Page 6 of 9

Re: MTX 512 (debugging thread)

Posted: 09 Jul 2020 16:17
by Dave
It's obviously been a year since you read Mark's post a couple of entries above this post.

Mark gave you some tips there on the -5V question, are you sure that the -V (-12V) line is OK?

The video RAM is the only user of -5VDC

regards
Dave

Re: MTX 512 (debugging thread)

Posted: 09 Jul 2020 17:31
by GozdniJezek
Sorry, I forgot to mentioned, that the ZD3 is the second diode that was replaced and R60 is also reading 1kΩ. Will check the voltage on D16 or D18 anode leads and for any other shorts. I'm sorry for asking stupid questions, should have referred to the previous post. Thank you for your input. :)

I will report on the results. :)

Re: MTX 512 (debugging thread)

Posted: 09 Jul 2020 17:44
by Dave
Hi,

the questions aren't stupid, it just helps to know that you have followed earlier advice, including checking for shorts at the locations mentioned

regards
Dave

Re: MTX 512 (debugging thread)

Posted: 28 Jul 2020 15:58
by GozdniJezek
Hello, as promised an update. I'm not so sure that the -5V issue was the real culprit, all of the voltages check out now. We've replaced all of the 4116 video ram with modified 4164 which should not use the -5v line or 12V (we changed the ZD3 diode (second time) with one that now gives, -5.3V without rams, I hope the voltage will lower under the load, but according to the datasheet its still ok) (during the limited testing the voltage did not collapse anymore, but need to do more testing with 4116). The results can be seen in the picture although it might be related to the 15ns ram (is this fast enough?). But the most pressing concern is that the image disappears after a second or 2. I suspect that the video board might have an issue, does this sound reasonable? Any additional pointers would be greatly appreciated. This is turning into a rabbit hole, and I'm starting to doubt everything. I might be missing something obvious, so any pointers are most welcome.
videoRams.jpg
videoRams.jpg (174.6 KiB) Viewed 17273 times
videoScrambled.jpg
videoScrambled.jpg (110.84 KiB) Viewed 17273 times
videoGone.jpg
videoGone.jpg (131.46 KiB) Viewed 17273 times
4164 rams.jpg
4164 rams.jpg (189.49 KiB) Viewed 17273 times
tl;dr:
So my rationale is:
-(image 1)the video rams are not working properly, maybe it's related to the 15ns timing, maybe its something else, need to try with 4116 ones (i did not had them at friends house).
-(image 2) scrambled and dis spearing video, my noob opinion is that this might be caused by a bad video board? I intend to try with RF output, to see if anything changes, I was also thinking on recapping the video board.

Any suggestions and observations are most welcome.

Re: MTX 512 (debugging thread)

Posted: 28 Jul 2020 19:02
by Martin A
Just for reference, those rams are 150ns, not 15. Back in the 80s, 15ns ram was the sort of stuff you'd find in a Cray mainframe. :o

When the MTX was built, 200ns for the rams would have been normal, so 150ns is faster than required definitely not too slow.

Re: MTX 512 (debugging thread)

Posted: 28 Jul 2020 20:28
by Dave
As you have changed out all of the VRAMs for sockets, there’s always the chance that there’s an issue there. If you haven’t already done so, I suggest that you remove the VRAMs and use a meter to verify connectivity between them and the VDP ss as per the drawing.

I.e., check for continuity as per the schematic and there are no shorts that shouldn’t be there

Regards
Dave

Re: MTX 512 (debugging thread)

Posted: 29 Jul 2020 13:05
by 1024MAK
Apart from the data in (D) and data out (Q) pins (2 & 14), ensure by continuity testing (or by using the 200 ohm resistance range) using a multimeter that all like numbered pins on each video DRAM socket are connected to all the other video DRAM sockets. So pin 3 on one video DRAM socket should have continuity to pin 3 on the other seven video DRAM sockets.

Also test the relevant data pins to the video chip. As well as all the address and control lines.

The RF/UHF signal is derived from the composite video signal, so although it is a good idea to try this, don’t be surprised if you get similar results.

In your pictures, it looks like a loss of synchronisation signals that is causing the picture to break up and disappear.

Mark

Re: MTX 512 (debugging thread)

Posted: 12 Aug 2020 18:41
by GozdniJezek
Hello,
so the progress and findings:
We've checked the all of the traces on all of the VRAMs sockets, from the VDP leg to the appropriate 4116 leg, all checked out and no shorts were found.
ZD3 still gives -5.3V, but I guess this is within specifications (or would you suggest swapping it out with more in speck one)?
In the mean time, I found lost 4116 DRAM chips and installed them.
The picture was as it should be for a second or two than it disappeared (went out of sync) and the result was black screen (the same as with modified 4164).
We measured the XTAL2 with the oscilloscope (probes on each leg of the crystal and on the VDP legs).
IMG_20200809_113918.jpg
IMG_20200809_113918.jpg (173.11 KiB) Viewed 17093 times
We also measured the XTAL1 and the result was:
IMG_20200809_114147.jpg
IMG_20200809_114147.jpg (151.97 KiB) Viewed 17093 times
The scope says 138KHz although it should be 4MHz, but my electronic engineer friend said it could be something with the capacitance of the scope probes or he doesn't have the proper probes for the job (honestly I did not understand what he was saying fully).

My reasoning now is, that since the XTAL2 does not give a nice oscillating signal as XTAL1 (although the frequency looks wrong) the culprit for missing image is bad/failing XTAL2 crystal. The image also flickered/changed when the probe was removed from the crystal2 leg.

The service manual also states that a possible cause for a black screen is a bad XTAL2.

Do my findings sound reasonable? Do you see any deep flaws in my logic? Would you suggest checking anything else?

Also would this crystal be a suitable replacement? I just looked for the 10.6875MHz one, with similar appearances, do I need to be careful about anything else or this is ok?
http://www.buy-transistors.com/categori ... 6875m.html

As always thank you for your help.

Re: MTX 512 (debugging thread)

Posted: 12 Aug 2020 21:40
by stephen_usher
I must admit that it doesn't look right from memory when I was testing my machine.

If it's the crystal then at least it's something cheap and relatively easily replaced.

Re: MTX 512 (debugging thread)

Posted: 13 Aug 2020 11:52
by Martin A
I probed the leg of the VDP crystal (using x10 probes and pin 16 if J7 for ground) and got the following waveform.
VDP crystal
VDP crystal
Ads00018.jpg (32.44 KiB) Viewed 17081 times
The scope didn't get a lock on frequency, but from the scale, you can see it's OK. So your 0v flat line does look very suspect!

Interesting asside the 10.6875mhz crystal on the MTX motherboard is a shade slow compared to the 10.7386mhz recomented on page 5-4 of the 1982 VDP databook. I couldn't see a reccomended crystal capacitance.

The 4mhz clock is probably best tested at pins 8,10 and 12 of the 74xx04 at position 9D (and with the scope on a somewhat shorter time per division 500us is kinda slow for a 250ns period clock). This trace happens to be from pin 10 (of the 74S04 fitted) at 100ns per division.
9D Pin 10
9D Pin 10
Ads00019.jpg (32.68 KiB) Viewed 17081 times