Hi Guys who know more about hardware faults than me!!
My 464plus (upgraded by bryce last year) has developed a fault.
Basically some pixels on screen are not being correctly refeshed or cleared when no longer needed.
I have confirmed it is the machine as when I run it with no expansions the issue is still present. however the most obvious way to recrete the issue is with the Symbos Starfield screensaver whereby some of the dots do not correctly clear..
It also happens when running disc software with no expansion attached.
It has no real interference with the machine, other than aesthetics, but I like my aesthetics.
Simply changing the screen mode (or cls) will correct the problem (or in symbos exitting the current app on screen) however it is quite annoying.
Any idea what might be on the blink? Hopefully nothing major, as this was destined to be my main CPC. :(
I can capture a You Tube video with the screensaver running if you think it will help diagnose.
Craig
Do you get only screen corruption or do you also have stability problems ?
If a mode setting or CLS clear the problem, then it sound like a Z80 read-modify-write rather than a CTRC/ASIC read issue.
I have attached a memory test that you can run. It will fill all extension ram as well a base ram 0x4000-0x7FFF and check it back wit two set of patterns.
Red mean bad, Green good.
I'd try re-seating the CPU and RAM ICs too.
Bryce.
Did you try another monitor, could it be a problem of the monitor?
Quote from: TFM on 22:41, 04 February 15
Did you try another monitor, could it be a problem of the monitor?
Yes... Swapped both the cm14's and expansions with my 6128 plus.
OK removed and reseated the CPU and extra ram... Made it worse.
Since the extra on board ram is not used anyway I have removed it all together, using just the xmem ram all is good.
Now where to get 2 more chips to pop in......
Anyone have any, and for what price.
Craig.
So it's the RAM... or a PSU problem of the Plus PCB (because the X-MEM has it's own PSU, so it's fine. Or I'm totally off). Hope you get it up and running again!!! :) :) :)
Seems so. Oddly with no expansion it does not happen. Something seems to go amiss when ram from 65k to 128k exists both internally and on the expansion.
So the problem only occurs when the x-mem is connected?
Bryce.
Or symbiface or dk tronics 256k expansion. In fact anything that also provides the extra banks up to 128k.
Since it works perfectly with the internal ram ic's removed I can live with that I guess. It would be nice to leave them in place and put a switch in to enable disable them. Is that possible ? And where should I put the switch?
It sounds like the RAMDIS signal isn't getting through. Might just be dirt on the expansion connector.
To disable the upper 64K you could add a switch across one end of R128. If R128 isn't connected the ASIC believes that the computer only has 64K.
Bryce.
Hmmm. What is the best way to clean a centronics socket?
Meths / rubbing alcohol I guess.
Yup. Could be a dodgy cable too. Are you using the same cable on all devices?
Bryce.
Quote from: Bryce on 23:26, 04 February 15
Yup. Could be a dodgy cable too. Are you using the same cable on all devices?
Bryce.
no, swapped the cable with the device. So that rules that out. Will give the connector some TLC with a cotton bud later. After I finish watching Moonraker.
OK, The connector was pretty clean anyway.. gve it a wipe down and clean with Meths.
Still the same issue tho. :(
I have pulled the 2 ram chips for now and screwed everything back together.
It's working fine. I can live with it I guess. But I'd like to get 2 new chips to test it and just see if that will resolve the issue
It's not all bad news however, Whilst I had it open I replaced the membrane with my last good spare, so the L key now works again LOL ;)
I have an x-mem and DK'troniks RAM here. I'll try them on a 6128+ as soon as I have time.
Bryce.
All the expansion kit works faultlessly on both of my 6128plus machines. It's just the 464plus that misbehaves. No worries I'll buy 2 new ram ic's and see if they are the issue. If not I might be interested in the Bryce repair service, to track down why the internal ram is not disabled correctly. If that is OK? Absolutely no rush however as it works fine with the chips removed and it will always have an x mem on it anyway!
RAMDIS is handled by the ASIC (Pin 127) and is pulled low by resistor R115 (2K2). You could check this pin to see what it's doing and check that the resistor is ok. If they are, it might point to an ASIC failure.
Bryce.
Oh hell I hope not. Will check the resistor tonight.
Worst case scenario I have a 664plus
Did you had a chance of running the memory test with all the RAM in the CPC ?
I had some troubles with the plus RAM when developping my extension :http://www.cpcwiki.eu/forum/amstrad-cpc-hardware/cpc-ram-extension/
In my test, this problem will show corruption of the base RAM while the extension is OK.
Quote from: gerald on 18:46, 05 February 15
Did you had a chance of running the memory test with all the RAM in the CPC ?
I had some troubles with the plus RAM when developping my extension :http://www.cpcwiki.eu/forum/amstrad-cpc-hardware/cpc-ram-extension/
In my test, this problem will show corruption of the base RAM while the extension is OK.
Yes and no. Let me explain...
When I have the ram in place and run memtest without the expansions all is good, no errors. no artifacts
When I have the ram in place and run memtest with the expansions, no error is returned but the artifacts appear on screen
WHen I remove the extra ram (64k on board) and the expansions all is good, no error, no artifacts.
So it seems that there is some wierdness with the 2nd bank of 64k when it is provided both internally and on the expansion. Suggesting as Bryce said that RAMDIS is pooched.
Unfortunately as he has advised this is a ASIC function in the plus machine so not good news, I will check the 2k2 resistor that is also part of the RAMDIS, and hopefully this is all that is needed...... I need another 2k2 first. Might as well order it with the 2 extra RAM IC's I guess
Here's hoping.
If this does not work, if I put a switch between +5v on the RAM socket and the chip, would this effectively disable the 2nd bank resulting in 64k without needing to pull the chips all the time?
Craig
Quote from: CraigsBar on 19:53, 05 February 15
When I have the ram in place and run memtest without the expansions all is good, no errors. no artifacts
When I have the ram in place and run memtest with the expansions, no error is returned but the artifacts appear on screen
WHen I remove the extra ram (64k on board) and the expansions all is good, no error, no artifacts.
That's all valuable information !
By the way, does the artifacts appears during the test of the internal RAM :
- base RAM on the 1st line is the : only the 0x4000-0x7FFF is tested.
- internal 64k extension on the last line, which is still avaliable even with the 512K of the XMEM.
Quote from: CraigsBar on 19:53, 05 February 15
If this does not work, if I put a switch between +5v on the RAM socket and the chip, would this effectively disable the 2nd bank resulting in 64k without needing to pull the chips all the time?
It would be better to keep them powered but have the CAS tight to 5V instead of beeing connected to the ASIC. If you remove the power and keep them in circuit, they may get damaged or perturbate the bus.
I've just put it all back together. Over the weekend I'll capture a video to you tube of all tests, will that be best?
Oh.... R115 is perfect so it's not *that* easy.
Quote from: CraigsBar on 20:28, 05 February 15
I've just put it all back together. Over the weekend I'll capture a video to you tube of all tests, will that be best?
It's up to you :D .
But I am mainly interested in the one with artefacts. If there is no artefact when the test is fill/reading the internal RAM (BASE or 7EC4), then your symptoms are the same as Remax ones.
OK with the internal 128k only, no artifacts.
internal 64k only, no artifacts
with internal 64k and xmem 512, no artifacts
with internal 128 and xmem 512, artifacts.
They appear pretty constantly, but with most ferocity when testing row 7fc4
Is that what you wanted?
7ec4 seems to never be tested regardless.
Quote from: CraigsBar on 21:07, 05 February 15
with internal 128 and xmem 512, artifacts.
They appear pretty constantly, but with most ferocity when testing row 7fc4
Artefact also on 7ec4 test ?
A video is welcome for the 128 internal + 512 external.
Quote from: gerald on 20:08, 05 February 15
It would be better to keep them powered but have the CAS tight to 5V instead of beeing connected to the ASIC. If you remove the power and keep them in circuit, they may get damaged or perturbate the bus.
OK so if I follow this correctly. I have to switch pin 16 CAS on the IC between CAS on the Socket and +5 v
When It's connected to CAS on the socket it'll be 128k
When connected to +5v it'll be 64k
Is that right? and I Obviously have to do that for both of the 2nd bank IC's.
Craig
Quote from: gerald on 21:11, 05 February 15
Artefact also on 7ec4 test ?
A video is welcome for the 128 internal + 512 external.
Ok 2 videos,
First up 128k all internal - no artifacts.
128k - YouTube (https://www.youtube.com/watch?v=XzR2hUyv5uQ)
Then all 576k with xmem - artifacts.
576k - YouTube (https://www.youtube.com/watch?v=6OeRmJ85Q-0)
Quote from: CraigsBar on 22:55, 05 February 15
Then all 576k with xmem - artifacts.
576k - YouTube (https://www.youtube.com/watch?v=6OeRmJ85Q-0)
It looks really different from Remax CPC sympthoms where artefacts are only interferences with the display scan. They disappear once there is no more access to the external RAM.
Here it looks like a real corruption of the video memory (and maybe other some location too). I only think about probing the RAM and the extension bus to understand the real problem.
Quote from: gerald on 09:21, 06 February 15
It looks really different from Remax CPC sympthoms where artefacts are only interferences with the display scan. They disappear once there is no more access to the external RAM.
Here it looks like a real corruption of the video memory (and maybe other some location too). I only think about probing the RAM and the extension bus to understand the real problem.
Well Good news folks, the issue seems resolved!!
OK The only changes I have made are...
1. Cartridge swap (Parados cart swapped for c4cpc)
2. Replaced the fan in my ATX power supply, now it is not so loud I am using that for the xmem and HxC instead of 2 Iomega Zip Power bricks
and that is it. I guess the ATX Power supply provides a better, cleaner, more consistant 5v 2a than the Iomega brick did.
I set Symbos running through the Starfield screensaver for 2 hours tonight and not a single artifact was present. I have now run the full memtest dsk again, All blocks are green and lso no artifacts.
My 464plus is back and healthy with 128k and no problems. I am a Happy Chappy :)
Craig
Cool.
As mentioned before (in several other threads), using two PSUs in parallel is a really bad idea. Although the MX4 has a separate power socket, you should use a PSU powerful enough for everything and connect it to both the CPC and the MX4 input.
Bryce.