Strange Risc PC video fault

discuss the archimedes & risc pc, peripherals and risc os/risc os on pi
User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Strange Risc PC video fault

Post by jms2 » Wed Jun 05, 2013 7:57 am

I've recently replaced my Risc PCs battery and all seemed to be going well. However, the machine now has a very strange intermittent fault. Bit difficult to describe but here goes:

- Initial "risc os 32m, acorn adfs" screen corrupted but just about recogniseable.
- Desktop useable but corrupted video with some grey colours altered (eg to pale yellow, but NOTthe bbc micro palette that you sometimes see when RO screws up)
- constant high pitched beep (not every time though)

If I reboot, this fixes it. Also, when using the corrupt desktop I was able to reduce the screen colour depth and this fixed everything too. Going back up to 32k colours then worked fine.

Any ideas? I suspect maybe I haven't quite put it back together right!

User avatar
TheCorfiot
Posts: 671
Joined: Mon Jan 08, 2007 5:22 pm
Contact:

Re: Strange Risc PC video fault

Post by TheCorfiot » Wed Jun 05, 2013 8:01 am

If you have Vram fitted remove it and try without...
It's probably worth reseating all the ram..

Check the psu connector is on properly, remove and refit...

Finally remove the podule bus card and see if it behaves without it...

TC :)

User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Risc PC video fault

Post by jms2 » Wed Jun 05, 2013 1:13 pm

Thanks for the tips. I've had my own idea as well. The machine has the "Mozart" 16 bit sound card fitted (no idea why I thought I needed that! :D ). I /think/ that since I reassembled the machine the sound isn't working, which suggests I've plugged in the Mozart card incorrectly.

The problem seems to occur on first boot, with subsequent boots OK. So maybe there is a capacitor on the Mozart card which, once charged up, cures the problem.

I'll have a play around with it and read the instructions on how to fit the card (typically, I put the machine back together without reading the instructions! :lol: ).

User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Risc PC video fault

Post by jms2 » Sun Jun 09, 2013 9:07 pm

I have not removed any VRAM or reseated the RAM yet, but I have done some simple checks and found the following:

1) Everything was installed correctly, including the Mozart card.
2) The sound isn't working at all.
3) If I remove the Mozart card, sound still isn't working and there are no improvements.
3b) Removing the podule backplane doesn't help either.
4) The actual fault is definitely something to do with the video system. On boot, the whole screen is duplicated left to right, ie it says "Risc OS 42M Risc OS 42M" etc. The ghost image is just as bright as the main one, and it appears at the midpoint of the screen.
5)If I try to change mode (from the desktop or from the command line) I just get a black screen now.
6) R-power on reset doesn't help.

I'm using the original AKF60 monitor.

dp11
Posts: 1135
Joined: Sun Aug 12, 2012 9:47 pm
Contact:

Re: Strange Risc PC video fault

Post by dp11 » Sun Jun 09, 2013 10:11 pm

Sounds like an Address Line fault to the VRAM to me.

User avatar
TheCorfiot
Posts: 671
Joined: Mon Jan 08, 2007 5:22 pm
Contact:

Re: Strange Risc PC video fault

Post by TheCorfiot » Sun Jun 09, 2013 10:40 pm

I would try running the system without the VRAM..

TC :)

User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Risc PC video fault

Post by jms2 » Sun Jun 09, 2013 10:47 pm

ok, I just tried removing the vram.

I get pretty much the same problem, only at lower resolution! On the start-up screen the ghosting of the text doesn't occur, presumably because the horizontal resolution is now much lower. On the desktop though, areas of the screen such as the icon bar are now 'crawling' ie moving around slightly as if covered by a swarm of ants!

Putting the vram back in reverts to the original setup.

Removed the pc card too. This doesn't help either. :cry:

User avatar
TheCorfiot
Posts: 671
Joined: Mon Jan 08, 2007 5:22 pm
Contact:

Re: Strange Risc PC video fault

Post by TheCorfiot » Sun Jun 09, 2013 10:58 pm

Sorry to hear it buddy...

If the system ram also proves ok then it looks like a pretty nasty board fault maybe the VIDC has failed......

steve3000
Posts: 2192
Joined: Sun Nov 25, 2012 12:43 am
Contact:

Re: Strange Risc PC video fault

Post by steve3000 » Sun Jun 09, 2013 11:59 pm

Possibly a VIDC fault, but when they go wrong, it's usually no display or a very colourful mess.

But the fact your computer boots at all suggests its not the ram either, or that would fail at POST.

TBH it sounds like interference/echoing is happening somewhere between the VIDC output and the monitor. Possibly damage tracks/circuity from a leaked battery?

I'd strip the computer back to basics. If no improvement, take the circuit board out for a good inspection - check the back too :)

Just before that, however - have you tried a different cable? Or confirmed you cable is good with a different computer? Next is the monitor, try a different one, and a different lead if possible - LCD TV with VGA input would be a good test.

Steve

User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Risc PC video fault

Post by jms2 » Mon Jun 10, 2013 7:06 am

I had considered the monitor being at fault, it is certainly a bit flakey and I had in fact obtained an lcd monitor to try out with this machine. The screen corruption looks more like it originates from the computer itself to me, but I could be wrong! It is just a bit of a coincidence that when I open up the machine, it suddenly develops a fault - would be strange if it were the monitor.

But yes, I have nothing to lose by trying the lcd so I will have a go. Also I will try removing my new battery pack just in case it's that.

User avatar
1024MAK
Posts: 9969
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Strange Risc PC video fault

Post by 1024MAK » Mon Jun 10, 2013 10:01 am

The most likely cause of an echo or ringing on a video picture is no termination (normally 75 Ohms) on the video input line (industial VDUs often have a termination selection switch).
If you are using a RGB or VGA monitor and there is no switch, loss of termination for all three colour signals is unlikely, so if white lines/characters are echoed it cannot be ringing on just one colour channel.

Another thought occurs to me. Are the supply voltages to the computer video circuits in spec? Are the decoupling capacitors okay? As supply and/or ground bounce can cause all kinds of weird faults.

Mark

User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Risc PC video fault

Post by jms2 » Mon Jun 10, 2013 10:49 am

All good questions Mark. Not sure of the answers but here is what I have found today:

- No visible damage on either side of pcb.
- changing monitor and cable has not helped.
- redrawing the desktop by pressing f12 then return actually does fix it for a while. Also, if I change mode, when the screen fades down (this is risc os 4) the interference goes away as it fades.
- it definitely looks like interference, the mouse pointer has two twinkly vertical lines next to it which follow it about.
- sound, strangely, remains completely dead.

I have not touched the psu at all, but maybe the decoupling caps are in trouble and it is pumping noise into the machine?

User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Risc PC video fault

Post by jms2 » Mon Jun 10, 2013 8:41 pm

Some internet research has revealed that VIDC20 controls both sound and video. Furthermore, an old post on one of the Risc OS forums describes a RiscPC which had a similar lack of sound (picture OK though) and the owner cured it accidentally by connecting up a different monitor!

I also read that dry joints can affect RiscPC sockets where they are soldered to the motherboard, so I might just try to reflow the joints to see if that helps.

steve3000
Posts: 2192
Joined: Sun Nov 25, 2012 12:43 am
Contact:

Re: Strange Risc PC video fault

Post by steve3000 » Mon Jun 10, 2013 10:49 pm

jms2 wrote:Some internet research has revealed that VIDC20 controls both sound and video.
Certainly does :)

You mentioned you have the 16bit sound upgrade though. I've not seen one of these before - my past computers having either had 16 bit built in or 8 bit only. So I'm not familiar how it interfaces with the VIDC20 (I assume it does?). Have you removed this, and does 8 bit sound work?

Thinking about video issues with the RPC reminded me of something. There was a modification which I carried out to my RPC circuit board shortly after buying it back in 1996 - to remove restrictive filtering from the video output. This involved cutting 3 PCB tracks and 'scrunching up' three SMT caps, iirc. The effect was great - much crisper video in high resolution modes. But if it had gone wrong, it would have been a disaster. Might be worth looking to see if any modifications like this have been made or attempted around the video socket of yours, just in case something was attempted? I have a feeling this mod was only useful on the later model PCBs though...?

Steve

User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Risc PC video fault

Post by jms2 » Tue Jun 11, 2013 7:11 am

Yes Steve, that mod is for Iss.2 and Iss.3 pcbs only. Mine is Iss.1, and I have never modified it having owned it from new.

The 16 bit card plugs in vertically to LK8 which is a 6 pin header. Its output then goes via a cable to LK14. Neither 8bit nor 16bit sound now works on my machine.

I have a new theory. The rear plate on the motherboard fits around the psu case and has a kind of spring loaded finger which presses against it. I assumed this was for mechanical security, but I now think it could be to provide an earth path. What it if I have broken this connection, could that be messing up the screening?

This machine was working fine until I took it apart to change the battery. I can't have done all that much damage, surely....? #-o

EDIT

I forgot to add - I did reflow the solder on the VGA connector and and LK8 as well. This didn't change anything.

I have been experimenting with using F12 to redraw the desktop. This seems to invoke a re-selection of the current mode and thus "reboots" VIDC. The corruption varies after each reboot. Usually, everything is perfect except for some sparkly pixels around the mouse pointer. Sometimes I get a completely black screen. Once I got the whole screen sheared so that vertical lines were at about 45 degrees.

One other observation on the "messed up colours" effect. Black and white are both unaffected. In fact, it seems to be only certain shades of grey which are affected.

I have tried making life easier for the VIDC by reducing resolution and number of colours. This has no effect, suggesting it is not related to overheating.

User avatar
1024MAK
Posts: 9969
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Strange Risc PC video fault

Post by 1024MAK » Tue Jun 11, 2013 11:23 am

Not that this helps much, just thinking out loud (well erm, thinking in written text :lol: )

Whatever the problem is, there is a variance in the symptoms. So the cause of the video problem itself is likely to be a poor connection, a partial short (that varies), crosstalk or interference.

That shielding you speak of, power off, get your meter (on resistance) and check it is connected to earth or to 0V.

Also have you done a tap test? This is where you "gently" tap various areas of the board with the top of a screwdriver handle to vibrate the board, looking for poor connections (including dry joints).

Mark

User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Risc PC video fault

Post by jms2 » Tue Jun 11, 2013 12:13 pm

I'm pleased you think it's probably a dry joint or something, because that suggests the problem will be fixable. It would certainly fit with the history, ie absolutely fine for years then the first mechanical disturbance causes the problem to appear!

I haven't tried tapping things yet, but I did try waggling the vga plug and LK8. This doesn't do anything.

Inside the PSU, everything looks fine. There are no obviously leaky caps anywhere (including on the board itself).

The only "defect" that I am certain I have introduced is to have interfered with the spring loaded finger which presses on the PSU casing. I need to examine this closely to see if I have perhaps broken it off or something. I'm currently running the PCB completely outside the case - taking it out hasn't changed the problem in any way, but then if the earthing was broken to start with, that's what you'd expect.

User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Risc PC video fault

Post by jms2 » Sat Jun 15, 2013 9:13 pm

This problem seems to be getting better all on its own! :D

As of yesterday the "double vision" effect on boot disappeared. The messed-up colours vanished some time previously to that. But the mouse pointer corruption was still present.

Having looked at the schematic, I started to suspect the "CleanVcc" signal which the video output makes heavy use of. There are two electrolytic capacitors used in the decoupling arrangements for CleanVcc. However they both look fine and tapping, prodding and waggling them made no difference.

At this point though, I started to notice that the interference around the mouse pointer was gradually disappearing after a couple of minutes. I don't think this is heat related, as an immediate reboot causes the corruption to reappear. Perhaps it is something to do with charge building up somewhere? :-k

Anyway, also from the schematic, I noted that the original configuration of the Issue 1 board was to have pins 1&2 and 3&4 of LK14 shorted by jumpers. Shorting these with a screwdriver generated an initial crackle from the speaker and then - 8 bit sound restored! So the sound does work. \:D/

So then I tried putting the 16-bit card back in. Nothing. Waggled it around a bit as well while playing sounds but made no difference.

Conclusion - the 16 bit card is fried. :( [I won't be replacing it, CJE want £69 for a new one! :shock: :shock: ]

However, my video fault is now completely gone! :D This applies seemingly irrespective of whether the 16 bit card is plugged in or not.

I now have a theory as to what has gone wrong. After rebuilding the machine when I changed the battery, I could have plugged in the cable between LK14 anf the 16 bit card incorrectly. Had I done this, pin 1 (left audio) would have been connected to ground. Presumably the start-up beep would then have done something nasty to the 16 bit card and generated the alarming squeal that I heard on the first few boots. Could this also have built up a charge somewhere, which has only just dissipated when I shorted the LK14 terminals out with my screwdriver?

I'd be interested to know whether this sounds plausible, or any other suggestions. Any explanation which involves me screwing up is always going to have a high level of credibility! :D

steve3000
Posts: 2192
Joined: Sun Nov 25, 2012 12:43 am
Contact:

Re: Strange Risc PC video fault

Post by steve3000 » Sun Jun 16, 2013 1:19 pm

jms2 wrote:However, my video fault is now completely gone! :D This applies seemingly irrespective of whether the 16 bit card is plugged in or not.
That's great news - really pleased to hear your Risc PC has recovered from its ill health :D

The CleanVcc/capacitor theory is a good one - old caps on this age computer could easily be approaching the limit of their spec by now... The 16bit sound card does sound a key suspect in this however. Just in case it is damaged, I'd keep it out of the machine until it's been checked out.
jms2 wrote:Conclusion - the 16 bit card is fried. :( [I won't be replacing it, CJE want £69 for a new one! :shock: :shock: ]
Cje do list the acorn 16bit sound upgrade for £20 incl. (I had an Acorn one bitd) I don't know the difference between the Acorn and the 'Mozart' one, but if you're after 16bit it might be the way to go?

Don't bin your 16bit card though, they are fairly simple circuits and it probably can be saved by some investigative work. (I'd be happy to take a look for you, as I'm intrigued what the differences are compared to the Acorn card, but have a lot on at the moment, so turn around wouldn't be fast).
Cheers,
Steve

User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Risc PC video fault

Post by jms2 » Sun Jun 16, 2013 3:00 pm

thanks steve.
I think the difference between the mozart card and the acorn one is that the former includes a cd audio mixer, which I now remember is the reason I bought the mozart card rather than the Acorn one. I just wanted to be able to play audio cds, I wasn't all that bothered about 16 bit sound - it sounds exactly the same to me!
Nothing on the card looks damaged, so I fear that one of the SMD chips is fried.

User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Risc PC video fault

Post by jms2 » Thu Jun 20, 2013 8:33 pm

I spoke too soon - although thanks to a new MDF, the machine is happily talking to my LCD monitor, on occasion the colour problems, "double vision" and flickery mouse pointer issues have reappeared. There's also an audible momentary crackle from the speaker when the machine initially boots, but that might have been there since day 1, I can't really remember.

I'll have to get a scope onto it I think.

steve3000
Posts: 2192
Joined: Sun Nov 25, 2012 12:43 am
Contact:

Re: Strange Risc PC video fault

Post by steve3000 » Thu Jun 20, 2013 9:50 pm

The start up crackle definitely suggests a bad capacitor.

My A4000 has developed a start up pop/crackle over the past 6 months, and during use I do get some background interference visible on the monitor in mode 31 (800x600) but its not too bad.

However the Risc PC uses a much more high powered video chip and also generates its pixel timing via a very different method, which could potentially be confused by a noisey power line.

...I'd definitely look for noise on your 5v line & check the quality of the pixel rate signal.

User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Risc PC video fault

Post by jms2 » Mon Jun 24, 2013 11:31 am

This isn't going too well..

I've tried looking for noise on the CleanVcc signal, and on the 5V line generally, and whilst there definitely are some ripples, I can't tell whether these are normal or not.

The machine continues to behave very oddly. This morning I switched it on and it was ok, and after a while the jittery mouse pointer settled down. I left it on, thinking that letting everything warm up was perhaps the key.

Once everything was settled I tried rebooting from software. This was fine - nothing changed.

Then I tried a hardware switch off. This led to the jittery mouse pointer returning and even if I left the machine to warm up again, it didn't go away. Then I found that repeatedly hitting F12 then return (to redraw the desktop) led to a different pattern of corruption every time! Sometimes OK, sometimes colours corrupt, sometimes black screen, sometimes whole blocks of the screen physically moved around.

I am wondering whether (as you have suggested) it is a PSU fault. I'd re-cap it, but there are lots of caps in there (all of which look OK). Which ones would be the favourite suspects?

User avatar
1024MAK
Posts: 9969
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Strange Risc PC video fault

Post by 1024MAK » Mon Jun 24, 2013 12:38 pm

If you can, renew every electrolytic capacitor. They may well look fine, but it does not mean they are okay.

Heat helps to kill them, so any that live in or near hot areas / heatsinks. Also any that act as ripple filters. Trouble is, in a PSU this means all of them #-o

For a switch mode power supply, use low ESR, 105C types.

Ripple on a +5V logic supply should be very low, less than 50mV. Note that this is not the same as noise, which you will also find. Ripple frequency depends on the frequency that the PSU works at.

Mark

dp11
Posts: 1135
Joined: Sun Aug 12, 2012 9:47 pm
Contact:

Re: Strange Risc PC video fault

Post by dp11 » Mon Jun 24, 2013 1:15 pm

Can you post some pictures of the screen ?

User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Risc PC video fault

Post by jms2 » Mon Jun 24, 2013 6:44 pm

I'm going to borrow another RiscPC quite soon, so I will swap PSUs and see whether this helps.

Here are some pictures:
RPC1008.jpg
RPC1008.jpg (140.71 KiB) Viewed 1534 times
The pointer - this is pretty typical. The vertical lines "sparkle".
RPC1009.jpg
RPC1009.jpg (128.11 KiB) Viewed 1534 times
Random corruption type 1. Obtained by pressing F12 and return to redraw desktop.
RPC1010.jpg
RPC1010.jpg (124.45 KiB) Viewed 1534 times
Random corruption type 2.
RPC1011.jpg
RPC1011.jpg (120.14 KiB) Viewed 1534 times
Type 3. Everything is blurred!
RPC1014.jpg
RPC1014.jpg (109.41 KiB) Viewed 1534 times
Type 4. This is the colour corruption mentioned before.
RPC1015.jpg
RPC1015.jpg (114.68 KiB) Viewed 1534 times
Type 5. Not sure how to describe this one.
RPC1017.jpg
RPC1017.jpg (134.21 KiB) Viewed 1534 times
This is the "double vision" reported earlier.
RPC1018.jpg
RPC1018.jpg (117.24 KiB) Viewed 1534 times
Another view of what this looks like.

I do have some more photos, but you get the general impression from the set above. The most common one actually is "blank screen", but you also get a perfectly good screen from time to time.

User avatar
paulv
Posts: 3854
Joined: Tue Jan 25, 2011 6:37 pm
Location: Leicestershire
Contact:

Re: Strange Risc PC video fault

Post by paulv » Mon Jun 24, 2013 7:02 pm

Having seen the photo's. I've had very extremely similar issues with an A3000 in the past. I know they're different beasts but I cured my machines ills by tightening up the spade connectors on the power connectors.

I know the RISC PC has a 6 way molex type plug thing for power on the motherboard so it'd be worth checking to make sure the power connectors are nice and snug fitting in the connection. It could well just be noise due to a faulty/slightly wonky power connector.

Paul

User avatar
1024MAK
Posts: 9969
Joined: Mon Apr 18, 2011 5:46 pm
Location: Looking forward to summer in Somerset, UK...
Contact:

Re: Strange Risc PC video fault

Post by 1024MAK » Mon Jun 24, 2013 7:03 pm

I wonder if you have either a faulty crystal or a dry joint on a crystal. You may want to reflow the solder joints and see if it makes any difference...

Mark

User avatar
jms2
Posts: 2536
Joined: Mon Jan 08, 2007 6:38 am
Location: Derby, UK
Contact:

Re: Strange Risc PC video fault

Post by jms2 » Mon Jun 24, 2013 8:10 pm

Two excellent suggestions - I will try them both out!

The power connector was obviously disturbed during my recent strip-down of the machine so that is a distinct possibility. It is a chunky item though, not easy to damage - in fact the actual metal pins/sockets themselves are quite hard to get at.

It is certainly an intermittent fault, so I would say that must mean it is external to VIDC. However, whatever is going on seems to be seriously messing with VIDC's composure!

I should add that all of the corrupted screens above are more or less stable, they don't move around. The double vision one flickers a bit.

Also, I tried waggling the VRAM. Doing this makes vertical stripes appear, but these seem to be additional to all the other problems, and not related to them.

dp11
Posts: 1135
Joined: Sun Aug 12, 2012 9:47 pm
Contact:

Re: Strange Risc PC video fault

Post by dp11 » Mon Jun 24, 2013 8:45 pm

The images are great.

Don't rule out the possibility of more than one fault.

I would be tempted to check the databus ( I only suspect the lower 8bits)to the VIDC20 . I wonder if you tried writing to the palette register the following patterns and check they read back correctly 0x55 0xAA.

I would also check the VCO circuitry which is clever.

Post Reply

Return to “32-bit acorn hardware”