CPC464 Sync problems

Started by imperious, 13:27, 02 July 22

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

imperious

This is a friends cpc I've got to try to fix. It's spent the last 7 years or so in His Shed, and this is QLD Australia so it has sufferered the heat and Humidity, especially the Humidity which I suspect has caused these problems.
Initially I could get this to work properly, but only the first time turned on after a few hours being off, but it's now got so bad I can't get it to work properly at all.
So I've done some poking around with multimeter and oscilloscope. The Ram seems ok and I have reseated The Gate Array which is a 40008 as well as tried another Z80.
What I am seeing is that the Vsync is either not there at all (no screen), or it's at different frequencies depending on what garbage is on the screen. hsync is ok at 15625. I'm thinking CRTC as according to various documents it supplies the Gate Array with H and V sync.  Any ideas?  Photo's attached.  The photo with AWA 64k microcomputer repeated was slowly scrolling up the screen endlessly.
 

Bryce

Looks more like a screen incompatibility problem than a CPC issue. Have you tried the machine on a different monitor?

Bryce.

imperious

Thanks Bryce.

It's not something as simple as that unfortunately. My CPC 464 works perfectly. I tried 2 different RGB (via scart) to HDMI adaptors, I have a cheap one that You can select 800x600, 1024x768, etc and a OSSC.
My friend also has both Colour and green mono Amstrad displays and they had no picture with it.

I'm mainly thinking CRTC as there is no vsync coming from it most of the time. I checked with an oscillloscope. When some sort of display does happen the vsync is not at the right frequency.  I have ordered some CRTC ic's from Aliexpress as well as a desoldering gun so will have another go when these arrive. 

eto

How can an incompatible screen could show a crips and clear text, but then repeat it after 8 pixels over and over? 

Does the startup-text on the third screen scroll?

And the second screen looks like mine, after playing with values in CRTC registers. 


Nworc

Quote from: etoHow can an incompatible screen could show a crips and clear text, but then repeat it after 8 pixels over and over?

Don't trust what you see on a digital screen, as there is a processor in the signal path that makes its own decision on what to do with the input.

And I personally wouldn't trust any system powered by some obscure third party power supply.

When checking the signals produced by the CRTC with a scope, I would do these checks without the LCD connected.

Bryce

Quote from: imperious on 15:21, 04 July 22Thanks Bryce.

It's not something as simple as that unfortunately. My CPC 464 works perfectly. I tried 2 different RGB (via scart) to HDMI adaptors, I have a cheap one that You can select 800x600, 1024x768, etc and a OSSC.
My friend also has both Colour and green mono Amstrad displays and they had no picture with it.

I'm mainly thinking CRTC as there is no vsync coming from it most of the time. I checked with an oscillloscope. When some sort of display does happen the vsync is not at the right frequency.  I have ordered some CRTC ic's from Aliexpress as well as a desoldering gun so will have another go when these arrive. 
Have you checked whether the solder joints on the monitor socket are good? This could cause that too.

Bryce.

imperious

Quote from: eto on 15:34, 04 July 22How can an incompatible screen could show a crips and clear text, but then repeat it after 8 pixels over and over?

Does the startup-text on the third screen scroll?

And the second screen looks like mine, after playing with values in CRTC registers.


Yes the 3rd screen scrolls upwards. I  couldn't post the video as the file is too large. It has crossed my mind that the Rom could be faulty.  I did mention that when turned on from cold occasionally it worked ok. I managed to play Nebulus and Defend or Die from my DDI5, but after warming up it progressively gets worse when turned back on. i also checked the reset line which is ok. It doesn't actually work properly anymore at all now and am lucky to get anything on the screen.

imperious

Quote from: Bryce on 17:17, 04 July 22
Quote from: imperious on 15:21, 04 July 22Thanks Bryce.

It's not something as simple as that unfortunately. My CPC 464 works perfectly. I tried 2 different RGB (via scart) to HDMI adaptors, I have a cheap one that You can select 800x600, 1024x768, etc and a OSSC.
My friend also has both Colour and green mono Amstrad displays and they had no picture with it.

I'm mainly thinking CRTC as there is no vsync coming from it most of the time. I checked with an oscillloscope. When some sort of display does happen the vsync is not at the right frequency.  I have ordered some CRTC ic's from Aliexpress as well as a desoldering gun so will have another go when these arrive. 
Have you checked whether the solder joints on the monitor socket are good? This could cause that too.

Bryce.

I replaced the Power and Monitor sockets as they were corroded badly.

imperious

Quote from: Nworc on 15:58, 04 July 22
Quote from: etoHow can an incompatible screen could show a crips and clear text, but then repeat it after 8 pixels over and over?

Don't trust what you see on a digital screen, as there is a processor in the signal path that makes its own decision on what to do with the input.

And I personally wouldn't trust any system powered by some obscure third party power supply.

When checking the signals produced by the CRTC with a scope, I would do these checks without the LCD connected.

The Power supplies i have are 2A ans 2.6A and work fine with my own CPC464 which has the same main board. Only differences are mine has 40010 and the faulty one has 40008.

eto


Quote from: Nworc on 15:58, 04 July 22Don't trust what you see on a digital screen, as there is a processor in the signal path that makes its own decision on what to do with the input.
We don't see anything that doesn't look like a "normal" CPC screen where some code went crazy. If the display is malfunctioning, it would break the screen at an unexpected spot - or bend the display - or do some crazy scaling - or simply not sync at all.

I'm not an expert, but I have never seen a screen malfunctioning, but still showing stuff as it would come from the real machine in several different ways that could also be caused by software in exactly the same way.



Nworc

Quote from: etoWe don't see anything that doesn't look like a "normal" CPC screen where some code went crazy. If the display is malfunctioning, it would break the screen at an unexpected spot - or bend the display - or do some crazy scaling - or simply not sync at all.

Can be that the signal was produced that way on the CPC, maybe very likely. What I wanted to say is: if I'm hunting a bug, I try to avoid as much side effects as possible.

imperious

I won't get the CRTC ic's for quite some time, so will replace the Rom IC first.
Faulty Rom ic's are quite common, especially in these 80's 8 bit machines. 
I'll update this thread after I've done that. I should get my desoldering gun today or tomorrow.

imperious

I replaced the ROM, made no difference. Put the original back in the socket.
Now the probably long wait for CRTC ic's from China.
If that doesn't fix it I'll give it back to my friend and He can sell it as not working. 

Nworc

retroclinic on *bay seems to have the chip in stock apparently and ships from UK.

imperious

I got the IC's today so will swap out CRTC sometime in the next few days.
Meanwhile I just found this thread  https://www.cpcwiki.eu/forum/hardware-related/amstrad-cpc-6128-weird-screen/msg214718/#msg214718
Well that definitely gives me hope the CRTC is the problem as the first photo has similarities to my 2nd photo.

imperious

The IC's I got are all HD46505SP-2  also called 68B45 and all 5 of them appear to be duds, that is unless someone more knowledgeable than me advises whether they are compatible or not?
I did see some being used on boards here and there which Is why I suspect they are duds or fakes.
I also got some IO 82C55 ics from the same seller which all look original, no reprinting or other shenanigans.

Bryce

Post a hi-res picture of one of the duds. I suspect you bought fakes.

Bryce.

imperious

They look pretty much the same as all the other ones with these numbers on them, but they were extremely cheap.
Buying from Aliexpress is a bit of a lottery unfortunately. Years ago I bought 20 Vic-1 chips for Vic-20's and they all work perfectly and some people at Denial forum wouldn't believe they were real. I digress though,here are photos

Bryce

That looks real, but salvaged. They may have fried them in the removal process though?

Here's an alternative source that I've bought from in the past and they all worked: https://www.ebay.de/itm/162423111533

Bryce.

imperious

Thanks Bryce. That had crossed my mind too. They are the most poorly soldered and straightened chips I have got off Aliexpress Yet.
I could remove the CRTC on my own CPC464 to prove it but don't really want to mess with a good board unnecessarily.

imperious

#20
Well I did what I didn't want to do which was desolder the HD46505SP in my own working CPC464. The PCB in these computer's really doesn't like too much work being done on it due to the not too great material used but only a few tiny marks on the board from desoldering.
So I put the HD46505SP from my own working computer in the faulty unit and it works perfectly.
All the 5 IC's I got from China are duds so I will put in a dispute with Aliexpress about this and purchase some more expensive ones, UM6845R looks good.
So to Sum up, my theory that the CRTC was putting out erratic V-sync and causing weird display effects was correct.
I'll update this thread again when I get some good replacement IC's. 

Powered by SMFPacks Menu Editor Mod