All RAM chips destroyed??

Nagging hardware related question? Post here!
martyn_hill
Aurora
Posts: 908
Joined: Sat Oct 25, 2014 9:53 am

Re: All RAM chips destroyed??

Post by martyn_hill »

Cristian wrote:Let me specify that I use a good quality RGB-Scart cable purchased from "Retrocomputing". Alternatively, I may use a normal antenna cable.
So can I assume it's fairly safe to try with the other ULA?
Not sure which ULA you're now speaking about, but presume you mean the ZX8301 :-)

Its _marginally_ possibly that you might fry a replacement 8301, even if the monitor/cable isn't causing excessive loading as the PAL components (+MC1377) also connect directly and _might_ be defective, but I personally wouldn't be be too concerned - especially as it's still possible to purchase another 8301 relatively cheaply.

Powering-up without the other ULA (8302) fitted should be your next test :-)


User avatar
Cristian
Aurora
Posts: 960
Joined: Mon Feb 16, 2015 1:40 pm
Location: Veneto

Re: All RAM chips destroyed??

Post by Cristian »

RWAP wrote:Hmm - I can't remember a QL where the HAL chip is not inserted in the socket
Yes, it looks so wrong but it's normal on iss5 boards. I had the same thought :-)


User avatar
Cristian
Aurora
Posts: 960
Joined: Mon Feb 16, 2015 1:40 pm
Location: Veneto

Re: All RAM chips destroyed??

Post by Cristian »

martyn_hill wrote: Not sure which ULA you're now speaking about, but presume you mean the ZX8301 :-)
Yes you’re right I used a shorter (but unprofessional) term to indicate zx8301.
So:
1) the resistor looks better in the reality than in photo. I measured it and the resistance matches.
2) I removed zx8302. Result: black screen. I put a different zx8302 from a working QL: the same white screen.
3) I also installed the original AH roms.
4) the oscillator pins were distorted but the solderings are OK (still, the oscillator may be faulty).
5) I took my (little) risk and installed another zx8301 from another working QL. Nothing changed: the same white screen.
6) Also I swapped IC 24 (bottom-right on the picture).
7) I noticed that the IC just under the RF modulator got relatively hot, so I swapped it, and nothing changed.
8) Maybe if I swap also the motherboard... :mrgreen: :|

I noticed that in the working QL the zx8301 stays almost cold, while in the white screen QL it becomes warm (not so hot).
Apparently the only chip getting clearly hot is the IC just under the RF modulator. The other chips behave as the ones in the working QL: some stay cold, other become warm (74LS245).


RWAP
RWAP Master
Posts: 2833
Joined: Sun Nov 28, 2010 4:51 pm
Location: Stone, United Kingdom
Contact:

Re: All RAM chips destroyed??

Post by RWAP »

martyn_hill wrote:
Its _marginally_ possibly that you might fry a replacement 8301, even if the monitor/cable isn't causing excessive loading as the PAL components (+MC1377) also connect directly and _might_ be defective, but I personally wouldn't be be too concerned - especially as it's still possible to purchase another 8301 relatively cheaply.
There are some ZX8301s left on ebay - I have run out... http://rover.ebay.com/rover/1/710-53481 ... =824&kw=lg


User avatar
Cristian
Aurora
Posts: 960
Joined: Mon Feb 16, 2015 1:40 pm
Location: Veneto

Re: All RAM chips destroyed??

Post by Cristian »

Thanks for your informatiion Rich, but fortunately my 8301 seems OK. I swapped it with other good QL and it works.
The minor problem of the grayscale UHF output should be related to the oscillator or IC28 (this might explain why IC28 gets got).
The main problem (white screen) remains. So, as next step I'll focus on oscillators...


martyn_hill
Aurora
Posts: 908
Joined: Sat Oct 25, 2014 9:53 am

Re: All RAM chips destroyed??

Post by martyn_hill »

Hi again Christian

We'll get there, I have faith!
Cristian wrote:1) the resistor looks better in the reality than in photo. I measured it and the resistance matches.
I assume you mean the zero-Ohm resistor/links JU2..4, measured at zero/short cct. Am I right? Or did you mean the resistor next to the CPUCLK oscillator?
Cristian wrote:2) I removed zx8302. Result: black screen.
Curious - I'll try this test myself shortly - had always thought you could get at least through the memory-test without the ZX8302, but could well be wrong about that!
Cristian wrote:I put a different zx8302 from a working QL: the same white screen.
There's something else at play then... I'm leaning more and more towards the 'stuck interrupt' possibility... I think I can test that as well - stand-by.
Cristian wrote:3) I also installed the original AH roms.
Were these EPROMs or ROMs? If EPROMS, you would have also needed to switch around the JU links and re-attach a 'flying-lead' to achieve any success with the picka-back EPROM arrangement that was typical when the AH was released.
Cristian wrote:5) I took my (little) risk and installed another zx8301 from another working QL. Nothing changed: the same white screen.
...I trust that the ZX8301 worked as expected when re-inserted in to the working QL... :-)
Cristian wrote:7) I noticed that the IC just under the RF modulator got relatively hot, so I swapped it, and nothing changed.
If that's IC28 - MC1377 - the PAL encoder, then yes, it's normal for this IC to get quite warm - more so than the other ICS...
Cristian wrote:8) Maybe if I swap also the motherboard... :mrgreen: :|
Its always an option (i always keep one 'spare'), but unless you have evidence of the replacement working before purchase, is likely to lead to more troubleshooting!
Cristian wrote:I noticed that in the working QL the zx8301 stays almost cold, while in the white screen QL it becomes warm (not so hot).
That matches my experience - whenever I've had a non-booting unit, the ZX8301 gets much hotter than usual and pretty rapidly! My hunch has always been that some 'bus-contention' is occurring in such a way that the ZX8301 is trying to drive a signal line (sink or source current) whilst another component is simultaneously trying to do the opposite...
Cristian wrote:Apparently the only chip getting clearly hot is the IC just under the RF modulator. The other chips behave as the ones in the working QL: some stay cold, other become warm (74LS245).
A warm bus-transceiver does sound suspicious - depends just how warm, of course - it could be normal. But now I'm leaning further towards bus-contention - both the ZX8301 and the bus-transceiver operate on the 'private' data-bus of the ZX8301 and need to be in-sync in terms of 'direction' - Pin 1 ('S') on the '245, Pin 4 ('RDWL') on the ZX8301 and all the other ICs connected to the CPU bus.)
Did you say that you had ever replaced the '245 bus=transceiver?

M.


martyn_hill
Aurora
Posts: 908
Joined: Sat Oct 25, 2014 9:53 am

Re: All RAM chips destroyed??

Post by martyn_hill »

...and another thing... :-)

I notice that this Iss5 board does not appear to have had one of the 'mandatory modifications' applied yet - namely the +5V by-pass line to the bank of DRAMs. Or am I going blind as I stare at your photo?

I've never been sure about how critical this really is, but understand that the DRAMs generate a lot of noise on their power & ground lines and so the modification is to help avoid sharing that noise with the rest of the board by tapping-off the +5V from near the 7805 regulator with a long, thick cable directly to the DRAM bank, connecting just under one of the keyboard-membrane sockets.

That's an easy (soldering) job, though you may need to get creative in soldering at the DRAM end - Sinclair inserted a pin through a via hole to make the connection.


User avatar
Cristian
Aurora
Posts: 960
Joined: Mon Feb 16, 2015 1:40 pm
Location: Veneto

Re: All RAM chips destroyed??

Post by Cristian »

1) R70, between 8301 and oscillator. Are you suggesting me to measure someone else?
3) OK, here my ignorance is greater, so to be sure I'll replace the MGI roms
5) yes, it still works :-)
7) in any case, I tried to swap also the PAL encoders
8) Yes sure, it would be the last desperate option


To be precise, the 8301 in the faulty QL becomes warm, certainly not hot. But in the woorking QL it stays cold.
Also I socketed and replaced the 74LS245 (IC21) with a new one. It becomes warm (not hot) like the one present in the working QL. Does your 74LS245 stay very cold?


User avatar
Pr0f
QL Wafer Drive
Posts: 1297
Joined: Thu Oct 12, 2017 9:54 am

Re: All RAM chips destroyed??

Post by Pr0f »

It really does sound like one of your DRAM chips is loading the bus. The fact that swapping the ZX8301 to another machine works, but it only gets warm in this machine... The piggy back method may not work in this case either, if the DRAM chip is holding the bus either high or low.

If you have a scope handy, you can look at the data lines on the 245 chip - the cpu data bus is on one side of the chip, and the DRAM/ZX8301 bus is on the other side. just by random chance, you would expect a varying cycle of 0 and 1 on each of the D0-D7 lines, so if one is always on or off, then you can narrow down to 1 pair of DRAM's - as it's possible the fault manifests even if the DRAM chip is NOT selected (RAS and CAS addressed) so could be in either bank - lower 64K (screen RAM area) or upper 64K.


User avatar
Cristian
Aurora
Posts: 960
Joined: Mon Feb 16, 2015 1:40 pm
Location: Veneto

Re: All RAM chips destroyed??

Post by Cristian »

Pr0f wrote:The piggy back method may not work in this case either.
The more I investigate this case, the more I feel it's "simply" a RAM issue.
I've just done a new test with Minerva Rom installed. I ran Ramtest_bas and entered the values. Which is the address value? The last one in the triplets? In this case, Ramtest_bas reports IC7 as faulty for all the triplets.
As you told, the piggy back method didn't work.
My electronics knowledge is too limited to understand and apply all your suggestions, but I might ask a friend.
On the other hand, my soldering iron is always hot and I still have one new Ram chip ready to install...

Here below the Minerva screenshots:

Image


Post Reply