N2MIY
January 6th, 2011, 05:02 PM
It's good that you cleaned the head, but there are two other good things to check in your SA800s. There is a little button of felt on the upper arm that is lowered by the head load solenoid (directly opposite the head, pressing on the other side of the floppy). Sometimes it falls off and needs to be replaced. If not missing, often it is filthy and needs to be replaced (I use a cork punch and a bit of self adhesive felt that craft stores sell for protecting wood furniture from the bottoms of trinkets).

Also, the sheet metal arm of the head load solenoid is supposed to have a rectangular piece of foam that should press down on the diskette jacket. Sometimes it is completely rotten and gone. I have replaced them with self adhesive weatherstrip foam.

The SA800 is a great drive otherwise, but the problems above I have had with my TM848s and RX01/2 also.

Lou

atod
January 8th, 2011, 01:49 AM
Yeah, I checked the CCS FDC Monitor code. Looks like it depends on sector seek for read and write. Well if the disk hasn't been formatted that operation will fail. I need to format by stepping motor to Track 0, then using the WRITE_TRACK command with a byte stream until the last track. I think I can use the CCSINIT.COM program from CCS to do this somehow.

Does anyone have any WD 179X assembly code that performs formatting?

Thanks

Chuck(G)
January 8th, 2011, 10:45 AM
Does anyone have any WD 179X assembly code that performs formatting?

There should be lots of code out there, but it's hard to know what you need, without specifics. Basically, you lay out a full track in memory with proper headers, using reserved codes as to where to put address marks and CRCs, padding the end generously to account for speed variations and issue a write track command.

Specifically, what code do you need?

Attached is a formatter written for a WD1781, but it's probably more complex than your need. I've got others in x86 code as well. Note that the attached formatter is DMA and interrupt-driven and performs a drive-speed test before formatting.

atod
January 27th, 2011, 12:38 AM
Wow, that's alot cleaner then the assembly I wrote/hacked up! Yup, I wrote the end padding routine to keep writing bytes until an interrupt.

I'll need to modify the IO ports in this. Not sure about the 8257 DMA controller don't see one in the FDC schematic. Otherwise, the only thing I'm wondering about is the "disk format constants." section. Also this code looks like it uses interleaving (factor of 3). I wasn't aware (at least from the CCS/WD docs) that interleaving was necessary.

I compared the gap lengths to that in the 2422B manual page 67 / A7 at :
http://www.bitsavers.org/pdf/ccs/2422/CCS_2422B_Floppy_Controller_May81.pdf

The lengths don't match that in the single or double density formats.

Thanks very much for sending this.