Fixing ZX Speccies... FIXED AT LAST (See last post)

[kevman3d]

Such a sad day - I bought up my old sinclair power supply, the Speccy with the DKT keyboard, and plugged in the $10 speccy first. I wasn't expecting it to work - but I did get a boot up, and then it sat on the black screen that usually happens at start up and didn't get to the Copyright text at all. Note that there's no keyboard membrane in here (since it was ripped, I removed it) - not that this should affect the machine?

Taking a peek at the board inside, its all very clean and physically doesn't appear to have any visible damage. I figure it may be fixable?

The other one I was concerned with - its been stored in a few garages now (since I've had nowhere to store stuff in the house), and one or two have had plenty of dampness. Alas, on powering up, I just get flashing squares and doesn't go any further... I'm not surprised, there's rust on the RF and I suspect that this may or may not be related to the rest of the board. Though apart from a brown rusty nut on the heat sink, theres just a little yellowing around some of the resistors by the RF - I don't know if that's rust, or just that discoloration you occasionally see.

I did a little research - according to a few sites, the blank screen could likely be faulty ram (low 16k) and the coloured mess on screen could be something similiar with ram - or the ULA. I also took a look at this guide, but either way, reading about it vs. knowing how to do something about it are two different things.

Does anybody have any tips on either of these issues - and is it possible to replace/repair/clean anything to see if they will live again? Or are these just trash-heap material? I'm kinda not holding out hope on the rest of the stuff in storage after this. Its a pity.

[kevman3d]

After a quick further look around in regards to the common online suggestion of the lower 16k of ram being a common fault, I was wondering if anybody here is familiar with the lower ram replacement product at zx.zigg.net? It looks promising as a way to replace old chips with more modern ones. He's also produced an interesting video on testing the spectrum for faults that makes it even easy to understand for a noob like myself... Though I need to get a multimeter of course.

[acsi]

Looks like Rich has them on sellmyretro as well

http://www.sellmyretro.com/offer/details/ZX-Spectrum-4116-%28lower-RAM%29-replacement-module.-2741

Are the chips socketed on your machine? I have some spare ones at the moment I could send you to test?

[kevman3d]

Hmmm. The 8 chips are soldered in, so I'd need to desolder and remove them somehow (its one of those daunting "I'm sure I'll break something" tasks ). The additional 32kb in both are socketed (as is the ULA, Z80, etc) - I recall upgrading mine a few months after owning it - and had to buy the upgrade keyboard because the space key started to fail making it hard to play Manic Miner.

I'm sure we have a multimeter at work - I might see if I can borrow it and use the zigg video to guide me through the process of checking stuff. Maybe it will reveal something... Interestingly the guy suggests in the video replacing all the blue capacitors as well - purely on an age thing and reliability. I dunno about that, but I think it may be a good learning exercise to give it a check and possibly help me raise the dead...

I'm just clutching at straws with this one - for all I know, this isn't the problem and it'll be something else. That said, I'm sure I can't make things any worse then the fact they're both not working in the first place. Well, maybe make them worse when they could be repaired...

The zigg board looks like something I'd grab if its the ram at fault - its not overly pricey and I can grab a couple for both speccy's if they're both suffering the same thing.

[kevman3d]

Here's my troublesome twosome... Until they're running, they'll sit here on the desk next to the Atari - its better then dumping them back in the 60l containers under the house.

https://www.dropbox.com/s/0wl5qfcfw2qe4mc/troublesome_twosome.jpg

[kevman3d]

Ok, I bought a multimeter the other day - and I'll apologise in advance that I'm going to ask if someone could assist me in my completely oblivious experience in Electronics. So, I've gone through both speccies with this multimeter following the steps in THIS VIDEO to look for faults. While I wrote the values down and tried to interpret them, either some of these are showing faults, or I have no idea what I'm looking at (probably both). I'm sure somebody here will know what's going on and be able to tell me if it looks like there's a fault??

Interestingly, the readings between the two speccies was pretty much the same (well, +/- a little difference but so close its almost the same). I set the resistance to 200K on the multimeter...

Voltage Regulator

Both measured 22.1 (which at 200k, I thought was read as 22.1k ohm? Maybe I'm wrong?). According to the video, the reading should be greater then 100k, or an open circuit. Fault? Hmmm...

Lower 16k 4116 chips

All chips I measured on the 12v, 0v and +/-5 v pins (as mentioned in the video). All chips in each machine had the same values - so I'm expecting that means that no single 4116 would be an issue. Both machines returned similiar results, though a small amount of difference between the two seemed to be noticeable. That said, as per images I've attached links to at the end of this post, the internals seem a little different so that could be the reason for the slight difference.

Pin readings:

• 0v = 0.0 (both)
• +5v = 0.4 (rubber) and 0.8 (DK'Tronics).
• -5v = 22.2 (rubber) and 21.4 (DK'Tronics).
• 12v = 2.7 (both)

Inductor

Both coils were read individually, and then across both to make sure the windings weren't damaged.

• Coil 1 = 0
• Coil 2 = 0
• Both = > 34 (a high value apparently should be expected, or an open circuit)

TR4 and TR5

From what I understand, TR4 is the one that is most important with regards to the circuits power. But I tested both using the diode test mode on the multimeter. The guide video indicated testing the base with both the positive and common when testing the collector and emitter pins. (I almost sound like I know what I'm talking about here). So I tested both of them since they were sitting close to each other anyway.

With TR4, Apparently the common as base should return an open circuit on the two side pins. A positive on the base should return a value between 0.6 - 0.9V

TR4

• With common on center pin, both side pins read open circuit (both machines)
• With positive on center pin, both side pins read 812 (rubber) and 804 (DKTronics) - which I think at a 1mV resolution is 0.8V???

TR5 - No idea if these values make sense (or if it was worth testing it), but...

• Common on center pin, both sides read 885 and 878 (rubber). On the DKTronics, it read 883 on both
• Positive on center pin, one side was open circuit. On the rubber the other pin read 1841. DKTronics was 1886 (approx 1.8v???)

So, is there a fault anywhere here? The transistors kinda read OK, the inductor apparently reads OK.
The voltage regulator and lower 16k have that odd value which to be honest is likely me not really understanding how to read the value on the multimeter properly. Any help for this noob would be obviously greatly appreciated!

Differences
I did note a few differences in the boards - that said, I'm guessing its an age difference in manufacturing that just means different components were used. The DKTronics board is my original one from 1983 - its got most the IC's soldered on board (ULA, Z80, Rom), where as the other is mostly socketted. The DKTronics one has a different RF block, and an extra wire/resistor all connected in. Its easier to see then explain... Images are on dropbox, and as it doesn't appear to let me embed the links into a post, the direct links are here...

That said, now I look at this macro piccy of the RF on the older DKT one, I can see corrosion on the resistors, etc that I didn't spot looking at it normally (I don't have the best vision when it comes to getting close). Thats probably likely the cause of the machine having a bad time (its got garbage on the screen - black, with vertical lines of flashing characters). The only reason I thought it was interesting is that there's the extra wire (compared to the rubber keyboard one) with plastic cover over it. Its not a problem as its how its always been since I got the machine as a kid, but more a curious thing.

Here's a picture with the tops taken offthe two RF blocks. Slightly different internally.

That said, they're almost the same (picture here). Obviously the rubber keyboard version is more serviceable with those socketed mounts, and it looks a little more modern. That said, the poor older one has had a bad life living in boxes and damp spaces. If it can't be saved, I'd still like to get at least one working spectrum in the end...

Any tips, suggestions welcome.

[kevman3d]

Just an update - I've swapped out the 7805 on the rubber Speccy - still no go. Threw in new 4164 32kb upper ram and Swapped out the Z80 with a new one. No go (but I have a horrid feeling its now worse as its got red squares on screen, etc. rather then the black with red edges. I might place the old Z80 back in and see if that changes in that case).

A few of the screens I'm getting on the Spec right now...
https://www.dropbox.com/s/lufswqqrvyyzs ... fdeath.jpg

Still no go - Next step to replace the two transistors on the board with new ones. (Actually, I should have done that BEFORE the ram and Z80 ideally).

After this, I'll be plum out of ideas. Have noted that none of the 16kb lower ram chips get hot at all, so unsure of whether that means they're ok or not. Argh! Meanwhile, have managed to burn myself twice with a soldering iron. This process has been more painful then anticipated.

Is there anybody in NZ (or on here) that has experience with Speccy repair that can offer hints or tips on what else I can look at/do/etc? I'm currently using a variety of resources online, my limited knowledge of electronics, and dangerous finger-scorching equipment to teach myself what to do. Its not terrible in terms of what I've managed to pick up so far and I like that I feel I'm coming to grips with stuff I could never really understand as a kid (I think a fresh enthusiasm for upskilling on those things I always wanted to know is my "mid life crisis" symptom - I'm definitely not inspired to get hair implants and buy a ferrari. )

I have 2 speccies that just aren't working - After I've tried all I can with the rubber speccy, I'll see if changing out the 7805 and transistors in the other make any change. I could try swapping ULA's and so on around if I feel one may work and the other not, but we'll see how things go. The other speccy I'm not really keen on touching too much - everything is mostly soldered on the board and I'm worried I'll completely screw it up with my lack of experience in that area.

Just thought I'd ask...

[TRS80]

Sounds like a challenge. What do you have in the way of test equipment? If you want to go much further than "keep replacing parts until something works" then it would help if you had the following:

Multimeter (initial voltage checks to make sure the regulators are working)
Logic Probe (allows you to check for activity on the various control lines to know if the CPU works/etc)
Oscilloscope (good for checking signals through the video chain)
Logic Analyzer (when everything gets really hard)

[kevman3d]

Just a multimeter as its pretty much all I can afford right now (well, that and the soldering iron basics (usual Jaycar special with solder sucker, etc)). I've tested the voltage on the regulator - it was a tiny bit over the 5v mark. Luckily I asked another Mark how to test and he recorded a "tip" video on how to measure that one. I've since replaced it anyway and its around 4.99 volts - that in itself was a mission with cleaning out the solder in the holes and trying to avoid burning anything (other then my fingers) at the same time.

Biggest issue right now is my rather limited knowledge of electronics (re: I'm learning as I go from various online sources). I'm thinking to type up and compile all of these resources into a simple Spectrum repair guide for noobs, but we'll see. There's a handful of useful informative pages and sites around... Not a huge amount, but enough to give me a rough guide of things to tweak (and how to at the same time).

I did get asked whether I'd checked the 5v and 12v rail's from the transistors respectively - but being a noob, I have no idea what I test (pins? tracks? Essentially which ones I use the multimeter against). I've been shown a few photo's for switching out the ZTX650 and ZTX213 with TIP31C and 32C transistors. It looks fairly straight forward and something I can do (as long as I follow the pin differences correctly) which will at least remove another possible problem.

Obviously having more experience and a lot more knowledge/gear with some guidelines what pins to check and how to use it to read it in the Speccy would help... As you can tell, I've probably bitten off more then I can chew. But its fairly interesting to learn as I go - just a little frustrating working in the dark and with a constant worry of destroying things.

This is where I'm thinking that Electrical engineering course at Manukau Polytech back in 1988 vs. my incompleted Comp. Sci degree at Auckland Uni may have been more useful (well, in this instance - but I never imagined I'd be even trying this stuff).

[tezza]

Stick at it Kev. Repairing these things can be frustrating but as you've mentioned...you tend to learn a lot!

[kevman3d]

Yup - I'm learning things like the fact skin tends to "stick" when it touches hot soldering irons, and how 10 minutes under the cold tap really does help stop the burn...

I'd rather I wasn't possibly destroying the Speccy - but I'm trying to be an optimist - having just ordered more cheap electronic parts from eBay (the other transistors I need). I was recommended the TIP31C/32C route over some old "new" stock of ZTX transistors I found online, as they were more reliable in general. This stuff only tends to take a week to arrive from Asia (where most of its made in the first place), so that's likely to be next weekends project (or the week after).

I'll post updates if I don't melt the rest of my fingers and find I can't type for a month.

[kevman3d]

FYI, in case anybody is interested...

The guy from zigg.net has posted up a few additional videos onto his Youtube channel

https://www.youtube.com/user/JoulesperCoulomb/videos

One bunch of tips on desoldering IC's from the PCB for anybody interested, and one on repairing a potential video output problem (which is full of snippets of advice - for me, quite educational). Theres a couple of others on fixing a reset switch and showing him soldering up his 32k replacement module as well (nice - now has replacements for both upper and lower RAM in a speccy)

He's definitely going to have a nice collection of Speccy repair tutorials if he keeps posting stuff like this...

I'll get back to my ones as soon as I grab some flux liquid... Pay day is next week... Its the only thing I need as I'm having problems getting copper wire to solder nicely to a transistor leg to replace the TR4 and TR5's with new TIP ones. Otherwise great to see more information appearing online about all the things I didn't know about.

[kevman3d]

Okely-dokely... Thought I'd post an update to my frustrating adventure into speccy repair...

(a) I have to say I'm almost soldering like a pro now that I have flux...
(b) Finally worked out how to make that solder sucker suck solder properly.
(c) Replaced the paired transistors TR4 and TR5 with newer TIP 31c and 32c
(d) Converted the RF to comp output successfully...

Some random piccies - https://www.dropbox.com/sh/7lhk3jgpes4qlzt/3jpbUf9sd_

However... While it still fires up and hasn't smoked or popped (yippeeee), I'm still getting black screen of death. I've also replaced the Z80, all of the upper 32K of ram and checked that there was no overheating on the 16k lower ram...

Hmmm... I have noticed that there are some scrapes on the underside where someone's scratched away possible solder splatters (they're quite deep). Under a magnifying glass, the tracks look "ok" and not broken (though perhaps a little scraped). Maybe those could be my issue all along. Anybody know what's the best approach is for these? They're a little too close to solder safely... Conductive Pen and draw across the lines perhaps?

Perhaps its also time to try my hand at the other speccy as well (the one without the scrapes and scratches) and see if the same basic repairs might bring that one back to life.

[tezza]

Sounds like you've given it a good go. I've had a few repair projects that have beaten me in the end,

However the skills learned are always valuable.

[kevman3d]

Sounds like you've given it a good go. I've had a few repair projects that have beaten me in the end,

However the skills learned are always valuable.

Yup, definitely - specially the confidence in soldering (my biggest concern to be honest).

I'll see if I can properly check the tracks (its like someone took a compass and scratched the crap out of the middle of two tracks, the other is strangely deep where the PCB has been cut right to the edges of two/three tracks - and they're clean - like a small chisel was used?)

I'll compare it with what's on the other speccy in case something was cut away that shouldn't have been. Maybe these tracks were the only issue in the end, which would bum me out a little after unsoldering and melting finger tips. The transistors are a little too big to close the case, but if it works I'll transfer it into the DKTronics case instead.

EDIT : Unless someone can tell me "No stupid, that's NOT how you check for broken tracks!", I used the continuity test on the multimeter to make sure that there was a connection - and nothing appeared to be broken from what all the beeping seemed to indicate... I noted a handful of bared tracks (the green had been scratched off them) - some very tiny, but nothing looked like it had a crack or break in it...

Will update this thread when I've got something else working.