Excellent video! I love these repair videos, you don't blindely replace chips or check for broken traces, you actually go from the software level to understand and control the problem, which lets you spot it on the hardware accurately, very informative, I love it !
When I was working as a self-employed IT guy, 30+ years ago (I used to do basically everything IT-related: networking, server and desktop deployment, development, etc) I made more money fixing other's peoples mistakes than taking jobs from the start :)
The CPC video memory makes way more sense when seen as a very long character line that wraps for 25 lines. It also helps to understand why 384 bytes are left unused : there is 48 bytes remaining at the end of each of the 8 pixel lines composing the characters. This memory can even be used to store a little program to restore the palette of a screen dump, by making it executable with an entrypoint located at the first lost memory zone, at &c7d0 ! Anyways, I would really enjoy a video about the CPC video memory, because there is no way i can explain the video ram better than you would.
Yes, that's the real reason for the video layout. I believe the CRTC was intended for text displays and so that's what happened? Don't even get me started on the mapping of the pixels in mode 0. That one I still don't understand why it was done that way to this day. They must have been able to save a gate or two somewhere.
I had a similar issue with an Amiga graphics (RTG) card a while back, where one of the address lines was disconnected due to bad solder job on the card, but it still fully worked, just the memory looped after every 512KB or so. It was quite fun fixing it, not always can one see this clearly how a computer's memory addressing works, but with an address line fault. Great job and video!
That 74LS153N for A13 is not from the same batch as the others. The printing is different and not batch code 8827. I guess it was replaced at some point… Ah, you eventually spotted it.
Noel, really nice to see you again in "old style" repair videos like the ones you used to do (which I much prefer to the new ones). I hope you will be able to stay so much longer in Spain then... 😁 Ps in football there is Mourinho: definitely in repairs you are our "Special One" 😎👏
Fantastic fault finding! Thanks for explaining the inner workings of the machine. Oh and definitely make a video on the repair of the other CPC. It's always fun to see a machine coming back to life again!
nice video, but there's something weird about the explanation, if you see 16:50 when you soldered pin 13 to ground, the output on the screen is mixed with some chunks of the text repeating the upper or the lower part of the characters. The first line repeated the lower part of the text until "Amstrad 64K Microcomput", then it continues repeating the upper part "er (v2)". On the second line, in "()1984" the lower part was repeated and then it starts repeating the upper part for "Amstrad Computer Electronics plc". Finally on the 3rd line the lower part was repeated until "and Locomotive Softwa", and the rest uses the upper part "re Ltd".
You're right! I was so focused on the Ready prompt that I didn't look closely at the rest of the screen. Hmm... and it's not changing from frame to frame, so it's not like a signal left floating that triggers either way. I don't think it has to do with the memory layout, but I suppose it could. Now I'm going to have to puzzle this out 😃
Beautiful video Noel... Welcome back to Europe for a bit! Love the amstrad content. I was half scared that the move to the states would see the end of it .. thankfully not!!😊
Fixing electronics is one thing, but fixing electronics that other people have previously tried to "fix" takes a lot of time and skill as there is no accounting for what actions they took. Very cool to see it done properly like this.
Great find. I can't fault anyone on lifting a trace. I've done it before. Usually though It's pretty apparent because of a new symptom magically appearing. One of the most important things I have learned from you is the importance of using quality solder and having desoldering gun.
Old PCBs from before the mid to late 80s are very prone to lifted traces, I swear. The ColecoVision's traces lift with minimal heat, meaning epoxy is pretty much a must.
Seeing that IC104 was already replaced did smell. But, there has to be a reason for that .. what if the trace was broken before (maybe a manufacturing defect which barely worked for a few years and then broke at some point), they noticed it was this multiplexer, replaced it .. and gave up when it didn't help?
Thank you Noel for all the effort you put in all the jobs... And by the way, Do the other video, all the videos are interesting! ;P I think you have said that, as clickbait!... I THINK, WE ALL WAIT FOR YOUR VIDEOS! Anyway. Thanks again for all the repairs and the research you do.
I haven't looked at it yet, but I will. Maybe it'll also end up making for a fun repair, but it's more likely just to be a bad Z80 or something boring like that.
Good one Noel!!! So you have your Spanish lab working... too bad I didn't know in advance. Maybe you could have had a look at a Mac which has a tendency to blow the power... :)
Omg! I had ZX Spectrum +2 with the keyboard bended like that. Even part of the case broken because someone put weight over it. Lucky the kayboard was working flawless, after I unbend it.
I love the 464 and its descendants - they looked a bit cut-price as did all Amstrad products but they did the job and time has shown they are reliable and not too hard to repair. its a shame they didnt get the software support that the C64 and Spectrum did, although a lot of the games there are good versions. At the time I never had a CPC but its on my list of things to look out for,
The strange video memory arrangement is most likely because it doubles as ram refresh, at least on the Spectrum it is that way and since the CPC video RAM space is also not linear I guess it is for the same reason. Another thing the Spectrum does is use page mode, RAS is set but is followed by 2 column addresses on CAS which saves time, reading 2 bytes with the need of setting up RAS only once. It is a special mode of the 4116 chips. Only this way the Spectrum could meet the timing constraints with those chips. Don't know if the CPC needed this, maybe 4164 chips are faster of Amstrad just used faster (more expensive) RAM? Would be nice to know.
I've seen a similar issue on a C64. The output was repeated, but offset. It was a bad MUX chip as well. Although in this case it was the trace not the chip.
This made me smile, as soon as I saw the effect I went, that's a faulty address line (most likely.) Done something similair on some of my homemade 6502 comps.
16:58 Something weird is happening there, because part of the characters appear as double-top while others remain double-bottom , and the strange thing is that it's consistent with time.
Just started to watch so my guess thats a memory issue, well see if my guess is correct. Probably 2 address lines from video controller are connected or something like that.
Nice to see an repair video from you agian 🙂 I noticed that with the v2 ROM the BASIC text says 'Amstrad 64K Microcomputer' and with the s3 ROM is says 'Amstrad 128K Microcomputer', is this just text in the ROM as the diagnostics show that there is no upper RAM available?
@2:17 my guess is that it looks like a high bit line is fried (always pulled low - broken would float and be even weirder)... Could be a character rom issue? It doesn't feel like a memory issue. Well, let's sit back and see!
It is so good to see you in action again, Noel! Egad, man! You are a freakin' bloodhound on the trail! So very well done finding that trace. You may have covered this before, but do boards exist to construct any CPC series systems from new parts? (That would be such a kickin' little project.)
Thanks! Yes, there are already several versions of CPC boards that you can build yourself. Check out this video: ruclips.net/video/e0eCOkU5HNA/видео.html
Wow, I didn't realize that amstrad wouldn't have been text mode with a character ROM. I didn't think the fully rendered text using bitmap modes started until Amigas, Atari ST, Macs, etc took over.
I didn't notice anything out of the ordinary and nobody has complained. Is it still happening for you? I always try to make sure audio is as clean as possible.
Other people's bad repairs??? Hey, don't look at *me*... (I think I my have already done a bad replacement on the NiCad battery in my Tandy Model 100...)
Excellent video!
I love these repair videos, you don't blindely replace chips or check for broken traces, you actually go from the software level to understand and control the problem, which lets you spot it on the hardware accurately, very informative, I love it !
Noel. Has a very strong understanding on the sequence of operation on these old computers. It's critical in properly diagnosing issues such as these
When I was working as a self-employed IT guy, 30+ years ago (I used to do basically everything IT-related: networking, server and desktop deployment, development, etc) I made more money fixing other's peoples mistakes than taking jobs from the start :)
The diagnosis based on first principles is impressive. Great detective work Inspector! See you in the next video!
The CPC video memory makes way more sense when seen as a very long character line that wraps for 25 lines. It also helps to understand why 384 bytes are left unused : there is 48 bytes remaining at the end of each of the 8 pixel lines composing the characters. This memory can even be used to store a little program to restore the palette of a screen dump, by making it executable with an entrypoint located at the first lost memory zone, at &c7d0 ! Anyways, I would really enjoy a video about the CPC video memory, because there is no way i can explain the video ram better than you would.
Yes, that's the real reason for the video layout. I believe the CRTC was intended for text displays and so that's what happened? Don't even get me started on the mapping of the pixels in mode 0. That one I still don't understand why it was done that way to this day. They must have been able to save a gate or two somewhere.
Video please, this sounds like a worthy entry in the long line of hacked-together 8-bits😀
It's great that you explained the Z80 bus before fix. I'd like you always do it.
Really enjoyed this video. I liked the process of fault-finding you took. Great stuff!
Awesome. Thanks! 👍
I had a similar issue with an Amiga graphics (RTG) card a while back, where one of the address lines was disconnected due to bad solder job on the card, but it still fully worked, just the memory looped after every 512KB or so. It was quite fun fixing it, not always can one see this clearly how a computer's memory addressing works, but with an address line fault. Great job and video!
That 74LS153N for A13 is not from the same batch as the others. The printing is different and not batch code 8827. I guess it was replaced at some point…
Ah, you eventually spotted it.
Haha, yes, the writing was totally different. I can't believe it took me that long to spot it! 🤪
RAS = row address STROBE, same for CAS
Love the graphic that shows where you are putting your probes!
I congratulate you Noel on another well manufactured video to attract watchers......😉☺
Noel, really nice to see you again in "old style" repair videos like the ones you used to do (which I much prefer to the new ones). I hope you will be able to stay so much longer in Spain then... 😁
Ps in football there is Mourinho: definitely in repairs you are our "Special One" 😎👏
Fantastic fault finding! Thanks for explaining the inner workings of the machine.
Oh and definitely make a video on the repair of the other CPC. It's always fun to see a machine coming back to life again!
You need to have another go at that amstrad gx4000 console
I do! I do! I have a whole stack of them sitting next to me. I will conquer that monster some day 😃
nice video, but there's something weird about the explanation, if you see 16:50 when you soldered pin 13 to ground, the output on the screen is mixed with some chunks of the text repeating the upper or the lower part of the characters. The first line repeated the lower part of the text until "Amstrad 64K Microcomput", then it continues repeating the upper part "er (v2)". On the second line, in "()1984" the lower part was repeated and then it starts repeating the upper part for "Amstrad Computer Electronics plc". Finally on the 3rd line the lower part was repeated until "and Locomotive Softwa", and the rest uses the upper part "re Ltd".
You're right! I was so focused on the Ready prompt that I didn't look closely at the rest of the screen. Hmm... and it's not changing from frame to frame, so it's not like a signal left floating that triggers either way. I don't think it has to do with the memory layout, but I suppose it could. Now I'm going to have to puzzle this out 😃
Your level of professionalism and understanding of operations is stellar. Thank you for all your hard work.
the S in RAS and CAS stands for 'strobe'.
Beautiful video Noel... Welcome back to Europe for a bit!
Love the amstrad content. I was half scared that the move to the states would see the end of it .. thankfully not!!😊
Oh no chance of that! I have two very interesting Amstrads in the US that need some work. You can expect more Amstrad content even from there!
Great fault finding, Noel.
Feels like it's been ages since the last CPC repair vid from you Noel. This was a pleasant surprise.
Noel. Would love to see the video of the 2nd amstrad repair as well. I love to see these machines rise from the ashes.
Funny how a single bad connection can cause such trouble.
Fixing electronics is one thing, but fixing electronics that other people have previously tried to "fix" takes a lot of time and skill as there is no accounting for what actions they took. Very cool to see it done properly like this.
And between those two, you have the issue of fixing your own mistakes many years later and forgetting what you were doing to begin with.
great video as usual. thanks.
Absolutely phenomenal diagnosis! I’m in awe 😮
Can you fix a ZX-Spectrum clone? 😉
Absolutely! ZX Spectrums (and clones) are pretty simple and definitely fixable.
Great! Those in depth troubleshooting videos are my favorites!
Glad to hear that! 👍
Great find. I can't fault anyone on lifting a trace. I've done it before. Usually though It's pretty apparent because of a new symptom magically appearing. One of the most important things I have learned from you is the importance of using quality solder and having desoldering gun.
Old PCBs from before the mid to late 80s are very prone to lifted traces, I swear. The ColecoVision's traces lift with minimal heat, meaning epoxy is pretty much a must.
Nice!
🐨Yeah nah NOPE!
You just got 'a stream us the 2nd Puter repair scope.
APU 👁 👉👍👈 Yo Video..
Really good repair video and especially diagnostics from a quite strange fault.
Seeing that IC104 was already replaced did smell. But, there has to be a reason for that .. what if the trace was broken before (maybe a manufacturing defect which barely worked for a few years and then broke at some point), they noticed it was this multiplexer, replaced it .. and gave up when it didn't help?
Thank you Noel for all the effort you put in all the jobs...
And by the way, Do the other video, all the videos are interesting! ;P
I think you have said that, as clickbait!...
I THINK, WE ALL WAIT FOR YOUR VIDEOS!
Anyway. Thanks again for all the repairs and the research you do.
Brilliant, Noel. Please make another video reparing the other 464. Very interesting,😊
I haven't looked at it yet, but I will. Maybe it'll also end up making for a fun repair, but it's more likely just to be a bad Z80 or something boring like that.
Good one Noel!!! So you have your Spanish lab working... too bad I didn't know in advance. Maybe you could have had a look at a Mac which has a tendency to blow the power... :)
Hì Ya & best wishes. Thanks for work. Be Happy. Sevastopol/Crimea.
Another magnificent video from Noel 😀. Those problems with "refurbished" computers are so common on eBay ones...
Omg! I had ZX Spectrum +2 with the keyboard bended like that. Even part of the case broken because someone put weight over it. Lucky the kayboard was working flawless, after I unbend it.
Excellent T-shirt 😊 my favourite game from my childhood
I love the 464 and its descendants - they looked a bit cut-price as did all Amstrad products but they did the job and time has shown they are reliable and not too hard to repair. its a shame they didnt get the software support that the C64 and Spectrum did, although a lot of the games there are good versions. At the time I never had a CPC but its on my list of things to look out for,
Even the boring repairs are informative !! please don't worry about making lots and lots of "the same" content.
The strange video memory arrangement is most likely because it doubles as ram refresh, at least on the Spectrum it is that way and since the CPC video RAM space is also not linear I guess it is for the same reason. Another thing the Spectrum does is use page mode, RAS is set but is followed by 2 column addresses on CAS which saves time, reading 2 bytes with the need of setting up RAS only once. It is a special mode of the 4116 chips. Only this way the Spectrum could meet the timing constraints with those chips. Don't know if the CPC needed this, maybe 4164 chips are faster of Amstrad just used faster (more expensive) RAM? Would be nice to know.
I've seen a similar issue on a C64. The output was repeated, but offset. It was a bad MUX chip as well. Although in this case it was the trace not the chip.
It's time to resurrect the ZX Spectrum baud load challenge.
Great video. The CPC464 gave me and my brother some great memories back in the day. Thanks for sharing your knowledge ☺
The last rom (s3) shows up as 128k not 64k. Would this be an issue having only 64kb onboard?
Noel’s Spanish location keeps on giving! Thank you for the interesting video.
That was a really interesting repair -- but *dang* that Sabre Wulf shirt looks so good
Great video! Plus for the edu part on muliplexers, kind of Ben Eater style ;)
This made me smile, as soon as I saw the effect I went, that's a faulty address line (most likely.) Done something similair on some of my homemade 6502 comps.
Welcome back, Noel. Your videos are always the best!!!
So great when you fix an Amstrad 🙂
They are very dear to my heart 😃
Having seen it before, my first impression was a stuck row address line in character rom.
It great when a fault can be diagnosed just by looking at the screen.
Somebody either bent the keyboard in rage or had it underneath a bunch of heavy stuff.
16:58 Something weird is happening there, because part of the characters appear as double-top while others remain double-bottom , and the strange thing is that it's consistent with time.
Just started to watch so my guess thats a memory issue, well see if my guess is correct. Probably 2 address lines from video controller are connected or something like that.
Like and comment so we get to see second amstrad fixed. :)
Welcome back from colonies.
Took me a few minutes to realize that you're not talking about Gatorade, but a gate array. 🙂
Welcome back to Spain!! Hope you have AC!!
So that doubling of the character set is actually how my eyes work.
Those are the jobs I like, the ones where other people have screwed up and given up, more fun when it's a challenge.
Did you find the any key yet, eh? I lost mine again. ;)
I was missing your repair videos. Thanks!
Wow an excellent fix! Cool to see the old lab.
Nice to see an repair video from you agian 🙂
I noticed that with the v2 ROM the BASIC text says 'Amstrad 64K Microcomputer' and with the s3 ROM is says 'Amstrad 128K Microcomputer', is this just text in the ROM as the diagnostics show that there is no upper RAM available?
Yes, that's correct. It's baked into the ROM since it was only for the 6128.
keyboard definitely met fist on that one :)
Welcomme to Spain i hope you enjoy the stay :)
@2:17 my guess is that it looks like a high bit line is fried (always pulled low - broken would float and be even weirder)... Could be a character rom issue? It doesn't feel like a memory issue. Well, let's sit back and see!
Ok, that is a really weird memory layout. Sorta like Apple II memory weird. I would have expected more randomness from a floating line. Interesting!
It is so good to see you in action again, Noel! Egad, man! You are a freakin' bloodhound on the trail! So very well done finding that trace.
You may have covered this before, but do boards exist to construct any CPC series systems from new parts? (That would be such a kickin' little project.)
Thanks! Yes, there are already several versions of CPC boards that you can build yourself. Check out this video: ruclips.net/video/e0eCOkU5HNA/видео.html
@@NoelsRetroLab Thank you so much! I will.
Great content. Thanks!
Jam in all new ceramic caps.
Wow, I didn't realize that amstrad wouldn't have been text mode with a character ROM. I didn't think the fully rendered text using bitmap modes started until Amigas, Atari ST, Macs, etc took over.
Yes, that's one of the reasons Amstrad graphics are so slow compared to other platforms.
It's also the same with the Spectrum: there's no actual text mode, and all character graphics are done via software routines in the ROM.
The joys of fixing something that someone else broke while "fixing" it... :S
Nice debugging !
Great stuff❤😊
Is it just me or is there a lot of crackle in this video?
I didn't notice anything out of the ordinary and nobody has complained. Is it still happening for you? I always try to make sure audio is as clean as possible.
❤ 👍
Schematics on the big monitor looks like something you'd see in a SciFi movie it feels like a vision of the future the way you'd want things to be
¡ñ!
Have a nice dat.
Other people's bad repairs??? Hey, don't look at *me*...
(I think I my have already done a bad replacement on the NiCad battery in my Tandy Model 100...)