Recapped ZX Spectrum (Issue 3B) and now the Colors are wrong?

User avatar
1024MAK
Posts: 5103
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: Recapped ZX Spectrum (Issue 3B) and now the Colors are wrong?

Post by 1024MAK »

You need to work through the advice from each person here, answering questions where relevant.

Remote fault finding is difficult at the best of times. It’s hard to help if we don’t know exactly what you have done or not done. These computers may be simple compared to modern computers, but they are still very complex systems. The more information you can provide, the more we have to go on.

Some of the solder joints still don’t look very good.

For a good solder joint, the surface of the parts to be soldered need to be clean. The tip of the iron needs to be clean, then ‘tinned’. The tip then needs to touch both items that need to be soldered (component pin/leg/lead) and the PCB pad. The iron needs to be hot enough and be powerful enough so that the component pin/leg/lead and the PCB pad heat up above the temperature that the solder melts (see note). Then apply flux cored solder to the joint. Allow the solder to flow into and around the joint. Then stop feeding solder in, and remove the iron.

Note, if the surface area of the PCB pad and surrounding track is large, then a more powerful iron will be needed to get enough heat into the joint.

For old computers like ZX Spectrum, these use 60/40 tin/lead solder. So as it is not good practice to mix different solders, I t’s best to use 60/40 tin/lead solder for repair, servicing or modification work.

If 60/40 tin/lead solder has been used, the solder should have a shiny surface, and the surface should be smooth with no lumps, bumps, holes, cracks or similar distortions.

Note that the factory made solder joints may not look shiny, as they were likely made by an automated system and hence will have a different appearance. It’s also been 37 or 38 years since the board was made.

Mark
ZX81 Variations
ZX81 Chip Pin-outs
ZX81 Video Transistor Buffer Amp

:!: Standby alert :!:
There are four lights!
Step up to red alert. Sir, are you absolutely sure? It does mean changing the bulb :!:
Looking forward to summer later in the year.
DistantStar001
Posts: 27
Joined: Sat Feb 29, 2020 2:07 am

Re: Recapped ZX Spectrum (Issue 3B) and now the Colors are wrong?

Post by DistantStar001 »

1024MAK wrote: Wed Oct 06, 2021 3:41 pm The membranes I recommend are those supplied by RWAP. E.g. this one on Sell My Retro. However, since brexit (where the U.K. left the European Union) and the change in how VAT (sales tax) is handled, many small U.K. businesses have stopped trading overseas. So I’m not sure if he still exports outside the U.K.

On one issue board, at least one electrolytic capacitor has the polarity incorrectly marked on the silk screen (printing on the PCB). Have you compared your board to pictures of others online, or did you make a record of the polarity of each capacitor before you started?

Is C65 the correct value (22µF) and polarity? Are the solder connections, pads and tracks all okay (visual inspection and continuity testing)?

As above for C25?

Note that as long as the DC voltage rating of a electrolytic capacitor is equal or greater than the component you are replacing, that’s fine. Small value electrolytic capacitors (e.g. 22µF and lower) are only available in a limited number of voltages unless you are a manufacturer buying in bulk.

Is the central area (paper area) white after power on / reset?
The colour output is determined by the analogue voltage on pins 15 and 16 of the ULA. With a white screen, what voltages do you get on these two pins?

What voltages do you get on pins 2 and 4 of IC14 (LM1889)?

Mark
I've checked and rechecked all of the capacitor values and orientations several times, and as far as I can tell, everything's good. Also, I replaced them one at a time, matching capacitor values and orientation as I went. So I'm pretty sure I got that much right. My soldering skills, on the other hand, are admittedly a work in progress. So it wouldn't surprise me if I missed something. I did use 60/40 solder, so I got that much right. However, no matter how much I go over it, I'm still getting the same results.

I tested the voltages on pins 15, 16 and 17 on the ULA, and this is what I got for each color:

1 Blue : Pin 15: 2.15 Pin 16: 2.06 Pin 17: 2.24
2 Red : Pin 15: 2.39 Pin 16: 2.07 Pin 17: 2.18
3 Magenta: Pin 15: 2.15 Pin 16: 2.06 Pin 17: 2.11
4 Green : Pin 15: 2.55 Pin 16: 2.07 Pin 17: 2.04
5 Cyan : Pin 15: 2.18 Pin 16: 2.08 Pin 17: 1.98
6 Yellow : Pin 15: 2.77 Pin 16: 2.05 Pin 17: 1.92

7 White : Pin 15: 2.25 Pin 16: 2.03 Pin 17: 1.86

0 Black : Pin 15: 2.28 Pin 16: 2.03 Pin 17: 2.31
Screen Shot 2021-10-09 at 11.41.06 AM.png
Here is the latest go over on the board:
Screen Shot 2021-10-09 at 11.58.32 AM.png
Screen Shot 2021-10-09 at 11.56.35 AM.png
Curious: Should I be expanding my search beyond the capacitors and solder joints? I'm wondering if the ULA might have been damaged do to years of aging out of spec capacitors? Or if other components responsible for color might have been stressed and replacing them with new ones will resolve the issue? I could be wrong, but it just feels like I've been doing the same thing over and over again, hoping for a different result. Maybe it's time to try something different?
User avatar
Tiger
Posts: 88
Joined: Tue Sep 01, 2020 10:37 am

Re: Recapped ZX Spectrum (Issue 3B) and now the Colors are wrong?

Post by Tiger »

Sorry that nothing of the hints was helpful for you. But as I told, the cap's seemed to be OK, so I believe that the second recapp was only wasting time.
The solution of your last pictures was very poor, so it's not possible to see if there are better solder joints. The essential part to produce the color-signal for the ZX is the ULA. Look if there is any shortcut in the pcb-lanes of pin 15 and pin 16 of the ULA (these signals are on 18A and 17A of the expansion bus). Check C69, C71 and C76. Is there +12V on IC 14 pin 14,15,16? If not: check R62 (15 Ohm) - is there +12 V at the plus of C44/C45 - if +12V on C44/45 is ok but on IC 14 not - there is something with the solder joints on them. If the 12V is ok at all - replace TR8 and TR9 (BC 549B).
User avatar
1024MAK
Posts: 5103
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: Recapped ZX Spectrum (Issue 3B) and now the Colors are wrong?

Post by 1024MAK »

If all the supply rails (+5V, -5V and +12V) are correct, as in, within specification and do not have excessive AC ripple or noise, then the only electrolytic capacitors that affect colour are the two that I mentioned in my earlier post.

I’m assuming based on what you have said, that the machine was working correctly before you started changing the electrolytic capacitors. Hence it is likely that either there is a poor solder joint, a short circuit due to a solder splash (which can be as fine as a hair) or you inadvertently disturbed something else while working on the board.

The ULA generates TWO colour difference signals. Which are then sent via transistors TR8 and TR9 to the PAL colour carrier encoding chip (IC14). As the transistors in positions TR8 and TR9 are normally VERY reliable, the possible causes are:
  • One of the colour difference outputs from the ULA have been damaged (the rest of the ULA is fine),
  • The analogue circuitry between the ULA and IC14 has a problem,
  • IC14 (LM1889) has been damaged,
  • Part of the support circuitry for LM1889 has a problem.
I will have a look at the results from your earlier post when I get some time.

Meanwhile, with the paper and the border both white, what voltages do you get on pins 2 and 4 of IC14 (LM1889)?

Mark
ZX81 Variations
ZX81 Chip Pin-outs
ZX81 Video Transistor Buffer Amp

:!: Standby alert :!:
There are four lights!
Step up to red alert. Sir, are you absolutely sure? It does mean changing the bulb :!:
Looking forward to summer later in the year.
DistantStar001
Posts: 27
Joined: Sat Feb 29, 2020 2:07 am

Re: Recapped ZX Spectrum (Issue 3B) and now the Colors are wrong?

Post by DistantStar001 »

1024MAK wrote: Wed Oct 13, 2021 3:26 pm If all the supply rails (+5V, -5V and +12V) are correct, as in, within specification and do not have excessive AC ripple or noise, then the only electrolytic capacitors that affect colour are the two that I mentioned in my earlier post.

I’m assuming based on what you have said, that the machine was working correctly before you started changing the electrolytic capacitors. Hence it is likely that either there is a poor solder joint, a short circuit due to a solder splash (which can be as fine as a hair) or you inadvertently disturbed something else while working on the board.

The ULA generates TWO colour difference signals. Which are then sent via transistors TR8 and TR9 to the PAL colour carrier encoding chip (IC14). As the transistors in positions TR8 and TR9 are normally VERY reliable, the possible causes are:
  • One of the colour difference outputs from the ULA have been damaged (the rest of the ULA is fine),
  • The analogue circuitry between the ULA and IC14 has a problem,
  • IC14 (LM1889) has been damaged,
  • Part of the support circuitry for LM1889 has a problem.
I will have a look at the results from your earlier post when I get some time.

Meanwhile, with the paper and the border both white, what voltages do you get on pins 2 and 4 of IC14 (LM1889)?

Mark
Unfortunately, I don't think I can test this at the moment, as I am quite certain that my ULA has failed (possibly fried).

I was testing the capacitor voltages suggested by Tiger, when all of the sudden, my screen glitched and then this was all I could get:
Screen Shot 2021-10-13 at 1.56.15 PM.png
Sometimes the little colored squares were different, but beyond that, this was the screen I was getting. So I looked into it and found that it was likely bad voltages on the lower RAM. Only problem was that other than the -5v being a bit low (-4.82v as I recall) the voltages seemed fine. Still, the site I found (https://www.retrotechlab.com/zx-spectrum-common-faults/) suggested four troubleshooting steps to take, so I went with it.

Step one was to check for shorts on TR4 and TR5 for shorts along the center pin. I found none.

I had already done step two by checking the volts at the RAM. As mentioned the -5 was a touch low.

Step three was to swap out the ULA. Unfortunately, I can't do that, as I don't have a spare.

So I moved on to step four: replace all of the lower RAM. This proved to be an issue, as a piece of debris managed to short the 12 and 5-volt rails between two of the RAM chips.

Long story short, before I realized the short, the ULA got exceedingly hot. The -5v rail has dropped completely (admittedly this is after I removed the ULA to prevent further damage, though I'm pretty sure it's dead). The 12v line reads at only reads about 10v. But! The 5v rail is near perfect at 5.04 volts! So there is something.

The picture on the screen with the ULA flashed between these:
Screen Shot 2021-10-13 at 1.56.31 PM.png
Screen Shot 2021-10-13 at 1.56.41 PM.png
I'm pretty sure I'm going to have to figure out what exactly is pulling the 12v line low, and restore the -5v line before I start repopulating the lower RAM or the ULA. The good news is that I have a local source for the 4116 RAMs, so if I damaged any of them, they won't be hard to replace. The bad news is that I have no idea where to get a replacement ULA. The best option I've found is from the Retro Spares Shop (retroleum.co.uk) which offers a programmable replacement module for about £21 plus shipping. I'd prefer an original ULA, just because I like to keep my equipment as original as possible, but that might not be economically feasible. As it is, I've already spent more on repairing this Spectrum than I did buying it in the first place.

All and all, I'd have to say that this recap has been a disaster. Although not entirely irreparable. I'm pretty sure now that the ULA had been the issue all along (as it had been showing some signs of glitching before all this started), but I believe that my attempts to diagnose have done some damage to the power lines that need to be addressed before I try to replace it. On a positive note, I've gone over all of the new RAM sockets I installed, and after I cleared that bridge, everything seems to be connected correctly! So I'm not giving up hope. I think this thing will work with a new ULA once I've corrected all of the voltage issues.

Here are some pictures of the board in its current state in case anyone notices any glaring issues. Sorry for the lousy resolution, but my iPad does not have the best camera. In the meantime, I'm open to any repair suggestions you might have.
Screen Shot 2021-10-13 at 3.00.23 PM.png
Screen Shot 2021-10-13 at 3.00.33 PM.png
User avatar
Tiger
Posts: 88
Joined: Tue Sep 01, 2020 10:37 am

Re: Recapped ZX Spectrum (Issue 3B) and now the Colors are wrong?

Post by Tiger »

I don't think that the ULA is the case or this newly fault. The solder joint at die lower RAM sockets are very poor, maybe there is a broken or shortened connection under the sockets. Test the sockets. only pin 2 and pin 14 of each socket is shorten (green in the example picture). No zero Ohm between the other pins on each separate socket. All other pins are connected with the next socket - e.g. 0 Ohm from pin 3 of IC 13 to pin 3 of IC 12, 11, 10, 9, 8, 7, 6 (yellow in the example). Only the DATA-lines (pin 2,14) didn't have a connection to the next lower RAM IC! If there is any 0 Ohm between the other pins (e.g. pin 4 - pin 5) you have shortened fine line directly under the socket (from the component side)! You have to replace TR 4 and 5. At the pcb-side there is a very bad solder joint between pin 8 (IC7) and pin 9 (IC6) that was the dead of the 12V line! :o You shortened 5V with 12V wich killed TR 4 and 5, maybe D15 and D17 also. :roll:
Sorry to say it, but the work doesn't look very skillful :?
Attachments
Example
Example
DistantStar001
Posts: 27
Joined: Sat Feb 29, 2020 2:07 am

Re: Recapped ZX Spectrum (Issue 3B) and now the Colors are wrong?

Post by DistantStar001 »

Tiger wrote: Thu Oct 14, 2021 6:34 am I don't think that the ULA is the case or this newly fault. The solder joint at die lower RAM sockets are very poor, maybe there is a broken or shortened connection under the sockets. Test the sockets. only pin 2 and pin 14 of each socket is shorten (green in the example picture). No zero Ohm between the other pins on each separate socket. All other pins are connected with the next socket - e.g. 0 Ohm from pin 3 of IC 13 to pin 3 of IC 12, 11, 10, 9, 8, 7, 6 (yellow in the example). Only the DATA-lines (pin 2,14) didn't have a connection to the next lower RAM IC! If there is any 0 Ohm between the other pins (e.g. pin 4 - pin 5) you have shortened fine line directly under the socket (from the component side)! You have to replace TR 4 and 5. At the pcb-side there is a very bad solder joint between pin 8 (IC7) and pin 9 (IC6) that was the dead of the 12V line! :o You shortened 5V with 12V wich killed TR 4 and 5, maybe D15 and D17 also. :roll:
Sorry to say it, but the work doesn't look very skillful :?
I've gone over all of the solder joints on the lower RAM and didn't find any unwanted shorts between pins, and so far as I can tell, everything that should be connected is connected. 2 to 14 on each socket, but not between sockets. All other pins are connected from socket to socket. I also checked the data lines to the ULA and Z80 which seem correct as well. To be sure I reflowed the solder on all pins and checked again, everything seems good even if my solder skills are lacking.

I did recheck TR4 and TR5. TR4 is reading a short, but TR5 did not. But the readings aren't the same as before I shorted the 12v line, so I'm going to replace both just to be on the safe side. I haven't found D15 and D17 yet, but I'll likely replace them as well.

I do agree that the ULA is not the cause for my current voltage issues, but I do think it was faulty. As I mentioned, it was glitching and crashing with increasing frequency before the recap. I just think that it probably gave up the ghost when I was testing the capacitors (I might have shorted something if my probe slipped), and me shorting the 12 and 5-volt lines was likely the final nail in the coffin.
User avatar
Tiger
Posts: 88
Joined: Tue Sep 01, 2020 10:37 am

Re: Recapped ZX Spectrum (Issue 3B) and now the Colors are wrong?

Post by Tiger »

:shock: As they say: shit happens. But I think you will solve this problem with a lot of patience and luck 8-)
DistantStar001
Posts: 27
Joined: Sat Feb 29, 2020 2:07 am

Re: Recapped ZX Spectrum (Issue 3B) and now the Colors are wrong?

Post by DistantStar001 »

OK. So quick update: I think I've identified the replacements for TR5, but want to be sure. TR4 on the other hand has proven harder to find. At least in the US. From what I can tell, the ZTX part numbers are European, where the US uses NTE. Me being in the states, this has proven a slight problem finding proper replacements.

So far this is what I have:

TR4=ZTX650=ZTX651=NTE???
TR5=ZTX213=ZTX751=NTE159
D15=BA157=NTE558
D17=Not present on this revision of the board

I thought that an NTE2363 would work for TR4, but the pinout is wrong (also the thing is about 3x larger than the original transistor. So really not a great fit).

Did I get TR5 right? If not, does anyone know the correct NTE equivalents for these parts? Or should I just import them?
User avatar
Tiger
Posts: 88
Joined: Tue Sep 01, 2020 10:37 am

Re: Recapped ZX Spectrum (Issue 3B) and now the Colors are wrong?

Post by Tiger »

I've got the ZTX 751 from Mouser Electronics Inc. - 1000 N. Main St., Mansfield, TX 76063. They send it from US to Germany ...
Post Reply