CPC-464 freezes on start-up with black background colour

[Tezz]

Hi all,

I bought a 464 about 10 years ago, it was the later shorter board version. I replaced the cassette belt and all was good. I did have a DDI and an FD-1 also. I was making use of a USB charger for power. One afternoon (without the DDI/FD-1 connected) whilst part way through loading a cassette the output suddenly went black and sadly the fault remained upon a cold start. I had some time last week to finally take a look at it again, the output gave a grey screen with a black border so the first port of call was to test the RAM, I don't have a diagnostics ROM but I do have a small USB RAM tester so I removed all the 4164 RAM with my desoldering gun and hot air station before installing sockets. Happily one of the RAM chips did test bad so I returned the other 7 back into the new sockets. I have some tested 150ns 4264 RAM taken from an Atari 800XL which should be fine along with the existing 4164 but upon power up I now get the expected 464 start-up screen but the background colour is black and the machine freezes. Any idea of the likely cause? Perhaps it's not happy with that single 4264 RAM or I need to check my soldering work with the sockets. It all looked visually fine after the desoldering, no lifted traces or damaged through holes but I should likely check that first unless the symptom of a black background/freeze points to something else causing the issue. I don't own a logic probe or oscilloscope, just a basic multimeter so if there's a guide to follow some basic testing that would be handy.

thanks all!

[McArti0]

photo?

[Bryce]

You probably still have a bad RAM. Those testers aren't perfect and will sometimes report a chip being good, even though it's not. This is because the tester tests the chip at a very low speed. Try blind swapping each original chip with one of the XL chips and test each time.

Bryce.

[Tezz]

Hi,

thanks for the reply. I've now tried swapping each one of the RAM with the 4264 testing the CPC as I change each one. Sadly no change. I've only got 7 spare 4264's so I had to leave one of the 4164's but I swapped a few to be sure. I have another known working 40010 gate array so I tried swapping that just in case but no change. I've put the original 4164's and 40010 back in now with the single 4264 replacing the known bad one. 

I've taken a couple of photos. One of the board and one of the output but there doesn't appear to be any way to upload photos to my post only to link an image to one uploaded elsewhere online.

I suppose I will need to be doing some checks with the multimeter to each of the sockets next?
Any advice on how to progress?

thanks again.

[McArti0]

photos you need add to post.

[eto]

I've taken a couple of photos. One of the board and one of the output but there doesn't appear to be any way to upload photos to my post only to link an image to one uploaded elsewhere online.

Click on "Reply" and then there should be the option to upload images. Or - if you clicked on "quick reply", you can click on "preview" and then the option to upload pictures appears. 

[Tezz]

Thanks, I've attached the two photos now. 

I've taken the board out of the case this afternoon and double checked my soldering which all appears to be fine. I did a quick continuity check between the pins although not thorough I didn't find an issue.

I'm stumped, any advice on how to progress appreciated! thanks again

[Bryce]

Does holding down the DEL key make a beeping sound?

Bryce.

[Tezz]

Does holding down the DEL key make a beeping sound?

Bryce.

Hi,

No, there's no beep when holding down the DEL key. 
Whilst I was testing however I accidentally changed my TV source to a different input and when changed back to SCART the CPC's output was in colour with the correct blue background and yellow text. Cold starting it displays with a black background with white text again until I change my TV from one source back to SCART.

[Tezz]

I also have a working long board 464 so I've just tested the keyboard just in case it was causing any issue but the keyboard is working fine with that board.

[Rabs]

Just so, I understand, originally the CPC screen turned black part way through a tape load.

On restarting you got a black border and grey screen, classic RAM failure, so checked the RAM and indeed found one faulty IC, so replaced this with a 4264. 

Now the CPC boots at least to a ready prompt but with a black background and white text and freezes?

You have retraced your steps and checked you RAM IC work. 

Odd so the CPC boots to the Ready prompt and but with the wrong colour and then hangs (not responding to keyboard). I would double check the RAM and retest.

[Rabs]

@Bryce, I think the MT4264 has a 256 cycle refresh. How does this work with the Z80 or is it not an issue? 

[eto]

I think the MT4264 has a 256 cycle refresh. How does this work with the Z80 or is it not an issue? 

On a Z80 system that uses the R register to refresh the RAM that would not work, however on the CPC the GateArray is responsible for the RAM refresh, which also updates bit 8 (thus uses 256 cycles). 

[Tezz]

Just so, I understand, originally the CPC screen turned black part way through a tape load.

On restarting you got a black border and grey screen, classic RAM failure, so checked the RAM and indeed found one faulty IC, so replaced this with a 4264.

Now the CPC boots at least to a ready prompt but with a black background and white text and freezes?

You have retraced your steps and checked you RAM IC work.

Odd so the CPC boots to the Ready prompt and but with the wrong colour and then hangs (not responding to keyboard). I would double check the RAM and retest.

Hi, yes that's right except that I noticed this afternoon that when it's hung on start-up with no keyboard response that if I change the TV source off SCART and back to SCART the CPC's output is then actually in the correct colours with blue background and yellow text so the TV must not be quite getting the color signal at the point it hangs which is more encouraging. 

I guess there must be an issue somewhere I've missed with my IC work. What would be the best approach to diagnose the RAM/sockets? I've done a basic continuity test between each adjacent pin for any shorts which were fine. I checked several connections between the RAM legs and the solder joints on the bottom of the board which were fine, I haven't checked anywhere near all 16 legs of the 8 IC's in this way however. The DRAM tester I'm using is the one created by Simon Raybould if anyone is familiar with that one.

[Rabs]

Is the CPC actually hanging or is it just not responding to any keyboard input?

Is there actually a second problem with tha AY-3-8912 or 8255 PIO when scanning the keyboard matrix?

[Tezz]

Is the CPC actually hanging or is it just not responding to any keyboard input?

Is there actually a second problem with tha AY-3-8912 or 8255 PIO when scanning the keyboard matrix?

It looks to be hanging rather than awaiting input as the display to the TV isn't quite initialised on power up. Yesterday I swapped the Z80 from my working long board to rule that out and no change although when I then swapped the TV source from and back to SCART input the frozen (now colour) display is shifted to the left. I did some more basic checks with the multimeter around the RAM following traces from the RAM legs to where the traces went to and all checked out. I checked continuity of most of the RAM legs to the solder joints on the bottom of the board and all looks fine there too. Perhaps one of the LS logic chips has failed? Any advice on how to progress with the diagnosis? Thanks

[Bryce]

My next check would be the 40009 ROM. Check if it has any stuck Data or Address bits.

Bryce.

[Tezz]

My next check would be the 40009 ROM. Check if it has any stuck Data or Address bits.

Bryce.

OK thanks, I'll look into that. I'm missing an oscilloscope to look at the current state but I did recently buy a TL866II Plus which I've not yet used so I could remove the ROM and check that it's ok along with the 74LS logic although from what I've read testing logic with the TL866II is not conclusive as to whether it is fully operational. It would be nice to have something like SLICE to use but £450 is a bit pricey. It's a shame there's nothing like the old Fluke 9010A functionality around today. I've always taken the lamers approach removing IC's until I find the culprit.

[McArti0]

Screen has "Ready" and currsor.  Firmware and proc in bank 2 ram work. Check joy port or Simulate press key beetwen AY and 147

[McArti0]

When you join LS145 pin 15 to expansion port pin 1 (sound) you hear in speaker brum 50hz scan keyboard. This is check that's interrupt and soft and z80 works fine.

Oscylosphone is in CPC :-D

[Tezz]

Screen has "Ready" and currsor.  Firmware and proc in bank 2 ram work. Check joy port or Simulate press key beetwen AY and 147

Thanks for the info, that's good to know. Could you explain which pin of the AY and where 147 is on the board to simulate a key press.

When you join LS145 pin 15 to expansion port pin 1 (sound) you hear in speaker brum 50hz scan keyboard. This is check that's interrupt and soft and z80 works fine.

Oscylosphone is in CPC :-D

Great, thanks I'll definately try that over the weekend.

[McArti0]

https://www.cpcwiki.eu/imgs/6/6d/Schaltplan_cpc_464.jpg
This is LS145 not 147. My error.
Sound is at left side resistor R114 too. Its nearest LS145.

On schem you can look how keyboard work.

[Tezz]

https://www.cpcwiki.eu/imgs/6/6d/Schaltplan_cpc_464.jpg
This is LS145 not 147. My error.
Sound is at left side resistor R114 too. Its nearest LS145.

On schem you can look how keyboard work.

Great, thanks very much for that. I'll look forward to making some progress with the board over the weekend.

[Rabs]

My next check would be the 40009 ROM. Check if it has any stuck Data or Address bits.

Bryce.

OK thanks, I'll look into that. I'm missing an oscilloscope to look at the current state but I did recently buy a TL866II Plus which I've not yet used so I could remove the ROM and check that it's ok along with the 74LS logic although from what I've read testing logic with the TL866II is not conclusive as to whether it is fully operational. It would be nice to have something like SLICE to use but £450 is a bit pricey. It's a shame there's nothing like the old Fluke 9010A functionality around today. I've always taken the lamers approach removing IC's until I find the culprit.

I use a Hantek DSO5102P oscilloscope. Makes investigating a lot simpler.

See https://www.amazon.co.uk/gp/product/B00I5EWF1U/ref=ppx_yo_dt_b_search_asin_title?ie=UTF8&psc=1

[Rabs]

This is me checking keyboard scan pulses out of the 8255.

You cannot view this attachment.