ZX Spectrum stuck before Sinclair copyright message.

[Muttley Black]

Hello friends!

I trying to repair an issue 3B *modified* board, that stuck before Sinclair copyright message. (white border with black screen).

Here what i did already with no success:

I have replaced all capacitors, the transistors TR4 /TR5, the D12 *(r79)* 2k2 was burned so i put new one, i swap ULA / CPU /ROM from a working board, but not any change on the screen. Both lower and upper ram was tested on andruino tester. Voltages (-5,+12,+5,gnd) are ok, see also below. All six 74LSxxx are tested on IC tester and they are in working order. The only thing that i am afraid and i going to try it, is to replace the sockets with turned pin ones, cause i am sensing loose connectivity in some upper ram ICs.

Tip with +5 volts on pin5 of 74ls32n do be able to start the zx in 16k mod also didn't post something new on the screen.

Voltages:

Lower ram
PIN (1): -5,1volt
PIN (8): +12,39volt
PIN (9): +4,97volt
PIN (16): 0,00volt


My test gear:

Multi-meter (not decent one)
IC tester
Crystal tester
Logic probe
esr meter
logic analyser (just got it yesterday)
Retroleum smart card




I am keeping my fingers crossed for succeed repair!

[1024MAK]

Using your Retroleum SMART card, can you run the Diagnostics ROM program from it? It can perform a RAM memory test. This should narrow down the area of the problem.

The documentation is available from http://blog.retroleum.co.uk/smart-card-for-zx-spectrum/

****************
* Diag_ROM.vxx *
****************

Used to test for faults in the Spectrum itself. See docs
folder for full info.

* DiagROM needs to go in SLOT B if you want to start
it with the switch on the SMART Card.

Or alternatively, you can use Brendan Alford's ZX-Diagnostics ROM, which can also be used from the SMART card:

Brendan Alford's ZX-Diagnostics ROM

V0.33

As this is a system diagnostic ROM, ideally it should
be installed in slot B and booted by setting the SMART
card to Diag Mode. (DIP switch 3 on V1 SMART Cards).

(This is so that the SMART Card boots it directly
and does not have to go via the GOTO mechanism which
requires working RAM on V1 SMART Cards, something a
Spectrum under test may not have).

Go here for more info: http://git.io/vkf1o

Brendan has kindly given his permission for me to
include his ROM in the SMART Card project archive.

Mark

[Muttley Black]

Here is what i get with retroleum smart card. With Upper ram out from the sockets.

[media]https://drive.google.com/file/d/1agPhOP ... sp=sharing[/media]

[Muttley Black]

By the way, i perform one more test with upper ram socketed and i got the same results.

[Ast A. Moore]

Looks like an intermittent fault with the D4 line. Could be IC10, could be the CPU itself. Since you say you’ve already replaced the CPU, I’d replace IC10.

I don’t know what the Retroleum test board does, so I could only further diagnose the issue if you post a video of a standard boot process.

Swap IC10 with another one, say, IC8. Do the red vertical lines during the boot change to yellow sometimes?

[Muttley Black]

I did the swap as you say but nothing change. Also I test them on other board and I get normal screen. I did the same with cpu and ula and all work as expected. I suspect the sockets or the traces under them. I didn’t put them and I don’t have a clue what happened under there.

[1024MAK]

Which sockets were fitted when you got the board?

Mark

[Muttley Black]

None. The turned pin sockets you see in lower ram, actually are plugged on slide type board sockets. I did that to secure the connectivity in case that was bad.

[1024MAK]

The idea of using the diagnostic ROM in the SMART card, is that (1) it replaces the Spectrum’s own ROM, (2) it does not need any fully working RAM in the machine and (3) it can give us an idea of the fault. But the SMART card does have to be configured to ‘boot’ from it’s own ROM.

I’m asking which sockets were fitted when you got the machine, at that will then give us a clue as to what work may have been done on it before you got it.

A solder splash, broken or cracked track, or damaged through-hole plating can cause all kinds of problems.

Mark

[1024MAK]

Here is a photo of a typical issue 3B board, as used in 48K ZX Spectrums. As you can see, only the ULA is in a socket.


ZX Spectrums sold as 16K machines used the same boards, but the control chips for the upper RAM and the upper RAM chips were not fitted, instead sockets were fitted.

This is an issue 2 board with no upper RAM fitted, it shows the manufacturer fitted sockets:


Some boards had the Z80 CPU and the ROM in sockets as supplied by Sinclair. But by the time of the 3B board, these were normally soldered directly rather than using sockets.

Mark

[Muttley Black]

I got the board fully socketed. The only parts i replace was capacitors. I did that before anything else cause many problems start just from faulty capacitors. Also i double check the solder side with magnifier and all looks good.

I was just checked the voltages on transistors and i get this results:

What i only found is that the voltages on TR7 are faulty. I read in a pdf repair manual that TR7 must have the following voltages: 0.0v - 0.23v - 0.23v
I got: 0.0v - 0.35v - 9.09v

TR1,TR2,TR3,TR6 have good voltages.

[1024MAK]

I was just checked the voltages on transistors and i get this results:

What i only found is that the voltages on TR7 are faulty. I read in a pdf repair manual that TR7 must have the following voltages: 0.0v - 0.23v - 0.23v
I got: 0.0v - 0.35v - 9.09v

TR1,TR2,TR3,TR6 have good voltages.

TR7 collector is connected to the +9V line (not the +5V as shown in most schematic diagrams), so 9.09V for this transistor lead is correct. It looks like there is also an error in the repair manual.

If all the DIL/DIP chips were already socketed when you got it, then the possibility of damage on the top side of the board is real.

But don’t desolder anything for now. Let me think...

Mark

[1024MAK]

While fault finding, I strongly recommend removing all the chips associated with the upper memory that provide for the optional extra 32K bytes of RAM.

Remove:
IC15 to IC22 (DRAM chips)
IC23 (74LS32)
IC24 (74LS00)
IC25, IC16 (both 74LS157)

Power up, take note of the colour of the border and take a photo of the screen. Power down again. Repeat at least 10 times.
Is the border colour always the same colour? Which colour?

Is the pattern on the screen always the same, or is it different every time?

Mark

[Muttley Black]

We have results.

I did the test 10 times as you said. In every test mostly got new border color. First yellow then cyan, blue, black. The center screen was about the same as always.

Some photos:




At the 6-7 test i got this distortion moving border followed by distortion loud speaker noise.

[1024MAK]

Okay, so now we know that the CPU is not managing to execute the ROM code correctly, but the CPU is doing something. Further, most of the ULA appears to be working.

Even if there is no working RAM, when starting up, one of the first things the ROM code does is set the border to white. Then a memory check is done. But in your machine, none of this is happening.

I suspect either one address line is faulty, or more likely, one data line is faulty.

Do you have a printer that you can use to print out the schematic?

If yes, print it out.
A copy of the schematic is here http://www.retroisle.com/sinclair/zxspe ... matics.gif

Mark

[Muttley Black]

I just printed the schematic. I wish i had A3 printer!

So what i need to do is to check the connectivity from CPU, ULA and ROM address /data lines to each other? Or you suspect just some of this (A) (D) lines?

[1024MAK]

Start with testing each data bus line (D0 to D7) between the CPU and the ROM. You need your multimeter switched to a low resistance (ohms) range, such as the 200 ohms range or equivalent. Or use a continuity mode/range. Mark each line on the schematic as you do it with a tick.

The resistance of each line should be less than 10 ohms.

If no fault is found, do the same with the address bus lines (A0 to A14). Don’t worry about A14 or A15 at the moment.

Mark

[1024MAK]

I forgot to say, you need to test between the legs of the relevant chips. NOT to the solder joints or to the pins of the sockets.

Mark

[Muttley Black]

I forgot to say, you need to test between the legs of the relevant chips. NOT to the solder joints or to the pins of the sockets.

Mark

Of course! To be sure that each line is totally connected to IC Leg. I check the connectivity holding each multimeter probe at the top of ic leg.

By the way. I take the continuity mode (beep sound) option, cause i don't trust so much the accuracy of my multi-meter. Is that ok?

[1024MAK]

After carrying out the continuity tests, the next step is to test for short circuits. This is a bit more complicated.

You need to switch your meter to the 2000 ohm / 2k ohm range.

At the CPU, test each of these lines to each other:
+5V, 0V, D0, D1, D2, D3, D4, D5, D6, D7, A0, A1, A2, A3, A4, A5, A6, A7, A8, A9, A10, A11, A12, A13, A14, A15.

So for example, first test D0 to +5V, then test D0 to 0V, then test D0 to D1, etc...

I suggest you draw up a grid on a A4 piece of paper, with the above lines as both the columns and rows. Tick as you test each combination.

Report back any results that are less than 400 ohms.

Mark

[1024MAK]

By the way. I take the continuity mode (beep sound) option, cause i don't trust so much the accuracy of my multi-meter. Is that ok?

Yes, as long as you are aware of the limitations. There is no standard as to the resistance value when a test instrument or multimeter should indicate electrical continuity. Similarly there is no standard as to the resistance value when a test instrument or multimeter should indicate an open circuit condition. So you need to read the instruction booklet or manual to find out what it says or specifies for your multimeter.

Mark

[Muttley Black]

Hello! Back with results.

You can see the continuity test on photo below. Lines that marked with red pen are tested and work well.






About short circuits test, none measurements saw me less than 400 ohm. But need to say what different s i saw.

CPU

All (D) with +5v measurement saw me around 895 ohm
All (D) with 0v measurement saw me around 640 ohm
All (D) with all other pins measurement saw me infinite ohms
---------------------------------------------------------------------------------------------

A0 / +5V saw me 712 ohm
A0 / 0v saw me 1681 ohm
A1,A2,A3,A4,A5,A6,A7,A8,A9,A10 / +5V saw me infinite ohms
A1,A2,A3,A4,A5,A6,A7,A8,A9,A10 / 0V saw me infinite ohms
A11,A12,A13,A14,A15 /+5V saw me around 1025 ohm
A11,A12,A13 / 0V saw me 605 ohm
A14,A15 / 0V saw me 691 ohm
All (A) with all other pins measurement saw me infinite ohms
-----------------------------------------------------------------------------------------------

All measurement done with IC4, IC23, IC24, IC25, IC26 and all upper ram removed, as you see in the photo below.
I didn't test ROM IC for short circuits. Tell me if i need to.





Thank you!

[Ast A. Moore]

Remove all RAM and reset the machine a few times. What do you see on the screen? Is the border always white? If not, with all the RAM still removed, replace the CPU, ROM, and ULA with known good ones (one at a time at first), and retest.

[1024MAK]

Some resistance values above 600 ohms are expected, because various lines have resistors connected to them. In addition, the test current from the multimeter will flow into or out of the power supply rails and via the internal resistive paths in some of the chips.

Mark

[Muttley Black]

Remove all RAM and reset the machine a few times. What do you see on the screen? Is the border always white? If not, with all the RAM still removed, replace the CPU, ROM, and ULA with known good ones (one at a time at first), and retest.

I will do. We did about the same test with 1024MAK but all lower ram was pluged. So next test follows with all ram off the board in a sort of time. Want me to keep IC4, IC23, IC24, IC25, IC26 out of the board or socketed?