May I just re-iterate what other people have said. The sound issue only affects machines fitted with the Z70830 motherboard. Due to the way this board was deisgned, it can either be populated with circuitry to enable the tape player in the +2A or floppy disk controller in the +3(A). Therfore the Z70830 board can be found in the +2A (Black) and earlier +3(A) machines. The later +2B and +3B machines had new re-designed boards that were non-interchangeable between +3 and +2 configurations. These later boards (Z70833, +2B) and (Z70835, +3B) were re-engineered with this fault rectified from the factory. I suspect the +2 you were comparing your +3 (evidently equipped with a Z70830 board) to, was a +2B (with a Z70833 board), or a mis-labelled +2A that actually contains the later Z70833 (+2B) board (good old Amstrad!). People watching your video need to confirm which board they have and only need to atempt a repair if they discover they have a Z70830 board, which can occur in +2As (black case) or +3(A)s.
It doesn't detract from the fact that it's a good, well explained video that talks you through the problem, unlike those that are just random prodding and poking (and hoping that the problem goes away!).
@@matthewhill5090 Your first comment is 100% correct, but i would like to add that some +2B boards were installed in +2A cases due to an excess of those cases when Amstrad released the new board; this may also confuse people who haven't looked at the actual boards inside the machine. In essence, the only way to be absolutely certain of the board is to open the case and look inside. P.S. Did you know that the Z70833 borad was released in three different revisions? They are called Issue 1, Issue 2, and Issue 4, respectively. I don't know if there was an Issue-3 board released for sale, i have never come across any mention of such a board.
@@BertGrink Sorry, I think it's badly worded, but I did make the remark that there's some +2As with +2B boards inside. I hoped that the overall feeling from my remark was that users should look inside any (black) +2Bs and any +3s, to see if they've got the board that ends '030'. I'm tempted to go and dig my +3 out. I thought it was one of the last ones, but as it was the shop's last one (their demonstrator), it could be an early one! As to the issues, I am not aware of an Issue 3 for the '033'. Maybe the design wasn't good enough and gave up with it.
@@matthewhill5090 Now that i re-read your first comment, i see that you did indeed mention about the +2B boards in +2A cases, therefore the duty to apologise is mine, not yours. ;) but hey, it must be better to have duplicate information than omitting it altogether, yes? Re. the missing Issue 3 board, my thoughts were that perhaps the number of corrections grew so large that Amstrad decided to skip that version and go directly to Issue 4 in a manner similar to Motorola skipping the MC68050 CPU in favour of the 68060; it's just a thought and i could be entirely wrong about that, hehe.
You just brought back the memory of my EE professor of 35 years ago. He always said emitter feedback via an emitter resistor was the best way to manage beta instability of a CE amplifier. Thanks for the happy memory!
Wow. Before I started watching this video I guessed a bad capacitator or audio chip. And it turned out to be a flawed original circuit design. Well done!
Absolutely fascinating. Thanks for talking us through the process of simulating the circuit and testing the changes in realtime. It's definitely something I'll be using for my projects in future. 👍
Excellent explanation of a fault documented for a few years. Your video will further improve people’s enjoyment of the +3. I applied mods to my +3’s about 2 years ago and nothing has gone wrong people!
I had a +3 back in the day, and always wondered why the sound was so bad compared to my friend's +2. I also discovered that the sound output through the Tape/Sound port was a little clearer (still not perfect though) so I often had my +3's sound being amplified by the cassette recorder rather than going through the TV.
I have one of these +3's. There were a couple of mistakes when these were made. Not only is the sound terrible, the video circuit presents a blue/yellow ghosting effect (can be seen on the spectrum's menu) due to noise in the circuit. Thankfully, there are pretty easy modifications to rectify both faults. Thanks Noel.
@@NoelsRetroLab Not sure about RGB but definitely composite. I did the modification from this video : th-cam.com/video/GMKGIP5Yuqs/w-d-xo.html. Starts about 18.25.
This was a great video! I have never seen the action of taking away sections of a schematic. I never knew that can do such a thing! That was a great lesson about circuit isolation, thank you!
You're welcome. I find that with really complex circuits, trying to isolate things helps a lot (if you can, because sometimes they're completely interconnected).
I thought all 128k spectrum music just sounded like that! :) Also love the nostalgia at 18:25 where the sound adds a buzzing noise of video interference when Cybernoid goes to the "hall of fame" screen!
Haha! Things you get used to when you grow up with them! 😃 I totally get the nostalgia even just from the buzzing noise (and that was me putting a noise filter on that as well, otherwise it was a lot more noticeable!).
Mister : only one word : respect !!! and total respect .... a very good idea !! I love your chanel and i'm learning many things with you knowlege ... Thanks for the share Eric from France ...
Great video, but your nomenclature may be confusing to those who are not in the know. What you refer to as a Spectrum +2 is, in fact, a Spectrum +2B-an updated version of the Spectrum +2A (the changes primarily concerned the sound output circuitry). There was a (rather rare) Spectrum +3B, which inherited the same changes. The Spectrum +2 is Amstrad’s first Spectrum, which was a reverse-engineered Sinclair’s 128K machine. (I’m sure you know all of this; it’s just that it might sound confusing to some people.)
Yes, you're totally right. I had first planned on doing a timeline of Spectrum models at the beginning, but I quickly ran out of time with everything else involved in this video (and even so, I released it a day late!). So thanks for adding that. It'll definitely help when some people are scratching their heads wondering why the +2 had the fixed version of the +3 circuit :-)
The +2A actually used the same circuit board as the +3, but without the ICs related to the disk drive, i.e. the 765 FDC, the SED9420 Data Separator, and a few "glue logic" ICs; instead it had a pin header for connecting the Datacorder. The +2B models had a smaller PCB with a jagged ( or "staircase") footprint with no provisions for the FDC circuitry. Oddly, though, the +2B boards still had solder pads for the floppy disk power cable. Great video by the way, Noel, you should give yourself a pat on the shoulder for a job well done. The circuit simulation was very helpful to get a better understanding of the goings-on. 😁 I'm certain i shall have a play-around with that simulator in the future, so thanks for linking to it. Greetings from Denmark. P.S. I hope you stay safe and healthy! Take care.
@@ElectronGordo It not only _could_ have the same problem, but to the best of my knowledge, it *HAS* that problem too - which is why Amstrad made the (slightly) cost-reduced +2B which not only fixed the sound issue, but also used less PCB material. I can illustrate this point by linking to a picture of the +2A PCB: hardware.speccy.org/imagenes/ordenadores/spectrum+2A-issue2-larga.jpg and one of the +2B: hardware.speccy.org/imagenes/ordenadores/spectrum+2A-issue1.jpg. At a guess, i would say this board configuration could save perhaps 20% PCB material. Around the same time, Amstrad also issued the (very rare) +3B: 2.bp.blogspot.com/-td3itcIQGSs/XIT1rC9IruI/AAAAAAAAQ4o/K0-AfR9NR90TAnaX6s7s97oS26-9myJegCLcBGAs/s1600/plus%2B3%2Bplaca1.jpg
@@ElectronGordo P.S. To illustrate how they could save PCB material, i have made this quick and dirty picture: bbeea579-a-62cb3a1a-s-sites.googlegroups.com/site/bertyfromdk/home/misc/2xspectrum%2B2A-issue2.jpg?attachauth=ANoY7cqOOP6ZHAo3_eiSqBuGO5tH3JEdL4fya3YKbKvnLUJl3fiZsiX_LSBbGc3yOwLMxU4MtlsunsS9Y9exWeDjAfcvEviSVEn0eTRp-F69H_XFKahVyiQ1VYxVJkAWsmV6Q9GlzcGxrrRweEpSNZ-Tg5qv26SNjs_SNlworAkD10klad5lP-g72p-21DOXQEDGZfmj07iSAe0WIRRAd-xCUvVEjtGg4NDWe5IDU6GMwKBV8ETAgMvlKbBQoT5HzDeHcEZLqn8_&attredirects=0
Thanks Noel for all your efforts. I have a lot of speccys, some +2a (short boards), +2b and a +3, and I never have noticed any sound distortion. Does it mean that i have one of the strange +3B? I just got a long board +2a and i plan to test it to see if it has that sound distortion. On the other side, I have heard that you need to add a capacitor to enable the tape/sound functionality of the +3 on the long board +2a. Could you tell me how? Thanks in advance.
Great video for me to start to understand the AY chip - i have an Amstrad 6128 on the way from the UK to me in the US, my first computer from the other side of the pond. I blame your videos for my purchase. lol.
@@IDPhotoMan Although I'm a Speccy guy myself, I am sure you will have a lot of fun with the CPC; in fact it has been on my own wishlist for quite a while. Oh and about a year ago i took the plunge and acquired a Commodore 128D (CR), a 1084S monitor, a Datasette, a 1750 REU and some other bits and bobs. That is a really nice machine too. 😊
@@BertGrink Awesome on the 128D! I still have my 128 (not the D) that i bought in high school when i sold my original C64 (DOH!), along with way too many other Commodores lol. I should have gotten a D a few years ago instead of waiting, as now they are about $400US for even not-perfect machines. Ah well. Now to find a Color Amstrad monitor some day (I have a few PAL LCD's i can use for now).
@@IDPhotoMan Well, i don't think getting a C128 instead of the C64 is such a bad exchange; after all the 128 has a C64 mode built in. I paid roughly US$320 for my 128D, and the seller had included all the system diskettes plus some GEOS disks and a game (Outrun Europe), a few cables, a Quickshot joystick, and as the _piece de resistance,_ a Commodore 1351 mouse! All in all a fair price in my opinion. Best of luck finding the Amstrad color monitor. 😊 P.S. Aside from the inevitable yellowing of the plastic parts, my 128 is in perfect condition.
Been watching many random videos from you lately to help spark my interest in analog electronics again because lately I've been doing VLSI/digital over and over and this video especially did the trick! I gotta say though, I miss the good old devices with reasonably-sized components, with my sausage fingers and unstable hands it's a nightmare handling or repairing the ones we have nowadays with resistors smaller than 1mm, not to mention those damn QFPs that are EVERYWHERE (these are what killed my interest originally)...
Glad those videos inspired you! I agree that older electronics are much easier to interact with than modern ones. And modern tech makes it easier to interface with older tech, so it's a big win. That's one of the reasons I love retro tech in general!
I noticed on the cybernoid music at the end, there's some sound interference from the ?video? circuit, which my spectrum has as well. I'm wondering if there is a fix for this as well?
Yes, you're right. It's very noticeable because it changes pitch when the screen changes. It could be an interesting project to try to figure out why that happens and how to solve it. Maybe some of the switching noise from the RAM makes it into the amplifier stage. Not sure.
@@NoelsRetroLab Yes, mine does exactly the same. The hum changes when the screen switches. I only notice it when not playing sound. It might be as you say something to do with RAM switching, as I'm sure you can hear it change when the computer boots and the screen goes from black to white.
@@WDeranged I've heard it's enough to remove just a single capacitor but its number depends on the particular PCB revision that you've got. On +2A it's C24 I think. This should de-couple audio from video but you won't be able to hear music through RF modulator output (which nobody uses anyway). Another thing that helps is to disconnect the tape deck from the main PCB. Even when I completely re-capped the tape deck PCB and re-did some of the oxidized solder joints, the annoying hiss persisted. Although this hack is only good if you use a mass storage interface.
Seems the Amstrad (or was it still Sinclair?) engineers skipped a class on proper BJT biasing :-) Which simulation software are you using here? This video is very instructional, great content and presentation. Keep up your good work!
Thanks! It was definitely Amstrad at this time. I put a link to the actual simulations in the description. The online simulation is great! tinyurl.com/vmtcsaq
Hi Noel. Fantastic video! The explanation of the circuit was really something. I fail to understand what changed in reality between the adding of the 220 ohm to TR5 and all the rest of the changes you did after that. I mean, what's different? I have a +2A that had the same issue (it's the same board as the +3) and I've added a 1K ohm resistor instead of a 220 ohm (I followed a guide from some place I can't recall), and the issue with the sound was fixed. I'm quite pleased What's the benefit of doing all the subsequent changes after adding that only resistor?
Thanks! Glad you liked it. The main change is that resistor. The rest is only if you want to match the circuit of the fixed +2A. I don't see any major difference. Maybe the volume levels will be a bit better with all the changes, but I think it's pretty minor. I just did it to be on the safe side :-b
Hi Noel, great video as always. I was wondering if you have an opinion on what the difference is between the simpler versions of this where just the tr5 resistor is added, then three resistors, then this many resistor version?
Thanks! I haven't done a full analysis to be able to tell the difference really well. The single resistor mod improves the sound hugely. The rest... I wasn't able to tell the difference, but I figured they were set up that way in the +2A by people who really knew what was going on, so maybe it's better biasing, or the way it mixes with the input sound or something. If you're going to do it, I'd go with the full resistor change if you cant just to be 100% sure.
It doesn't matter much, but in general you should use 10x whenever possible(so with any signal higher than a few millivolts). 1x reduces your bandwidth significantly (usually down to a few MHz) and may have you miss some interference.
What about a digital solution where the audio signal get digitized via a ADC then processed in a DSP chip or a suitably programmed FPGA then output to the DAC
Thank you for your quick answer. Going back to the Spectrum+3 sound issue I have this wonderful interface bought at Byte Delight, it includes a jack output to connect a pair of loudspeakers www.bytedelight.com/?product=zx-ay-external-stereo-audio-interface-for-all-zx-spectrum-models
swapping R and C around in a series coupling makes absolutely no difference - neither theoretically nor in reality. It could be that it was purely because of board layout reasons (it's easier to pass traces under a resistor than a capacitor).
Can you add a track list so those of us unfamiliar with the music can find it? Great videos, BTW - it's helping me out doing component repair on other electronics!
@@xvenomtype0x I keep forgetting I didn't have that one in the credits at the end. I just added it to the description. Same author as the ending song opengameart.org/content/theme-song-8-bit
I LOVE to see PROs like you working! ;) I wish I was a PRO like you too! LOL. Anyway, THANKS a lot for the GREAT video and explanations! One question from someone who doesn't understand much of this: why would Amstrad change the "circuitry" from +2 to +3 if the result is so much worse?? What happened on their minds to mess so much with the sound?!
Haha thanks! Look at the pinned comment. I failed to mention the correct sequence of fabrication in the video. Basically, they released a version of the +2 and the +3 at the same time with the sound bug in it. Then they released an updated version of the +2 with it fixed (and supposedly a fixed version of the +3 but that's less common I think). So yeah, the fixed +2 came later.
I'm going to do this mod as per Noel, but I wonder how it compares to the mod on Spectrum for Everyone? Line below, Noel's has less work :) spectrumforeveryone.com/technical/fixing-audio-distortion-spectrum-3/
They're pretty much the same. We're just copying the "fixed" circuit from the ZX Spectrum +2 that came out later. It looks like there they're changing a couple extra resistors related to the MIC and EAR lines (only useful if you're going to be loading from tape). If you want to be extra sure, do them all. Otherwise, if you want the least work, just add that single resistor and don't worry about changing the rest.
May I just re-iterate what other people have said. The sound issue only affects machines fitted with the Z70830 motherboard. Due to the way this board was deisgned, it can either be populated with circuitry to enable the tape player in the +2A or floppy disk controller in the +3(A). Therfore the Z70830 board can be found in the +2A (Black) and earlier +3(A) machines. The later +2B and +3B machines had new re-designed boards that were non-interchangeable between +3 and +2 configurations. These later boards (Z70833, +2B) and (Z70835, +3B) were re-engineered with this fault rectified from the factory. I suspect the +2 you were comparing your +3 (evidently equipped with a Z70830 board) to, was a +2B (with a Z70833 board), or a mis-labelled +2A that actually contains the later Z70833 (+2B) board (good old Amstrad!). People watching your video need to confirm which board they have and only need to atempt a repair if they discover they have a Z70830 board, which can occur in +2As (black case) or +3(A)s.
Thanks for the clarification. That's very useful!
It doesn't detract from the fact that it's a good, well explained video that talks you through the problem, unlike those that are just random prodding and poking (and hoping that the problem goes away!).
@@matthewhill5090 Your first comment is 100% correct, but i would like to add that some +2B boards were installed in +2A cases due to an excess of those cases when Amstrad released the new board; this may also confuse people who haven't looked at the actual boards inside the machine. In essence, the only way to be absolutely certain of the board is to open the case and look inside.
P.S. Did you know that the Z70833 borad was released in three different revisions? They are called Issue 1, Issue 2, and Issue 4, respectively. I don't know if there was an Issue-3 board released for sale, i have never come across any mention of such a board.
@@BertGrink Sorry, I think it's badly worded, but I did make the remark that there's some +2As with +2B boards inside. I hoped that the overall feeling from my remark was that users should look inside any (black) +2Bs and any +3s, to see if they've got the board that ends '030'. I'm tempted to go and dig my +3 out. I thought it was one of the last ones, but as it was the shop's last one (their demonstrator), it could be an early one! As to the issues, I am not aware of an Issue 3 for the '033'. Maybe the design wasn't good enough and gave up with it.
@@matthewhill5090 Now that i re-read your first comment, i see that you did indeed mention about the +2B boards in +2A cases, therefore the duty to apologise is mine, not yours. ;) but hey, it must be better to have duplicate information than omitting it altogether, yes?
Re. the missing Issue 3 board, my thoughts were that perhaps the number of corrections grew so large that Amstrad decided to skip that version and go directly to Issue 4 in a manner similar to Motorola skipping the MC68050 CPU in favour of the 68060; it's just a thought and i could be entirely wrong about that, hehe.
You just brought back the memory of my EE professor of 35 years ago. He always said emitter feedback via an emitter resistor was the best way to manage beta instability of a CE amplifier. Thanks for the happy memory!
Wow. Before I started watching this video I guessed a bad capacitator or audio chip. And it turned out to be a flawed original circuit design. Well done!
Absolutely fascinating. Thanks for talking us through the process of simulating the circuit and testing the changes in realtime. It's definitely something I'll be using for my projects in future. 👍
Excellent explanation of a fault documented for a few years. Your video will further improve people’s enjoyment of the +3. I applied mods to my +3’s about 2 years ago and nothing has gone wrong people!
I had a +3 back in the day, and always wondered why the sound was so bad compared to my friend's +2. I also discovered that the sound output through the Tape/Sound port was a little clearer (still not perfect though) so I often had my +3's sound being amplified by the cassette recorder rather than going through the TV.
Hi Noel great video and easy to follow instructions on how to fix the +3 sound problem after a few hours work I now have a fully working +3
Glad it helped! It makes such a huge difference, doesn't it? Shame on Amstrad for shipping it that way! 😃
I have one of these +3's. There were a couple of mistakes when these were made. Not only is the sound terrible, the video circuit presents a blue/yellow ghosting effect (can be seen on the spectrum's menu) due to noise in the circuit. Thankfully, there are pretty easy modifications to rectify both faults. Thanks Noel.
The video ghosting effect you're getting has to be through RF or composite out, right? I can't imagine the RGB out having that kind of artifact.
@@NoelsRetroLab Not sure about RGB but definitely composite. I did the modification from this video : th-cam.com/video/GMKGIP5Yuqs/w-d-xo.html. Starts about 18.25.
I knew the +3 had bad sound, but I did'nt know it was that bad. Nice fix.
Thanks. Yeah, I was blown away when I repaired this machine last time how bad it sounded. I also thought it wasn't that bad.
I. cannot. stop. watching. videos. by. Noel.
Glad to hear that!
This was a great video! I have never seen the action of taking away sections of a schematic. I never knew that can do such a thing! That was a great lesson about circuit isolation, thank you!
You're welcome. I find that with really complex circuits, trying to isolate things helps a lot (if you can, because sometimes they're completely interconnected).
Adding the emitter resistor is intended to reduce the variation in circuit gain due to the variation in gain of transistors in production.
I thought all 128k spectrum music just sounded like that! :) Also love the nostalgia at 18:25 where the sound adds a buzzing noise of video interference when Cybernoid goes to the "hall of fame" screen!
Haha! Things you get used to when you grow up with them! 😃 I totally get the nostalgia even just from the buzzing noise (and that was me putting a noise filter on that as well, otherwise it was a lot more noticeable!).
@@NoelsRetroLab , how can the buzzing noise be reduced? It is something quite annoying.
Just found your channel. Very well explained subject matter. Great work. Subbed!
Thanks. Glad you liked it!
@@NoelsRetroLab I will have to catch up on your other videos now. 👍
I really like how you explained this definitely worth a thumbs up.
Mister : only one word : respect !!! and total respect .... a very good idea !! I love your chanel and i'm learning many things with you knowlege ... Thanks for the share
Eric from France ...
Thank you very much! Really appreciate it!
Great video, but your nomenclature may be confusing to those who are not in the know. What you refer to as a Spectrum +2 is, in fact, a Spectrum +2B-an updated version of the Spectrum +2A (the changes primarily concerned the sound output circuitry). There was a (rather rare) Spectrum +3B, which inherited the same changes. The Spectrum +2 is Amstrad’s first Spectrum, which was a reverse-engineered Sinclair’s 128K machine.
(I’m sure you know all of this; it’s just that it might sound confusing to some people.)
Yes, you're totally right. I had first planned on doing a timeline of Spectrum models at the beginning, but I quickly ran out of time with everything else involved in this video (and even so, I released it a day late!). So thanks for adding that. It'll definitely help when some people are scratching their heads wondering why the +2 had the fixed version of the +3 circuit :-)
The +2A actually used the same circuit board as the +3, but without the ICs related to the disk drive, i.e. the 765 FDC, the SED9420 Data Separator, and a few "glue logic" ICs; instead it had a pin header for connecting the Datacorder.
The +2B models had a smaller PCB with a jagged ( or "staircase") footprint with no provisions for the FDC circuitry. Oddly, though, the +2B boards still had solder pads for the floppy disk power cable.
Great video by the way, Noel, you should give yourself a pat on the shoulder for a job well done. The circuit simulation was very helpful to get a better understanding of the goings-on. 😁 I'm certain i shall have a play-around with that simulator in the future, so thanks for linking to it.
Greetings from Denmark.
P.S. I hope you stay safe and healthy! Take care.
@@BertGrink So... the Spectrum+ 2A could has the same audio problem than +3?
@@ElectronGordo It not only _could_ have the same problem, but to the best of my knowledge, it *HAS* that problem too - which is why Amstrad made the (slightly) cost-reduced +2B which not only fixed the sound issue, but also used less PCB material. I can illustrate this point by linking to a picture of the +2A PCB: hardware.speccy.org/imagenes/ordenadores/spectrum+2A-issue2-larga.jpg
and one of the +2B: hardware.speccy.org/imagenes/ordenadores/spectrum+2A-issue1.jpg.
At a guess, i would say this board configuration could save perhaps 20% PCB material.
Around the same time, Amstrad also issued the (very rare) +3B: 2.bp.blogspot.com/-td3itcIQGSs/XIT1rC9IruI/AAAAAAAAQ4o/K0-AfR9NR90TAnaX6s7s97oS26-9myJegCLcBGAs/s1600/plus%2B3%2Bplaca1.jpg
@@ElectronGordo P.S. To illustrate how they could save PCB material, i have made this quick and dirty picture: bbeea579-a-62cb3a1a-s-sites.googlegroups.com/site/bertyfromdk/home/misc/2xspectrum%2B2A-issue2.jpg?attachauth=ANoY7cqOOP6ZHAo3_eiSqBuGO5tH3JEdL4fya3YKbKvnLUJl3fiZsiX_LSBbGc3yOwLMxU4MtlsunsS9Y9exWeDjAfcvEviSVEn0eTRp-F69H_XFKahVyiQ1VYxVJkAWsmV6Q9GlzcGxrrRweEpSNZ-Tg5qv26SNjs_SNlworAkD10klad5lP-g72p-21DOXQEDGZfmj07iSAe0WIRRAd-xCUvVEjtGg4NDWe5IDU6GMwKBV8ETAgMvlKbBQoT5HzDeHcEZLqn8_&attredirects=0
Again, great job and nice teaching. Please, May I ask you about the simulator software that you have used? Thank you
Thanks! The links are in the description. It's a really nice simulator, check it out. tinyurl.com/vmtcsaq
@@NoelsRetroLab Thank you, I thought it was a Schemma instead the link to the simulator software, shame of me :)
I was amazed the Spectrum could play music that good when I saw GliderRider for the first time...
Thanks Noel for all your efforts.
I have a lot of speccys, some +2a (short boards), +2b and a +3, and I never have noticed any sound distortion. Does it mean that i have one of the strange +3B? I just got a long board +2a and i plan to test it to see if it has that sound distortion.
On the other side, I have heard that you need to add a capacitor to enable the tape/sound functionality of the +3 on the long board +2a. Could you tell me how? Thanks in advance.
Great video for me to start to understand the AY chip - i have an Amstrad 6128 on the way from the UK to me in the US, my first computer from the other side of the pond. I blame your videos for my purchase. lol.
Haha, blame away. That's one of the best compliments I've gotten :-)
@@NoelsRetroLab Been a Commodore guy since i was a kid, finally gave in. :-)
@@IDPhotoMan Although I'm a Speccy guy myself, I am sure you will have a lot of fun with the CPC; in fact it has been on my own wishlist for quite a while. Oh and about a year ago i took the plunge and acquired a Commodore 128D (CR), a 1084S monitor, a Datasette, a 1750 REU and some other bits and bobs. That is a really nice machine too. 😊
@@BertGrink Awesome on the 128D! I still have my 128 (not the D) that i bought in high school when i sold my original C64 (DOH!), along with way too many other Commodores lol. I should have gotten a D a few years ago instead of waiting, as now they are about $400US for even not-perfect machines. Ah well. Now to find a Color Amstrad monitor some day (I have a few PAL LCD's i can use for now).
@@IDPhotoMan Well, i don't think getting a C128 instead of the C64 is such a bad exchange; after all the 128 has a C64 mode built in.
I paid roughly US$320 for my 128D, and the seller had included all the system diskettes plus some GEOS disks and a game (Outrun Europe), a few cables, a Quickshot joystick, and as the _piece de resistance,_ a Commodore 1351 mouse! All in all a fair price in my opinion.
Best of luck finding the Amstrad color monitor. 😊
P.S. Aside from the inevitable yellowing of the plastic parts, my 128 is in perfect condition.
Been watching many random videos from you lately to help spark my interest in analog electronics again because lately I've been doing VLSI/digital over and over and this video especially did the trick! I gotta say though, I miss the good old devices with reasonably-sized components, with my sausage fingers and unstable hands it's a nightmare handling or repairing the ones we have nowadays with resistors smaller than 1mm, not to mention those damn QFPs that are EVERYWHERE (these are what killed my interest originally)...
Glad those videos inspired you! I agree that older electronics are much easier to interact with than modern ones. And modern tech makes it easier to interface with older tech, so it's a big win. That's one of the reasons I love retro tech in general!
I noticed on the cybernoid music at the end, there's some sound interference from the ?video? circuit, which my spectrum has as well. I'm wondering if there is a fix for this as well?
Yes, you're right. It's very noticeable because it changes pitch when the screen changes. It could be an interesting project to try to figure out why that happens and how to solve it. Maybe some of the switching noise from the RAM makes it into the amplifier stage. Not sure.
@@NoelsRetroLab Yes, mine does exactly the same. The hum changes when the screen switches. I only notice it when not playing sound. It might be as you say something to do with RAM switching, as I'm sure you can hear it change when the computer boots and the screen goes from black to white.
These specific interference noises are one of my big annoyances with the Spectrum. Oh how I'd love a fix.
@@WDeranged I've heard it's enough to remove just a single capacitor but its number depends on the particular PCB revision that you've got. On +2A it's C24 I think. This should de-couple audio from video but you won't be able to hear music through RF modulator output (which nobody uses anyway). Another thing that helps is to disconnect the tape deck from the main PCB. Even when I completely re-capped the tape deck PCB and re-did some of the oxidized solder joints, the annoying hiss persisted. Although this hack is only good if you use a mass storage interface.
Seems the Amstrad (or was it still Sinclair?) engineers skipped a class on proper BJT biasing :-) Which simulation software are you using here? This video is very instructional, great content and presentation. Keep up your good work!
Thanks! It was definitely Amstrad at this time. I put a link to the actual simulations in the description. The online simulation is great! tinyurl.com/vmtcsaq
This is perfect 👍
Hi Noel. Fantastic video! The explanation of the circuit was really something.
I fail to understand what changed in reality between the adding of the 220 ohm to TR5 and all the rest of the changes you did after that. I mean, what's different?
I have a +2A that had the same issue (it's the same board as the +3) and I've added a 1K ohm resistor instead of a 220 ohm (I followed a guide from some place I can't recall), and the issue with the sound was fixed. I'm quite pleased
What's the benefit of doing all the subsequent changes after adding that only resistor?
Thanks! Glad you liked it. The main change is that resistor. The rest is only if you want to match the circuit of the fixed +2A. I don't see any major difference. Maybe the volume levels will be a bit better with all the changes, but I think it's pretty minor. I just did it to be on the safe side :-b
The purpose of all the other resistor changes is to match the output of the BEEPer audio (which is much higher) to AY level
As always, top content - amazing work, even though I didn't understand half of what was going on, LOL. :D
Hi Noel, great video as always. I was wondering if you have an opinion on what the difference is between the simpler versions of this where just the tr5 resistor is added, then three resistors, then this many resistor version?
Thanks! I haven't done a full analysis to be able to tell the difference really well. The single resistor mod improves the sound hugely. The rest... I wasn't able to tell the difference, but I figured they were set up that way in the +2A by people who really knew what was going on, so maybe it's better biasing, or the way it mixes with the input sound or something. If you're going to do it, I'd go with the full resistor change if you cant just to be 100% sure.
The purpose of all the other resistor changes is to match the output of the BEEPer audio (which is much higher) to AY level
When you testing sound output with osciloscope, do you set probe to 1x or 10x? Or it doesn't matter? Great modification and explanation. Cheers! S
It doesn't matter much, but in general you should use 10x whenever possible(so with any signal higher than a few millivolts). 1x reduces your bandwidth significantly (usually down to a few MHz) and may have you miss some interference.
En el minuto 18:33 al cambiar las letras a blanco se escucha un ruido de interferencia, en el mío sucede lo mismo, a que se debe eso? se puede quitar?
Very nice job!
Thanks!
I have no experience with these machines but seeing a dc offset on audio signals is always worrying. :) Is there a blocking crapacitor in the circuit?
I don't remember the specifics anymore, but I believe there is further down.
That was very interesting
Bravo.
What about a digital solution where the audio signal get digitized via a ADC then processed in a DSP chip or a suitably programmed FPGA then output to the DAC
Absurdly expensive and complicated. It's like breaking butterflies upon a wheel.
Great video. Where did you buy that T-shirt?
Thanks! I think I got this one here: www.latostadora.com/web/sinclair_zx_spectrum_-_logo/1078007
Thank you for your quick answer. Going back to the Spectrum+3 sound issue I have this wonderful interface bought at Byte Delight, it includes a jack output to connect a pair of loudspeakers
www.bytedelight.com/?product=zx-ay-external-stereo-audio-interface-for-all-zx-spectrum-models
@@pjrodriguezl I am definitely going to get that stereo mod in the near future too. :)
swapping R and C around in a series coupling makes absolutely no difference - neither theoretically nor in reality. It could be that it was purely because of board layout reasons (it's easier to pass traces under a resistor than a capacitor).
Can you add a track list so those of us unfamiliar with the music can find it? Great videos, BTW - it's helping me out doing component repair on other electronics!
Good idea! Adding it right now. Thanks for the suggestion.
@@NoelsRetroLab what is the song at 16:53? I notice you have it on your other videos too, it's dope!
@@NoelsRetroLab Yeah i would also love to know what song you used at 16:53min, please, the song is so good and goes well with your type of videos ^^
@@xvenomtype0x I keep forgetting I didn't have that one in the credits at the end. I just added it to the description. Same author as the ending song opengameart.org/content/theme-song-8-bit
Here you go. Forgot it wasn't in the credits. Just added it: opengameart.org/content/theme-song-8-bit
You made a mistake on the board the 390 ohm resistor goes from the tr5 leg to the tr5's hole. you fitted a 10k resistor instead
I need to do this to my plus 3sounds way better .
You could just find a better suited value for the 220R emitter resistor then you don't need output attenuator....
I LOVE to see PROs like you working! ;)
I wish I was a PRO like you too! LOL.
Anyway, THANKS a lot for the GREAT video and explanations!
One question from someone who doesn't understand much of this: why would Amstrad change the "circuitry" from +2 to +3 if the result is so much worse?? What happened on their minds to mess so much with the sound?!
Haha thanks! Look at the pinned comment. I failed to mention the correct sequence of fabrication in the video. Basically, they released a version of the +2 and the +3 at the same time with the sound bug in it. Then they released an updated version of the +2 with it fixed (and supposedly a fixed version of the +3 but that's less common I think). So yeah, the fixed +2 came later.
@@NoelsRetroLab Is there any +3 issue with the sound distortion "bug" fixed?
Yes, although I don't think they're very common. The bug is only present in Z70830 boards, so you can open it up and look.
I'm going to do this mod as per Noel, but I wonder how it compares to the mod on Spectrum for Everyone? Line below, Noel's has less work :)
spectrumforeveryone.com/technical/fixing-audio-distortion-spectrum-3/
They're pretty much the same. We're just copying the "fixed" circuit from the ZX Spectrum +2 that came out later. It looks like there they're changing a couple extra resistors related to the MIC and EAR lines (only useful if you're going to be loading from tape). If you want to be extra sure, do them all. Otherwise, if you want the least work, just add that single resistor and don't worry about changing the rest.