News:

Printed Amstrad Addict magazine announced, check it out here!

Main Menu
avatar_Velociraptor

CPC464 LED on, no output

Started by Velociraptor, 21:25, 15 July 19

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Velociraptor

Quote from: Bryce on 07:52, 25 July 19
You can try that, but you'll discover after it takes you about 30 minutes to remove the first IC, that it's just not worth it and the chance of damaging the board is still high. There can be a lot of solder behind the pin (underneath the IC) that the braid won't have removed and is impossible to get to.


My curiosity is outweighed by my trust in you :)
[/size]
Quote from: Bryce on 07:52, 25 July 19
There's no real guide, just because so many things can cause problems. When you have moved and are ready to repair the PCB, come back and we can go through the options in stages.

Bryce.


Sure! In the mean time I'll watch whatever youtube vids I can and do a bit of sniffing in the forums.

MikeSDK

I want to say.. if you absolutely, positively must must must remove the ICs in an intact form you need a proper vacuum desoldering station.. even then you can still damage traces if you're not careful, but I would almost guarantee if you try it with solder wick you end up having to do more repairs than when you began.

Velociraptor

Quote from: MikeSDK on 18:45, 25 July 19
I want to say.. if you absolutely, positively must must must remove the ICs in an intact form you need a proper vacuum desoldering station.. even then you can still damage traces if you're not careful, but I would almost guarantee if you try it with solder wick you end up having to do more repairs than when you began.


Ok, at some point with something, if I stay interested in this I'll want to try it I'm sure. But I won't be trying it with this 464 unless it's dead and I'm trying to salvage components or something like that.


I'm curious in that I want to give it a shot, but I'm not daft enough to ignore advice from experts.


I have the hot air station, the soldering station and I haven't bought the vacuum gun thingy yet but I intend to.

Bryce

Go to a flea market, buy some random broken 286/386 Mainboard from the 80's / 90's for a few cents and practice on that PCB until you are confident that you can remove IC's without frying them. Avoid devices like telephones and radios, because they are usually single sided PCB's. If you can successfully remove a TTL chip from an 8 layer PC PCB with internal VCC and GND plains, then you're there.

Bryce.

Velociraptor

That's a cracking idea. I'll give it a punt.

Velociraptor

Time for an update :


I now have Duratool tools - a hot air rework station, a soldering station and a desoldering station. I also have an Oscilliscope, a better multimeter and an ESR meter.



I wasn't able to find any 286/386 boards. Even broken ones are going for a fair amount. I did however get someone to send me some boards from a Sky box and some of those were similar. I've removed several components including ICs from them. I've done some SMD soldering to assemble a DIN to VGA for an Atari ST and a basic USB powered board. I've assembled the serial port board for the CPC too. So while I'm hardly an expert on soldering I now at least have done some. Solder wick is the most disappointing thing on the planet!


I now have the broken 464, and a working 6128 and a working 664. I also have a CTM644. I mentioned in passing to someone that I might need a gate array and before I knew it I had a donor 464 board, it lacks the YM IC but that's all.


The 6128 and 664 work fine on the CTM644 (although I will recap, or at least test the caps at some point and clear out the dust/filth).


The 464 still doesn't work, which is no surprise because I've not changed anything. However when I plug it into the CTM644 I do get output. I get raster bars across the whole screen like you'd see when something was loading.


Because the donor board has the gate array and Z80 socketed I've swapped them over and it has no effect.


I'm now ready to try to diagnose the problem - where do I get started? :)

Bryce

Are they raster bars with a border around them, or do the bars go to the very edge of the CRT?

Bryce.

Velociraptor


https://i.imgur.com/4V850j3.jpg


I can take a video if you think that would help?

Bryce

Ok, my guess would be a failed CRTC. Have you tried swapping the Gate Array for a good one yet, or do you not have a spare to try?

What you can measure now is pins 39 and 40 of the CRTC. They should both show regular pulses on the scope. Either of these missing could cause the effect you're seeing and also mean that the CRTC is dead.

Bryce.

Velociraptor

I have tried swapping the gate array and it made no difference.


So next step is for me to get the scope out and find where the CRTC is, and then with the power on measure it. The scope hasn't been out of it's box but I'm sure I can work all that out!


Can the keyboard be disconnected for this? So I can have the 464 case open hinged on it's right hand side so that the connector is still in which would allow me to use the power switch to turn it on?

Bryce

The CRTC is the big 40pin soldered chip with 6845 written on it. The keyboard doesn't need to be connected. When I repair 464's I have a jumper to short pins 1 and 3 on the cassette connector, which is the connection the power switch makes, so the PCB can be completely removed from the case.

Bryce.

Velociraptor

Yeah I found it through the power of google and realising that if I need to measure 39 and 40 then it has to be a big chip :)


I'll see about a jumper for 1 and 3 on there because I don't want to put strain on stuff.

Velociraptor

#37

https://www.youtube.com/watch?v=dFB9JJjwJuU



https://i.imgur.com/awOTWDZ.png


A big ominous cross!

Edit : I have desoldered the CRTC from the donor board. I thought even if I don't need to swap it out, then it won't do any harm to have it. I'm really pleased with how neatly I got it out.


https://i.imgur.com/xXSNZeO.jpg

A better result than that other guy's attempt you warned me about Bryce, I think :)

Gryzor

You do realise other people try to achieve this output for years, right? :D /sorry

Bryce

Ok, the "big ominous cross" is because the scope is set up wrong. You need to set the vertical to around 1V/Div and the horizontal to around 20ms/Div. Then you should see something completely different :)
If I have time later I'll connect up a scope and make screen shots of what you should see and what settings I used.


Bryce.

Velociraptor

Ah ok. Well.


In the meantime I desoldered both CRTC from the 464 I want to repair and the donor board, and then fitted a socket on the repair board and put the other CRTC in and it still gives the same output.


I'll go get the oscilloscope and try to work out how to set it up.

Velociraptor


gerald

Interresting case  :D

I grabbed the following picture from the video you posted.
We can clearly see a proper VSYNC (the black bar at top of screen), but the screen did not sync on it
Then the beginning of a proper screen with border, with garbage on the screen section, but the screen sync properly on the HSYNC
The on screen garbage may hint a RAM failure.Then at the bottom, everything goes bad.

Did the screen lost HSYNC, did the CPC really send garbage ?

I would
- 1st check that we have a proper 5V supplies (also, do you have a way to know how much you CPC is consuming ?)
- Checking that the various clocks are OK
  - 16MHz clock is OK at gate array input  - 4MHz to the Z80 (it may have a bit of jitter)
  - 1MHz to the CRTC


Velociraptor

I know that we have the right supply because the 664 and 6128 both work from it. I don't have a way to check the power consumed.


Is there a way and do you want me to check the voltage while it's on across two points?


How do I check the 3 clocks you've asked me to check?

Velociraptor

#44
So I thought I'd socket and replace all the ram to see what would happen. I didn't cut the legs, I very carefully desoldered. It was a little more difficult than the other ICs as it seems solder had gone up the legs of the ram ICs so I had to resolder and desolder some of them up to 3 times.

I'm now getting a different screen - the screen is a sort of green and purple chess board. I don't know if this is progress or not.



https://i.imgur.com/19xYBEe.jpg


Ran out of time tonight but there's lots more chips I could desolder, socket and then bring across from the donor board to see what happens.

I think tomorrow I might want to put the old ram chips back in to see what happens and then try swapping out one at a time.

gerald

Quote from: Velociraptor on 00:45, 14 September 19
So I thought I'd socket and replace all the ram to see what would happen. I didn't cut the legs, I very carefully desoldered. It was a little more difficult than the other ICs as it seems solder had gone up the legs of the ram ICs so I had to resolder and desolder some of them up to 3 times.

I'm now getting a different screen - the screen is a sort of green and purple chess board. I don't know if this is progress or not.



https://i.imgur.com/19xYBEe.jpg


Ran out of time tonight but there's lots more chips I could desolder, socket and then bring across from the donor board to see what happens.

I think tomorrow I might want to put the old ram chips back in to see what happens and then try swapping out one at a time.
Did you replace the RAM or did you just left the socket empty ?
If you have new RAM in, then the next suspect would be the address multiplexers (4x74ls153  IC104 IC105 IC109 IC113)

Velociraptor

I replaced the ram after I added the sockets. Next up is those 4 guys then!


Looks like I might end up with the whole 464 socketed. :)

Bryce

Quote from: Velociraptor on 16:35, 13 September 19
https://cdn.discordapp.com/attachments/444419003419721750/622092622772568072/Screenshot_20190913-163307_Video_Player.jpg

So that does look how it should I think.


Meaning the problem lies elsewhere?

Woah, you really need to sit down and learn how a scope works first. You haven't connected the GND'S to anything (those two crocodile clips). They need to be connected to the GND of the CPC. The waves you are seeing there are 50Hz, ie, the stray frequencies coming from the mains via you.

Here's what the two pins should look like. Use the settings I've used (as seen in screenshots) and set the trigger to about 0.5V.

Bryce.





Velociraptor

I know I should work out the oscilloscope, and I will, but I was on a bit of a desoldering run so I've desoldered the entire donor board.


In addition to desoldering and socketing the ram yesterday I've now done the same to the 153 chips and once I'd put the donor ones in I got a black screen, no output.


Turns out the power supply from the monitor has died. However when I use a different PSU. I'm thinking it's the cable, the coil was very reluctant to unwind and maybe something has snapped inside it.



https://i.imgur.com/BOTlALQ.jpg




Bryce

Using desoldered chips, without knowing whether they survived the desoldering is a bad idea. The parts aren't "known good", so you may be removing good parts and testing with bad parts.

Bryce.

Powered by SMFPacks Menu Editor Mod