View Full Version : ZX Spectrum Power Transistors

September 19th, 2014, 02:25 PM
So after careful selection of a power supply (I found one rated for 10v 3a which was the closest to the original 9v 1.2A I could get) I thought I was all set but no, Sinclair decided it was a great idea to reverse the pins on the DC jack and not include a protection diode so after several cautious seconds watching my display to show something there was a pop and the 7805 blew.
I replaced that and switched the polarity on my AC adapter and now while the new 7805 has a healthy +5 out all I get is a white screen. Unlike Commodores where there's piles of documentation on how to troubleshoot everything there's almost nothing I can find for the Spectrum. I can't even get straight answers on where good voltage test points are. Best I can be told is that either the ZTX650 or the ZTX213 has also blown but I can't determine from testing if they are healthy or not, or if my problem is something completely different.

In diode check mode my DMM reads nothing between collector and emitter both ways, a reading between of .057 between collector and base both ways (that MUST mean it's blown) and nothing both ways between base and emitter on the 650.

The 213 reads nothing between collector and emitter both ways, a reading of .680 one way between collector and base and a reading of .681 one way between base and emitter.

September 19th, 2014, 11:07 PM
From your readings I would say that the ZTX650 has gone to the great repair shop in the sky - and the ZTX213 is OK.

First of all - have you obtained the correct schematic for the Spectrum? if not - try http://zxspectrum.it.omegahg.com/doc/48ksm.pdf - this contains the service manual...

If you look at the schematics you will find an internal +12, -5 and -12V power supply derived from the +5V rail. This is where your two transistors are. I would check these voltages with your DMM. If they are correct - then your transistors are probably OK despite your testing. If not - the ZTX650 is probably a good candidate for replacement.

Note that there may be some errors in the manual. I have found different schematics list +12V or -12V for one of the derived rails!


September 20th, 2014, 08:16 AM
Having had a quick read of the service manual - page 4.4 paragraph (h) identifies a significant problem with the failure of TR4 (your ZTX650) and recommends a circuit improvement for issue 1 boards. There are another host of 'improvements' identified...


September 22nd, 2014, 04:46 AM
Also expect a failure of the lower RAM (the bottom left row of 4116). One of more ICs must be fried...
Beware the transistor used as replacement of TR4, it needs to support a pretty high current (2Amp); so you can easily blow the new one. The best replacements are from Zetex (obviously), the ZTX651 seems to be still in production.

September 22nd, 2014, 04:47 AM
I'm sorry about the "pop". I always double-check on the machine, Internet and if it can't be determined, by opening it up and check traces on the motherboard to identify polarity. Voltage is harder to determine based on trace thickness. :)

For reference, the ZX Spectrum, Nintendo Famicom (*), Sega Master System, Genesis/Mega Drive and a few more consoles (NEC PC Engine I believe, but don't quote me on that) all take centre negative polarity, about 9-10V DC and approximately the same number of ampere. Thus within that group, the power supplies tend to be fairly swappable.

Also, a friend of mine yesterday brought out 8 various ZX Spectrum 16/48K from his collection. We got 2 of those to boot, of which one with malfunctioning keyboard and the other with a horrible RF picture, far worse than the other Speccys that only displayed static garbage using the same antenna cable and TV. So you shouldn't be too surprised it broke down, they seem remarkable fragile (or cheap design perhaps).

(*) The US/European NES on the other hand prefers 9V AC but just like many (but not all) devices that take AC it can be fed with DC, at least for a shorter while. It has a rectifier inside which over time supposedly gets hot when the current never alternates.

September 22nd, 2014, 08:03 AM
The machine ran and worked before it was boxed and shipped out. Any problems now will because of the voltage reversal. That I can simply tell with a scope and DMM right now (and the two transistors removed) we have a clock, the speaker and its transistor is not blown, the system is reading from ROM and the ULA jumps two different states before sitting there, possibly because the first bank of ram is not powered and has no idea what to do next.

Seriously. F**k Sincalir. Why is British engineering so garbage? What the hell would they of achieved by reversing the plug polarity compared to everything else out there?? I'll be replacing the two ZTX transistors anyways with much better TIP31C and TIP32C transistors. They aren't directly pin compatible but they can just be glued somewhere on the board and wires run between them and the holes.

September 22nd, 2014, 11:08 PM
Seriously. F**k Sincalir. Why is British engineering so garbage? What the hell would they of achieved by reversing the plug polarity compared to everything else out there??
You had better ask that question of many other 'portable' makers as well.
Take a look at the list [here (http://support.radioshack.com/support_accessories/doc63/63880.htm)].

September 23rd, 2014, 01:24 AM
Interesting list, and while it'd be off-topic for this discussion I have also asked myself which is better: positive or negative polarity. There ought to be a reason, other than confusing for the buyer. I suppose the surface area of the tip is smaller than the barrel, but do some applications need a bigger area for ground/return while other need it for incoming voltage?

Besides, if you ever come across an Oric-1 or Atmos, it has center positive polarity so there you have a case of British engineering that adheres to your expectations... at least when it comes to the power supply, perhaps it would be a let-down in other aspects.

September 23rd, 2014, 10:04 AM
Coming from the UK it is not British Engineering that is garbage - but the penny-pinching bean counters that insist that we don't spend a penny/cent more on the product than they think we should.

I am afraid Sinclair's engineering left a bit to be desired - but he did come up with some cost-effective computer kit (ZX80, ZX81 and the Speccy) that caused a whole generation of kids to become interested in computers in one form or another when a vast majority of the equipment was just too expensive for parents to buy their kids for Christmas. But (as everyone remembers) they suffered from the wobbly RAM pack syndrome and (as NeXT has found out to his cost) a lack of inbuilt protection that should have been there really.

One of the basic problems is that there is no accepted IEC/ISO Standard for power supply polarity. If there was - everyone would adopt it. There isn't - so everyone 'does their own thing'. It is interesting to note from the referenced list that HP is one polarity whereas Compaq is the other - but HP owns Compaq...

There is no 'better' way. The electrons flow from the negative pin of the power supply to the positive pin so the same current flows through both the centre pin and the outer sleeve. The current density may be different (due to the cross-sectional area of the metal) - but this will be the same irrespective of whether the centre pin is positive or negative.


September 23rd, 2014, 11:25 PM
One of the most useful mods that can be done on a Spectrum is the protection diode. I can't remember the details; but it simply involved cutting a track in the back side and installing the diode.

The Spectrum is a cheap computer indeed; but as daver2 was a revolution in its own. IMHO, even worse than the polarity flaw it's the keyboard membrane, extremely prone to fail.

October 22nd, 2014, 08:51 PM
Yes. The diode has now been installed so that this will never happen again.


I've also gone ahead an installed the substitute TIP31C and TIP32C transnstors and wired everything up. The happy result so far before my PAL adapter pooped the bed was the black square, before seeing the copyright so we should be good to go again.