ZX Spectrum +2 Repair (No. 4). Plus a look at methods of improving picture quality.

[@buddhistsympathizer1136]

10 years now since you made this video, but I must say I LOVE your style. Clear, concise and beautifully presented.Well done!

[@jaycee1980]

Came back to this video today for the video filtering stuff.. I realised that if you re-route the way the audio carrier is coupled into the signal, you actually dont need a bandstop filter like this, a simple low pass filter will do. I found that soldering a 150pF capacitor across R31 works very well in this regard! Hope you're safe and well, and maybe you'll start making videos again soon :)

[@EgoShredder]

By the way I love your style; very calm, evenly metered and informative unlike the other videos on KZbin that are aggravating, attention seeking, loud mouthed and full of annoying moronic music throughout, making it impossible to focus on the content.

[@richardjames712]

Agreed. So relaxing to watch (though I don't understand much of it!). Reminds me of UK public information films, or Open University programmes. I mean that in a good way!

[@wettuga2762]

Beautiful detailed explanation about the Spectrum's video signal! I'll definitely make use of your knowledge to improve my picture quality, ghosting is definitely an annoying "feature" of lower quality signals.

[@hadesmcc]

You're videos are excellent and they help me learn stuff. I've only started dipping my toes into electronics and sometimes it's a bit overwhelming. Seeing you explain how and why you diagnose things is a great help. I've recently brought back to life ZX Spectrum +2, without much skill and with a bit of instinct and a lot of luck, I figured the issue was the RAM, I ordered a full set of new ICs and I got it to work properly again. I've made the modifications for Peritel RGB and I get a picture on every TV now, using a Scart RGB cable, before I would only get a picture on certain TVs. Nevertheless, I have an issue with my video signal. I appear to no have green I believe. On the menu, in the bar where the colours are displayed, I don't get yellow and I don't get green, everything else is working fine apart from that. I booted up a game, Count Duckula and the only things in the game not displaying properly are the green elements in the UI, which are dark and the yellow clock counter, which is red and should be yellow. This also happens if I use RF. If you have any advice on how to diagnose and fix this, I'd be very grateful.

[@stevesstuff1450]

I know this video is now 6 years old, and is terrific for correcting a number of issues with this model of Spectrum. However, I have a +2 and use a Din-to-Scart connector, and there is absolutely no problem with the picture quality.... the RGB output (yes, both connector, and my current LG FHD LED TV handle RGB video input) is absolutely rock-solid and clean; no ghosting, no colour artifacting, and using the audio-out patch/loop-back cable on my scart cable, I get very clean, noise-free audio too. I haven't tried it on my old Philips CRT with RGB scart input as yet, but it's something to think about! Be interesting to see how my other Spectrums (48K, and 48K+) handle it; both with composite mods.... I hadn't thought about trying the CRT TV till now..... but I think I will!! Meanwhile, the RGB output is looking fab on my LG LCD... ;-)

[@lawrencemanning]

Very interesting and educational. Love the style, content and production values. I wish I'd known, circa 1986, that the dreadful ghosting could be fixed on the +2s... Keep up the excellent work; can't wait for the next video....

[@stripwax]

Why would TR8 have been installed the wrong way around? Surely, if that was a prior repair, it would have been somewhat obvious that the repair didn't work. Or is the case that it might have worked "for a bit" and the subsequent failure not realised related to the prior TR8 replacement? Excellent investigation and fix of course!

[@BarryManatee]

@David Hooper I'd love to know the answer to this too. I'm having a horrible time with a +2 that keeps blowing TR9. I get no display at all, and TR9 pops within about 5 seconds. I haven't tried reversing TR8 yet, but it would be really helpful to know if the computer you fixed could ever have worked with TR8 installed incorrectly. My board has the same orientation of TR8 as shown here, but I have no idea if someone else has replaced it incorrectly, or if it was like that from the factory. Thanks for any thoughts on this, my head hurts!

[@miklanattila7227]

What a nice idea it is to repair a ZX Spectrum with a Spectrum analyzer :)

[@FireballXL55]

Transistors will work in reverse like this, especially in the emitter follower configuration. So some of those may have worked and hence not noticed during production.

[@jaycee1980]

I was recently reading the AD725 datasheet, as I am connecting this to my +2. The AD725 datasheet specifically mentions a "luma trap" and recommends 68uH + 18pF for it. It also specifically mentions it works against a 1K resistor internal to the AD725. When I read that, I immediately thought "Aha, that might be why JoulesperCoulomb's filter on the TEA2000 works". Indeed, it would seem that this is exactly what you have implemented :) Using LTSPICE i found that your 15uH + 87p combination gives 52dB of attenuation at 4.4MHz, and about 45dB at 4.43MHz. With 90pF it's -34dB at 4.43MHz, which is undoubtedly good enough. It's similar with 82pF too. There'll also likely be differences due to the tolerance of the parts used, specifically the inductor you show is 10% tolerance. In my own +2 I used 10uH + 120pF, and that works fairly well too.

[@faumnamara5181]

I thought I was back in the 80s watching open university before the cartoons started !

[@davidlunchvhs]

I went with a 68pF and 20pF in parallel with the 15mH for the mod. Also did the C31 change. Picture is now excellent, yellow and blue ghosting is gone.

[@GadgetUK164]

I've just done the mod there to add the inductor & cap filter circuit and it works perfectly! Thanks very much for your work with the Spectrum Analyser there - awesome, really glad you found the cause of the ghosting and provided a simple fix! One interesting point - I also tried the mod with cap 31 (under the board on my +2), ie. removing the side that goes to pin 8 of the TEA2000 and connecting that to the emitter of TR2 (exactly as you showed) - I've had to revert that change as it results in horizontal noise of some sort (at a slight angle), and actually brings back the ghosting on my +2 - very strange... Can't see how removing the audio from the composite could in anyway cause those symptoms?!?

[@JoulesperCoulomb]

I've not ever encountered the latter problem. Did the noise appear as soon as the audio carrier connection was removed or only when it was re-applied to the composite feed to the modulator? Were you feeding the display from the composite output or the modulated RF output? Is the audio carrier correct and present when added to the RF output (either as per the original circuit or the modified version I described), i.e. does a TV connected to the RF output receive the Spectrum's sound?

[@GadgetUK164]

Hi, I will check again later but I think the banding appeared when the relevant side of C31 was connected to TR2 emitter. The same thing happened (I think) just by removing C31 from its normal position. It could be related to my LCD TV in some way? (Philips). I am using the composite video out, and the audio sub carrier is present after doing the C31 mod - so routing sound through the modulator works. It's just the weird banding and the return of the ghosting that doesn't make sense to me? I've previously swapped the orientation of the transistor you've shown which is related to composite video out, with regards to collector and emitter as they weren't right on this rev 3 board. So I do get crystal clear video now with no ghosting (except on the cyan bar on the 128 boot menu), so I guess no further mods are required, but I just thought the behaviour I was seeing is strange with regards to C31. When I get time I will upload a video so you can see what I mean. I've got a 60Mhz scope with some video related triggering features but not sure how I could use that to assist me at the moment - I might just have a tinker at some point. I guess this is where a Spectrum Analyser really helps, or maybe a scope with storage etc.

[@JoulesperCoulomb]

Most displays aren't affected by the audio carrier being present on the composite video signal, so you could just leave C31 connected as it was originally. If you want to spend the time to work out why the problem you've encountered occurs, it'll be important to determine whether the problem is triggered simply by disconnecting C31 or only appears when a new connection is made with a capacitor between pin 4 of IC11 and TR2's emitter. If the former is the case, something very odd is going on. If the latter, I would think interference with frequency components around 4.4 MHz is present on the audio carrier. It could be worth trying a different capacitor for C31 or a different position in case it is picking up the interference. With the audio carrier taken directly to TR2 instead of via the TEA2000, it would probably be possible to use a low pass filter consisting of just a capacitor to filter the luminance signal in place of the notch filter. It might even be more effective, though I think using the filter demonstrated makes the picture about as good as it possibly can be.

[@GadgetUK164]

Thanks for the detailed reply btw! Sorry I've been away and had problems with PC and internet on return. Will get back to you as soon as I can.

[@MrSlipstreem]

@@GadgetUK164 Are you back yet? I have what seems like it may be the same problem. Most strange!

[@jamhamster]

Hi, these have been very interesting and your relaxed tone is perfect for these videos. You are highly revered in Spectrum circles and have solved a problem that has been around for decades with the ghosting fix! It's a shame that there are no fixes for the vertical Jailbars on the 128k Toastrack via RGB, Mine has it and I've recapped it, cleaned it and have checked every solder joint, apparently nearly all of them do it. I would try putting some 1uf ceramic capacitors between the 5v and ground on each ram chip if I knew it had a chance of working but I'm too scared to damage mine. If only I had your knowledge! :-)

[@jamhamster]

If anyone is interested in the ToastRack fix, it is here: spectrumforeveryone.com/technical/fixing-vertical-jailbars-48k-128k-spectrum/

[@jamhamster]

I have come up with a Jailbar fix for the +2 issue 3. Swap C52 from a 22uf capacitor to a 100uf Capacitor. Add 8 1uf ceramic capacitors in positions C12 - C19. I have a guide on my channel.

[@HACK3R2K]

The tweak for the ghosting effect should work for +2A , +2B, +3 as well !?

[@JoulesperCoulomb]

+Raducu Daniel Ciocarlan I haven't tried it, but they use the TEA2000 and in a similar way, so I'd think fitting the same filter in the same way might well help.

[@HACK3R2K]

+JoulesperCoulomb Thx. Im going to try and let you know.

[@HACK3R2K]

+JoulesperCoulomb 15uH inductor and 82pf capacitor in series should have the same effect ?

[@pedroferreira2367]

Thank you very much for sharing this. Definitely very useful. I would like to make an observation, if I may, about the composite video out option. While testing a video cable I made myself, using a standard RCA plug to connect the +2 to the TV I got the following results: Pin 1 on the +2 RGB port connected to center pin on RCA and pin 2 connect to ground on RCA shows the speccy menu on black and white. However, when I tried composite pal + composite sync (Pins 1 and 4 on the +2 RGB port) to center pin on RCA and pin 2 to ground RCA, I got the speccy menu in colour. So, two questions arose: 1- Why did I get the black and white picture using composite pal? and 2- Is the composite pal + composite sync solution safe on the long run or will it end up ruining the TEA chip or other components? Sorry for the long post. Regards from Portugal,

[@JoulesperCoulomb]

The situation you describe is a bit strange and suggests something isn't quite right inside the +2. The wiring you first used is the correct way to do it. I don't think connecting pins 1 and 4 together will do any harm, but I wouldn't recommend it and it shouldn't be necessary. It sounds as if the chroma carrier is missing from the composite output, but is present on the composite sync signal. It's not intended to be there at all on the latter, but perhaps just enough leaks through the TEA2000 back to its /CS input for your display to be able to detect it. You should check that link LK4 is in place and that LK2 is not. You should also make sure transistor TR4 is not faulty and is installed the correct way around; in fact, I'd probably replace it even if I couldn't find anything obviously wrong with it.

[@pedroferreira2367]

JoulesperCoulomb Thank you very much for your quick and very detailed reply. I had my suspicions about TR4.I managed to get a couple of days off work for Easter and I'll definitely check TR4 and LK 2 and 4. I'll drop you a line when I get round doing it. Thanks again and keep up the good work with your videos!

[@frazzleface753]

This reminds me of 'EXPERIMENT' on ITV Schools from the 70s. Even the background colour is right. :)

[@seakinggr]

Hi, very good video, is the ghosting fix mod useful for the 128k + toastrack? Regards

[@JoulesperCoulomb]

I'd expect it to work. It uses the same RGB to composite conversion circuit.

[@alrod6228]

I made the c31 mod (attach to tr2 leg) but now only have composite video! Very good composite video but Rf not work anymore. It's normal? Thanks

[@BarryManatee]

Does anyone have an idea why TR9 would keep going pop within about 5 seconds? I've got no display at all. The voltage regulator is working correctly and I've recapped the board. 5v is fine. I'm getting no 12v at the TEA2000 and it's breaking my brain.

[@1337Shockwav3]

Could you just shortly confirm that 100uH & 13pF for the LC-Circuit would be close enough (result is 4.4142MHz)? I currently have a gray +2 here that severly needs that fix. Also, really nice series with a good and in-depth narration.

[@JoulesperCoulomb]

I should think that would work well; give it a try.

[@1337Shockwav3]

JoulesperCoulomb Just gave it a try, using a 12pF capacitor (since I didn't have any additional 1pF at hand) as well as trying a variable capacitor going from 10-60pF (I think at least) results were okay, but nowhere near as good as on your +2. Also seems like my +2 has another issue resulting in a vertical "noise band". Picture comparing all 3 here: i.imgur.com/FIzV4zf.jpg Are you on the world of spectrum forums? Thinking about opening a thread there.

[@1337Shockwav3]

JoulesperCoulomb Seems like lifting C31 did wonders for me :) Picture is now almost as clean as I remember it from a 48k

[@GeoGrig]

Hello! I have a +2 grey case and i can't get the image on a modern TV using RGB signals. I can get Composite image, but very flickery, the same for RF output. and when using RGB on a scart plug, i get black screen. I checked the switching voltages on PIN 8 and 16 on the scart, they are correct. As soon as i connect the composite sync signal to pin 20 on the scart, the picture turns from "no signal" to black. I suspect the TEA 2000 to be faulty. Also, on an old crt TV the picture is ok`ish on RF and composite, but very flickery on my LG led TV. What is your opinion on this problem? Thank you!

[@JoulesperCoulomb]

Unfortunately, I do not know the cause of the problems you are experiencing. If you can obtain a picture via the RF or composite output, the TEA 2000 probably is not faulty. Note that the TEA 2000 is not involved in the generation of any of the signals needed for RGB output. I assume you have examined the video output option wire links in your +2 and checked for any of the video related transistors having been installed incorrectly. I recommend consulting previous threads in the forums at the World of Spectrum web site. There have been many discussions on video output from +2 Spectrums there.

[@EgoShredder]

I have a ZX Spectrum+ 128K 'Toaster' model which worked perfectly when I received it from Ebay in 2010. However one very exhausted evening, I stupidly connected the PSU with the polarity the wrong way around. I believe this is a common thing to happen and usually affects certain components, but a little bit of advice would be appreciated, or maybe you could repair it for me? I'm in West Yorkshire, UK.

[@JoulesperCoulomb]

It probably hasn't suffered especially extensive damage, though very occasionally this type of event destroys most of the digital logic components if the power supply is then re-connected with the correct polarity. This happens when the voltage regulator has failed with a short between its input and output and the external supply voltage gets applied to the 5 V rail. Most likely the regulator and/or the voltage converter circuit that generates the 12 V supply required by the composite video encoder have been damaged. If you want to attempt to repair it yourself, the above information, along with a look at the circuit diagram should give you a good idea of where to start. Alternatively, I can offer a repair service; please contact me via the form at zx.zigg.net if you want to know more about this.

[@trevorrobinson7614]

I've bought a vga interface it costs £50 but it worked fine my 128k spectrum had video problems

[@kotm3]

Hi just wondering if you maybe able to look at my spectrum? Many thanks.

[@JoulesperCoulomb]

Please go to zx.zigg.net and send me a message with a description of the fault or whatever it is you'd like to be done.

[@GadgetUK164]

Love the vids and the channel! Subbed!

[@LindaSunshine]

Ghosting will degrade the sharpness of fonts looks blurry I was playing with this so long time and decided to do new Modul instead of TEA2000.

[@isfiyiywafibc6qaiiiiiiiiii570]

It seems awfully common for people to mount the transistors backwards.

[@ImpossiblyRetro]

Very well done! Narration is excellent indeed. In your video you used a 15uH coil and a 87pF ceramic capacitor but if I try those values on my +2 I get a fuzzy image. In my case I get a better picture if I use 18uH and 47pF ceramic (this was done using a trial and error approach). I understand that according to the formula my values are not the ideal ones but my eyes tell me the other way :/ Coul there be small diferences or slight adjustements made elsewhere on different motherboard revisions of +2 that cause this? Please have a look at: www.worldofspectrum.org/forums/showthread.php?t=47244 I think I will try the variable capacitor solution to check if it is possible to "tune" the image so that it looks the best possible. What do you think about the hint given that is better to use higher inductance and lower capacity to prevent clipping high frequencies in the video signal?

[@JoulesperCoulomb]

The components you used give the filter a resonant frequency rather higher than I'd think would be optimal, but I'd stick with them if it works in your Spectrum. Perhaps there's something about the components you're using or the way you've installed them that gives the filter a low Q and so with a lower centre frequency it cuts some wanted picture information. Generally, I wouldn't think the component values used in this filter are particularly critical so long as the centre frequency lies in approx. 4 - 5 MHz. The best combination might well vary from Spectrum to Spectrum, depending on the precise nature of the interference on the luminance signal. So long as a decent improvement is made, I wouldn't bother trying to tune it too exactly. If you want the very best picture, use the RGB output instead. However, in the Spectrum featured and after the modifications, the picture from the composite output was almost as good as when using the RGB output.

[@ozw1z5rd]

well done

[@azshaw123]

Geek Vision Plus!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!1