News:

Printed Amstrad Addict magazine announced, check it out here!

Main Menu
avatar_Charlie

Blast! Dead CPC6128 already!

Started by Charlie, 21:48, 23 January 11

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Charlie

Hello all.
As usual it's been a busy few weeks, so sorry for not being about.


I had a bit of time this evening so thought I'd wire 'my' RGB-SVideo adaptor in to my 'new' 6128.


Here's said adaptor in the flesh just to prove it exists:



Hmm, funny. Rubbish output and the screen is yellow on red rather than the usual...
...check connections - ok
...check TV - ok


Another hmm - what if my CPC was a bit jiggered and that's why I couldn't get it to work properly with SCART on a telly that's never borked on anything I've fed it..?


Well, hard to test that without another CPC so plough-on:
Let's take the power + GND from the CPC rather than an external PSU + reduce wire length in case my initial, um, botch-up was causing interference.
So I opened it up and wired a short length of ribbon cable to the back of the RGB and power connectors...
...put it all back together, flip the power and - NOTHING! Not even a friendly power light - D*MN AND BLAST IT!!!


-At this stage the adaptor wasn't connected so there was nothing electrically new attached-


Open it up again to check everything - all seems good, and no 'hot chips'. Check the PSU - working.
Put it all back together and still nothing, just a dead lump of plastic.
Plug in the adaptor just in case, not even a flicker from the telly to suggest that got power. Hmm, check the CPC's power connector - that's ok too.


So now my CPC is a lot jiggered before any of my plans have been realised.


Years of fiddling and this could be the first retro-box I've killed outright. I hope not.
CPC's are mostly TTL aren't they? I would have thought they are a bit more robust than this?


Sadly I'm so new to hacking these machines I'm wholly unaware of their nuances.


Can anyone give me some happy news or failing that a few pointers?


Thank you!

Charlie.

Are you pondering what I'm pondering? The Qube Server

arnoldemu

If you can, check the WAIT signal on the Z80, and also see if the Z80 outputs are pulsing.
I don't know if the z80 is that well protected on the cpc.
My games. My Games
My website with coding examples: Unofficial Amstrad WWW Resource

Charlie

Hi,
Thanks for that, I'll see what I can rig up (no access to an oscilloscope).


A question:
How's the power circuitry (inc the power LED) rigged-up in a CPC?


ie:
Power to 'electrical bits' inc the power-LED and then on to the 'electronic bits'.
Dead = something popped or a bad connection near the supply, possibly fixable


Power to the 'electronic bits' which toggle the power-LED as part of the system's start routines.
Dead = Toasted IC(s) will need finding and replacing, possibly not fixable


Another question:
How important are the various connectors with regard to the system showing signs of life?
I had to disconnect everything to get the mobo out for access to the underside - apart from the obvious power supply ones are there any less-obvious connections fundamentally important for a CPC to start?
eg:
Some retro-systems won't show any sign of life if there's something wrong with the connection to the keyboard even though it's not a 'primary system'.
I believe CPC's do have rather delicate membrane connectors and maybe this is all I've done. Also if I'm testing for faults it will be a hell of a lot easier to do with the keyboard unplugged if the CPC's design will allow it to start without a keyboard connected.


Thanks again.


P.S.
As an aside, looking though the Wiki it seems I have a fairly early motherboard but a late-type keyboard (membrane). Signs of past repairs I presume.
Charlie.

Are you pondering what I'm pondering? The Qube Server

Bryce

Hi Charlie,
         the first thing you need to do is undo whatever changes you made, you may have shorted the power supply with a stray wire somewhere. The power supply inside the CPC is pretty simple: 5V (regulated in the Monitor) supplies every IC in parallel, with an electrolytic capacitor at the start and local 100nf caps at each IC. The CPC will boot normally without the keyboard attached. (old version motherboards with new foil connectors were produced for a while, this isn't a repair / upgrade, mine is that way too).
The easiest way to check that the power on the motherboard is OK, is to coneect the power and measure the voltage across any capacitor or the power pins of any TTL IC (pin 7 and 14 on the 14pin ones or 8 and 16 on the 16pin ones). If you're not reading 5V here, then A) Your power supply isn't supplying anything, B) The power switch doesn't work, C) you've a short-circuit somewhere.
Speaking of important connections: I assume you have reconnected the blue 2way connector on the motherboard, otherwise the LED won't go on even if the CPC works.

Bryce.

arnoldemu

Did you try pressing delete or clr key?
from basic this makes an audible beep (turn up the speaker ;) ), or connect an amp and speakers to the audio out.
If it beeps changes are the cpcs working fine, just not showing anything.

My games. My Games
My website with coding examples: Unofficial Amstrad WWW Resource

Charlie

Thanks Chaps :)

@Bryce:
Thanks for the info, it will ease the checking process a lot. It's odd that there is such an eerie silcence from the machine, a lot like there's no power - especially as the LED doesn't light...
...I guess as you say a short (hopefully) is near the top of the list of possibilities.

@arnoldemu:
Delete or CLR key, eh? This shows up my lack of experience with these machines - I've kept pressing CTRL+G last night thinking 'B*m, it can't even be persuaded to beep...'

Right, I have some better ideas:
-Re-recheck the power & LED connections
-Desolder the ribbon cable & see if I accidentally shorted it
-Start testing voltages across a few caps/IC's
Charlie.

Are you pondering what I'm pondering? The Qube Server

Charlie

A further question:
Does this forum hand out any prizes for stupidity?


The reason I ask is having had a further quick fiddle this evening I noticed something a bit, well, basic:
You know how 6128's have a 12v & a 5v input?
Well, I got the 12v and power switch plugs the wrong way round - Oops, that will explain why the motherboard wasn't powering up when I switched it on.
:-[ :-[ :-[


So here's what my CPC does via my RGB->Video board when it's actually got power.

Apologies for the awful image quality - the output is actually clean 'n crispy. I had to use the camera's low-light mode because of the stupid energy saving light bulb in the room...
...oh, and the dark band on the right side of the screen? The telly started doing that a few days ago after it decided to 'auto-upgrade' it's software - no option to 'downgrade', grrr!


Hands up all those who see the problem here?


Two possible culprits:
-My board needs a re-design...
...surely not! I'll go check it with a known-working computer when I have time.


-The CPC has a problem with it's video output...
...I'm pretty sure this is the case as getting a red background (apart from cr*p image quality) is the reason I abandoned using SCART in the first place.


Yet more questions I'm afraid:
Assuming it is the CPC and not my rubbish PCB design skills where's the most likely place to go looking for such a fault?
I'm eyeing up the caps just behind the RGB connector at the mo' but as they are ceramic (rather than electrolytic) and look undamaged I'd be slightly surprised if one of them is on the blink.
If this is most likely to be 'cap-rot' and I need to go replacing them are all the ceramic ones 100nF? (Hopefully the electrolytic ones all have legible markings on them)


Thanks again.
Charlie.

Are you pondering what I'm pondering? The Qube Server

robcfg

As the text can be read, i'd say that you have the red and blue signals swapped.

Bryce

Good Morning Charlie,
          Due to the fact that it's a rather perfect red, I'd tend to agree with robcfg on this one, you probably have the red and blue signals swapped. Have you tried changing the border colour in basic to red to see if it's then blue?

If the problem isn't a swapped Red/blue signal,  then I'd check the capacitors you mentioned, C139 - C141 (all values are shown on the schematics), but it's unlikely that these are the problem, because if they were damaged, they wouldn't cause this effect, rather the colours would be dark and off colour, not such a nice red.

The next thing you should check is the resistors R131 - R133, if these are shorted you could have the blue signal leaking to the red line.

Unfortunately, if these resistors / capacitors are all fine (which is likely, because they rarely break), then it's the CRTC that's broken, unfortunate.

Bryce.

arnoldemu

Quote from: Bryce on 09:41, 26 January 11
Unfortunately, if these resistors / capacitors are all fine (which is likely, because they rarely break), then it's the CRTC that's broken, unfortunate.

Bryce.
I would say more the gate-array than crtc.

The CRTC doesn't generate the pixels or colour, the gate-array does that.
Looking at the image, the text is good, screen size is ok, the crtc seems ok to me.

I'm hoping it's just a swapped colour ;)
My games. My Games
My website with coding examples: Unofficial Amstrad WWW Resource

Charlie

Thanks yet again chaps. :)

Quick look at my board's schematic:
I labled RGB as BGR on the input header! Doh!
Thankfully no redesign required, just change the lables.

So do I get the 'stupid idiot' prize today?

Will hopefully have some fiddling time this evening - I hope that's all it is.
(For - Next loop with INK 0,<value> showed all 27 colours properly? Not sure)
Charlie.

Are you pondering what I'm pondering? The Qube Server

Bryce

Unfortunately there's no idiots prize, for the simple reason that we'd need thousands of them, because everybody has done something like this at some stage, and those who claim they haven't are liars. Mistakes are good, we learn from them.

Btw, the PCB board looks very good, although I would have chosen different (more reliable) contacts to connect the signals. Can you post a high-res picture of the screen, I'd like to see how well the AD724 performs. Although I can't find a supplier for that part here in Germany :(

Bryce.

Charlie

#12
Kind of you Bryce.


I had some trouble sourcing the AD724's. I eventually bought mine off eBay from a chap in Hong Kong.
I'll gladly take a better photo for you - it may be the weekend before I can take a camera-friendly daytime shot.
FWIW: I suspect the board could do with a little tweaking - though the IC is PAL/NTSC compatible things I've read suggest a generic board like mine doesn't give best quality in PAL mode...
...I read somewhere the 15pF cap in parallel with the crystal can be a hindrance - odd! I may well try without once I've tweaked a few obvious things like the long wires my S-Video plug is currently hanging off.

Speculation:
Another thought is 'my' board is really internal only for the CPC as it needs a 5v line. It should draw very little power...
...I wonder there's enough on the sync/lum pins to use a cap+diode as a 'charge pump' to gain that 5v..?
+Could then be a wholly plug-in adaptor for any who prefer
-May draw too much current from the CPC for comfort
-Dirty power may play merry hell with the output
Charlie.

Are you pondering what I'm pondering? The Qube Server

Bryce

Although the AD724 is a relatively low power device, there's not a chance that you'll get enough power out of the Sync + Lum to power that circuit. I also doubt that removing the 15pf Capacitor will improve the signal quality, but give it a try if you like.

You could still use your circuit externally by adding a pass-through from the 5V input on the CPC, it adds a plug and socket to your BOM, but it's then a "no-brainer plug and play" peripheral.

Bryce.

Powered by SMFPacks Menu Editor Mod