Get a scope and check the color oscillator is actually running. Also (with it on) try placing a finger gently over the pins of the XTAL, both pins, then each pin in turn, if it suddenly pops into full color then the series capacitors around the XTAL are the wrong value for its resonant mode.
I prefer to bend the leads in towards each other and twist them together to hold the component in place, and it makes snipping the extra easier since the pieces don't go flying off. It also usually keeps the trimmed ends out of the way of other vias / solder points.
Check the video crystal frequency. Color is very dependent on that. Once I managed to buy a wrongly marked batch of 4.433619 crystals, they were about 1KHz off (all 5 of them!) and the picture was just like this, very sharp, but no color.
For all those screaming NTSC! To get a spectrum like this to display an NTSC compatible signal, you’d need a special variant of the ULA, and two different crystals. And even then… in the UK most TVs have been able to display PAL and NTSC for decades. If it’s capable of S-Video, it’ll handle both. Saying that… trying it on a different TV seems sensible I’d also give swapping chips on a regular speccy… I’m sure you have one. The reality is that the ULA is king here, and you’ve kind of tried the LM1889. TR1 and TR2 are the other components to check… with R48/49/50
Be careful with switching voltage regulators, as the switching is high frequency and can introduce noise into the system. Try with a linear one instead to see if it helps. No idea if it’s that, but I know you can’t happily replace the linear regulators without being sure that the circuit can take the noise.
@@TheRetroShack You should worry if the switching regulator is in the rail powering the video circuit. Try looking with an oscilloscope to see if the rail is noisy. If it’s not too noisy, don’t worry. But if you see a lot of noise, maybe try with a 7805.
I know it's stating the obvious but does anything display a colour image on that monitor? The number of calls I attended in my PC support days where someone had turned the brightness down .
Im 100% invested now. Can't wait for part 2! Board is looking amazing btw. I have an issue 2 ZX Spectrum 48k that was upgraded and recapped by a close friend in the UK. I absolutely love this computer and I'm so excited to see how yours turns out when its finished.
I would strongly recommend Andrea Vavassori's "Issue 4V" board rather than this issue 3B clone. This one is a perfect replica, there is absolutely nothing wrong with it, but issue 4V is based on original Issue 4B, which by me (and many others) is considered to be the most stable/reliable 48K board of them all. It has got a better DC-DC circuit (where the -5V and +12V is generated) and Andrea added some very clever options like to be able to use easier to find 4164 DRAMs in lower 16K instead of hard to find and very delicate 4116s, optional footprints to use very commonly available radial caps instead of harder to find (and usually old stock) axial ones, and some more options.
Had this problem developing my reproduction D32 board. Every time it happens the first thing to check is the oscillator signal to the lm1889. No clock means no colour and from experience it was always the lm1889 at fault. These things like to play Russian roulette with a fully loaded gun and don’t need an excuse to fail. There are of course other failure modes and the most likely is no continuity between the modulator the point at which you pick up the signal to inject it in the output. It could of course be the modulator not receiving any colour information from the ULA
You correctly said you were going to solder everything from the lowest profile first and work your way up. So why did you do the power stuff and the tall standing transistors before the sockets? The whole point of the process is so you can still lay it flat helping keep things like the sockets steady and flat while you solder. With the taller power stuff and transistors in place you will have it wobble and the weight of the board wont be helping keep the sockets in place.
I know this sounds silly but have you checked the monitor? When I used to repair Spectrums when I was young the other lads used to think it was funny to turn down the colour on the TV or even pull the aerial from the back.
When I started my retro-tinkering career I watched a lot of videos with recommendations to get the turnpin sockets. I thought to myself: I want to have good electrical contact also for vintage ICs with desoldered and cut legs, so I will buy new turnpin sockets. My mistake was to get cheap ones. They caused me nothing but grief. No electrical contact in some cases and broken pins on vintage extremely expensive ICs in other cases. After inspecting these turnpin sockets with a microscope I saw the barrels were of extremely poor quality, and would either not fit or not clamp the pins very well. I cursed a lot, went back to dual leaf sockets and never had another issue. I probably would have had a better experience if I ponied up the cash for some high quality turnpin sockets, but then again as in this video they have a very tight tolerance, and could cause issues as well. KISS principle... Just go with the standard dual leaf sockets. I now own 7 C64's and 3 spectrums, and never have these dual leaf sockets given me any grief.
the LM chip only generates the colour signal, not the whole video, the spectrum will work black n white with it removed , there is also a 'similar' TI chip that can be used as a drop in equivalent, but its supposedly a bit unreliable , and also obsolete
@ 10:30 - you really should have tested the power rails at this point before inserting any ICs - if the rails are wrong, that's an awful lot of expensive PHUT BANG Phizzzz to come.
Rubbish leaf sockets I bet ! lol! Nothing constructive to add other than sillys like colour on the monitor ok (turned up) and is the 5v clean with that DC DC converter in it as opposed to the venerable 7805 i'd scope it ? I had a flick through the data sheet for th LM1889 and would check values of components around there because the chip can be configured to produce a monochrome output which yours is doing and doing it well ! are all the clocks running at the right speed and running cleanly. Dunno I know nothing about the 'speccy' but I hope (and believe) you crack it soon and it's probably something daft !...cheers.
Have you tried a soldering microscope? I too am visually impaired (20/200 in my left eye, and for all intents and purposes completely blind in the other) and also thought that soldering was beyond my ability. On a whim I decided to buy a cheap microscope (not even a particularly fancy one, just a cheap $40ish one I found randomly on Amazon) and it literally made a world of difference. It takes a little while to get used to basically operating your hands at a distance (not staring at them directly, but instead through a screen) but once you can wrap your brain around that, it's like magic. I can even manage those ridiculously small SMD parts, if my hands aren't too shaky because I've had too much caffeine. :)
I know when we first bought a ZX Spectrum back in the day we had a problem with it. It would be black and white for the first hour or so but once it warmed up it was fine.
what values did you use for r43, r44, and c72, theyre critical ,also try feeding it with an external +12v supply, maybe up to +12.5 or so, as your +12 is a bit under, some of these TV circuit chips are very fussy with supply voltage , datasheet lists it as 12v 'minimum'
you said you tried another LM chip, are these definitely ok? tried in another machine? also try another ULA as it may be faulty, not generating the U and V colour difference signals.. , or try this one in another known good machine to test
I had a similar problem when I built the Harlequin, b/w picture. In my case I'd bridged two pins of the little SMD Analog Devices chip. I know there isn't one on the traditional board though so not sure here!
I wish they would just integrate a simple bridge rectifier on the DC supply input so polarity on the DC adapter doesn’t matter. This was a common tactic used back in the days of CB radio so an accidental reverse polarity didn’t fry your board.
The next best thing to populating a pcb is to watch somebody populating theirs :). One thing I noticed, if you didn't pick it up yourself, was an 'incomplete' solder joint on one of the big caps at the bottom of the board ... it just needs to be reflowed.
Truly enjoy your builds, Swap out the ULA to test, If memory serves there are some other mods required for the ULA+ in an issue 3 ( its been a while ) Looking forward to part 2 =) Z.
AFAIK, Charlie Ingley tests each vLA82 as he makes and ships them out. Of course, it may got somehow damaged on the way from New Zealand, but highly unlikely. Also vLA82 is not a ULA+, so no mods needed, it is a direct replacement for 6C001E-7 ULA.
i can't help but wonder if TR1 & TR2 are the wrong way round ? but this would depend on the pinout of the transistor and are the markings reversed from the original board ? i have some of these boards in my to do pile so looking forward to part 2.. shame that video chip cant be bullied into taking 5 volts
TR1 and TR2 are correct on the silk screen for the original transistors which was not a TO92 package. A common mod is to use a BC549C which is a TO92 package and has the the pins in reverse order when looking at the flat side. So nothing wrong, just not using the original component...
I’m very curious what the problem might be as well. Seems like a longshot, but does that VLA82 FPGA have any configuration options or need a core update?
many black pcb's might look nice but are pain to troubleshoot when you can't see the traces for prototyping it is much easier to get yellow or translusent green or red soldermask black soldermask work nice for identyfing chiped soldermask or damaged traces for bga, but also dark blue with aded bonus of seeing the traces but that is only a silly opinion of a chicken who fixes stuff
Get a scope and check the color oscillator is actually running. Also (with it on) try placing a finger gently over the pins of the XTAL, both pins, then each pin in turn, if it suddenly pops into full color then the series capacitors around the XTAL are the wrong value for its resonant mode.
I prefer to bend the leads in towards each other and twist them together to hold the component in place, and it makes snipping the extra easier since the pieces don't go flying off. It also usually keeps the trimmed ends out of the way of other vias / solder points.
My theory, based on the kind of stupid thing I'd do, is that you have the colour setting turned down to zero on your display.
Check the video crystal frequency. Color is very dependent on that. Once I managed to buy a wrongly marked batch of 4.433619 crystals, they were about 1KHz off (all 5 of them!) and the picture was just like this, very sharp, but no color.
For all those screaming NTSC! To get a spectrum like this to display an NTSC compatible signal, you’d need a special variant of the ULA, and two different crystals. And even then… in the UK most TVs have been able to display PAL and NTSC for decades. If it’s capable of S-Video, it’ll handle both.
Saying that… trying it on a different TV seems sensible
I’d also give swapping chips on a regular speccy… I’m sure you have one. The reality is that the ULA is king here, and you’ve kind of tried the LM1889.
TR1 and TR2 are the other components to check… with R48/49/50
Be careful with switching voltage regulators, as the switching is high frequency and can introduce noise into the system. Try with a linear one instead to see if it helps. No idea if it’s that, but I know you can’t happily replace the linear regulators without being sure that the circuit can take the noise.
Can't hurt to stick a 7805 in and test :) I'll give it a go :)
@@TheRetroShack You should worry if the switching regulator is in the rail powering the video circuit. Try looking with an oscilloscope to see if the rail is noisy. If it’s not too noisy, don’t worry. But if you see a lot of noise, maybe try with a 7805.
That trico regulator has extremely low ripple, but it can't hurt to try
I know it's stating the obvious but does anything display a colour image on that monitor? The number of calls I attended in my PC support days where someone had turned the brightness down .
Im 100% invested now. Can't wait for part 2! Board is looking amazing btw. I have an issue 2 ZX Spectrum 48k that was upgraded and recapped by a close friend in the UK. I absolutely love this computer and I'm so excited to see how yours turns out when its finished.
I would strongly recommend Andrea Vavassori's "Issue 4V" board rather than this issue 3B clone. This one is a perfect replica, there is absolutely nothing wrong with it, but issue 4V is based on original Issue 4B, which by me (and many others) is considered to be the most stable/reliable 48K board of them all. It has got a better DC-DC circuit (where the -5V and +12V is generated) and Andrea added some very clever options like to be able to use easier to find 4164 DRAMs in lower 16K instead of hard to find and very delicate 4116s, optional footprints to use very commonly available radial caps instead of harder to find (and usually old stock) axial ones, and some more options.
Had this problem developing my reproduction D32 board. Every time it happens the first thing to check is the oscillator signal to the lm1889. No clock means no colour and from experience it was always the lm1889 at fault. These things like to play Russian roulette with a fully loaded gun and don’t need an excuse to fail. There are of course other failure modes and the most likely is no continuity between the modulator the point at which you pick up the signal to inject it in the output. It could of course be the modulator not receiving any colour information from the ULA
You correctly said you were going to solder everything from the lowest profile first and work your way up. So why did you do the power stuff and the tall standing transistors before the sockets? The whole point of the process is so you can still lay it flat helping keep things like the sockets steady and flat while you solder. With the taller power stuff and transistors in place you will have it wobble and the weight of the board wont be helping keep the sockets in place.
I know this sounds silly but have you checked the monitor? When I used to repair Spectrums when I was young the other lads used to think it was funny to turn down the colour on the TV or even pull the aerial from the back.
When I started my retro-tinkering career I watched a lot of videos with recommendations to get the turnpin sockets. I thought to myself: I want to have good electrical contact also for vintage ICs with desoldered and cut legs, so I will buy new turnpin sockets. My mistake was to get cheap ones. They caused me nothing but grief. No electrical contact in some cases and broken pins on vintage extremely expensive ICs in other cases. After inspecting these turnpin sockets with a microscope I saw the barrels were of extremely poor quality, and would either not fit or not clamp the pins very well. I cursed a lot, went back to dual leaf sockets and never had another issue. I probably would have had a better experience if I ponied up the cash for some high quality turnpin sockets, but then again as in this video they have a very tight tolerance, and could cause issues as well. KISS principle... Just go with the standard dual leaf sockets. I now own 7 C64's and 3 spectrums, and never have these dual leaf sockets given me any grief.
the LM chip only generates the colour signal, not the whole video, the spectrum will work black n white with it removed , there is also a 'similar' TI chip that can be used as a drop in equivalent, but its supposedly a bit unreliable , and also obsolete
@ 10:30 - you really should have tested the power rails at this point before inserting any ICs - if the rails are wrong, that's an awful lot of expensive PHUT BANG Phizzzz to come.
Rubbish leaf sockets I bet ! lol! Nothing constructive to add other than sillys like colour on the monitor ok (turned up) and is the 5v clean with that DC DC converter in it as opposed to the venerable 7805 i'd scope it ? I had a flick through the data sheet for th LM1889 and would check values of components around there because the chip can be configured to produce a monochrome output which yours is doing and doing it well ! are all the clocks running at the right speed and running cleanly. Dunno I know nothing about the 'speccy' but I hope (and believe) you crack it soon and it's probably something daft !...cheers.
Looking forward to part 2! I am visually impaired and soldering is impossible, so I get my satisfaction from watching videos like this!
Have you tried a soldering microscope? I too am visually impaired (20/200 in my left eye, and for all intents and purposes completely blind in the other) and also thought that soldering was beyond my ability. On a whim I decided to buy a cheap microscope (not even a particularly fancy one, just a cheap $40ish one I found randomly on Amazon) and it literally made a world of difference. It takes a little while to get used to basically operating your hands at a distance (not staring at them directly, but instead through a screen) but once you can wrap your brain around that, it's like magic. I can even manage those ridiculously small SMD parts, if my hands aren't too shaky because I've had too much caffeine. :)
I know when we first bought a ZX Spectrum back in the day we had a problem with it.
It would be black and white for the first hour or so but once it warmed up it was fine.
I had the exact opposite problem. Would start up in colour and as it warmed up (after about an hour) it would go black and white.
what values did you use for r43, r44, and c72, theyre critical ,also try feeding it with an external +12v supply, maybe up to +12.5 or so, as your +12 is a bit under, some of these TV circuit chips are very fussy with supply voltage , datasheet lists it as 12v 'minimum'
It's not going to be something as silly as the TV trying to run NTSC colour mode when the unit is trying to use PAL? or Vice Versa.
I jammed a TSR-1 2450 in my TS1000 and it works great.
At 5:40, regarding your comment about neat and tidy - what happened to that one blue resistor near the middle of the PCB?
Oh yes! I’ll sort that one out :)
you said you tried another LM chip, are these definitely ok? tried in another machine? also try another ULA as it may be faulty, not generating the U and V colour difference signals.. , or try this one in another known good machine to test
It could be a compatibility problem with the ULA. Some issues of the Speccy don't like certain ULA specifications. (?)
Watcha ! RIP Calculon
@@andymouse Hi-ya 🧀
RIP Calculon.
Nothing beats an oscilloscope on the chroma cables, working backwards :)
Really nice soldering work. I also have been too lazy to print up one of those lead bending tools and just do it ad-hock. Turns out fine.
I had a similar problem when I built the Harlequin, b/w picture. In my case I'd bridged two pins of the little SMD Analog Devices chip. I know there isn't one on the traditional board though so not sure here!
I wish they would just integrate a simple bridge rectifier on the DC supply input so polarity on the DC adapter doesn’t matter. This was a common tactic used back in the days of CB radio so an accidental reverse polarity didn’t fry your board.
Suggest you scope the xtal colour oscillator, that it's running, and is at the right frequency?
Anything with PAL vs NTSC?
The next best thing to populating a pcb is to watch somebody populating theirs :). One thing I noticed, if you didn't pick it up yourself, was an 'incomplete' solder joint on one of the big caps at the bottom of the board ... it just needs to be reflowed.
Have you got the wrong frequency for the crystal? You may have configured it as NTSC.
What type of transistor did you use for TR3?
Does your TV support S-vid natively? Or did you use a scart adapter? The TV has to be able to change to S-vid.
Check how may lines are drawn, NTSC/PAL color issue ? Check oscillograms on color ;)
Expecting any color when C connected to a big electrolytic capacitor?
Truly enjoy your builds,
Swap out the ULA to test, If memory serves there are some other mods required for the ULA+ in an issue 3 ( its been a while )
Looking forward to part 2 =) Z.
AFAIK, Charlie Ingley tests each vLA82 as he makes and ships them out. Of course, it may got somehow damaged on the way from New Zealand, but highly unlikely. Also vLA82 is not a ULA+, so no mods needed, it is a direct replacement for 6C001E-7 ULA.
@@borayurt66 - Ahhh thank you I thought it was a ULA+ upgrade - I am un aware of the vLAB2 so off I go to find out...
@12:57 Please tell me those aren't ripped out PCB pads still stuck to the desoldered sockets.
Nope, just solder build up :)
Great video. Where did you get the new coil from?
Looking forward to part 2.
i can't help but wonder if TR1 & TR2 are the wrong way round ? but this would depend on the pinout of the transistor and are the markings reversed from the original board ? i have some of these boards in my to do pile so looking forward to part 2.. shame that video chip cant be bullied into taking 5 volts
TR1 and TR2 are correct on the silk screen for the original transistors which was not a TO92 package. A common mod is to use a BC549C which is a TO92 package and has the the pins in reverse order when looking at the flat side. So nothing wrong, just not using the original component...
Nice video. How long did the soldering take?
i am pretty sure it is the monitor, a B/W mode isn't that easy to establish on a spectrum.
Ask Alfred (the Butler)? 😋👍
Is it a Pal/NTSC problem?
Great episode, looking forward for the 2nd part :-)
I’m very curious what the problem might be as well. Seems like a longshot, but does that VLA82 FPGA have any configuration options or need a core update?
Being a swiss, I'd like to remind you, that the UK is still a European country. It's geography, not related to the EU and the Brexit. 😁
5:45 Erm... Those are signal diodes! 1N4148's
haha nice work mat😁
Colorburst not working?
"Have you already tried turning it off and on again?"
great video
Glad you enjoyed it
Looks like its outputting an ntsc signal to a pal tv that cant handle it.
TR1 and TR2 are reversed (180 degrees) from what I can see.
many black pcb's might look nice but are pain to troubleshoot when you can't see the traces
for prototyping it is much easier to get yellow or translusent green or red soldermask
black soldermask work nice for identyfing chiped soldermask or damaged traces for bga, but also dark blue with aded bonus of seeing the traces
but that is only a silly opinion of a chicken who fixes stuff
not a scoob mate......not helpful in the slightest but honest at least👍🙂
So... Let's get the obvious out of the way. Are you sure the screen you're using is capable of colour?
Turned pin sockets are not designed for dip ic's. They are designed for turned pins. They will just cause problems over time.
Well I’ve been using them and specifying turned pin sockets for DIP packaged IC’s for 39 years without any problems. I’m not the only one either.
Frustrating.
It's not brand new, this clickbait trash deserves punishment, I hoped for a harlequin build.
Why doesn't the lower ram board populate the outer pins of the outer sockets?
it only needs one set of address lines for the 8 chips it replaces and one data line from each (plus a couple of other signals, ras,cas,wr + 0v,5v)