Troubleshooting a failed TI Floppy Controller

[Cory]

I'm looking for suggestions, advice, and mentoring as I try and diagnose and ultimately repair a failed TI floppy controller. Initially it was my only card and was therefore somewhat desperate for a repair, however, a generous TI community member was kind enough to loan me one of theirs so I could be back in operation. I'm now turning this into a learning exercise and would appreciate your thoughtful advice and patience as I ask some potentially revealing and dumb questions.

 

Symptom:

The FDC is no longer recognized by the TI.

Disk Manager 2 reports disk error 0.

Setting CRU bit 0 to 1 via Mini-Memory turns on the LED, however, when I inspect the memory at >4000 it shows 00.

 

Troubleshooting so far:

 

Visual Inspection: Burn marks around the 12V regulator. Everything else looks good - to my eye.

Tracing - I've tested continuity throughout the board at all ground points and voltage points. I've traced address lines etc.. So far, I have not located an issue.

Voltage Test - I've tested voltage at Vcc for each chip and it that appears to be good - to my knowledge.

Resistor and Caps - I've desoldered one lead from each Resistor and Cap and have tested these. I'm not certain about the tolerance for the .01uf caps on the board. Here's the readings for the caps around the buffers:

 

.01uf CAP measurements: 10.57nf, 10.56nf, 9.20nf, 10.38nf, 11.36nf, 11.27nf, 9.52nf, 8.60nf, 11.05nf, 8.99nf, 9.45nf, 9.33nf

 

I'm using an AMPROBE digital meter with the following ratings for capacitance tests:

 

Range Accuracy Resolution

40nF +/- 3%+10 10pF

400nF +/- 2.5%+5 100pF

4uF +/- 2.5%+5 1nF

40uF +/- 2.5%+5 10nF

400uF +/- 5%+10 100nF

4000uF +/- 20%+20 1uF

 

I'm not clear what the +5, +10, +20 number means in the Accuracy column. Can someone help educate me. I've been reading the manual inside and out and don't get it.

 

So that's where I'm at.

 

Any thoughts?

Edited May 11, 2011 by Cory

[matthew180]

Just a follow up to the PM message I sent, I would suggest:

 

1. Testing the ROMs to make sure they contain valid data. Does anyone have the binary dumps online anywhere?

 

2. Desoldering the three 74LS244 buffers in turn, add a socket, and swap in a known good 244. Test each one and note any differences. Alternatively you can bench test the 244 (my EPROM programmer will do this for me, which is a nice feature and one of the reasons I'm still using the ancient thing.)

 

3. Do the same as #2 but replace the 74LS245 right next to the three 244 chips.

 

Also, take your time with the desoldering, be gentle, and clean the heck out of the pins before starting.

 

Alternatively, if you get a logic analyzer, you can drop the "swap troubleshooting" method and start testing the data path. Start with the logic analyzer on the 244 buffers and make sure valid data is passing through them, along with the 245.

[Cory]

I'll move forward on your suggestions and post an update once complete. I'm expecting my EPROM programmer and the Logic Analyzer to arrive in the next day or so. In the meantime, I'll replace the regulators, electrolytic caps, and the 244s and 255.

 

Do the .01uF cap measurements I posted in my last post seem within reason?

[matthew180]

The caps are probably fine since they are are just being used as small batteries to supply fast response on-demand power to a chip when needed (i.e. during fast switching.) Caps to get old though, so replacing them won't hurt anything, but I don't think it will fix the problem.