CPC 472 disaster. Please help

[angelcaio]

Hi. A year ago I found an Amstrad CPC472 with an English keyboard in an ebay auction, I pushed and I took it. It was quite dirty, but it worked properly. I cleaned it, including the membrane, and it has worked well until today.

Also have a CPC464 with high keys, it is my first computer with me since 1985. To this I have connected a MX4 from Toto with the X-Mem, X-Mass, Minibooster also from Toto, an edge adapter with a lightpen from Electric Studio and a clone of the DDI1 of Zaxon with two floppy drives (3.5 "and 5.25") connected running this whole set for years. The floppy drives take the power of two psu: 5V 4A and 12 V 2 A and the equipment with all expansions with the GT65 monitor.

Testing, because the CPC464 did not start me, I connected all the expansions to the 472 that did not start either. I then tried to power the 472 with the 5V adapter but the monitor began to blink intensely and I turned it off immediately.
Since then, when I turn on the 472, the keyboard light comes on and only this is seen:         
https://youtu.be/v64oBZc9KyM
I am almost certain that I have connected it to the 5V source because I have marked the connectors of the chargers with red-black 5V and yellow-black the 12V but it is possible that I had connected it by mistake to the 12V because I had the two cables within reach and this possibility terrifies me a lot.
I am very interested in repairing the 472 because it is a somewhat rare machine and I  was looking for one like this for many years with the 664 ROM, Basic 1.1 and the daughter board included.
I will appreciate any comments that help me recover it although the truth i am not very skilled with the welder ....
I find the post in the forum
https://www.cpcwiki.eu/forum/amstrad-cpc-hardware/(help)-cpc6128-black-screen-after-12v-accidental-'injection'/msg161909/#msg161909
but I don't know how to start

Can someone help me with this? It would be great if Bryce could check it out  
Many thanks in advance

[llopis]

Isn't the 12V connector female, just like the one coming out of the monitor? If so there's no chance of accidentally plugging it in the 5V socket. That's what I do with mine anyway.


As for what's happening, it's hard to say. It doesn't look like the normal bad RAM. It could be a bad GA, but I can't tell just from that.


If you don't get any good leads I could look at it. It's 100% reparable, even if we need to swap out a couple of chips.

[Bryce]

Hi angelcaio,

unfortunately it does look like the computers have experienced an over-voltage. You may be lucky and only a single chip fried, but you may also have fried half the board. Even with good soldering skills, there's a lot of fault finding required here and they are probably not easy fixes.

You can of course send me the PCBs if you like and I will see if I can get them back working. Send me a PM.


Bryce.

[TotO]

It looks like you do not have sync (black strips are border), but the GA and CRTC works.
Probably it is a RAM issue. You can try to program the X-MEM with the gerald's RAM test from an other CPC to test it.


http://www.cpcwiki.eu/forum/amstrad-cpc-hardware/quick-and-dirty-ram-test-for-cpc/

[angelcaio]

Isn't the 12V connector female, just like the one coming out of the monitor? If so there's no chance of accidentally plugging it in the 5V socket. That's what I do with mine anyway.


As for what's happening, it's hard to say. It doesn't look like the normal bad RAM. It could be a bad GA, but I can't tell just from that.


If you don't get any good leads I could look at it. It's 100% reparable, even if we need to swap out a couple of chips.

Hi Llopis. No, it is not the 12v socket of the monitor.
I have two adapters with the 5.5 x 2.1 connector and I put two female connectors, one in each pair of wires from a branch of a molex PC source. The two 3.5 "and 5.25" floppy drives take power from these psu and I don't need the PC PSU.
My problem is that when I did not start my 464 (my old 464) with all the cards, I tried the 5V 4A (or 12V  ) source to see if there was a lack of power from the GT65 monitor's PSU.

[angelcaio]

It looks like you do not have sync (black strips are border), but the GA and CRTC works.
Probably it is a RAM issue. You can try to program the X-MEM with the gerald's RAM test from an other CPC to test it.


http://www.cpcwiki.eu/forum/amstrad-cpc-hardware/quick-and-dirty-ram-test-for-cpc/

Thanks Toto, But I just packed the motherboard before reading this

[angelcaio]

Hi angelcaio,

unfortunately it does look like the computers have experienced an over-voltage. You may be lucky and only a single chip fried, but you may also have fried half the board. Even with good soldering skills, there's a lot of fault finding required here and they are probably not easy fixes.

You can of course send me the PCBs if you like and I will see if I can get them back working. Send me a PM.


Bryce.

Thank you very much Bryce. I just sent you a PM

[VincentGR]

It goes to good hands.
The same pattern was on a 464 that I bought in a "working" order...


It was the basic rom and a ram IC.

[Bryce]

The boards arrived last night and I'll take a look at them during the week. If they are interesting fixes I'll post some info and pictures about them.

Bryce.

[VincentGR]

The boards arrived last night and I'll take a look at them during the week. If they are interesting fixes I'll post some info and pictures about them.

Bryce.

Oh come on, I need to know the faults 

[Bryce]

You'll find out soon enough. I have 7 other computers that need to be finished first. Tonight there's a 27in iMac on the bench and then a few Sega Master II and an Interton VC4000 to be looked at. These CPC's are planned for Thursday.


Bryce.

[Gryzor]

How does a 27" Mac come before a VC4000, even?

[Bryce]

Normally I go by the first-in first-out rule. But in this case it's a machine which is in daily use at a company and they need it back urgently. It's a simple fix, so I hope to have it back to them by tomorrow morning. The VC4000 isn't all that urgent.

Bryce.

[Gryzor]

What a load of poppycock. Just fix the CPC for them

[Bryce]

I'll be starting on the CPC's sooner than expected. The iMac is a bigger repair than expected, so I transferred their data to another machine for them and will leave it until the weekend. Bench now being prepared for the CPC repairs.


Bryce.

[Bryce]

Soooo.... This morning the 472 got some attention. Here's how it went and short description of how I go about diagnosing something like this.

1 - Connected the board to a current limited supply set to 1.5A (The supply turns off if the CPC tries to take more current). I do this to avoid frying tracks off the board in case there really is a complete short. The result was normal, the CPC was pulling around 830mA which is just marginally higher than what I would have expected.

2 - Time to connect it to a screen as see what we get. Completely black screen in this case.

3 - This is where I get slightly lazy and decided to view the board through a thermal imager. Tip: Leave the board untouched for 30 minutes before you do this, otherwise you'll just see your own thermal fingerprints all over the board. The results can be seen below. The bottom left RAM (D0 as far as I can remember) could fry burgers, so I immediately removed it to make sure it wasn't messing up one of the buses. The second thermal picture shows the CPU, CRTC and 8255. The ROM and all other chips weren't any hotter than I would expect them to be.

4 - I connected up my LowerROM board with @gerald RAMTest installed. Still no picture, so I swapped the CPU and 40010 for known good ones, but still no picture on the RAM Test.

5 - I removed the 472 sub-board and retried geralds RAM Test. This time I got a result: 7 green bars, plus one red bar because I hadn't installed a new RAM chip yet. So I installed a new RAM chip before I continued. Tip: On early Amstrad boards the chip footprints are slightly wider than they should be, so if you try to install a turned-pin socket the holes are too far apart for the pins so you need to seperate the socket down the middle to get it to fit. Very annoying and confused the hell out of me the first time I encountered this.

6 - I safely removed the ROM IC (40022) for testing, but my EPROM reader said it was completely dead Tip: If you are testing an EPROM that you just desoldered, don't install it directly into the EPROM reader. Use a chip socket in between, otherwise the uneven soldered pins will report false failures, because the pins aren't making good contact.

7 - Wrote the 40022 image to a new EPROM and re-soldered everything back in place. Yeaa!! It boots.

8 - Connected up a keyboard and speakers to check that the AY was doing its stuff - All good.

So the result was a bad RAM and a dead ROM. A lot less than I was expecting. Unfortunately the "special" part of the 472 is no longer 100% original, but that couldn't be avoided. The sub-board / extra RAM is still where it should be, which is the best that could be done. I will also send @angelcaio the original 40022 for keepsake. I also didn't use a socket on the new EPROM as I wasn't sure whether there's enough height in the case to fit one.

I was actually expecting 2 boards as angelcaio mentioned a dead 464 too, but there was only one board in the box, so that's all for now...

Bryce.

[VincentGR]

Great job, thanks for the detailed info Bryce.


I should play a lotto 

[angelcaio]

WOW!!!!!
you're my idol, Bryce!
I have no words to describe excellent work.
Many Thanks.

[Gryzor]

You fry your burgers at 38 degrees?

Thanks that was very informative and entertaining!

[Bryce]

You fry your burgers at 38 degrees?

Thanks that was very informative and entertaining!

Ok, smart arse, let call it simmering. Either way, 38°C for a RAM that's been powered up for less than 60 seconds in a 22°C room is not a good thing.

And anyway, it's fixed and that was the goal. More importantly, I wanted to give some insight into how to properly diagnose and fix a CPC. There have been several threads recently where people have been blindly swapping out chips for no reason other than the fact that they hadn't understood the root of the problem. This will of course eventually fix the problem, but it's a brute force method and unnecessary, so I wanted to give some tips on how to make sure you are only swapping what needs to be swapped. Less work, less cost, less chance of damaging the PCB.


@angelcaio: I hope you're not too disappointed that the original 40022 had to go, but there was no way around this. I've tried to keep everything as original as possible as I know that a 472 is not a plain old 464 or 6128. The sub-board is still connected with the original flat cable and on the original foam block, that's as close as I could get to keeping it unchanged.

I noticed that the non-functional RAMs actually have some pins connected. It would have been extremely funny if a failed "non-RAM" had actually caused the problem. It would have been the first time that a purposely non-functional part had caused a non-function of the system!


Bryce.

[angelcaio]

HI @bryce, I'm not disappointed at all.
I really would have liked to keep the original 40022, but only if it works.
It's great that you fix it so fast.

If you also send me the damaged 40022, I will store it in a bag and fix it inside the case 

[TotO]

I have 40022 if required.

[Bryce]

I've just pointed this out to @angelcaio . He will contact you about sending the part to me (you probably still have my address, I haven't gone anywhere), so that I can install it and get the machine even closer to its original state.

Bryce.

[angelcaio]

I've just send message to TotO
thank you for both

[angelcaio]

I just bought the 40022 from TotO.