exxos's Atari, Amiga & retro forum

stephen_usher wrote: 24 Sep 2023 09:35 The voltage is only "reversed" relative to today's standards. All the equipment I have with an external power supply with a similar plug from the 70s and early 80s has centre negative.

I'm guessing that it was changed so that you couldn't accidentally short the power if the cable was dropped onto something metal.

Yeah, I have a mix of PSUs from the era, some of it is centre positive, some centre negative, some centre positive and some Atari specific that are 5v/9v AC output. Then of course there was the ZX81 PSU that had a 3.5mm plug that was super easy to short.

Well, I swapped out D17 with the one from spares board, and the problem remains with the + board, and the spares board runs just fine with the original.

That means the only components left are the transistors. Resistors all test fine, electrolytic caps are new, C49 has been removed and the value checked, diodes have all been replaced or swapped with the spares board. I suppose my 1N4148's could be dodgy, but I've used them in other projects with no issues.

Both TR4 and TR5 are new and work fine with the spares board, so all that's left for me to do is do another continuity check between all the components to see if I've missed something.

The boost converters just take too much current, though they do work well with an adequate supply.

The spare board is working to a point. I was able to load a 16K game, but stability seems poor, but I've got a diag cart coming so fingers crossed that will move me on with that.

Time to warm up the multimeter in continuity mode :)

@stephen_usher

Yes, centre negative is normal for Spectrums and other oldies...

As you mentioned, the centre is positive now for safety reasons.

@rubber_jonnie

Just to be sure, did you check the polarity with a multimeter, as sometimes the previous owner may have swapped the polarity internally, i.e. in the PSU itself by reverting leads ?

HigashiJun wrote: 24 Sep 2023 14:09
@rubber_jonnie

Just to be sure, did you check the polarity with a multimeter, as sometimes the previous owner may have swapped the polarity internally, i.e. in the PSU itself by reverting leads ?

Yes, I have already checked polarity, plus the PSU works on my 48K rubber key Spectrum and it works for my spares board, so I know it's fine.

It isn't the PSU that is giving 70v, it's the power circuit on the Spectrum itself, it should convert the PSU input to 12v & -5v. I get -5v, but get 70v on the 12v circuit. It's not a PSU issue, as I said, that works just fine

Ok, so I have fixed the overvoltage problem , we're back to a straight 12v and not 70v and the machine is running stable, bar some issues with the colours being muddy, even when I fit a new LM1889. But that's fine, the main thing is that it's it's running unaided again.

And what was the problem? TR5 was the problem. I replaced the A733 transistor that was fitted with another A733. Simple right? Same orientation and all should be good, right?

Well no. Despite looking at the datasheets, and fitting the new part in the exact same orientation, the new part had a different pin layout to the datasheet I had and the collector and base were transposed, so TR5 wasn't working and TR4 was running wild and free, generating 70v.

I came to this conclusion, since despite replacing all the electrolytic caps with new, checking all the resistors were in spec, and using the spares board to swap out various diodes, I went back and buzzed out all the point to point connections and compared the pinout of the transistors I took using my parts tester, to the PCB connections using a simple multimeter.

Once I had the genuine pinout of the part in my hand, I soon realised that the base and collector were transposed, so I removed the part, and fitted a new one, transposing the pins. Suddenly I have 12v back and everything is good in the world :)

I learned a lesson though, that datasheets and the part in your hand may simply differ depending on who made the part. I've been fixing stuff a long time now, but mostly it's been IC's so theres never really been much in the way of transistor replacement, and usually it's been a 100% direct swap, so it isn't a snafu I've hit before.

From now on I shan't be trusting the datasheets to match the part in my hand without using my component tester to verify, as it's been invaluable, along with my multimeter, in getting to the bottom of this puzzle.

Just be aware folks, transistors can be tricky, and it's easy to get caught out by differences in pin layouts, it coaught me out twice during this adventure.

:sigh:

:dizzy:

Right, next I need to sort the Spares board, as it's almost a runner, I'm pretty sure it's memory problems. Fingers crossed the diag cart I have coming will shed some light on what's wrong and that one will be running just fine soon.

:bravo: :dizzy:

I've never known same transistor numbers having different pinouts. Could the replacement ones be a counterfeit part ?

Glad you have sorted this issue out !

:bravo:

exxos wrote: 24 Sep 2023 23:39 :bravo: :dizzy:

I've never known same transistor numbers having different pinouts. Could the replacement ones be a counterfeit part ?

Me either. I suppose counterfeit ones are always a possibility, but it does work as expected, and I bought them from a supplier I've used a lot, so it's a bit of a puzzle.

The spares board had a ztx213 fitted and that does have the same layout as the board, so I may just replace the A733 with one of those and then both transistors are to spec on the circuit diagram, but I know one thing for sure, I'll be checking any new transistors with my tester:


At least then I can be sure that I know which pin is which.

HigashiJun wrote: 25 Sep 2023 08:53 Glad you have sorted this issue out !

:bravo:

Thanks, bit of a puzzler that one.

In other Spectrum news, this arrived in the post this morning:


And here it is fitted to the original machine that started this whole post off.


Very happy with the results, new keyboard surround, new keyboard membrane, composite mod and case cleaned up.