Issue 4A, TR4 gets hot, coil damaged?

[shred]

Hi!

I'm trying to repair an Issue 4A board. When I got it, 5V was OK, but at the 12V and -5V lines were zero volts.

I checked the coil in-circuit, and found a resistance of a few ohms between the windings. So I unsoldered the coil and checked it, but then it seemed ok. I measured almost 0 ohms at the windings, and nothing between the windings.

So I soldered it back in again, and replaced TR4, TR5 and D16. For TR4 I used a ZTX651, and for TR5 I used a BC557B (facing in the opposite direction). On other Speccys with a broken power converter, these replacement types worked fine.

Now I'm getting 9V on the 12V line, and about -4V on the -5V line. Better, but still not good. Also, the new TR4 is getting pretty hot.

I suspect that the outer coil might be short, and drawing too much current from TR4. But before I unsolder and rewind the coil for nothing, I would like to hear your opinion. Did I miss something?

Thank you!

SOLUTION: One of the lower RAMs got very hot, too. After replacing it, TR4 stayed cold and 12V was back. After replacing two more broken lower RAMs, and soldering in a missing C47, -5V also reached -4.1V. It is still not an optimal value, but it is acceptable for me.

[1024MAK]

It’s more likely that one or more of the 4116 (or equivalent) DRAM chips (IC6 etc.) is damaged and drawing more current from the +12V rail than it should.

Are any of them running hotter than normal?

Mark

[shred]

Yes... IC8 got hot.

I replaced it. The 12V is back now, and TR4 isn't getting hot anymore.

On the -5V line I still get -3.8V though. No other lower RAM chip is getting warm, so maybe it's the DC converter this time.

[1024MAK]

What are you getting across diode D11?

And what are you getting across the 5.1V zener diode (D19, maybe labelled as D16)?

Also test the voltage across resistor R29 or R79 (depending on which issue board you have)?

Mark

[shred]

It's my first Issue 4A board, and I struggle with matching the components to the schematics. For example, the zener diode is populated at R54.

D11 is at the correct position, and I measured 6.2V there

I found D18 is at the position of R55. There I measured 7.5V.

I have already replaced the zener diode because I always do that on a broken power converter, so I assume it is good. It's placed at R54, and I measured the -3.7V there that I also do on the RAM chips.

I assume that R29/R79 is populated at the position of D12 (labelled "12" on the board). I'm getting a voltage drop of almost 10V there.

I was hoping that D11 or D18 were broken, but a diode checker says they're good. So now it looks like there is another broken lower RAM that draws too much current on the -5V line. They're not socketed, so I guess I'll have to remove them one by one until the -5V is back.

[1024MAK]

Ah, sorry, I meant to return and post this link which shows the final version of the DC/DC circuitry. Sinclair changed various parts over the various board versions. Hence the schematic, silk screen (layout) and what was populated on the board were not always the same… (compare the layout for an issue 3 with an issue 5 for example). Layouts available here but also various other sites.

I do recommend that where members can, that they upgrade the circuitry to the version in the link above. But it does mean working out what they actually have on their board version to start with.

The resistor that is the series resistor feeding the zener (well, for the later versions of the circuit), R29/R79/D12, what value is it? is it 2.2kΩ?

The current used by each 4116 on the -5V should be fairly low (200µA max.), as this is just a bias supply. If you click the link in my signature below for ZX81 Chip Pin-outs, one entry in that topic is about 4116 DRAM chips.

I do recommend upgrading the DC/DC converter circuitry as per the link above, as the earlier versions do not generate a within specification -5V rail voltage (even when the computer was brand new).

If this does not help, then if you don’t want to remove each 4116, the other method is to cut both the -5V AND the +12V pin on each chip (pins 1 and 8) . If you are careful, you can do this in such a way, that you can join the cut/break with solder if that particular chip is not a problem. It’s still a process of elimination. With the pins isolated, a resistance test on pin 1 to GND and pin 8 to GND may pick up a faulty chip.

Anyway, -3.8V on the DRAM is enough voltage to leave them powered up for short periods say two to five minutes) for testing purposes. Does the computer get to the copyright screen? Is the border always white after powering on (test at least ten times)?
If you don’t get ro the copyright screen, and the border is always white, what image do you get? I ask, because the pattern on screen may help to work out which 4116 chips are still faulty.

Do you have a diagnostic card or a diagnostic EPROM chip or other expansion/interface that contains a diagnostic ROM?

Mark

[shred]

First of all, thank you very much for your help!

Since the voltages were more or less on spec, I ran the diagnostic module. It said that IC9 and IC10 were faulty, so I replaced them as well. I later checked the extracted 4116s with a chip tester. They are indeed completely dead.

The RAM test of the diagnostic was green after that. On the -5V line, I am reading -4V now. Since the other RAMs seem to be good, I wonder if the -4V comes from the DC/DC converter that is not modified according to your link yet, or if it's worth unsoldering the remaining 5 RAMs.

The next problem seems to be a broken ROM. Diagnostic says "Unknown or corrupt ROM... AD01". When I boot the machine, it just shows a black paper on a white border. Even the red lines from the initial RAM test are missing. Unfortunately the ROM is soldered directly to the board, so I can't just swap it with one from a working machine. But that's a different problem.

[shred]

Oh, I forgot...

The resistor that is the series resistor feeding the zener (well, for the later versions of the circuit), R29/R79/D12, what value is it? is it 2.2kΩ?

Yes... It's 2.2kΩ. I checked it.

[1024MAK]

In the service manual, the official tolerance for the -5V rail is -5.5V to -4V. This is because of the strange earlier design of the circuitry.

ROM chip failure is not very common, but does sometimes occur (especially if the machine suffered from edge-connector abuse in the past).

It’s a good idea to check that the ROM links match the type/make of ROM fitted. These are the four positions marked H and N (left hand side of the board near the speaker). H links are made for Hitachi (or equivalent) ROM chips. N links are made for NEC (or equivalent) ROM chips.

A test that can be done, that may help, is to disable the ROM and see what happens. The data bus should float high, hence the Z80 should execute 0xFF which is RST 38h. This results in this display https://youtu.be/ve5zWArAmi8

To disable the ROM, connect edge-connector /ROMCS pin (25 lower/underside) to +5V. This is most easily done on the appropriate leg of resistor R33 (680Ω) but do double check you have the correct end! It’s shown on the schematic as the left leg, but you should definitely confirm this first.

BTW a good copy of the service manual is here (from the spectrumforeveryone.com website).

Mark

[shred]

I checked your link with the DC/DC converter modification, and found that it has already been done on that board. Except of C47, which was unpopulated.

I soldered in a 1µF cap there, and now the -5V rail gives about -4.1V (with a peak of up to -4.3V).

In the service manual, the official tolerance for the -5V rail is -5.5V to -4V. This is because of the strange earlier design of the circuitry.

Okay, so I think I will keep it like that. I don't want to take the risk of damaging the RAMs or the PCB while unsoldering the remaining ones.

ROM chip failure is not very common, but does sometimes occur (especially if the machine suffered from edge-connector abuse in the past).

I unsoldered the ROM and soldered in a socket. Unsurprisingly, the ROM still doesn't work in the socket. However if I press on the bottom half of the chip, the red lines for RAM initialization appear, but then the machine freezes again. So it seems to be a mechanical issue.

It’s a good idea to check that the ROM links match the type/make of ROM fitted. These are the four positions marked H and N (left hand side of the board near the speaker). H links are made for Hitachi (or equivalent) ROM chips. N links are made for NEC (or equivalent) ROM chips.

As the ROM was soldered in, I assume that the links are correct. It's an "AMI 356-201802 AUSTRIA" type, and the links are set to NEC. I found another thread with that ROM, and a photo there also shows the NEC configuration, so it seems to be correct. The ROM stays cold, by the way.

The Diagnostic Board and my ZX Dandanator work perfectly, so I guess the ROM is the last problem to solve. I will probably do an EPROM mod and use an EPROM.