php1800 sidecar disk controller no worky

[Ed in SoDak]

Hokay! Had a spare moment today so I tried the sidecar/PEB combo. First off, I withdrew the PEB controller just enough to disconnect it. Daisy-chained the PEB off the sidecar, I powered the sidecar, PEB and console in various orders. If the PEB was powered, the console showed only a blue screen, no title screen, locked before booting. Just turning off the PEB "freed" the console to proceed to the title screen.

 

Leaving the PEB off till I had the title screen, then powering it, the console was again locked up.

 

Next I tried also unseating the PEB's RS232 card. That worked! and I was able to boot up and show the 32k in the PEB with a SIZE command.

 

So, on to try Mizapf's XB test code. It worked and I got

 

0,131,224,0 as the result.

 

I see these are different numbers than what the PEB's TI controller produced (170,2,0,0) in my earlier test above. But not zeros! Does this tell you disk gurus anything? Does this absolve the ROMS, perhaps? And still leaving the 244/245 as suspects?

 

I also tried activating DSK1 from XB and this time it produced error 00 instead of the 50 I was getting before. I've had the boards out and flipping them over for inspection. Don't think I bodged anything, but that's always a possibility.

-Ed

Edited April 3, 2018 by Ed in SoDak

[marc.hull]

You cannot diagnose that issue with software.

[Stuart]

Error codes first. I believe XB will give you error code 00, and plain TI BASIC will give you error code 50. So if you were previously testing with TI BASIC, there's no problem with what you're seeing.

 

Try the test using the MiniMem cartridge, and see if you get the same results. In hex, what you're seeing through XB is >00, >83, >E0, >00, and >83E0 happens to be the first two bytes of the console ROM ...

[+mizapf]

@Ed: Would you mind to type in this program:

 

 

10 DATA 4,224,131,196,2,12,17,0,29,0,2,0,64,0,2,1,208,0,2,2,1,0,204,112,6,66,22,253,30,0,4,91
20 CALL INIT
30 FOR I=0 TO 31
35 READ A
40 CALL LOAD(-16384+I,A)
50 NEXT I
60 CALL LOAD(-31804,192,0)
100 HEX$="0123456789ABCDEF"
110 FOR I=-12288 TO -12032 STEP 6
120 CH$=""
130 FOR J=0 TO 5
140 CALL PEEK(I+J,A)
150 HEX1=INT(A/16)
160 HEX2=A-16*HEX1
170 PRINT SEG$(HEX$,HEX1+1,1);SEG$(HEX$,HEX2+1,1);" ";
180 IF (A>31) AND (A<128) THEN CH$=CH$&CHR$(A) ELSE CH$=CH$&"."
190 NEXT J
200 PRINT "   ";CH$
210 NEXT I

 

Just hacked in quickly (getting hungry, dinner waits to be prepared... ), a small hexdumper.

[Ed in SoDak]

@ Marc My hopes were to show signs of life in the sidecar ROMs, not sure if these software experiments will disclose anything or not, but it's non-invasive at least, so it can't hurt to try.

 

@ Stuart, Thanks for clearing up my error code confusion. It's been awhile since I've been playing on the iron and I've forgotten a lot!

 

@ Mizapf, Thanks for your efforts too! My supper's on the fry right now too, so I'll try your test and Stuart's MM test later tonight.

-Ed

Edited April 4, 2018 by Ed in SoDak

[marc.hull]

True but it tells you nothing in this situation. Since there is no front panel LED you cannot determine if the card ROM is even being activated.

 

I'll bow out now so as not to contribute to the noise. Good luck.

[Ed in SoDak]

Marc, that's a valid point. I'd mentioned the absent LED, and hoped for a disk LED instead, but obviously the sidecar ain't gettin' that far. Funny TI didn't provide one. What would it take, a dime's worth of parts? I mean, the thing's built like a tank and couldn't have been easy nor cheap to produce with those tri-level boards.

 

It wasn't a total waste of time trying, since I learned a bit about using MiniMem and XB to poke around. Probably the first time I used MiniMem for this purpose.

 

Here's an interesting tidbit. At first I guessed the PEB's controller card had to be extracted, then found it was the RS232 the sidecar was conflicting with. I was able to reinsert the contoller card and still boot with the sidecar controller attached as well. This made me wonder if the sidcar's CRU was actually 1100 or something else. But changing CRU using Minimem made no difference that I could tell. And the PEB controller was not recognized until the sidecar was removed from the chain. Merely turning it off didn't do it.

 

On to the tests! Using both Stuart's MiniMem method and Mizapf's XB program gave all zeros. Just before the XB program ended, it displayed EE 00.

 

Repeating both tests with sidecar out and PEB controller in produced the expected DSR codes. With the MiniMem, the PEB's controller card lit up and stayed on as soon as I entered the 1. I was able to turn it back off with 0.

 

With that, it's time to start pulling chips while I wait for the replacements to arrive. Or try a logic probe first as another learning exercise. I'll get more out of this sidecar as a learning test bed than I'd expect to have a use for it if I actually fix it!

-Ed

[+mizapf]

One final point to rule out: Is the controller located on CRU >1100?

[+mizapf]

@Ed: Suggestion: Locate the value 17 in the DATA line; change it to 16, then iteratively 18, 19, ... 31 until you get some reasonable values.

 

If you are interested what the DATA lines do: This is a machine language program that first clears the interrupt hook (be called only once), then sets the CRU base to >1100 (hence the 17,0), sets bit 0 to one, copies the contents from >4000 - >40FF to >D000 - >D0FF, clears bit 0, and returns.

 

Watch out for readable text in the right column. You should find DSK1 at some time.

[Ed in SoDak]

Thanks as ever, Mizapf! I'd read in another controller thread (there seem to be several right now) about changing the 17 to a 16, etc. I had tried that a bit on my previous testing but not with the XB version last night. But last night, I did try changing the CRU using MiniMem from 1000 to 1500. And all were zeros. I thought I might come across the second controller when both were mounted or the 32k. I probably didn't search high enough.

 

With the working PEB on and sidecar removed, I did use the XB test to see the DSK... and other info as I had seen in my reading of these threads. So my guess was the sidecar's problem is blocking something. But yet the 32k is recognized, so that theory has a hole in it. And the two controllers not conflicting like the RS232 and sidecar did is of note. To this layperson, anyway!

 

And no, I don't know what CRU the sidecar is supposed to reside in. Which was another reason for trying other CRUs. And both controllers at the same time. I'll carry on awhile and maybe learn more.

-Ed

[Ed in SoDak]

And on Stuart's link to the MiniMem code is great! Right above that on the page is his logic probe tests of a TI controller set up as a "sidecar." It's starting to sink in on how to go about that process. Thanks, Stuart!

-Ed

[+mizapf]

@Ed: one more version, slightly updated:

 

 

100 DATA 4,224,131,196,2,12,17,0,29,0,2,0,64,0,2,1,113,0,2,2,1,0,204,112,6,66,22,253,30,0,4,91
110 FOR I=0 TO 31
120 READ A
130 CALL LOAD(28672+I,A)
140 NEXT I
150 CALL LOAD(-31804,112,0)
160 HEX$="0123456789ABCDEF"
170 FOR I=28928 TO 29183 STEP 6
180 CH$=""
190 FOR J=0 TO 5
200 CALL PEEK(I+J,A)
210 HEX1=INT(A/16)
220 HEX2=A-16*HEX1
230 PRINT SEG$(HEX$,HEX1+1,1);SEG$(HEX$,HEX2+1,1);" ";
240 IF (A<32) THEN 280
250 IF (A>127) THEN 280
260 CH$=CH$&CHR$(A)
270 GOTO 290
280 CH$=CH$&"."
290 NEXT J
300 PRINT "   ";CH$
310 NEXT I

 

This code runs in TI BASIC with plugged-in MiniMemory, without memory expansion. As for the CRU base, the same applies as above: Try different values from 16 to 31 for the 7th value (17). The contents of the MiniMemory from >7000 to >71FF will be changed.

[Ed in SoDak]

That'll be great and save me from switching back and forth from sidecar to sidecar+PEB32k to just PEB and around again. Usually I also forgot to turn something back on, which caused me a few retries along the way, when I didn't realize my oversight till after I'd switched it to the other method.

 

Really glad I have such long cables for the drives and my firehose extender. My PEB is currently on the floor on its side, like a tower pc. So popping the lid and sliding a card or two out or in and getting at cables is simplicity, albeit tedious after a dozen iterations or so. It was tempting to drag out a second console and set one up for MiniMem and the other for XB/PEB. I may yet, though it'll be "cozy" in the man cave, surrounded by TI and Mac, with hardly room left for my beer and TV remote.

 

I did hook up cassette, so I can at least spare retyping the program each try.

-Ed

[Ed in SoDak]

All zeros from 16-31. The ones I let run to completion ended with 51 98 "Q"

-Ed

[+mizapf]

The last two bytes are not from the disk controller, since I copied 256 bytes, but the loop increments by 6, so the last iteration is 252 ... 257. These two bytes are from the previous MiniMemory contents at that location.

 

OK then, bad or good news, as you take it. The DSR cannot be turned on, obviously. Continue from there.

[Ed in SoDak]

I'm calling it good, since I can move on from dissecting my setup to dismembering the sidecar - gently, of course! Your programs will help verify if the repair worked.

-Ed

[Ed in SoDak]

I breadboarded three different logic probes. A two-transistor, a three-transistor and an IC version using a 4001 that claims to show a tri-state output. Both of the transistor probes lit the "0" logic LED constantly till the probe was connected to a high signal. I was able to modify both to show tri-state of low, high and infinite (open-circuit). I figured that would be the most useful troubleshooting the sidecar.

 

The IC circuit hasn't worked yet, but I'm just using chips from my "recycle bin" so they may or may not function or I have a wiring error. Date codes on most of these are from 1974, which makes reading the chip number confusing at times! Meanwhile I found a basic tri-state probe kit on Digi-Key (looks like it's an Elenco unit) for under 10 bucks and small shipping.

 

Tonight while browsing another computer forum, I noticed an ebay link to this little gadget. For about $30 it'll test a good range of 74xxx and 4xxx chips. It looks like a good tool to have if I'm going to continue messing around with learning about ICs and can check many of those in my stash. As usual, there's several sellers offering a couple variations. I have a similar gadget that checks all sorts of non-IC through-hole parts. While not perfect, it's pretty useful to both ID and test parts.

 

https://www.ebay.com/itm/IC-Tester-74-40-45-Series-lC-Logic-Gate-Tester-Digital-Detection-IC-Series-Meter/182642756392

 

Looks pretty cool! But, DAMNIT, it won't test the 244 nor 245!!

-Ed

Edited April 13, 2018 by Ed in SoDak

[+Ksarul]

One note, a lot of the EPROM burners we use will also test logic chips. I use my TL866CS to do that all the time. . .and it will test 244s and 245s.

[Ed in SoDak]

Small update. I tried my own DIY probes with inconclusive results. It seems I don't know what the heck I'm doing in the circuit or with my probes, or if those even work properly.

 

My usual method is to get a "known good" state to compare with, so I made that order with Digikey and it arrived on Thursday. I bought a couple 244 and 245 and that $9.30 logic probe plus a few other trinkets. I'm impressed with the entire order. Ordered online with paypal on the 7th, it's here, deep in flyover country on the 10th. Shipping was $8 and change. I could've gone cheaper with First Class for less than half that cost, but I opted to pay more to help protect the circuit board.

 

The packing can only be described as extreme for a few parts and a slim circuit board. Makes me wonder how they afford it, but their higher up-front packing cost no doubt earns them plenty of goodwill. Meanwhile, Amazon will use a huge box with one item bouncing around inside with almost nil for padding. My guess is their slave labor shippers are paid an extra pittance by pallets filled, and what better way to burn space or save a few seconds to ram another one out the door?

 

If that all sounds like a glowing review of Digikey, well, it is!

 

Here's the probe: https://www.digikey.com/product-detail/en/twin-industries/TW-DIY-5024/438-1077-ND/735276

 

The pic shows red and black clips. Mine has a single double-ended green that will be half the length when cut. No probe tip, I'll scavenge an old meter probe. Nit picks. Who doesn't have more clips or probe wire? It's certain to be better than my breadboard mockups!

 

Looks like I'm finally trying to learn ICs and logic. It's mostly why I bought that RS/Elenco trainer some two decades ago. Instead of mere curiousity, this time it's with a defined goal in mind to start me off!

-Ed