A comment regarding memory testing: Back in the day,... say around 1980ish - I was working with several others on a digital oscilloscope product - we needed a memory test and another engineer wrote a test that was essentially the same as the one you show. All tests passed, and we then later discovered that the VCC was not connected to the RAM! It passed the test because the data lines capacitance 'saved' the data on the wires. Therefore a better test writes random data to each location and then reads back the memory with the same random seed. In fact it may have been charging the RAM thru the data and address wires.
When you are stopped, I believe what you have ruled out is your only way forward…. Is there handshaking with the front panel display???? I know you believe that’s not the problem but perhaps it’s not acknowledging something causing the micro firmware to time out and reset. ????
OK :-(. A friend of mine and I were debugging a server CPU upgrade last night, we were SURE that the upgrade wasn’t causing the memory failures… We were SURE! But we got out the microscope to examine the Xeon sockets only to find bent pins on the socket on the mobo which destroys any usefulness of that server… It’s a painful outcome, but moved everything to a spare server and it’s up and running now…I hope the root cause on your DVM doesn’t end up being as bad…
I assume that by part 7, you will reveal the problem,... but still I am curious if your adding all the wires to measure the signals has caused the flashy lights to occur?
I wonder if a clue is that the unit was probably dropped? I had a similar nightmare trying to repair an Atari 1050 floppy disk drive which took a big hit (thanks DHL) during delivery. It was very intermittent. I wasted a lot of time testing chips - just like you. Eventually I spotted the problem. The large main power cap had a dry joint on one pin, either caused or exacerbated by the drop. A simple,re-flow fixed the problem. So I wonder as to which parts might be “jolt” sensitive. Is the micro clocked by a xtal?
I do not believe it as a mechanical shock problem, yes there was a xtal and that part of the uP worked fine. it seemed as though the uP was just a bit sick and did not reset correctly.
Seems bit odd that the reset is fired after 8 - 9 clock pulses while being loaded with count 16, could the down count be defect (only half of it is working, causing it to reset premature) ?
If your EPROM programmer is a TL866 II + you can write your own template to test chips that aren't already defined in the list , or you can use and existing profile with bigger capacity (maybe you'd have to wire your chip to match the pinout) and see if it errors out inside the addressable space of your chip or not. Did this when I had no other choice.
@@IMSAIGuy I use an existing one ( 8255 is a good starting point ) then rework it to suit my needs.It's a matter of creating a new logic device and setting each step in the testing sequence by setting input and output states according to the symbol definition table shown in the logic test window. Well used it only once and can't remember how I did it but it worked.Xgecu pro can be powerflul tool
At this point I'd pull out the big guns and read the manual of the IO chip, writing down the initialization commands and look for them using the LA to troubleshoot why is is not outputting data.
every part in a socket is suspect. A close visual inspection is in order. Use a can of freeze spray or heat directed gun. never liked socketed devices ... and connectors that are 40 years old ....
Why not read the content of both EEPROMs several times to get sure, their contents reads always the same? I had some EEPROMs in the past, which had weak behaviour. After replacing those with fresh burned binaries, the unit works fine again. Just my two cents.
YUP!! arduino testers!! good 1...the good, bad. & indifferent!! changed it's dipper! U realize TI has a chip 4 this? why ask why? bud dry!! a hotel 4 engineers? runs like a top, don't flop!
A comment regarding memory testing: Back in the day,... say around 1980ish - I was working with several others on a digital oscilloscope product - we needed a memory test and another engineer wrote a test that was essentially the same as the one you show.
All tests passed, and we then later discovered that the VCC was not connected to the RAM!
It passed the test because the data lines capacitance 'saved' the data on the wires. Therefore a better test writes random data to each location and then reads back the memory with the same random seed. In fact it may have been charging the RAM thru the data and address wires.
Ah, but you need several buses that can be charged up an hold data, and then you can save on RAM! :)
wahhh. A gray beard has spoken.
I can't count how many times you called the 6522 an 8255 :D
the 8255 is burned into my memories
@@IMSAIGuy Just like an EPROM. 😁
When you are stopped, I believe what you have ruled out is your only way forward…. Is there handshaking with the front panel display???? I know you believe that’s not the problem but perhaps it’s not acknowledging something causing the micro firmware to time out and reset. ????
there is no handshake, one way to the front panel.
OK :-(. A friend of mine and I were debugging a server CPU upgrade last night, we were SURE that the upgrade wasn’t causing the memory failures… We were SURE! But we got out the microscope to examine the Xeon sockets only to find bent pins on the socket on the mobo which destroys any usefulness of that server… It’s a painful outcome, but moved everything to a spare server and it’s up and running now…I hope the root cause on your DVM doesn’t end up being as bad…
I assume that by part 7, you will reveal the problem,... but still I am curious if your adding all the wires to measure the signals has caused the flashy lights to occur?
I wonder if a clue is that the unit was probably dropped? I had a similar nightmare trying to repair an Atari 1050 floppy disk drive which took a big hit (thanks DHL) during delivery. It was very intermittent. I wasted a lot of time testing chips - just like you. Eventually I spotted the problem. The large main power cap had a dry joint on one pin, either caused or exacerbated by the drop. A simple,re-flow fixed the problem.
So I wonder as to which parts might be “jolt” sensitive. Is the micro clocked by a xtal?
I do not believe it as a mechanical shock problem, yes there was a xtal and that part of the uP worked fine. it seemed as though the uP was just a bit sick and did not reset correctly.
Seems bit odd that the reset is fired after 8 - 9 clock pulses while being loaded with count 16, could the down count be defect (only half of it is working, causing it to reset premature) ?
It needs a MK20 eyeball close inspection of the sockets and all solder joints. MK20 is the best instrument in the shop.
If your EPROM programmer is a TL866 II + you can write your own template to test chips that aren't already defined in the list , or you can use and existing profile with bigger capacity (maybe you'd have to wire your chip to match the pinout) and see if it errors out inside the addressable space of your chip or not. Did this when I had no other choice.
do you have instructions on how to create a template
@@IMSAIGuy I use an existing one ( 8255 is a good starting point ) then rework it to suit my needs.It's a matter of creating a new logic device and setting each step in the testing sequence by setting input and output states according to the symbol definition table shown in the logic test window. Well used it only once and can't remember how I did it but it worked.Xgecu pro can be powerflul tool
At this point I'd pull out the big guns and read the manual of the IO chip, writing down the initialization commands and look for them using the LA to troubleshoot why is is not outputting data.
You obviously solved the issue since this is 7 parts. My guess is c126 is causing the unit to randomly reset. Otherwise something with U122.
nope
every part in a socket is suspect. A close visual inspection is in order. Use a can of freeze spray or heat directed gun.
never liked socketed devices ... and connectors that are 40 years old ....
As someone else said - this is part 4 of 7, so are you teasing us ?? :) Freezer spray might tell you something ?
it might have, I'm not sure now.
I think the Retro Chip Tester Pro does these chips too? Pretty sure, going of the top of my head here as I don't have the list in front of me.
I've run out of ideas. It's far too odd.
Why not read the content of both EEPROMs several times to get sure, their contents reads always the same? I had some EEPROMs in the past, which had weak behaviour. After replacing those with fresh burned binaries, the unit works fine again.
Just my two cents.
these are uV erasable PROMs EEPROMs were not invented yet
YUP!! arduino testers!! good 1...the good, bad. & indifferent!! changed it's dipper! U realize TI has a chip 4 this? why ask why? bud dry!! a hotel 4 engineers? runs like a top, don't flop!