CPC6128 power supplies - must both 12v and 5v be connected?

Started by ibisum, 21:07, 07 May 18

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

ibisum

Yeah, the service manual doesn't provide much useful info imho .. but I've since found a source for RAM chips and have ordered them, in case that is the issue (cross-fingers), and even if its not I figure it'd be good to have some spare .. so I got the MT4264-15's, which I've heard around the 'net are compatible ..


Anyway, I'm still working on recovering the machine - will let you guys know if I make progress.  Tonight I'm just doing research; I'll break out the screwdriver some time tomorrow... hope I can feel a warm RAM chip, LOL!

tjohnson

What does the seller say about this?  I find it strange that you were sold a working machine yet it doesn't work and now you are buying ram chips.  Is there something you haven't told us?

Sent from my E5823 using Tapatalk


ibisum

I trust that the seller sent me a working Amstrad - it just didn't work when it arrived.  I'm prepared to do a bit of work to get it working - maybe all I need to do is re-seat a chip or two .. as far as getting the RAM chips already, they're useful for other systems too, don't forget. ;P


And before I throw this machine back to the seller, I really do want to try to recover it and get it working again.  That's part of the fun of the retro scene, isn't it?  And no, there is nothing I'm not sharing here - the machine simply hasn't powered on for me... yet.  As far as I know, I'm not responsible for it not powering on properly (I did double-check all my power supplies with a meter before plugging them in...) I'm hoping I'll get it working with your help, though.

tjohnson

Quote from: ibisum on 20:55, 14 May 18
I trust that the seller sent me a working Amstrad - it just didn't work when it arrived.  I'm prepared to do a bit of work to get it working - maybe all I need to do is re-seat a chip or two .. as far as getting the RAM chips already, they're useful for other systems too, don't forget. ;P


And before I throw this machine back to the seller, I really do want to try to recover it and get it working again.  That's part of the fun of the retro scene, isn't it?  And no, there is nothing I'm not sharing here - the machine simply hasn't powered on for me... yet.  As far as I know, I'm not responsible for it not powering on properly (I did double-check all my power supplies with a meter before plugging them in...) I'm hoping I'll get it working with your help, though.



I guess a component could have come loose, on my 6128 the cpu and the 40010 gatearray are socketed but seems unlike they would dislodge in transit without the machine getting damaged in the process.   I assume you didn't try and power it with a higher voltage than 5v by mistake, you seems clued up enough not to have done that.

ibisum

I'm quite familiar with the power issues that we retro-enthusiasts have to deal with at times, so yeah .. the suspicion that I've fried it with a dodgy power supply, while discomforting to me personally, is still something I have to contend with.  But I did measure the 5V supply before plugging it in the first time, and it just has not powered on.


So today I open the case and start the burnt-finger testing of chips to see if anything is particularly warm. 


In the Oric/Atmos world (and elsewhere) we have a few docs that describe test-point procedures - i.e. look for this voltage X on pin Y, etc.  No matter how I search, I don't seem to be able to locate a similar set of docs for the Amstrad world - the Service manual is woefully useless it seems, with a flow chart that just says ''replace these chips if you don't get video output'' .. but I'm wondering if any of you other enthusiasts have encountered a more detail test-point procedure for narrowing down the issues?  I'm all for burnt-fingertip testing, lets see if it provides some insight into what is going on .. but if there is a more technical test-point procedure I'd love to know it.


Anyway, my next action will be: open the case, inspect for obvious damage (blown caps, are they a thing?), leave it on for 20 minutes, look for overheating chips, replace chips.  Repeat until it boots.


Bryce

My tips:

Check that the socketed chips are properly seated.
Caps: Not really an issue in the CPC world. There is only one critcal electrolytic on the power rail (just behind the power socket), it rarely leaks as it wasn't really put under any stress in its lifetime. Swap it anyway.
Check that you are definitely getting voltage to all the ICs. I usually check the power pins of IC105 (74LS153 beside the CRTC) This is one of the furtherest points from the power input. If it's anything below 4.8V then there may be something partially shorted somewhere.
Regarding RAM: Possibly defect, but there are more likely candidates if you are getting absolutely nothing on the screen. The Z80 and the CRTC would be the first things to check. The Z80 is easy to check because it's socketed, the CRTC means lots of soldering, but as an Oric owner, you probably have some experience swapping those rather dodgy ULA chips :D

Bryce.


ibisum

Haha - my solution to the ULA problem on Oric has always just been: find another Oric. ;). (I have 7 of them now...)


Thanks for the tips - this is precisely what I was hoping for .. I'm going to open 'er up later this evening (Central European time) in the peace and quiet, and start probing around.  I sure hope its not the CRTC or CPU .. those haven't been so easy to find online.. but I did just get confirmation that the RAM has shipped, so lets see ..


Either way, thanks again, I'll dutifully report back my findings.  I know what its like when you see a new member get an old machine up and running again, hopefully you guys will share my joy when it happens ...

Bryce

Both the CRTC and CPU are still available to buy. Regarding the Orics, if you want them repaired then let me know. I have a source for the ULAs. I'm also the proud owner of both an Oric 1 and an Atmos. I even managed to load a tape on the Oric 1 once (or was that just a dream?) :D

Bryce.

ibisum





Good to know there's another Oric fan here .. well actually all my machines except for one Atmos, are now functioning .. I have the full suite, Oric-1 (x2), Atmos (x5), and even a Telestrat .. I'm an active contributor to the scene over on defence-force, and I have a couple Cumulus boards, as well as the brand-spanking new Twilighte+Sittler card which I just received last week, and if it weren't for the CPC6128 I'd be playing with Orix .. in case you haven't seen that yet, check it out - its bonkers: http://orix.oric.org. Basically, a Minix-like OS for Oric, with bank-switching features and other great peripherals, like USB, mass-storage, serial .. a second 6502 .. etc ..


(And if you are interested in my non-working Atmos as a fun project, I'd be quite happy to trade it for something Amstrad'ish of similar value/working-condition .. I think its just an ULA replacement, but .. with my CPC6128, who has time for that noise ..)


Anyway, we can talk Oric stuff elsewhere, glad to know I'm not the only one here with that particular lust ..


Bryce

We can continue this discussion in a PM. I know someone who's interested in getting an Atmos, so I may take you up on that offer.

Bryce.

ibisum

Just to follow this up .. We had a RetroSession at Metalab here in Vienna last night (https://metalab.at/wiki/RetroSessions) and I invited MaV, from this forum, to join us - and he did!


MaV very kindly assumed responsibility for case-opening duties - I guess curiosity as to the state of things overcame him, and after a brief intro and looking away to other attendees, I returned to find MaV with the CPC6128 in intimate embrace over the electronics workbench .. he'd gotten 'er opened up, and was sitting there cooing away at all the things that could be wrong. 


It didn't take long.  Every single RAM chip was heating up.  :(


So, the conclusion is: probably I zapped the system when I first powered it on.  The seller I obtained it from is trustworthy and provided me video of it booting prior to shipment - but, I guess I must have done something wrong with the 5v and 12v power supplies I'd prepared... the theory that MaV and I formulated after observing very hot RAM, was maybe I'd plugged in the 12v after all.


So, the system is fried.


That said, I've ordered a full replacement-chip kit for the CPC6128, as well as a full set of RAM chips for it - they should arrive next week, hopefully, and I will proceed  - slowly and carefully, under controlled conditions of the Metalab electronics workshop - to start removing the dud RAM, putting in sockets, and so on.


So, as I make progress over the next few weeks, I'll keep you updated ... I'm committed to getting this system working again, and  I am grateful for all the help - especially MaV and Bryce, who warned me I might have fried things all along.  Ah well.


At the very least I can provide some full-chip-replacement-olympics amusement as I plod along.




MaV

I might add:
The main chips all had 4,6V at the Vcc pin, which seems ok if a little low. I suspect other components might have been fried as well, i.e. resistors and such, capacitors probably not because most of them on the board can take 12 volts, AFAIK. I didn't check on the glue logic and the FDC, the latter is important for stage 2, once the system shows the basic screen. The 74-series chips will need a bit of investigation after the main components are replaced, I presume.

The screen was black all the time, no flickering at all, which leads me to suspect that the CRTC and/or Gate Array are gone to meet the maker. The keyboard backspace key did not lead to the expected beep and I assume that the 8255 and AY-chip are compromised as well (alternatively: they might be ok, but the ROMs are fried, so the Z80 cannot initialise them properly).
I also checked the Z80 pins mreq and halt and they seemed to do their job. We'd need further investigation with a scope, but at the state the CPC is in I think ibisum should start with replacing the RAM chips first (and put sockets in first). The GAL, the Z80, and the GA are socketed, as is usual I guess, some things therefore can be tested more easily.

The question now is: Is it safe to replace the chips one by one in a certain pattern, then check in between whether the system starts to work or not, or will it need the full replacement anyway and ibisum is having to face a lot of work ahead?
Black Mesa Transit Announcement System:
"Work safe, work smart. Your future depends on it."

tjohnson


Oh dear not good, I did ask a while ago if you managed to put a higher voltage into the machine which caused you some discomfort.


Re. resistors I'm sure someone more knowledgeable than me may chip in but they tend not to be voltage specific but rated on current e.g. 1/8w.


I think it would be ok to swap individual components until you find the ones that have blown.

ibisum

Yeah, so I'll start the RAM socket installation some time next week, once I receive a replacement RAM set .. I seem to recall someone saying its possible to sit a replacement chip on top of the old one, without soldering it, to serve as a means of detecting bad RAM - i.e. with new, good chip on top, the old one functions .. wonder if that is worth the effort or if I should just rip them all out?


MaV and I noticed that every single RAM chip seemed to have heating issues - they were very hot after 10 minutes of power being applied.  I wonder -could they all be bad, or is something further upstream possibly bad .. hmm .. well anyway, it can't hurt to socket the chips, and I've got good desoldering equipment at the  lab anyway ..

gerald

Quote from: MaV on 11:13, 26 May 18
I might add:
The main chips all had 4,6V at the Vcc pin, which seems ok if a little low. I suspect other components might have been fried as well, i.e. resistors and such, capacitors probably not because most of them on the board can take 12 volts, AFAIK. I didn't check on the glue logic and the FDC, the latter is important for stage 2, once the system shows the basic screen. The 74-series chips will need a bit of investigation after the main components are replaced, I presume.

The screen was black all the time, no flickering at all, which leads me to suspect that the CRTC and/or Gate Array are gone to meet the maker. The keyboard backspace key did not lead to the expected beep and I assume that the 8255 and AY-chip are compromised as well (alternatively: they might be ok, but the ROMs are fried, so the Z80 cannot initialise them properly).
I also checked the Z80 pins mreq and halt and they seemed to do their job. We'd need further investigation with a scope, but at the state the CPC is in I think ibisum should start with replacing the RAM chips first (and put sockets in first). The GAL, the Z80, and the GA are socketed, as is usual I guess, some things therefore can be tested more easily.

The question now is: Is it safe to replace the chips one by one in a certain pattern, then check in between whether the system starts to work or not, or will it need the full replacement anyway and ibisum is having to face a lot of work ahead?
With the screen black all the time, there is for sure something else wrong than RAM. Failing "RAM" would just give you a grey screen with black border. I quoted RAM because the symptom can come from anything related to RAM access.
Full black screen can be caused by Z80 / ROM / CRTC / GA .
So after removing the heating RAM, try to get to that grey screen with black border first. Then mount the new RAM.

MaV

Quote from: geraldWith the screen black all the time, there is for sure something else wrong than RAM. Failing "RAM" would just give you a grey screen with black border. I quoted RAM because the symptom can come from anything related to RAM access.
Full black screen can be caused by Z80 / ROM / CRTC / GA .
So after removing the heating RAM, try to get to that grey screen with black border first. Then mount the new RAM.
Grey screen, black border. Thanks for the info.
I wouldn't know that because my CPCs are all in good working condition. /me knocks on wood.

Quote from: tjohnsonRe. resistors I'm sure someone more knowledgeable than me may chip in but they tend not to be voltage specific but rated on current e.g. 1/8w.
Yes. I was just guessing that some of the blown components might even cause a short somewhere, which in turn causes resistors to malfunction due to running out of spec.

Quote from: ibisumthey were very hot after 10 minutes of power being applied
More like after 10 to 15 seconds, really. While probing the pins on the other chips I made sure to switch the CPC off after every two, three attempts just to make sure the broken RAM chips are not getting too hot and damaging other parts.
Black Mesa Transit Announcement System:
"Work safe, work smart. Your future depends on it."

ibisum

I added pics in case its of interest - perhaps some of you Amstrad gurus can take a closer look at the chipset and let me know of any gotchas/interesting details I don't know yet:


https://metalab.at/wiki/RetroSessions




Bryce

If most ICs are measuring 4.6V on VCC then one or more ICs are shorted inside. I'd remove the RAM first and power it up again to measure VCC. Then start removing the AY (the AY is very voltage sensitive) and then the TTL chips one by one a re-measure VCC each time until it gets back to 5V. When you've got to 5V you can start putting the new ICs in and see if it boots. It's a long repair ahead of you though. Don't start removing the CRTC or 765 until you've removed the AY/TTLs and confirmed that VCC is still being pulled low. The analogue ICs in the amplifier section are probably ok, they could take 12V, but everything else could potentially die at anything above 5.25V. That's what the specs say.

Bryce.

ibisum

Cool, thanks for the advice Bryce - I'll be doing this surgery some time later in the week, will keep you posted - and I'll be sure to follow your advice very closely.  Great info!

Bryce

If you're really lucky, sometimes the RAM will have shorted the 12V so fast that it protected the other ICs from damage. I've seen this happen a few times (not just on CPCs). A fully shorted IC can pull the rail back down to 5V or lower.

Bryce.

ibisum

Interesting!  I'll be sure to follow the test procedure rigorously after I get the RAM removed ..


ibisum

Just an update .. this chip kit arrived today, seems to have a majority of what I might need to replace:


https://www.ebay.com/itm/Zilog-Z80-Amstrad-CPC-6128-Homebrew-IC-Kit/151327810599?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2057872.m2749.l2649


Still waiting for the RAM replacement kit, however .. as soon as that arrives, I will attempt resuscitation...

Bryce

No idea why the seller associates this bundle with the CPC 6128. The majority of the ICs are not used in the Amstrad.
Bryce.

tjohnson

Smart seller definately not China prices tapping into the retro revival

Sent from my E5823 using Tapatalk


MaV

Well, taking a glance at the CPC 6128 schematics and taking the RAM chips already in account, this leaves ibisum short of at least one of each:
74hcuo04
73ls32
74hc132 or 74ls132
74ls145
74ls153
74ls244
74ls273 or 74hc273
74ls373

PAL 16L8AC
Gate Array 40010

Since Bryce proposed checking on the TTL chips first, the batch above will be absolutely necessary for repairs.
Black Mesa Transit Announcement System:
"Work safe, work smart. Your future depends on it."

Powered by SMFPacks Menu Editor Mod