Outstanding video. Clear and informative information. I have the same issue 2 model. Puzzle: Wondering why Sinclair used a spectrum case foot under the Zx80 chip on the machine like in your video!
ok thanks. i did not realize the 9v was fed further back into the spectrum. i might try a dc to dc booster. so upscale the usb 5v to 9 and feed 9v into Tr4. thanks for your vids. busy replacing my capictors just now
Thanks for the video, You never mentioned which way the additional capacitor to R58 goes, ie positive or negative on R58, and it is not clear from the video,, I've gone through it and paused it multiple times and I still don't know which way I need to put it.
This is one awesome video and i like how you installed radial elec. caps to mobo :) I have working Issue 4A board. Though i need to replace keyboard membrane but caps are on my No, 1 list these days. Do you have any advice for issue 4A mobo or can you point me out some web site with enough info? I can see that my mobo is somewhat different than your issue 2 mobo esspecially where voltage regulator is. Cheers mate Long live specky
I was looking at mine and C46 seems to be in the wrong way round. Yours is the same so I`ll put it back as it came out. It`s got a + on the wrong side on the board. OK I got to the end of the video lol. Thanks for clearing that up.
I've replaced all electrolytic capacitors by Tantalum capacitors in my Spectrum +3 and it works perfectly. Later, I did the same to the Spectrum 48K and again no problems, the video even got better.... I wonder why more people do not do this, instead of replacing the electrolytics every 30 years or so... The only problem I've had so far was a static discharge from my PC that ruined a +2B board when I was connecting a 3.5 stereo jack to the soundboard of my PC... Even I felt the shock, some of the old PCs had a bad grounding (in some cases not ground at all)... I heard that Tantalum capacitors have a bigger impedance but, after measuring a lot of these I've found that the values are much more closer than the rated. It's not uncommon to see 100uF electrolytics with 130 or more...
tants have problems too, the main one being they fail to a short, electrolytics when they fail go open circuit so are safer. i design and build detectors, and have tried various things to make circuits less noisey, including tants, the results of my tests showed the quality of components was far more important, metal can transistors and ic,s, even ceramic chip variants perform better than cheap plastic packages, and i only use japanese caps.
+tim bo Probably not. The problem would be that the voltage converter circuit that generates the 12 V and -5 V supplies uses the 9 - 11 V input. Though I haven't tried, I suspect it would not work properly if its higher voltage input were connected to 5 V instead. Regulated 9 V supplies are available at low cost and so I recommend using one of those instead. Do look for one of reasonable quality; the market is flooded with dangerous small power supplies that pose a serious risk of fire or electrocution.
Hello Joules. Congrats for the video! Looking forward for recapping my Spectrum+ (issue 4S) , I'd like to ask know if modification to voltage converter circuit (resistor change out + additional cap) is even recommended/needed. Also I planned to replace the original 7805 with a dedicated switching regulator making the speccy actually cooler and more efficient. Thanks in advance!
Someone told me the issue i have in my spectrum 48k is this : 1)Lower memory fault. Changed IC9 and IC10 4116 chips. 2)Tested TR4 and TR5 (OK). Changed all electrolytic capacitors. Do you have skype to help me ?
Hi. I'm finding your videos very helpful. I bought a selection of capacitors from a local supplier and they are quite a bit smaller in physical size to the originals. Does this matter? I went for higher voltage ratings generally. I also noted on one lot, they are for audio application which was probably an oversight on my part when ordering. Can I use the audio rated capacitors? Thanks.
So long as their ratings are appropriate, their physical size isn't very important. Certainly, electrolytic capacitor technology has improved since the 1980s, allowing them to be smaller now. In general there isn't any property of an electrolytic capacitor that makes it especially suitable for use in audio circuits. It's usually a label added by some manufacturers/distributors as an excuse for making them more expensive.
Very Useful Video. Thank you! I am about to carry out the caps replacement on a non working Spectrum 48k. I have a number of capacitors of the same capacitance value but have different voltage ratings.(all above 16v) Does this make any difference when replacing?
It doesn't matter so long as the voltage rating matches or exceeds that of the capacitor being replaced. In principle, those with a higher voltage rating will have a longer life (and be physically larger), but there will likely be greater differences between different brands than between the different voltage ratings. Note that because electrolytic capacitors usually fail due to drying out, it's best to use relatively new components and avoid old stock.
JoulesperCoulomb Thanks. I changed out the caps today. Also changed out the voltage regulator and did the composite mod while I was at it and now have a video signal but it has vertical black and white bars with flashing colour squares on screen. I assume I have bad memory somewhere but I will check that for a later date. Least im halfway there considering I have not picked up a soldering iron in many years!! :) Thanks again!!
Are the vertical stripes wide (several character squares) or narrow? Can you post a picture of the display? Are all three power supply voltages present and correct at the lower RAM ICs? Given that there is colour in the display, I suspect they will be, but they should be checked.
JoulesperCoulomb They are wide vertical black and white stripes. Posted Pic Here: tinypic.com/r/mvomz9/8 I checked the the 3 supply voltages on all of the lower RAM IC's IC6 to IC13 and all seem to be present and correct. It's an Issue 3B Board. Have printed out the repair guide for world of spectrum to check others.
The non-white border suggests at least one of the following: 1. Faulty CPU. 2. Faulty ULA. 3. Faulty ROM (very rare). 4. Something interfering with communication between the above components; very often a faulty upper RAM IC (common). It is also possible for a faulty lower RAM IC to interfere with communication between the CPU and ULA and so prevent the border from being set to white at initialisation, but then the central area of the display is usually more disorganised or shows a pattern of narrow vertical stripes. Wide vertical black and white bars are also a common symptom of a faulty CPU. So, I think the problem is most likely 4 or/and 1. It might be possible to further diagnose this type problem by using an oscilloscope to examine activity on the control and data bus at the CPU.
Great video. I just got what i believe to be a ZX Spectrum +, and I'm gonna do this cuz I plugged it to a 9v 1000mA and nothing happened except a low crackling. Do you know were I can get a PSU for this spectrum? Thanks
A 9 V, 1 A capable power supply is a good choice for a Spectrum, so long as you don't have any power hungry peripherals connected. Don't forget that the centre terminal of the Spectrum's power input socket must be connected to the negative side of the supply. Getting this wrong will likely damage the Spectrum. Replacing all the electrolytic capacitors may or may not solve your problem, but it's a good thing to do anyway. Does the computer produce any video output? It should be even if it has suffered many of the most common faults. Try the tests described in the ZX Spectrum Initial Tests video. If you want to obtain an original Sinclauir power supply for a Spectrum or are in need of a repair service, contact me via zx.zigg.net or post in the forums at the World of Spectrum web site; a number of regulars there also offer ZX Spectrum related services.
JoulesperCoulomb I don't have any peripherals at the moment. About the center of the power input on the spectrum being connected to the negative side of the supply, I have to find a way to see if that is the case. When I fixed the transformer to the spectrum I made sure the white/black wires matched and so the all black wires. All I got was some low crackling and the 7805 heating a lot in just a few seconds. I haven't plugged it since, being to busy watching your videos :) and learning a thing or two. The machine gives no video, at least nothing changes on the tv. I may need to try to find the signal, (it finds my master system, but maybe the spectrum is some were else.) but with the overheating I don't feel safe searching for the signal. Can you see my last video, I showed the spectrum, I think its the 46k because it says ZX Spectrum +.
DarthNinjaCode If you've connected a power supply with incorrect polarity, it's possible the 7805 regulator has been damaged along with components in the voltage converter circuit, most likely TR4/TR5. If you've since re-applied power with the correct polarity without having corrected any faults, one or more of the lower RAM ICs may have been damaged. Note that unless any modification has been made, the ZX Spectrum outputs a modulated UHF signal of the type that should be connected to the aerial socket of an older PAL television. Send me a message at zx.zigg.net should you be in need of a ZX Spectrum repair service.
When you plugged it in, did you use a power supply with the negative pin in the centre? You've probably fried the 7805, they're quite common to find however.
Joules, i have issue 4A board and lower ram chips are installed in orange sockets!? Did someone tamper with board or this was usual from Sinclair in that time? Thanx for reply
I haven't seen orange sockets in a ZX Spectrum before and so far as I know, the lower RAM ICs were always soldered directly to the board, so I suspect the sockets have been added later. However, all sorts of oddities turn up from time to time, so I can't say for certain they aren't original. Comparing their solder joints to others on the board might offer a clue.
They are probably added at a later time coz solder joints are brownish from flux residue. I tried to power up this board today and got "1982 Sinclair Research Ltd" on screen but in a while i am getting black and white raster with white border. I can hear clicking sound from speaker and when it do click the white border also tik in a second. I thought that my ULA chip is damaged but then i check PS voltage and i am getting more than 16 volts dc unloaded. Some 14 volts under load. When i unplug my power supply and discharge remaining voltage from capacitor inside PS unit and then plug it again it works again for short time. This is strange? Is my ULA really failing or input voltage is way higher than it should be? Do you have some advice for me? I dont want to burry this specky yet :) Cheers
The voltage from your power supply is a bit higher than is ideal and such a voltage can damage some versions of the voltage converter circuit in ZX Spectrums (quite a few variations of the circuit were used over the years). However, I don't think that has happened in your case. If you can, I recommend using a lower voltage supply (perhaps use a modern regulated 9 V one) or put two or three diodes in series with the output of the one you are using. Don't forget that the centre terminal of the Spectrum's power input socket must be connected to the negative side of the supply. One possibility is that your Spectrum has some faulty upper RAM that either isn't fully detected by the initialisation RAM test or RAM that interferes on the data bus at times when it shouldn't. If the computer will run for long enough, try loading the Memtest program (available at zx.zigg.net) and tell it to test the upper RAM. If the Spectrum isn't stable enough to let you load that program, try using a jumper wire to connect pin 5 of IC23 to the 5 V rail at power on; this connection can be removed one the computer has initialised. Take care not to let this wire contact anything else. This disables the /CAS signal to the upper RAM and should force the computer to initialise as a 16 K model; you might then be able to load the Memtest program. However, if the fault is due to one or more upper RAM ICs interfering on the data bus, then this won't work and diagnosing the fault will be more difficult, but certainly possible. I would use a logic analyser, as I demonstrated in another video.
Thank you for your kind response Joules,. I will first try with lover voltage. Polarity is ok as i changed cracked cables on power supply few hours ago. Center pin is negative. If i find some 9v reg PS i will give it a try. Currently only one i have at hand is 500mA max :( I doubt this is enough? Right? I can try with original PS and some diodes in serie. Since this is my first time into ZX i need additional info on diagnostic and trouble shooting :/ Do you cover that in some of your videos maybe? I dont have logic analyser unfortunately :( Where i need to poke my probe to check voltage supply at lower ram? Thanx again
Hello JoulesperCoulomb! I think I heard you say in the beginning of the video something like this: "...unlike some people I don't advocate automatically replacing all electrolytic capacitors in any piece of electronic equipment that is more than a few years old..." I agree with you but I would also like to hear your advice regarding this subject. I mean, when is to be expected a complete recap of a ZX spectrum motherboard? 10 years? 15 years? 20 years? Some people I talk to tell me that some capacitors used were realy good so it's best just to leave them (nichicon, Gelec) but since I don't own a capacitance/ESR meter I feel this could be like walking on thin ice. Just how real is this advice I hear that japanese capacitors are the best to buy, seconded by tawainese makers while chinese makers are complete rubbish and should be avoided? Perhaps capacitors used close to their working limits are expected to fail no matter their brand and therefore this caused a huge hype or foul myth regarding some brands over time? Best Regards
The main enemy of an electrolytic capacitor is heat, whether while in operation or storage. This is one of the reasons the loft is not usually a good storage place for electronic equipment (get your Spectrum out of there now!) Without knowing the full environmental history of a capacitor it's probably impossible to make any useful prediction about its remaining life span (even assuming it hasn't been subjected to any electrical or physical abuse). If the history were known, any prediction would most likely be inaccurate. There can be quite a bit of variation between different batches of the same capacitor (at least in the past, I'm sure manufacturers will claim this doesn't happen now) and even the most reputable manufacturers have had some bad designs and periods of sloppy standards. There are some tests that can give a guide to the health of an electrolytic capacitor, but to carry them all out properly requires removing the component from the circuit. Unless the capacitors are expensive or difficult to obtain (which they aren't in the case of the ZX Spectrum), it's probably quicker and easier just to replace them at that point. So, while there are probably some Spectrums out there with their original capacitors that'll go on for another 30 years, I'm certain there are far more that'll die when next switched on. I can say from experience that failed or failing electrolytic capacitors are a very common cause of non-functioning ZX Spectrums. This suggests that on average and after being subjected to average use and storage, the types originally used are at or approaching the end of their lives. The failure rate is probably even higher than I suspect as when I get a Spectrum to repair, I usually replace all the electrolytic capacitors before I power it on. I've had a fair few that I've been told were faulty, but worked perfectly after I replaced the capacitors.
Excellent video. Thanks for taking the time to create it.
Outstanding video. Clear and informative information. I have the same issue 2 model. Puzzle: Wondering why Sinclair used a spectrum case foot under the Zx80 chip on the machine like in your video!
Great videos!
ok thanks. i did not realize the 9v was fed further back into the spectrum.
i might try a dc to dc booster. so upscale the usb 5v to 9 and feed 9v into Tr4. thanks for your vids. busy replacing my capictors just now
+tim bo The 9 V should be connected to the correct end of one of the inductor windings - follow the PCB trace from the 7805's input..
Thanks for the video, You never mentioned which way the additional capacitor to R58 goes, ie positive or negative on R58, and it is not clear from the video,, I've gone through it and paused it multiple times and I still don't know which way I need to put it.
hi joules
a wonderful video , i really enjoy watching your videos, do you have any spectrum 4a yellow boards buy any chance
warm regards
This is one awesome video and i like how you installed radial elec. caps to mobo :)
I have working Issue 4A board. Though i need to replace keyboard membrane but caps are on my No, 1 list these days.
Do you have any advice for issue 4A mobo or can you point me out some web site with enough info? I can see that my mobo is somewhat different than your issue 2 mobo esspecially where voltage regulator is.
Cheers mate
Long live specky
I did the R6 mod and got 13.2 volt on the 12 volt rail. I think I used a too lower wattage resistor. I used a 1/4 watt. Should they be 1/5 watts?
I was looking at mine and C46 seems to be in the wrong way round. Yours is the same so I`ll put it back as it came out. It`s got a + on the wrong side on the board. OK I got to the end of the video lol. Thanks for clearing that up.
Can you do one in refurbing a TS 1016.
I've replaced all electrolytic capacitors by Tantalum capacitors in my Spectrum +3 and it works perfectly. Later, I did the same to the Spectrum 48K and again no problems, the video even got better.... I wonder why more people do not do this, instead of replacing the electrolytics every 30 years or so...
The only problem I've had so far was a static discharge from my PC that ruined a +2B board when I was connecting a 3.5 stereo jack to the soundboard of my PC... Even I felt the shock, some of the old PCs had a bad grounding (in some cases not ground at all)...
I heard that Tantalum capacitors have a bigger impedance but, after measuring a lot of these I've found that the values are much more closer than the rated. It's not uncommon to see 100uF electrolytics with 130 or more...
tants have problems too, the main one being they fail to a short, electrolytics when they fail go open circuit so are safer.
i design and build detectors, and have tried various things to make circuits less noisey, including tants, the results of my tests showed the quality of components was far more important, metal can transistors and ic,s, even ceramic chip variants perform better than cheap plastic packages, and i only use japanese caps.
can i power a zx spectrum48 with a 5v usb 2amp charger.i.e. remove regulator and feed 5v direct to spectrum
+tim bo Probably not. The problem would be that the voltage converter circuit that generates the 12 V and -5 V supplies uses the 9 - 11 V input. Though I haven't tried, I suspect it would not work properly if its higher voltage input were connected to 5 V instead.
Regulated 9 V supplies are available at low cost and so I recommend using one of those instead. Do look for one of reasonable quality; the market is flooded with dangerous small power supplies that pose a serious risk of fire or electrocution.
Hello Joules. Congrats for the video!
Looking forward for recapping my Spectrum+ (issue 4S) , I'd like to ask know if modification to voltage converter circuit (resistor change out + additional cap) is even recommended/needed.
Also I planned to replace the original 7805 with a dedicated switching regulator making the speccy actually cooler and more efficient.
Thanks in advance!
The modification to the voltage converter circuit applies only to issue 2 ZX Spectrums.
Perfect! Many thanks!
Someone told me the issue i have in my spectrum 48k is this : 1)Lower memory fault. Changed IC9 and IC10 4116 chips.
2)Tested TR4 and TR5 (OK). Changed all electrolytic capacitors. Do you have skype to help me ?
I've only just seen your message. If you haven't got your Spectrum working yet, can you give me a description of the symptoms of its fault?
Hi. I'm finding your videos very helpful. I bought a selection of
capacitors from a local supplier and they are quite a bit smaller in
physical size to the originals. Does this matter? I went for higher
voltage ratings generally. I also noted on one lot, they are for audio
application which was probably an oversight on my part when ordering.
Can I use the audio rated capacitors? Thanks.
So long as their ratings are appropriate, their physical size isn't very important. Certainly, electrolytic capacitor technology has improved since the 1980s, allowing them to be smaller now.
In general there isn't any property of an electrolytic capacitor that makes it especially suitable for use in audio circuits. It's usually a label added by some manufacturers/distributors as an excuse for making them more expensive.
This is very informative as I recently aquired a spectrum for proper recapping and modding, cheers.
Very Useful Video. Thank you! I am about to carry out the caps replacement on a non working Spectrum 48k. I have a number of capacitors of the same capacitance value but have different voltage ratings.(all above 16v) Does this make any difference when replacing?
It doesn't matter so long as the voltage rating matches or exceeds that of the capacitor being replaced. In principle, those with a higher voltage rating will have a longer life (and be physically larger), but there will likely be greater differences between different brands than between the different voltage ratings.
Note that because electrolytic capacitors usually fail due to drying out, it's best to use relatively new components and avoid old stock.
JoulesperCoulomb
Thanks. I changed out the caps today. Also changed out the voltage regulator and did the composite mod while I was at it and now have a video signal but it has vertical black and white bars with flashing colour squares on screen. I assume I have bad memory somewhere but I will check that for a later date. Least im halfway there considering I have not picked up a soldering iron in many years!! :) Thanks again!!
Are the vertical stripes wide (several character squares) or narrow? Can you post a picture of the display?
Are all three power supply voltages present and correct at the lower RAM ICs? Given that there is colour in the display, I suspect they will be, but they should be checked.
JoulesperCoulomb
They are wide vertical black and white stripes. Posted Pic Here:
tinypic.com/r/mvomz9/8
I checked the the 3 supply voltages on all of the lower RAM IC's IC6 to IC13 and all seem to be present and correct. It's an Issue 3B Board. Have printed out the repair guide for world of spectrum to check others.
The non-white border suggests at least one of the following:
1. Faulty CPU.
2. Faulty ULA.
3. Faulty ROM (very rare).
4. Something interfering with communication between the above components; very often a faulty upper RAM IC (common). It is also possible for a faulty lower RAM IC to interfere with communication between the CPU and ULA and so prevent the border from being set to white at initialisation, but then the central area of the display is usually more disorganised or shows a pattern of narrow vertical stripes.
Wide vertical black and white bars are also a common symptom of a faulty CPU. So, I think the problem is most likely 4 or/and 1.
It might be possible to further diagnose this type problem by using an oscilloscope to examine activity on the control and data bus at the CPU.
Great video. I just got what i believe to be a ZX Spectrum +, and I'm gonna do this cuz I plugged it to a 9v 1000mA and nothing happened except a low crackling.
Do you know were I can get a PSU for this spectrum? Thanks
A 9 V, 1 A capable power supply is a good choice for a Spectrum, so long as you don't have any power hungry peripherals connected. Don't forget that the centre terminal of the Spectrum's power input socket must be connected to the negative side of the supply. Getting this wrong will likely damage the Spectrum.
Replacing all the electrolytic capacitors may or may not solve your problem, but it's a good thing to do anyway.
Does the computer produce any video output? It should be even if it has suffered many of the most common faults. Try the tests described in the ZX Spectrum Initial Tests video.
If you want to obtain an original Sinclauir power supply for a Spectrum or are in need of a repair service, contact me via zx.zigg.net or post in the forums at the World of Spectrum web site; a number of regulars there also offer ZX Spectrum related services.
JoulesperCoulomb
I don't have any peripherals at the moment. About the center of the power input on the spectrum being connected to the negative side of the supply, I have to find a way to see if that is the case. When I fixed the transformer to the spectrum I made sure the white/black wires matched and so the all black wires.
All I got was some low crackling and the 7805 heating a lot in just a few seconds. I haven't plugged it since, being to busy watching your videos :) and learning a thing or two.
The machine gives no video, at least nothing changes on the tv. I may need to try to find the signal, (it finds my master system, but maybe the spectrum is some were else.) but with the overheating I don't feel safe searching for the signal.
Can you see my last video, I showed the spectrum, I think its the 46k because it says ZX Spectrum +.
DarthNinjaCode
If you've connected a power supply with incorrect polarity, it's possible the 7805 regulator has been damaged along with components in the voltage converter circuit, most likely TR4/TR5. If you've since re-applied power with the correct polarity without having corrected any faults, one or more of the lower RAM ICs may have been damaged.
Note that unless any modification has been made, the ZX Spectrum outputs a modulated UHF signal of the type that should be connected to the aerial socket of an older PAL television.
Send me a message at zx.zigg.net should you be in need of a ZX Spectrum repair service.
When you plugged it in, did you use a power supply with the negative pin in the centre? You've probably fried the 7805, they're quite common to find however.
Are your replacement caps also electrolytic?
Yes.
Joules, i have issue 4A board and lower ram chips are installed in orange sockets!? Did someone tamper with board or this was usual from Sinclair in that time?
Thanx for reply
I haven't seen orange sockets in a ZX Spectrum before and so far as I know, the lower RAM ICs were always soldered directly to the board, so I suspect the sockets have been added later. However, all sorts of oddities turn up from time to time, so I can't say for certain they aren't original. Comparing their solder joints to others on the board might offer a clue.
They are probably added at a later time coz solder joints are brownish from flux residue. I tried to power up this board today and got "1982 Sinclair Research Ltd" on screen but in a while i am getting black and white raster with white border. I can hear clicking sound from speaker and when it do click the white border also tik in a second. I thought that my ULA chip is damaged but then i check PS voltage and i am getting more than 16 volts dc unloaded. Some 14 volts under load. When i unplug my power supply and discharge remaining voltage from capacitor inside PS unit and then plug it again it works again for short time. This is strange? Is my ULA really failing or input voltage is way higher than it should be? Do you have some advice for me? I dont want to burry this specky yet :)
Cheers
The voltage from your power supply is a bit higher than is ideal and such a voltage can damage some versions of the voltage converter circuit in ZX Spectrums (quite a few variations of the circuit were used over the years). However, I don't think that has happened in your case. If you can, I recommend using a lower voltage supply (perhaps use a modern regulated 9 V one) or put two or three diodes in series with the output of the one you are using. Don't forget that the centre terminal of the Spectrum's power input socket must be connected to the negative side of the supply.
One possibility is that your Spectrum has some faulty upper RAM that either isn't fully detected by the initialisation RAM test or RAM that interferes on the data bus at times when it shouldn't. If the computer will run for long enough, try loading the Memtest program (available at zx.zigg.net) and tell it to test the upper RAM.
If the Spectrum isn't stable enough to let you load that program, try using a jumper wire to connect pin 5 of IC23 to the 5 V rail at power on; this connection can be removed one the computer has initialised. Take care not to let this wire contact anything else. This disables the /CAS signal to the upper RAM and should force the computer to initialise as a 16 K model; you might then be able to load the Memtest program. However, if the fault is due to one or more upper RAM ICs interfering on the data bus, then this won't work and diagnosing the fault will be more difficult, but certainly possible. I would use a logic analyser, as I demonstrated in another video.
P.S. Just to make sure the voltage converter circuit hasn't been damaged, check for the presence of the -5 V and 12 V supplies at the lower RAM.
Thank you for your kind response Joules,.
I will first try with lover voltage. Polarity is ok as i changed cracked cables on power supply few hours ago. Center pin is negative. If i find some 9v reg PS i will give it a try. Currently only one i have at hand is 500mA max :( I doubt this is enough? Right?
I can try with original PS and some diodes in serie. Since this is my first time into ZX i need additional info on diagnostic and trouble shooting :/ Do you cover that in some of your videos maybe? I dont have logic analyser unfortunately :( Where i need to poke my probe to check voltage supply at lower ram?
Thanx again
Hello JoulesperCoulomb!
I think I heard you say in the beginning of the video something like this:
"...unlike some people I don't advocate automatically replacing all electrolytic capacitors in any piece of electronic equipment that is more than a few years old..."
I agree with you but I would also like to hear your advice regarding this subject.
I mean, when is to be expected a complete recap of a ZX spectrum motherboard? 10 years? 15 years? 20 years?
Some people I talk to tell me that some capacitors used were realy good so it's best just to leave them (nichicon, Gelec) but since I don't own a capacitance/ESR meter I feel this could be like walking on thin ice.
Just how real is this advice I hear that japanese capacitors are the best to buy, seconded by tawainese makers while chinese makers are complete rubbish and should be avoided?
Perhaps capacitors used close to their working limits are expected to fail no matter their brand and therefore this caused a huge hype or foul myth regarding some brands over time?
Best Regards
The main enemy of an electrolytic capacitor is heat, whether while in operation or storage. This is one of the reasons the loft is not usually a good storage place for electronic equipment (get your Spectrum out of there now!)
Without knowing the full environmental history of a capacitor it's probably impossible to make any useful prediction about its remaining life span (even assuming it hasn't been subjected to any electrical or physical abuse). If the history were known, any prediction would most likely be inaccurate. There can be quite a bit of variation between different batches of the same capacitor (at least in the past, I'm sure manufacturers will claim this doesn't happen now) and even the most reputable manufacturers have had some bad designs and periods of sloppy standards.
There are some tests that can give a guide to the health of an electrolytic capacitor, but to carry them all out properly requires removing the component from the circuit. Unless the capacitors are expensive or difficult to obtain (which they aren't in the case of the ZX Spectrum), it's probably quicker and easier just to replace them at that point.
So, while there are probably some Spectrums out there with their original capacitors that'll go on for another 30 years, I'm certain there are far more that'll die when next switched on.
I can say from experience that failed or failing electrolytic capacitors are a very common cause of non-functioning ZX Spectrums. This suggests that on average and after being subjected to average use and storage, the types originally used are at or approaching the end of their lives.
The failure rate is probably even higher than I suspect as when I get a Spectrum to repair, I usually replace all the electrolytic capacitors before I power it on. I've had a fair few that I've been told were faulty, but worked perfectly after I replaced the capacitors.
why 5 dislike this ? must be zx81 owners ;-)