News:

Printed Amstrad Addict magazine announced, check it out here!

Main Menu
avatar_llopis

6128 glitching out (memory issues?)

Started by llopis, 22:18, 06 February 17

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

llopis

The other day I was happily playing Boulder Dash when, in the middle of a level, the screen "glitched out" for a couple of seconds and then froze. By glitching I mean the screen was half-corrupted: you could see some of the original graphics, but also corrupted with what looked random memory contents.


I rebooted the computer and it wouldn't even boot up. I would just get a blank screen, or sometimes a solid brown/green one. It wasn't just a display issue since there was no sound from pressing the delete key.


I left the computer alone, came back the next morning, and it booted up fine. Today I started doing some stuff with it, and after a few seconds, a similar thing happened, this time from the BASIC screen. Rebooting it will sometimes give me blank screen, and sometimes I'll get glitches on the BASIC screen itself (and no input). See the image I attached.


Any idea what might be going on? My first guess would be some memory issues (the memory itself or the bus), but that's just a guess. Is there something I can replace, so am I better off getting a new CPC.


Thanks.

VincentGR

Try a different monitor/psu.

llopis

Interesting. I hadn't considered the power supply issues. I was even using a 5V power splitter for the floppy disk emulator, so maybe that's why. I'll check it out, thanks.

Bryce

Yup, my first guess would be a dying PSU that craps out as soon as it's warm. If a new PSU doesn't solve it then it's either RAM or the CPU.

Bryce.

rpalmer

lets not forget the capacitors may have issues?

Bryce

Quote from: rpalmer on 11:38, 07 February 17
lets not forget the capacitors may have issues?

Extremely unlikely, as the 6128 only has 1 electrolytic capacitor on the voltage rail and the CPC works fine in most cases even if it's missing completely.

Bryce.

llopis

Quick update on this: After a few days of intermittent failures like I described above, now every time I turn it on it's consistently just a screen of solid color (the border is still black, but the main part of the screen is a brownish-grey). No responses to pressing any keys or anything.


I did check the power out from the monitor and it seemed like a perfectly solid 12V and 5V (although the 5V would only be for the disk drive).


So I guess that leaves me with some kind of failure in the main board, whether it's the CPU, memory, or bus. At this point I'm looking for a replacement computer and I'll keep this one for spare parts when the other one starts failing.

VincentGR

12V reversed is for the disk drive.

gerald

Quote from: llopis on 15:17, 11 February 17
Quick update on this: After a few days of intermittent failures like I described above, now every time I turn it on it's consistently just a screen of solid color (the border is still black, but the main part of the screen is a brownish-grey). No responses to pressing any keys or anything.
There are good chance that on RAM chip died. The black border / grey screen is the main symptom for this.
The RAM can be changed, but need de-soldering.

llopis

Quote from: gerald on 17:46, 11 February 17
There are good chance that on RAM chip died. The black border / grey screen is the main symptom for this.
The RAM can be changed, but need de-soldering.
Resurrecting this old thread since I might tackle this repair project soon.


Do you have any suggestions on how to go about testing the RAM chips and isolating the bad one? Are there any tests we can do with a probe or oscilloscope? Or are we stuck in the "replace it and see if it works" mode?

Bryce


llopis

I spent some time looking into this CPC again and spent many hours re-reading a lot of the threads on this forum. Here's a summary of what happened and what I've done. Hopefully someone can suggest a next step.

       
  • Every time I turn it on, I get a gray/beige screen and a black border (see screenshot). It seems a bit weird that the solid color part is not quite centered. I wonder if that's a hint of what might be happening.
  • The glitching/corruption I used to get only happened a few times at the beginning. Now it's just 100% that screen on start.
  • No IC gets particularly hot. The PAL one is the hottest, but I can hold my finger on it (from what I read that's normal). The AY chip is the next warmest one, but it's very mild. RAM chips are totally fine.
  • Voltages are fine everywhere I measured (including the IC105).
  • I removed the PAL and hardwired the connectors to use the second RAM bank (as described here). No changes whatsoever.
  • Just for grins, I turned on the CPC with the PAL removed and no wires. No difference.
  • I tried measuring the resistance on the RAM pins (as described by Bryce here). The results were kind of inconclusive. Resistance would be the same across most ICs, but then it would change on different ICs for different pins. So I wasn't able to isolate anything obviously wrong.
One accidental thing that happened that might help us narrow things down: When I started measuring resistance across pins, I had the computer on. Oops! I quickly realized I wasn't supposed to do that. However, as I was measuring on some of the pins on the bottom left RAM IC (IC126) the screen became garbled with colors and would change as I would touch the IC again.


So yeah, probably a bad idea, but I realized I was probably triggering the RAM chip and changing values, but it was interesting that they were being displayed. So that probably helps to remove the Gate Array as being defective since it was "correctly" sending video signals?


Any suggestions on what I should try next while I wait for my X-MEM to arrive and run the memory test?

Bryce

Have you checked the entire PCB for dry joints?

What did Geralds RAMTest report?

Bryce.

llopis

Quote from: Bryce on 12:59, 18 June 18
Have you checked the entire PCB for dry joints?
Not yet. I'll look at it tonight.

Quote from: BryceWhat did Geralds RAMTest report?
Haven't ran it yet. I'm waiting for an X-MEM to be able to run it.

VincentGR

The CPC I'm using now had the same screen as yours.
My problem was the two top ram chips.


=== ===
----- -----
----- -----
----- -----
----- -----
----- -----
----- -----
----- -----

llopis

Quote from: VincentGR on 13:04, 18 June 18
The CPC I'm using now had the same screen as yours.
My problem was the two top ram chips.
How did you narrow it down to those? Gerald's X-MEM test?

VincentGR

They were hot.
But on other machines I've found also bad ram chips that they were just fine.
Also voltage was OK too and then I did the resistance trick from pin to pin while the cpc was off.
That showed me the problematic chip.

llopis

Quote from: VincentGR on 13:16, 18 June 18
Also voltage was OK too and then I did the resistance trick from pin to pin while the cpc was off.
That showed me the problematic chip.
Interesting! So that trick works. Which pairs of pins were you measuring? 1-16, 2-15, etc or a different combination?

llopis

Motivated by the comment from VincentGR, I decided to measure the resistance across opposite pins more carefully. This is what I found out:

       
  • Pairs 3-8 and 14-9 were the same on all of them.
  • Pair 2-15 had no connectivity in the bottom 5 rows, and had some resistance in the top 3 rows (red).
  • Pair 1-16 had no connectivity in any of the chips except for one (yellow).
I looked at the pinout in the logic diagram and it the pins are:

       
  • 1 - No idea. It shows as an X. Does it mean it's not connected?
  • 2 - Address data in
  • 15 - CAS
  • 16 - Ground
Because of this, I had a particularly close look at the area with problems, and I noticed some kind of corrosion/build up. I'm not sure what it is. What's the best way to clean that up? Stiff plastic brush and alcohol on a cotton tip?


VincentGR

The trick was by Bryce as far as I remember in a different post.
I would have clean them with some vinegar, alcohol, resolder and clean again  ;D

Bryce

Isopropanol and the wifes toothbrush. Then reflow any pins that look dodgy.

Some things I should have mentioned when doing this resistance test: The meter needs to be a pretty accurate meter with lots of resolution. Also, make sure that you have the probes the same way around each time, ie: if you measure pair 3-8 with the red probe on 3, then don't accidentally measure the next 3-8 pair with the black probe on pin 3.


Yes, Pin 1 is not connected or used.


Bryce.

llopis

I gave it a good scrubbing with a toothbrush and alcohol and it's really clean now. I don't think it was corrosion but a bit of junk that had gotten stuck to some extra flux.


I turned it on and... exactly the same as before  :(


I measured resistances again (exactly the same way as I did before) and now the second row has normal resistance across the 2-15 pin, and the 1-16 pin in that yellow chip earlier is back to no connection. So it's all consistent except for pins 2-15 in the top and third rows.


That seems really weird, doesn't it? That it a) Changed and row 2 is back to normal and b) That the resistance is different in rows instead of individual chips.


I took the PCB out and gave it a good look and some minor brushing, and it looks to be in great shape. I don't see any dry joints or corrosion.


Any suggestions on what to do next? Replace those 4 chips? Wait for X-MEM to arrive (it could be a while)? Something else?

Bryce

You had the computer turned on minutes before, so all the capacitors will still hold a charge and can cause spurious resistance measurements.

I'd wait for the X-Mem, before doing anything further.

Bryce.

00WReX

I had similar symptoms to yours on a Tandy CoCo 1 a few years back.
After about 5 minutes form turning the computer on it would 'glitch' and the computer would freeze.
This happened over a few days until finally the computer would freeze the moment I turned it on.
It was tracked to a single RAM chip that finally failed, all good when that single RAM chip was replaced.

Cheers,Shane
The CPC in Australia...
Awa - CPCWiki

llopis

Update on this case (for those of you following along):


I decided to use the logic probe to see if I could detect a RAM chip that wasn't working like the others. It turns out pin 14 (Dout) of both chips in the third row (which was already one of the suspects after the resistance test) is always high. All the other chips show normal activity in pin 14.


The rest of the pins look like you'd expect: CAS is off for bank 2 but on for bank 1, Din flickers everywhere...


So the evidence mounts against row 3. The thing that still surprises me is why the whole row is acting that way. If this were a normal RAM failure I would expect a single chip to fail right? (On the other hand, it explains why hardwiring the PAL to use bank 2 made no difference).

Powered by SMFPacks Menu Editor Mod