[48k 4A] Missing sync / empty screen

Post Reply
happycactus
Posts: 24
Joined: Tue Sep 28, 2021 6:16 pm

[48k 4A] Missing sync / empty screen

Post by happycactus »

Hello everybody!

My name is Federico and I'm from Italy. After many years I finally own a 48k 4A, I bought somewhere from ebay, but it definitely requires some fix.

This is the first time I fix a Spectrum, I have a M.Sc. in electronics so I know the basics... but of course, diagnosing and fixing a board is yet another movie.

Anyway. I powered the board on with a 9VDC wall PSU, it's a decent one and the voltages are pretty ok, well within the required ranges.

First I had to bypass the RF modulator, because the output was grounded and indeed I wanted to make the Composite mod anyway. So I disconnected +5V and Yin from the RF mod, desoldered the Resistor on the central pin of the output, and bypassed it with a 1000uF/63V capacitor (it's all I found actually). The image is probably present but it can't sync. Even without the 1000uF capacitor the image is similar.
FAXTaqGXMAQYCwN.jpeg
I checked the outputs of the ula at pin 15/16/17 (Luma / Chroma / sync) and they seems correct. I even can see the flyback signal at 50Hz:
Schermata 2021-09-28 alle 09.17.02.png
Schermata 2021-09-28 alle 09.17.29.png

I checked the clock and some of the bus signal, addresses seems ok and even \IRQ, \CS and other signals. The only thing I found is that the Data signals seems a bit slow rising. The falling edge is always ok but the rising edges seems slow and sometimes they don't even reach the VHmin level.

For this reason I suspect that there is at least one chip in the lower DRAM that is KO. Still i can't explain the missing sync on the Component output. I think I should at least see a border, but I can't.

So: how can I test the LM1889?

Side question: I was wondering if by removing the ULA and properly programming some pattern on the bus, I can test the different components (Hi/Lo DRAM, ROM, Address selectors....) Is there any project for something like that? Would it worth to have one?


Thank you in advance!

Federico
dessony
Posts: 342
Joined: Tue Oct 29, 2013 2:26 pm
Location: Indiana, USA

Re: [48k 4A] Missing sync / empty screen

Post by dessony »

I encourage you to find a schematic and a service repair manual for your computer. That is how I found my TS2068 service repair manual (SAMS Computerfact CC19). This manual is already posted here somewhere in the forum as a PDF file.

Thank you very much in advance.

Cheers,

DesSony
happycactus
Posts: 24
Joined: Tue Sep 28, 2021 6:16 pm

Re: [48k 4A] Missing sync / empty screen

Post by happycactus »

Hi DesSony!
Thank you for your suggestion, I already have both of them with a lot of other material.
I made some progress in the meantime.

First, I wanted to remove the huge heatsink, so I changed the LM7805 with a new switching module, same form factor. I also changed both the input and output capacitors with newer electrolitics (radial not axial unfortunately). Also, because of the switching nature of the new module, an higher ripple was present in the 5V output, so I added a smaller 100nF filter in parallel with the INPUT capacitor (yes, input).

I replaced a few other capacitor in the Power supply zone, so I was able to make a better 12V.

Finally, because the UYV component signal was present in pins 15/16/17 of the ULA, but the component output of the Component modulator module was probably very noisy, I thought it was worth trying to change the two amplifier transistors TR1 and TR2 with newer BC549 samples. In case this didn't fix, my plan was to change the LM1889 modulator chip.

After changing 2 times the TRs (because there was a 90% probability of mistakenly confusing Emitter with Collector!), it DID the fix.

Here is the result.
IMG_20211001_160040.jpg
This is an almost perfect border, but the black square and the fact that every 5 or 10seconds the board seemed to reset (no reset signal was triggered, though), I tried to isolate the expansion memory part by cutting the power supply trace as suggested in the service manual (IIRC). This is the NEW result:
IMG_20211001_160159.jpg
Next steps: socketing the ROM, trying with a diagnostic ROM to identify the failing RAM chips, if any.

Cheers ;-)
happycactus
Posts: 24
Joined: Tue Sep 28, 2021 6:16 pm

Re: [48k 4A] Missing sync / empty screen

Post by happycactus »

Ok a quick update.
I did a lot of tweaking, also due to my Daewoo monitor that definitely has some problem with CVBS signals, it can't properly sync but you must first select another input and then switch back to CVBS.
LabNation_Screenshot6.png
Anyway lately I changed TR4 (ZTX651), TR5 (BC557), TR8 and TR9 (BC550), and some capacitor. I found a very strange fix someone tried to do, C49, a 560pF cap to filter out high frequencies / noise from the -5V / -12V power rail, was changed with a 1uF capacitor. I'm waiting for some correct sample, in the meantime I changed it with a 100nF. Not great improvement. What I saw was a bad oscillation in the 12V. It's somehow reduced, who knows. Anyway.
LabNation_Screenshot0.png

After all this, the biggest thing I did was to modify the board to use a modern EEPROM (AT28C256) instead of the original ROM. I burned the Diagnostic Rom v56 from Retroleum.

That's the result. I don't know how to interpret. I can't hear anything from the speaker.
IMG_20211013_152257.jpg
Playing with the oscilloscope I found that D0-D9 are not good. They are very slow raising, I suspect all the lower bank ram is to change.

I'm planning to change it with the SRAM module.

Any comment, suggestion, hint? Should I give up?

Thanks in advance.
User avatar
1024MAK
Posts: 5102
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: [48k 4A] Missing sync / empty screen

Post by 1024MAK »

Give up :?: Definitely not. The ULA is working and so is the analogue video.

The photo you posted does not look like the output from the diagnostic ROM. Have you taken account of the fact that the mask ROM chip has three control inputs compared to EPROM/EEPROM chips which only have two control inputs?

Before you did any work on the ROM, when powering up, was the border always ending up white (sometimes looks light grey), or did you sometimes end up with random colours?

Does the Z80A have the correct clock signal at pin 6?

If yes, what are all the Z80 control pins doing?

Mark
ZX81 Variations
ZX81 Chip Pin-outs
ZX81 Video Transistor Buffer Amp

:!: Standby alert :!:
There are four lights!
Step up to red alert. Sir, are you absolutely sure? It does mean changing the bulb :!:
Looking forward to summer later in the year.
happycactus
Posts: 24
Joined: Tue Sep 28, 2021 6:16 pm

Re: [48k 4A] Missing sync / empty screen

Post by happycactus »

Hi 1024Mak, thank you for your response!

Well: sometimes, say 1 time over 8, the border is black instead of white. This happens seldom. But I have some update.

To implement the ROM upgrade I followed the instruction in the retroleum's blog. Unfortunately though I noted on my paper that AT28C256 required a different connection than the AT27C256 mentioned in the blog post, then afterwards I forgot these modifications and implemented entirely the AT27C256 scheme.

The difference is: on the EPROM pin 1 Vpp is A14 on EEPROM; and pin 27 A14 is really \WE. So for this EEPROM, 1 must be tied to GND instead of VCC and 27 must be tied to VCC instead of GND.

For this reason the whole system was STUCK.

After fixing this the system boots in diagnostic mode.
IMG_20211014_170658.jpg
What I see now is:

1) DiagRom doesn't find any Lower RAM failure.
2) DiagRom finds a problem in the UPPER RAM IC13
3) The screen seems to blink somehow, suggesting that there's something changing in the RAM, not sure how to interpret this
4) from previous measurement, there were noise in the DATA signals (all of them), signals seems to go high slowly and may trigger any invalid state, as if there were some chip that was trying to driving the level up. Tomorrow I'll capture some more signal.
5) there seems to be a lot of noise in the video as well, more than when the system was stuck; this seems to suggest that there is some noise injected from the power rail from the RAMs. Tomorrow I'll try to add 100nF capacitors to the bypass capacitors of the lower ram

I'm still waiting for a replacement keyboard to be delivered, then I'll perform all the other tests and see what happens. I'm really excited because I was pretty sure all the lower ram was dead, but it seems it's ok instead. I'll also try to remove IC13 and see if things improves.

Thank you! Kind regards.
User avatar
1024MAK
Posts: 5102
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: [48k 4A] Missing sync / empty screen

Post by 1024MAK »

The photo you posted clearly shows that there is a lower RAM fault. The lower RAM is chips IC6 to IC13, which are type 4116 or equivalent DRAM chips.

The screen display is taken from the lower RAM, so if any of the chips in positions IC6 to IC13 are faulty, it will affect the screen.

Each of IC6 to IC13, provides ONE BIT of each 8 bit byte.

If by blinking you mean that parts of the display flash regularly (and all at the same time), then that is the affect of the flashing attribute. This is bit 7, so it may be IC13 that is faulty.

Mark
ZX81 Variations
ZX81 Chip Pin-outs
ZX81 Video Transistor Buffer Amp

:!: Standby alert :!:
There are four lights!
Step up to red alert. Sir, are you absolutely sure? It does mean changing the bulb :!:
Looking forward to summer later in the year.
happycactus
Posts: 24
Joined: Tue Sep 28, 2021 6:16 pm

Re: [48k 4A] Missing sync / empty screen

Post by happycactus »

Uh, you are right, I mistakenly read "lower RAM error encountered: 0" instead of "[Lower RAM error encountered!]".
That makes perfectly sense.
IC13 is for D7 indeed.
Thanks, I'll try a swap.
happycactus
Posts: 24
Joined: Tue Sep 28, 2021 6:16 pm

Re: [48k 4A] Missing sync / empty screen

Post by happycactus »

So Ok, I received two 4116.

Removed IC6, socketed... but...

Everything went crazy.

Just to be sure, I tested the board before removing IC6 and it worked as usual.

Then when I put the new one... everything went nuts. The all the other dram chips started failing. I removed the new one as well as 3 other chips (IC11 to 13) and the diag rom now tells that the other 4 chips are failing EXCEPT the 4 removed one.

I measured the voltages Pin 1 is -4.8V (ok), Pin 9 is 4.8V (meh -- cheap DCDC modules, I just ordered some tracopower module) but pin 8 is around 10.2V (not well).

The strange issue is that on my other board I'm fixing the 12V PS is around 9.8V (!!!)... I guess either my multimeter is going nuts, or there's something I can't catch.
happycactus
Posts: 24
Joined: Tue Sep 28, 2021 6:16 pm

Re: [48k 4A] Missing sync / empty screen

Post by happycactus »

Update:

I checked the connection of IC6 and found that differently form all the other chips, pin3 (\WR from ULA) is not connected to pin3 of IC7-13. All pins are connected one-to-one, except for DI and DO (pin 2 and 14) that are connected to a Dx of the data bus. Likely while removing the chip the via was removed as well. it went unnoticed indeed (I checked visually but not tested with the multimeter).
But pin 3 of IC6 is connected to both pin3 of IC7 AND pin 4 of ULA, so while removing the chip all the other chips were isolated.

I reconnected Pin 3 of IC6 with Pin3 of IC7 and the defect was gone! Hurrah!

Now my 48k is working PERFECTLY. I reassembled everything, powered on and all the tests were OK.

By the way, with the good DRAM I noticed the screen image is much better! maybe the failing dram was injecting noise in the power rail.

last thing to do, change the DCDC converter (a cheap switching module) with a Recom/Tracopower one, and put back the original ROM image.

Thank you. Next Tuesday I'd receive the power supply and I'll post a screenshot of it!

Bye!
Post Reply