News:

Printed Amstrad Addict magazine announced, check it out here!

Main Menu
avatar_JonB

CPC6128 white screen on startup

Started by JonB, 12:04, 18 June 13

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

JonB

Hi All

I'm new here! I just acquired a 6128 from ebay (uh-oh..) sold as untested -to be fair the seller listed it as "for parts or not working". I thought I'd try it as he claimed not to have any power supply to test it with (I guess this is a common problem with these machines what with people chucking CRTs away).
Anyway, once I'd got it out of its inadequate wrapping, I connected it all up using a power supply that is identical to the one sold by RetroComputerShack. It's a cheap Chinese HDD supply - gives +12v and +5v at 2A each. I've wired the 5v as +ve tip and the 12v as -ve tip per all the reading I've done (and the RetroComputerShack picture which shows these polarities too). The 6128 has a dicky power switch that needs a fiddle before it will close, but on successful connection I see the power light come on but it is not booting. The screen is white with a black border. It's connected to an LCD TV via a SCART lead; another RetroComputerShack product.
So... what does this indicate and how can I fix it? I had a look at the WIKI's service manuals page and there doesn't seem to be one for a 6128 - the link takes you to a 664 manual and this has a flowchart that suggests Z80, ULA or PCB bad.
Can anyone advise please?

Thanks

JonB

Bryce

Try swapping the Z80. It likely that it's broken and it's socketed, so easy to swap.

Bryce.

JonB

#2
Thank you for your quick reply. I tried it just now with a Z80A out of my Memotech MTX512 (sacrilege but it's the only machine with a socketed Z80 I have). I still have a white screen with no boot.

There is also a slight whine (interference?) coming from the loudspeaker which responds to the volume wheel.


One thing - the power supply I was using just blew up and took the mains supply down, so I am only giving it 5v. As I understand it, that should allow me to boot into BASIC but obviously the floppy is not going to work. That's a bit moot because I have no discs and it needs a drive belt anyway!

With the Z80 out of its socket I get no video signal at all - it must be doing something. How can I tell if it is the ULA and if so, can replacements be bought?

Regards

JonB

ralferoo

My feeling is that if you have a border at all, the Z80, ROM and gate array must be working, so it's probably just the RAM.

JonB

Hmm, that's the hardest bit to fix, all soldered straight to the board like that... 2 banks of 4164s? Can they be tested at all?

Bryce

As mentioned in other threads... Stick your finger on each one and see if any are getting hot.

Bryce.

JonB

Thanks Bryce, good point.


This morning I have left it running for an hour or so and checked the chips. Those that are warm are all roughly the same temperature and none of them are so hot I couldn't touch them. How hot would a broken chip generally be? As in "too hot to touch"?


Another thing occurred to me - is it a real black border or is my 16:9 TV set to the wrong aspect ratio? So I checked, by forcing it to 4:3. There's a definite border here and I would say the screen is white but one might say grey if the brightness / contrast were set lower. There is another thread where the OP talks about a grey screen and black border that is currently active. Perhaps we have the same problem...


Regards


JonB

JonB

#7
I have done a small test on the CPU by pulling the reset line low. It is resetting back to the white screen, but I think this is showing the CPU is working. Quick check of the address lines with a logic probe show them pulsing, so the CPU is getting a clock signal and trying to do something.


Interestingly, the two ROMs are cold and so are the four 74LS153 chips. Is this normal?

JonB

There's a modification on the PCB, is it normal? See picture. The white wire is connected to the floppy drive cable pins, and a track has been cut (circled in orange).

Bryce

Looks like an ABBA switch mod, but you've zoomed in a bit too close, which IC is that?

Bryce.

JonB

Try this..


It's the underside of IC206, a 74LS38 logic chip. Pin 6.. Actually, looking at the PCB layout in the service manual, the link is restoring the broken track. Odd..


Oh, by the way, the CPCWiki's link for the 6128 service manual is incorrect. Following it loads the 664 manual.


Regards


JonB

Bryce

Looks like someone had an ABBA switch there at some stage and reverted back to the original configuration later.

Gryzor has just updated the service manuals lately because they were corrupted, looks like he's mixed something up.

Bryce.

JonB

Thanks Bryce. What about the temperature test of the chips? Should they be too hot to touch if blown?

Devilmarkus

If they are already blown, they perhaps don't even get warm.

Another step is to measure all IC's input voltages.

Perhaps somewhere is a power missing?
When you put your ear on a hot stove, you can smell how stupid you are ...

Amstrad CPC games in your webbrowser

JavaCPC Desktop Full Release

Bryce

Yes, they would be noticably hotter than other the others, possibly too hot to touch.

Regarding the ROMs and other ICs, they will always be cold, the ROMs aren't accessed enough to get them even slightly warm, same with most of the 74xx logic.

Bryce.

gerald

Quote from: JonB on 11:28, 19 June 13
Actually, looking at the PCB layout in the service manual, the link is restoring the broken track. Odd..

That's an AMSTRAD original patch  :)
I have one 6128 with the same patch (MC0009B). That version of the PCB has two flaws
  - the drive select for internal drive is connected on Pin 6 instead of Pin 4. The track cut and wire fix this
  - the ready signal from internal drive is unconnected. There is a wire that goes from floppy centronics footprint pin 2 to the internal drive connector pin 26

The funny thing is that the service manual shows a MC0009A (older ?) PCB, which is correct.

JonB

Ah, good... that's a relief. Now just need to get some new RAM for it..

robcfg

Guys, could any of you take high quality pictures/scans of both board's sides?


The MC0009B model is missing from our motherboards wiki page.


Your help is appreciated!  :D

JonB

#18
@robcfg : I will try and get round to it.


Meanwhile, following some of Bryce's other advice, I tested the RAM chips with a multimeter using the method in this post (I tried to paste the URL into my post but it does not work (due to the brackets, I suspect) - please cut & paste into your browser URL bar):-


http://www.cpcwiki.eu/forum/amstrad-cpc-hardware/(hw-problem)-cpc464-with-'poked'-screen/msg50859/#msg50859


The results are a bit confusing. These measurements are for the left hand bank and all are in Mega Ohms. There doesn't seem to be many patterns there, perhaps they are all zapped? Has anyone got a set of measurements from a working 6128 for comparison?




PIN| IC126 | IC125 | IC124 | IC123 | IC122 | IC121 | IC120 | IC119 |
1  | 0     | 0     | 0     | 0     | 0     | 0     | 0     | 0     |
2  | .8    | 0     | 0     | 0     | 0     | 0     | 0     | 0     |
3  | 0     | .47   | .78   | 1.37  | 1.52  | 1.68  | 4.45  | 1.42  |
4  | 0     | 5.13  | .526  | 0     | 1.62  | 0     | 0     | 0     |
5  | 0     | 0     | 0     | 0     | 1.62  | 0     | 0     | 0     |
6  | 1.5   | 0     | 0     | 7.43  | 6.76  | 6.79  | 6.55  | 0     |
7  | .454  | 0     | 0     | .462  | .462  | .462  | .465  | .460  |
8  | 0     | 0     | 0     | 0     | 0     | 0     | 0     | 0     |

gerald

Quote from: robcfg on 16:41, 19 June 13
Guys, could any of you take high quality pictures/scans of both board's sides?

Best picture I can  :D
My scanner is not big enough !!!

Bryce

Quote from: JonB on 17:15, 19 June 13
@robcfg : I will try and get round to it.


Meanwhile, following some of Bryce's other advice, I tested the RAM chips with a multimeter using the method in this post (I tried to paste the URL into my post but it does not work (due to the brackets, I suspect) - please cut & paste into your browser URL bar):-


http://www.cpcwiki.eu/forum/amstrad-cpc-hardware/(hw-problem)-cpc464-with-'poked'-screen/msg50859/#msg50859




The results are a bit confusing. These measurements are for the left hand bank and all are in Mega Ohms. There doesn't seem to be many patterns there, perhaps they are all zapped? Has anyone got a set of measurements from a working 6128 for comparison?




PIN| IC126 | IC125 | IC124 | IC123 | IC122 | IC121 | IC120 | IC119 |
1  | 0     | 0     | 0     | 0     | 0     | 0     | 0     | 0     |
2  | .8    | 0     | 0     | 0     | 0     | 0     | 0     | 0     |
3  | 0     | .47   | .78   | 1.37  | 1.52  | 1.68  | 4.45  | 1.42  |
4  | 0     | 5.13  | .526  | 0     | 1.62  | 0     | 0     | 0     |
5  | 0     | 0     | 0     | 0     | 1.62  | 0     | 0     | 0     |
6  | 1.5   | 0     | 0     | 7.43  | 6.76  | 6.79  | 6.55  | 0     |
7  | .454  | 0     | 0     | .462  | .462  | .462  | .465  | .460  |
8  | 0     | 0     | 0     | 0     | 0     | 0     | 0     | 0     |

You measured from pin to ground?

@Gerald: Very interesting, I didn't know there were production corrections on any CPC board (Spectrum was famous for them).

Bryce.

gerald

Quote from: Bryce on 20:37, 19 June 13
@Gerald: Very interesting, I didn't know there were production corrections on any CPC board (Spectrum was famous for them).

When I first saw this, I had the same reaction you had, ABBA switch gone wrong ;)
But I the soldering looked too clean and the ready signal fix involved an additional drill in the PCB.

It seems the layout engineer messed up when moving the diode and resistor from bellow to above the internal floppy connector.
These with a re-routing of the 12V input seems to be the only changes from MC0009A to MC0009B

Bryce

What was wrong with the previous layout? Are there known issues that justify a layout revision?

Bryce.

JonB

No, from pin to opposite per your advice in the linked-to thread I mentioned above.

Quote from: Bryce
Ok, with a multimeter you can do a quick test to check the RAMs without having to remove them:

With the power turned off, measure the resistance across oposite pins of each IC. ie: Resistance between pins 1 and 16, resistance between pins 2 and 15, 3 and 14, and so on. Write the value down for each IC in a table. If an IC is damaged, it will be obvious from the table, because the value of one of the measurements will be a lot different from the others.

If no RAM IC seems to be damaged, then I am guessing that it's one of the RAM related logic ICs that's broken, but this could be any of the following: IC104, IC105, IC109, IC113, IC114 or IC115 and this could be difficult to diagnose without further equipment (unless you spot one that's getting hot.

Bryce.

Perhaps my measurements indicate they are all blown? But of course, as they are in situ, we are also measuring the resistance of the board and all other connected components.

gerald

Quote from: Bryce on 21:50, 19 June 13
What was wrong with the previous layout? Are there known issues that justify a layout revision?

Bryce.
I guess it is an accessibility reason. The moved component were under the internal floppy cable.

Anymone has time to check differences between the 10 revisions of MC00020  :o

Powered by SMFPacks Menu Editor Mod