Help needed with a unstable +2 grey

gseb
Posts: 18
Joined: Mon Jun 08, 2020 1:22 am

Help needed with a unstable +2 grey

Post by gseb »

Hello,

I am attempting to repair a +2 grey motherboard issue 3 that is not running stable at all.....
It boots to the selection screen, I am able to choice "load from tape", it load games from tape but always end up with an error.
Same with the divmmc, it also always end up with error when loading games and there is graphics glitches in the divmmc selection menu.
I suppose the 2+ has one or some broken memories ICs.

The motherboard is not really in a good shape. What I did till now :

- The rom, CPU are on Sockets. The sockets were so bad that I replaced them with new one,
- I did the same with power jack,
- I had to deeply cleaned the edge connector to get the divmmc to work,
- Also revert TR4...

I am thinking to put a GAL instead of the old PAL, I do not know if it can be the problem.... But I think it's a good things to do.

I am using the +2 with a new PSU from here :

https://www.thefuturewas8bit.com/sinclairpsu.html

I did some measurement on the power rails :

- Directly on the regulator, 4.87V and 8.53V
- the 12V rail has 11.55V, the -5V rail has -5.13V and the -12V one has -13V

I tried to load a program to test the ram from the divmmc but it failed. It's the test ram.tap from Brendan Alford.
It failed with an error at the start probably because it's starting it code from memory. See picture. You can see some graphical problems, some bigs blue dots...

As my rom is on Socket,I would like to start the test program as a replacement rom but I could not find how to prepare a 27C256 with it.
I just programmed a 27C512 with the testrom.bin from te package (I doubled the files as I only have 27C512) but the +2 does not boot the test program, instead it shows the same typical screen as if you boot the spectrum without a rom inserted.

Could someone explain me how to prepare a working Rom with testrom.bin and tell me if the power rails value are OK ? I am not sure about the -13V !

Thank you for reading and help me.
Grey.jpg
User avatar
1024MAK
Posts: 5103
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: Help needed with a unstable +2 grey

Post by 1024MAK »

The voltages are all within specification.

The +5V should be within +/-5% (so +4.75V to +5.25V).
The rest should be within +/-10%

The +2 grey uses a single 32k byte ROM chip or equivalent EPROM (27256 or 27C256).

You should program a 27256 or a 27C256 with the test ROM image from 0x0000 to 0x3FFF. Then the same test ROM image 0x4000 to 0x7FFF.

If using a 27512 or a 27C512 it should be programmed 0x8000 to 0xBFFF and from 0xC000 to 0xFFFF.

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.
gseb
Posts: 18
Joined: Mon Jun 08, 2020 1:22 am

Re: Help needed with a unstable +2 grey

Post by gseb »

Thank you very much for the reply.

I did as you wrote and I got the testrom working in the rom socket.

I first tried it without a keyboard and because the testrom was not able identify the Spectrum model I have, it automatically started the test as a 48k machine. See first picture. It passes the test successfully....

Then, I connected a keyboard and choose +2 model. Now it gives mistakes. See pictures.
I think it's time to replace the PAL by a GAL... Anyway, I wanted to do it. Will start with the original equation equivalent for the GAL first. Give it a try, then if the problem is still there I will replace the 174.

Is there something broken with the (/M1 bad ?) result from the testrom ? I read in internet that this signal from CPU is needed to get the dimmc working. The divmmc is working on my +2.....

Those graphic corruption are not a sign of a bad ULA ? I will also replace the ULA socket I think, just in case....

Here still one more picture from the motherboard...

IMG_20200608_213842.jpg
IMG_20200608_221931.jpg
IMG_20200608_011433.jpg
User avatar
1024MAK
Posts: 5103
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: Help needed with a unstable +2 grey

Post by 1024MAK »

Please ignore the /M1 report. The diagnostic test was originally designed to run on an external diagnostic board which includes some hardware so that the Z80 /M1 line can be tested. If the diagnostic board is not present, it prints the report you see, telling you that it can’t detect this board.

The screen corruption is indeed indicating that there is a problem. It could be one of three things, a RAM chip that is malfunctioning, a connection problem between the relevant chips, or a faulty ULA (although at the moment I don’t suspect the ULA chip itself, it doesn’t ‘smell’ like a ULA fault).

The 128K paging tests failed could also be due to a RAM chip failure, a connection problem as well as IC6 or IC7.

The above symptoms could be caused by a single fault, or may be due to multiple problems, it’s hard to say :(

Before changing anything else, using a multimeter on the 200 ohms resistance range, or on the continuity test range, test the address lines DMA0 through DMA7 on ULA (IC14) pins 4 to 11) to DRAM IC25 to IC32 address pins A0 to A7. Then test to resistors R70 to R76 (all 470Ω), and R67 330Ω) for DMA7. Then test from resistors R70 to R76 to the relevant IC10 pins. And from R67 to IC16.

Next test all the IC7 pins to their destinations.

Other possible causes are IC10 and IC16.

This fault could be a bit tricky to track down :(

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.
User avatar
1024MAK
Posts: 5103
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: Help needed with a unstable +2 grey

Post by 1024MAK »

Will the machine run BASIC programs?

If yes, can you please try this:

10 CLS : BORDER 2 : INPUT “speed”,s
20 FOR y=0 TO 21
30 FOR x=0 TO 31
40 PRINT AT y,x;CHR$ 134
50 PAUSE s
60 PRINT AT y,x;CHR$ 137
70 PAUSE s
80 PRINT AT y,x;CHR$ 32
90 NEXT x
100 NEXT y
110 GO TO 20

When you RUN it, enter a speed value between 1 and 100 (higher is slower) and see if the screen corruption appears and more importantly, does it change as the program runs.

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.
gseb
Posts: 18
Joined: Mon Jun 08, 2020 1:22 am

Re: Help needed with a unstable +2 grey

Post by gseb »

Tank you very much for the help.

I already replaced the PAL by a GAL16V8-15 before I saw your message. I finally used the new gal with the wire going from gal pin 11 to CPU pin 28. I thought it's done and won't have to do it later....

Does the romtest always work and get the same "valid" result with this change ?

I did more repeatedly testing after to have changed the PAL by the GAL. I should have done this before too.

I ran the romtest in soak mod. Only in 48K mode available it seems. Sometimes it passes all test, sometimes not. See picture.
I ran the romtest several times in +2 mode. Sometimes it shows the same error as on the picture in my previous post and sometimes it gives another error screen. See 2nd picture.

I will do the measurement you asked with the gal installed if it's not a problem. And also test if I can get the Basic program to run.

Here the last testrom pictures.

IMG_20200609_111317.jpg
IMG_20200609_110418.jpg
gseb
Posts: 18
Joined: Mon Jun 08, 2020 1:22 am

Re: Help needed with a unstable +2 grey

Post by gseb »

I checked all. Except one thing that I can't find. The PAL pin 9 goes to a (19B)... I can't locate (19B)
Otherwise all connections are there and all resistors have the good value.

I still did not tried the basic program. I am fighting with the keyboard... I never used the Spectrum to write something, I only played with my Plus....
The automatic writing of command is annoying :cry: On the 2+ all those automatic things are not written on the keys..... I have to learn how it's working first.

I could remove IC6 from the board and test it with my TL866 (I think it can test it) but I think the testrom should be first OK in 48K mode.
Or do you think a "broken" IC6 could produce mistakes in the 48K test ? I could also test IC15 and 16 even replace them by new one. It cost almost nothing and it can be quickly done.... Hopping IC10 in nit the bad one....

This may be better to start with this, it's less work than to replace all the 48K memories...
User avatar
1024MAK
Posts: 5103
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: Help needed with a unstable +2 grey

Post by 1024MAK »

gseb wrote: Tue Jun 09, 2020 10:44 pm The PAL pin 9 goes to a (19B)... I can't locate (19B)
IC7 (PAL) pin 9 is an input, it’s the /WR (also written ‘WRL’) line from the Z80 CPU (IC9) pin 22. This line goes to various other places including the edge connector contact 19 on the component side (which is what 19B could be referring to). Although the references to the edge connector are inconsistent in that the component side is sometimes referred to as the ‘A’ side and sometimes as the ‘B’ side.

Amstrad (who bought the computer section of Sinclair Research Ltd) did not print all the keywords on the +2 keyboard because (1) it was expected that the majority of users would only be playing games, and (2) if you did want to use BASIC it was thought that users would prefer the 128 BASIC editor where everything is typed character by character.

If trying to use 48K BASIC, I suggest having a photo of a ZX Spectrum+ or of a ZX Spectrum 128 keyboard in front of you and reading the original ‘rubber key’ manual.

The program should work in either 128K BASIC or 48K BASIC. However it’s not vital to run it, I was hoping if you could get it to run, it may give a clue if the corruption on screen was related to the other data on the screen. If you can’t enter it, or can’t get it to run, don’t worry.

IC6 (74LS174) is not likely to be the source of all these problems. If you test the output pins (2, 5, 7, 10, 12, 15) while the computer is in 48K BASIC, they should all be stable. That is, they should all be at valid logic levels and not changing. Although I have not tested this myself, I think pins 12 and 15 should be high with the others being low.

In any other mode, these pins will change state depending on which part of RAM and ROM are being used. However pins 10 and 15 should be low.

I think it’s unlikely that IC15 or IC16 is faulty. If IC15 was faulty I would not expect the machine work as well as it does. However, it’s possible that any of IC6, IC15 or IC16 do have faults.

The symptoms displayed do indicate that there are either multiple faults, possibly including poor connections (dry solder joints, cracked PCB tracks, bad socket contacts) or something strange going on with the RAM.

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.
gseb
Posts: 18
Joined: Mon Jun 08, 2020 1:22 am

Re: Help needed with a unstable +2 grey

Post by gseb »

Thanks very much for all the details you explained.

128 mode is totally unstable, unusable. It reboot on 48Ko... But I will try again.
It's why I went to 48Ko. I will get the program basic in. I will type it on my plus and save it on divmmc to load it on +2
If it does not work like this, I will type it on the +2 with the plus near to me to have a look at the keys.

I will replace IC 15,16 and 6 just to be sure this is not coming from there, that's something quickly done. Also check the output pin of IC 6.
If not OK, I will put socket on the ram and try with others. I have ram here already from other project so I have no problem to do it.
If this does not work, I may refresh all solder point. If always not working I will look for a board replacement, a cack problem will be too hard to solve.....
It's easier when something does not work at all than this kind of problem :-(
Anyway, this motherboard unfortunately does not look really nice.
gseb
Posts: 18
Joined: Mon Jun 08, 2020 1:22 am

Re: Help needed with a unstable +2 grey

Post by gseb »

IC6, IC15 and IC16 tested on my programmer and they are OK.
As I have no CPU, ULA nor 40058 to swap and see if it would be better, I will next solder sockets and replace all the rams... Last try before I start to look for another motherboard.....
Post Reply