128k +2 Mysterious Videos issue.

A place to discuss the ZX80's and ZX81's younger brother - the Sinclair ZX Spectrum
User avatar
Erm
Posts: 187
Joined: Thu Jun 20, 2019 7:07 pm

Re: 128k +2 Mysterious Videos issue.

Post by Erm »

With my 'scope out I had a look at the ram address lines. The top row (IC17 - 24) all have nice square activity (the top row in the pic below). The bottom row (IC25 - 32) shows weirdness, not proper square waves and some that peak in the middle somewhere, neither high nor low (bottom line in the pic below).
+2OSC.jpg
So would I be correct in suspecting something is wrong with that bottom row. I found this schematic (which is not quite the right issue but is by far the clearest and largest I could find):

https://z00m.speccy.cz/images/ZXS128k-Plus2-sch.png

There is a 74LS157, which seems to be the only multiplexer that has anything to do with the bottom row of ram. Would it be a good idea to pull that and test/replace it? I think I'll try that, unless you or anyone thinks I'm barking up the wrong tree.
I'm Chris, a Brit in Germany (Chr$ on the QL forum).
User avatar
1024MAK
Posts: 3375
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: 128k +2 Mysterious Videos issue.

Post by 1024MAK »

The multiplexer for all the DRAM chips is IC10 (which goes by one of these part numbers depending on when it was made: PCF1306P, ZX8401, 40058).

Unfortunately new chips are very rare now. But there are PCB modules that perform the same function.

To make sense of the signals going to/from the DRAM, you need to include some control signals. The address and data signals are only required to be at valid logic levels at certain times. The rest of the time they may float or have irrelevant data on them.

For the DRAM, the control signals are /RAS, /CAS and /WR (sometimes /W). Note that on some diagrams and in some text L is added before or after the signal name to indicate that they are active low signals. That’s also what the forward slash is indicating.

DRAM requires the address in two parts, the row address (/RAS - meaning Row Address Strobe) and the column address (/CAS - meaning Column Address Strobe). For a read, the system puts the first part of the address on the DRAM address lines (in this case 8 bits of the address) and then takes /RAS low. Then it puts the rest of the address bits (another 8) on to the address lines and takes the /CAS line low. If the /WR (write) line is high, the DRAM will after a short delay, put the data held at that address onto the data lines.

But that is just one of the a available cycles. There is also a refresh cycle from the Z80 and of course there is a write cycle as well. Also don’t forget that one set of the DRAM is being read by the ULA. This does a normal read for the first byte, but then does another read while keeping /RAS low. It changes the column address and then takes /CAS low again. This is quicker than doing two separate read cycles, it’s known as page mode reading.

Mark
User avatar
1024MAK
Posts: 3375
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: 128k +2 Mysterious Videos issue.

Post by 1024MAK »

Although it is a different version DRAM chip, this datasheet has useful English descriptions of the various control signals and access cycles.

I need to point out that there are two types of 4164 DRAM chips, those with 7 bit/128 cycle and those with 8 bit/256 cycle refresh. Z80 systems normally use 7 bit/128 cycle types because the Z80 only produces a 7 bit/128 cycle refresh address.

However some later DRAM chips can self refresh...

Mark
User avatar
Erm
Posts: 187
Joined: Thu Jun 20, 2019 7:07 pm

Re: 128k +2 Mysterious Videos issue.

Post by Erm »

1024MAK wrote: Mon Nov 23, 2020 1:33 pm The multiplexer for all the DRAM chips is IC10 (which goes by one of these part numbers depending on when it was made: PCF1306P, ZX8401, 40058).


Mark
Aha. I do actually have a broken issue 6A 48k board that had an issue and I've been using as spares. It has it's PCF1306P present, but obviously I don't know it's 100% good. I'll remove that, remove the 40058 in the +2 (pop a socket on) and then try the PCF1306P in its place. If that does nothing then I have it socketed and the option of getting a modern replacement.

And after that I shall try and get my head round all that other stuff you just wrote :o :?
I'm Chris, a Brit in Germany (Chr$ on the QL forum).
User avatar
Erm
Posts: 187
Joined: Thu Jun 20, 2019 7:07 pm

Re: 128k +2 Mysterious Videos issue.

Post by Erm »

Fixed! It totally was the 40058. It now has the PCF1306P fitted... which is of course even older than the original, but that's fine, there's a socket in case that dies.
I'm Chris, a Brit in Germany (Chr$ on the QL forum).
User avatar
1024MAK
Posts: 3375
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: 128k +2 Mysterious Videos issue.

Post by 1024MAK »

Good to hear that you are up and running :D

Mark
User avatar
Erm
Posts: 187
Joined: Thu Jun 20, 2019 7:07 pm

Re: 128k +2 Mysterious Videos issue.

Post by Erm »

Yes, it's nice to have my main Speccy going again with perfect picture etc. I even oiled the tape mech, it's not perfect but loads and saves well.

I like the grey +2 because:

1. They're cheap compared to the original 128k.
2. Microdrive compatibility (unlike all the later versions).
3. Cassette deck built in (saves space if you ever need to use a cassette but does have the drawback that you can't load from other audio sources without modding).
4. 128 Basic option, so you don't have to remember key combinations and cuts down on finger gymnastics.
5. It works with my QL-VGA so I can just switch leads between QL/+2 and use the same monitor, which again saves space.
6. Uses same easy to find/buy PSU type as every other previous Spectrum.
I'm Chris, a Brit in Germany (Chr$ on the QL forum).
Post Reply