39 replies to this topic

[OLD CS1]

marc.hull, on Sun Nov 14, 2010 10:45 AM, said:

These combo drives have been debated in the past and the conventional wisdom is that they do not work, something about them being the 1.2 meg format I believe. I don't know if anyone has tried them and reported on it recently though. Hopefully you can get some definitive data for us.

To put the 3.5" drive issue to rest. Any TI controller can use 3.5" drives but the largest format the TI can utilize is DSDD. HD drives work provided the Floppy is a DD one or one of the workarounds (tape, glue jumpering etc) is done.

I will be giving this a try sometime over the next few weeks. For the time being, I am very pleased with my latest purchase, thank you.

At this point, I have a working DSDD floppy controller and drives. Also, I seem to have fixed the data problem I was having with my WHT 'E' SCSI card and the drive; termination can be a SCSI killer! The termination packs on my Quantum 75MB drive apparently do not work properly. I found a cable with built-in termination and all appears to be well, now. This does not, however, address the SCS1 (device 0) problem I mentioned before, and I will work on getting it off to Richard Bell at some point (probably after the holidays so as not to impose.)

But as of now, my TI system is as complete as I need it to be: 32K, serial/parallel, SCSI, and DSDD floppy. Maybe later I will consider a RAM disk, but for now I think I am done building on this box.

Edited by OLD CS1, Thu Nov 18, 2010 12:01 AM.

[OLD CS1]

So I played a little bit today with the combo drive I could find, a TEAC FD-505 Series 102. The HD 5.25" drive, as many people probably already know, does not work as the rotation speed in HD mode is 360 rpm. Interestingly enough, this drive apparently has dual-speed spindle, capable of running at 300 rpm, as well. It is apparently triggered by pin 2 which, according to the manual, is in a high state for HD mode. I have not pulled out the multimeter yet to determine if the controller sets this pin high, or if it is held high by the drive to be pulled low by the controller. (If I am correct, it appears the actual control line is not connected to the interface so bridging that pad gap may do the trick.)

Now, the 3.5" side of the unit works fine. It formats 40 tracks DSDD just fine. I decided to play a little bit as I noticed the head only traveling half-way across the surface. I formatted 80 tracks DSDD and the format completed as expected, but during verification the head only traveled half-way before starting to step backwards at sector 720.

I am not sure why that happens. It could be a hard-coded setting in the controller which expects a 40-track device when calculating sector locations. Even so, I find it interesting that it would allow the drive for format 80 tracks. I will dig through the BwG manual to see if I find any hints.

I dislike the idea of only using half the disk's capacity, but I more like the idea of having 3.5" drive available. I will work on the 5.25" 360/300rpm bit and post my discoveries. I have probably exhausted my hobby time for the weekend.

[OLD CS1]

Last night I found the "I" strap which sets dual-speed mode, 360rpm for HD and 300rpm for DD/SD. It works, but the controller does not pull pin 2, the density select, low, so the drive runs in HD mode. I shunted pin 2 and 1, thereby grounding the density select line; in retrospect, I could have also simply set the "LG" strap which changes the density select from active low to active high.

Irrespective of all that work, the system still reports a "DISK ERROR 16" when trying to use the 5.25" drive for anything, including formatting. I have not done much research on this, but I speculate that it is possible one of the lines from the drive may not match what the controller expects, such as *READY, etc. I will dig some more into that later.

So, for now, I have come the the same conclusion as others: the HD 5.25" part of the dual floppy does not work. I plan to continue working on this over the next couple of weeks. The way I look at it, it works in DD mode for PCs, there should be away to get it to work the same for the WD1773 controller -- though newer FDC controller may "speak" differently to HD drives while still allowing for DD disks. (There are other straps on the drive which do not show up in the data sheet, so there may still be hope for this project.)


EDIT: I have found what might be the issue: pin 2 on the drive is the density select, but the original floppy interface documents for the Shugart interface, which I assume the BwG would follow, has pin 2 as "DCD," Disk Change Detect. On the newer interface diagram I see pin 34 as ReaDY/DiskChangeDetect (dear, God, this reminds me of the Amiga floppy interface issue.) Anyone know anything about these signals in terms of HD versus DD drives?

Edited by OLD CS1, Sun Nov 21, 2010 9:12 PM.

[InsaneMultitasker]

OLD CS1, on Fri Nov 12, 2010 7:25 PM, said:

I am not sure what the exact problem is. I formatted and initialized a Quantum 75MB drive I have, and that completed fine, but I could neither save nor write with the drive until I moved it to SCS2. Ernest mentioned as much in his auction, and Richard Bell has agreed to try to work on it, I just need to ship it to him.

Silly question time: what SCSI ID is your card set to? Is it by any chance configured as ID 0? I ask because the data bus is used during arbitration and ID selection, which seems to infer that a ID selection problem would also manifest itself in communication errors between the card and any device, not just ID0.

If your Card ID DIP SW2 is set to an ID other than 0, use a program to verify the card's ID s not incorrectly being read as ID0 (I think Fred's DU2K has this option). U9, to the left of SW2, is an LS251 responsible for 'reading' the DIP setting. This chip, the DIP switch, or pullup resistor might be suspect. My REV-E card has DMA 4-25 (U13) and SCSI Address Rev 2.0 installed.

I'd be willing to send you my 'E' card at some point if you want to do a comparison test. Just so long as I get it back I use it for DSR compatibility testing even though most cards have been upgraded.

[OLD CS1]

InsaneMultitasker, on Sun Nov 21, 2010 5:33 PM, said:

OLD CS1, on Fri Nov 12, 2010 7:25 PM, said:

I am not sure what the exact problem is. I formatted and initialized a Quantum 75MB drive I have, and that completed fine, but I could neither save nor write with the drive until I moved it to SCS2. Ernest mentioned as much in his auction, and Richard Bell has agreed to try to work on it, I just need to ship it to him.

Silly question time: what SCSI ID is your card set to? Is it by any chance configured as ID 0? I ask because the data bus is used during arbitration and ID selection, which seems to infer that a ID selection problem would also manifest itself in communication errors between the card and any device, not just ID0.

The problem with reading/writing the drive after format turned out to be a termination issue. The terminator packs on the hard drive seem to be bad. I used a cable with a built in terminator and all is well. I have the DIP switches set for device 7, but DU2K does not show me any device in the chain other than the hard drive.

The SCS1 issue still appears to be present, though.

Edited by OLD CS1, Sun Nov 21, 2010 9:36 PM.

[snume]

OLD CS1, on Sun Nov 21, 2010 2:33 PM, said:

Last night I found the "I" strap which sets dual-speed mode, 360rpm for HD and 300rpm for DD/SD. It works, but the controller does not pull pin 2, the density select, low, so the drive runs in HD mode. I shunted pin 2 and 1, thereby grounding the density select line; in retrospect, I could have also simply set the "LG" strap which changes the density select from active low to active high.

Irrespective of all that work, the system still reports a "DISK ERROR 16" when trying to use the 5.25" drive for anything, including formatting. I have not done much research on this, but I speculate that it is possible one of the lines from the drive may not match what the controller expects, such as *READY, etc. I will dig some more into that later.

So, for now, I have come the the same conclusion as others: the HD 5.25" part of the dual floppy does not work. I plan to continue working on this over the next couple of weeks. The way I look at it, it works in DD mode for PCs, there should be away to get it to work the same for the WD1773 controller -- though newer FDC controller may "speak" differently to HD drives while still allowing for DD disks. (There are other straps on the drive which do not show up in the data sheet, so there may still be hope for this project.)


EDIT: I have found what might be the issue: pin 2 on the drive is the density select, but the original floppy interface documents for the Shugart interface, which I assume the BwG would follow, has pin 2 as "DCD," Disk Change Detect. On the newer interface diagram I see pin 34 as ReaDY/DiskChangeDetect (dear, God, this reminds me of the Amiga floppy interface issue.) Anyone know anything about these signals in terms of HD versus DD drives?

Yeah, I was looking at a way to use a TEAC FD-55GFR with the stock controller and it is supposed to be able to work at 300 RPM as well. There are a lot of jumpers and someone even put masking tape over the first 4 pins (first 2 on each side of card edge) but that info wasn't for the TI-99/4A and I suspect there is more to it as you are discovering here as well. If you come up with anything, I will be attempting to adapt my drives in the same way. Keep it up, you're work is very inspiring.

Steve

[marc.hull]

OLD CS1, on Sun Nov 21, 2010 2:33 PM, said:

Last night I found the "I" strap which sets dual-speed mode, 360rpm for HD and 300rpm for DD/SD. It works, but the controller does not pull pin 2, the density select, low, so the drive runs in HD mode. I shunted pin 2 and 1, thereby grounding the density select line; in retrospect, I could have also simply set the "LG" strap which changes the density select from active low to active high.

Irrespective of all that work, the system still reports a "DISK ERROR 16" when trying to use the 5.25" drive for anything, including formatting. I have not done much research on this, but I speculate that it is possible one of the lines from the drive may not match what the controller expects, such as *READY, etc. I will dig some more into that later.

So, for now, I have come the the same conclusion as others: the HD 5.25" part of the dual floppy does not work. I plan to continue working on this over the next couple of weeks. The way I look at it, it works in DD mode for PCs, there should be away to get it to work the same for the WD1773 controller -- though newer FDC controller may "speak" differently to HD drives while still allowing for DD disks. (There are other straps on the drive which do not show up in the data sheet, so there may still be hope for this project.)


EDIT: I have found what might be the issue: pin 2 on the drive is the density select, but the original floppy interface documents for the Shugart interface, which I assume the BwG would follow, has pin 2 as "DCD," Disk Change Detect. On the newer interface diagram I see pin 34 as ReaDY/DiskChangeDetect (dear, God, this reminds me of the Amiga floppy interface issue.) Anyone know anything about these signals in terms of HD versus DD drives?

I would suspect that the drive is also wanting to operate with 80 tracks per side IE 720K. Is there a way to make this drive double step like a 3.5" when reading and writing? Remember that the BwG controller can only support 40 tracks per side for a max of 360K.

[OLD CS1]

snume, on Sun Nov 21, 2010 11:31 PM, said:

Yeah, I was looking at a way to use a TEAC FD-55GFR with the stock controller and it is supposed to be able to work at 300 RPM as well. There are a lot of jumpers and someone even put masking tape over the first 4 pins (first 2 on each side of card edge) but that info wasn't for the TI-99/4A and I suspect there is more to it as you are discovering here as well. If you come up with anything, I will be attempting to adapt my drives in the same way. Keep it up, you're work is very inspiring.

Thanks. I will keep this thread updated as I make (non-)progress on this. The FD-55GFR is the 5.25" part of the FD-505. Mine will have a different board layout, but it should have the same straps for the most part. For instance, my main board is missing the "IU" strap. There are several other strap points on the board which are undocumented, like S1-S5, and E1 and E2 (IIRC, I am not looking at it right now.)

marc.hull, on Mon Nov 22, 2010 7:41 AM, said:

I would suspect that the drive is also wanting to operate with 80 tracks per side IE 720K. Is there a way to make this drive double step like a 3.5" when reading and writing? Remember that the BwG controller can only support 40 tracks per side for a max of 360K.

That is a worthwhile thought to be certain. I do not think the 3.5" drive in this combo double-steps, as I noted earlier that the head only travels half-way across the disk when formatting. It is too bad the BwG does not support 80 track devices.

And in regards to the 80/40 track bit, I have read in more than one place, including some manufacturer docs, that you should use an 80 track device to write to 40 track media. This makes sense as 40 track media written by a 40 track device will have tracks wider than the 80 track head, thus causing all sorts of issues once you take the media back to a 40 track device. This alone may make this project entirely moot as you may wind up with unsharable 5.25" media even if one can get these units to work.

This problem does not seem to affect 3.5" media, as the only difference between 720k and 1.44MB 3.5" disks is the data rate and amount of data contained within a track. 3.5" DD drives are difficult if not impossible to find in the PC world, whereas HD drives are plentiful. A big benefit of the 3.5" drive is the hard detection of media density, a second microswitch which finds the HD hole. In Amiga HD drives this causes the spindle to spin at 150rpm rather than 300rpm so that the data rate to the floppy controller remains the same for both densities; obviously this is not the case in the PC realm as the data rates for DD and HD are 250bps and 500bps, respectively.

If I had more skill available, I might take on a project of working out an 80 track version of the BwG DSR. I see that Mainbyte has an article about an 80 track upgrade for the stock TI controller.

All considered, my goals are as follows in regards to the TEAC FD-505 (FD-55GFR and FD-05HF):

1) Read 40-track 5.25" DSDD (360k) media*
2) Read and write 40 track 3.5" DSDD (360k) media with head double-stepping**

* This may wind up being impossible as the FD-55GFR is listed as a dual-mode 1.2/1.0 device, so its ability to recognize 360k media is now in question. I need to slap together a PC to do some testing there.

** I need to see how the system works with other 3.5" drives. Can anyone confirm double-stepping in their 3.5" drives?

Ideally, I would like to read and write 80 track 3.5" DSDD (720k) media, but this would require an upgraded DSR for whichever controller is to be used.

As I am creeping closer to final exams, I will probably have to shelve this for a few weeks and get back to it. So do not expect too many updates

BTW, these were difficult to come by, so I post them here for added posterity. Any PDFs I found I ran through OCR.

3fd0020a.pdf TEAC FD235HF Floppy datasheet
5fd0050a.pdf TEAC FD55GFR Floppy datasheet
5fd0010a.pdf TEAC FD505 Combo floppy datasheet

wd1773_datasheet.pdf Western Digital 177x floppy controller IC datasheet
MitsubishiDrives.pdf Mitsubishi MF501, 503, 504 drive datasheet (derived from original RTF document)

Attached Files

•  3fd0020a.pdf   781.06K   6 downloads
•  5fd0050a.pdf   437.76K   7 downloads
•  5fd0010a.pdf   539.63K   5 downloads
•  wd1773_datasheet.pdf   928.58K   6 downloads
•  MitsubishiDrives.pdf   122.27K   5 downloads

[snume]

So, do you have documents that list the pinouts for both the Floppy Cable interface and the drive you are trying to use? Maybe a jumper cable can be made that flips the right wires around to have the drive be recognized (which I believe it isn't) and still lock it down to 300rpm.

[OLD CS1]

snume, on Mon Nov 22, 2010 2:29 PM, said:

So, do you have documents that list the pinouts for both the Floppy Cable interface and the drive you are trying to use? Maybe a jumper cable can be made that flips the right wires around to have the drive be recognized (which I believe it isn't) and still lock it down to 300rpm.

Not a FULL pinout, but enough to see that pin 35 on the Mitsubishi DSDD 5.25" drive is used for READY, while pin 34 on the TEAC FD505 is disk change for the 5.25" and 3.5" devices, then notes "Others: The same as the conventional FDDs." I am not convinced the READY line is the problem as the controller will operate the 3.5" drive just fine.

I have it running at 300rpm now by setting the "I" strap and the "LG" straps: "I" which, contrary to the documentation, sets dual-speed mode, and "LG" which inverts the density select on pin 2 from HD on low to HD on high. I was able to directly manipulate the spindle speed using pin 2.

I also set the "RY" strap which is supposed to set READY on pin 34. That did not make a difference.

[snume]

OLD CS1, on Mon Nov 22, 2010 5:05 PM, said:

Not a FULL pinout, but enough to see that pin 35 on the Mitsubishi DSDD 5.25" drive is used for READY, while pin 34 on the TEAC FD505 is disk change for the 5.25" and 3.5" devices, then notes "Others: The same as the conventional FDDs.".

Pin 35 a typo, right? The documentation on pinouts.re shows pin 2 and 34 as well as pins 10-16 all can be slightly different between the original Shugart and modern PC floppy interfaces. The test setup used on this page seems like it would be the best way for us to figure this out conclusively, if it is even possible. There is mention that a certain resistor (though which one is still up for debate) needs to be cut or removed but that seems scary.

[OLD CS1]

snume, on Tue Nov 23, 2010 10:14 AM, said:

OLD CS1, on Mon Nov 22, 2010 5:05 PM, said:

Not a FULL pinout, but enough to see that pin 35 on the Mitsubishi DSDD 5.25" drive is used for READY, while pin 34 on the TEAC FD505 is disk change for the 5.25" and 3.5" devices, then notes "Others: The same as the conventional FDDs.".

Pin 35 a typo, right? The documentation on pinouts.re shows pin 2 and 34 as well as pins 10-16 all can be slightly different between the original Shugart and modern PC floppy interfaces. The test setup used on this page seems like it would be the best way for us to figure this out conclusively, if it is even possible. There is mention that a certain resistor (though which one is still up for debate) needs to be cut or removed but that seems scary.

Yeah, type-o. I would like to have a floppy drive break-out box. May have to build one while I work on this project.

EDIT: That page has some links to some good information, thanks; I could have saved a good bit of time. Funny thing is, I had already visited there... I must have closed the tab by accident before browsing through it.

Edited by OLD CS1, Tue Nov 23, 2010 8:21 PM.

[InsaneMultitasker]

OLD CS1, on Sun Nov 21, 2010 9:35 PM, said:

InsaneMultitasker, on Sun Nov 21, 2010 5:33 PM, said:

OLD CS1, on Fri Nov 12, 2010 7:25 PM, said:

I am not sure what the exact problem is. I formatted and initialized a Quantum 75MB drive I have, and that completed fine, but I could neither save nor write with the drive until I moved it to SCS2. Ernest mentioned as much in his auction, and Richard Bell has agreed to try to work on it, I just need to ship it to him.

Silly question time: what SCSI ID is your card set to? Is it by any chance configured as ID 0? I ask because the data bus is used during arbitration and ID selection, which seems to infer that a ID selection problem would also manifest itself in communication errors between the card and any device, not just ID0.

The problem with reading/writing the drive after format turned out to be a termination issue. The terminator packs on the hard drive seem to be bad. I used a cable with a built in terminator and all is well. I have the DIP switches set for device 7, but DU2K does not show me any device in the chain other than the hard drive.

The SCS1 issue still appears to be present, though.

Interesting. SW2 DIPs 1,2,3 control the ID IIRC; if they are set opposite of what is expected, the card would respond to ID0 instead of ID7, and a conflict with SCS1 (device on ID0) would arise. I thought for sure Fred's program interrogated the bus and displayed all devices including the card. Might have to fire up my system and have a look-see tomorrow.

[ti99userclub]

Hi To all...

i have a question also is is not a SCSI problem...
with the normal controller for ti99 can be used a normal 3.5" floppy diskette at 720kb o i need only a Flppy drive HD 1.44Mb ?

Thankyou.

[humeur]

I have many scsi card broken, i search a schematic have you this ??

thank

Jean Louis