News:

Printed Amstrad Addict magazine announced, check it out here!

Main Menu
avatar_ShadowTD

6128plus Troubleshooting

Started by ShadowTD, 15:08, 21 September 22

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

ShadowTD

Well, after 30 years I finally own a 6128 plus.  Of course it's got some problems, because I wasn't prepared to pay silly money for it!  The listing described the text as being 'crushed up' to the side, which I thought would be a video issue but the machine never actually manages to initialise.  I've cleaned the cartridge slot and all my carts - this is what I get:

Supplied green system cart: Originally would look like it was trying to boot, now just gives a corrupted graphic screen
Red Burning Rubber card: Shows distorted version of title screen, then locks and corrupts
C4CPC: With CPselect, it's a black screen.  Switching to a bank with the Arnold V test cart gives the correct green edging with blue back ground and a yellow A flickering in the top left corner for a few seconds before it becomes solid.

The main board looks fine, no evidence of any butchery although there's two factory bodge wires near the RAM.  The Z80 is also socketed, and has what looks like a blob of solder on pin 38.  It was entirely missing all the screws from the base though, which does make me worry.

Screenshots can be provided if required.  Any help is massively appreciated.  Oh, and the disk spins constantly with all 3 carts.

Bryce

Most likely a dodgy connection on the cartridge port, but screenshots would be helpful.

Bryce.

ShadowTD

As requested, images:


The Arnold V diagnostic cart (on my C4CPC).  The A flickers for a few seconds and then goes solid


The system cart.  Note that this has got worse since I got the machine and cleaned both the cart and the slot.  It used to show the blue screen and a few words of the ROM text before it all got 'smushed' (technical term) up to the left had side.  Now just garbage



My GX Burning Rubber.  Note that by switching it on and off a few times you get a few different results.

All three carts work perfectly in my GX4000.  All have been cleaned, and the cartridge port itself has been thoroughly scrubbed with ISO.  I've also beeped out the port to the board.  The power jack seemed a little finicky, so I reflowed the solder.  When putting a multimeter on it, the voltage seems to rise from about 2.5 up to 5 when testing, which strikes me as the system isn't pulling the full amount of power.  The only chip that seems to be getting even slightly warm is the ASIC.
Finally, the previous owner reached out to me and noted that the only thing he'd tried was to swap the Z80 out with a known good one.

Bryce

Looks like there's a bad connection somewhere.

Bryce.

ShadowTD

OK, thanks Bryce.  When I was fixing my original 464 there were plenty of videos (Noels Retro Lab, Jan beta etc etc) with step by step troubleshooting - 'you're looking for 5v here, clock signal there, data to the memory here' etc etc which helped me diagnose that my RAM was bad.  Can't seem to find anything for the Plus machines - does anyone have any links they could share?

I beeped out the A and D lines from the ASIC to the CPU, all good.  There's good contact between the CPU and its socket as well.

ShadowTD

OK, so I've got myself an oscilloscope to try and help with this project.  Looking at the clock signal on the CPU, it doesn't look particularly clean - is this OK?

Bryce

A clock signal doesn't have to be clean. It just has to be there, symmetrical and without much jitter.

Either way, the actual clock will be "cleaner" than you show there. A lot of the rounding etc you see is due to the bandwidth of your scope.

Bryce.

pelrun

One will also see this if the probe compensation hasn't been calibrated properly - it's necessary to connect the probe to the calibration point on the side of the scope and adjust the trimpot on the probe until the square wave is as square as you can get it.

ShadowTD

Thanks both.  I'm operating at the very edge of my knowledge here :)

@Bryce - I bought a 40Mhz scope on the advice of a friend - 5x the frequency I'm measuring, so I should have no issues with 4Mhz?
@pelrun - Thanks, that was news!  Tried it and the calibration signal is as clean as can be.

Bryce

Your friend was correct, however, you didn't say where you were measuring the clock signal. The base clock going to the GA is 16Mhz, so 5x would be 80Mhz.

Bryce.

ShadowTD

Ahhhhh, right - again, thanks!  That clock was measure on the CLK pin on the CPU.

ShadowTD

Update: I'd had a few people say they thought it was the RAM, so I ordered some new chips and and sockets.  The sockets arrived first so I thought I'd crack on with getting them installed, and maybe play switcheroo with the RAM chips in case I could get it running with two chips.
Well, that made no difference.  I was sitting looking at the board and the thought occurred to me that I hadn't tried a CPU swap.  The previous owner said that the only thing they'd tried was cleaning and a CPU swap with a known good, but I figured 'what the hell' and ripped by GX4000 apart.

Well, now there's a thing.  I forgot the golden rule of IT - try everything, believe nothing.  Arnold V test cart runs fine, and the machine seems fine.

The disk drive isn't working, but I'm reckoning it just needs a new belt, which is on order.  I did notice that when mashing on the AZERTY row, I saw a few '*break*'s coming up - is that normal?

I'm wondering whether to 're-plot' the keyboard as I realised when loading a GB cart using C4CPC that the layout was now 'correct' - can I just pull the keycaps off and swap them around?

Going to experiment with my first go at Retrobrighting tomorrow.  The top case is a bit yellowed, and the keycaps need doing as well.  Another Plus lives!

As for the 'faulty' CPU, it works 'better' in the GX as in I can get the Arnold V test ROM up, but the IRQ test causes screen corruption and it soon crashes.  Burning Rubber doesn't work.  Is there anything special about these Z80s?  Will this do for a replacement?
https://www.ebay.co.uk/itm/225089488910

Again, thanks for the help!

eto

Quote from: ShadowTD on 22:13, 28 September 22The disk drive isn't working, but I'm reckoning it just needs a new belt, which is on order.  I did notice that when mashing on the AZERTY row, I saw a few '*break*'s coming up - is that normal?

I'm wondering whether to 're-plot' the keyboard as I realised when loading a GB cart using C4CPC that the layout was now 'correct' - can I just pull the keycaps off and swap them around?

Breaks should not be normal. Have you cleaned the keyboard membrane? 

Unfortunately you can't swap keycaps around. I tried on a French keyboard too, but the caps on each row have a slightly different form.

pelrun

Quote from: eto on 09:38, 29 September 22Breaks should not be normal.

I disagree - in the keyboard matrix Esc is on the fourth corner of the square with A,Z,E, so the CPC will see it as a ghost key press if you're mashing the other three keys simultaneously.

eto

Quote from: pelrun on 10:00, 29 September 22
Quote from: eto on 09:38, 29 September 22Breaks should not be normal.

I disagree - in the keyboard matrix Esc is on the fourth corner of the square with A,Z,E, so the CPC will see it as a ghost key press if you're mashing the other three keys simultaneously.

Ah right! I forgot the ghost keys.


ShadowTD

Any thoughts about a replacement CPU?

zebulon158

Hello,
I had already a similary issue on my 6128+ ; of course, not sure it's the same problem but...
When I started some games, I had stranger pixels on the screen.
I thought it was a RAM issue, but it was a power issue : the 5V power from the screen was "tired" et not enough for the computer. I now use an external 5V power, my 6128+ works well now...

eto

Quote from: ShadowTD on 17:11, 02 October 22Any thoughts about a replacement CPU?


Make sure it's an NMOS CPU (Z8400A), not CMOS (Z84C0004). 

ShadowTD

I'm beginning to think this thing is cursed.

So I had the working CPU from the GX in it.  Ordered a new drive belt, which I installed very carefully, cleaning up the crud on the brass capstan.  Straight away it seemed to be trying, but would barely read a couple of disks.  The disk test seemed to indicate a speed issue, so today I was about to try and get one of the speed adjusting programs over to it (somehow!) and try to sort that out but now it's saying there's no disk in the machine!  The wiki isn't clear about how the drive detects the presence of a disk, but I've definitely not lost the write protect pin!

As an aside, eBay CPU showed up and installed in the GX.  All fine on that front.

Bryce

The write pin is not involved with the disk missing error, but the speed is. Did you definitely clean all the old gunk of the wheels before installing the new belt?

Bryce.

eto

Just as a hint: "all the old gunk" really means all. I once had a tiny little bit left under the rim(? not sure if that's the right word), which was really hard to spot. And only after I recognised and cleaned that too, it finally worked. 

ShadowTD

Well, very useful information from everyone.  I think what has happened is I've used the wrong belt - presumably it's a different item compared to the older drives.  It did seem rather tight when I was installing it.  Unfortunately, I think that has put the old motor under too much strain - even with nothing on it it's not spinning.  So now I need a new motor and the right belt.  There's some motors on eBay but I presume they're the 12v type rather than the 5v ones I need.  

ShadowTD

Any suggestions? I can't seem to find a 5v motor anywhere, only 12v ones.

tjohnson

Quote from: ShadowTD on 17:23, 14 October 22Well, very useful information from everyone.  I think what has happened is I've used the wrong belt - presumably it's a different item compared to the older drives.  It did seem rather tight when I was installing it.  Unfortunately, I think that has put the old motor under too much strain - even with nothing on it it's not spinning.  So now I need a new motor and the right belt.  There's some motors on eBay but I presume they're the 12v type rather than the 5v ones I need. 
If I recall the correct belt size is 72.0 x 3.5 x 0.5mm

eto

We had that discussion about 9 months ago and I think Bryce mentioned the belt size as 3mm wide, 0.6mm thick and a total length of 230mm but is often also advertised as 115mm.


Based on that information I bought some belts which were a proper fit and I have documented the size here: https://www.cpcwiki.eu/index.php/Changing_the_drive_belt

Powered by SMFPacks Menu Editor Mod