Hi Noel. That VDP TMS9929 seems to be a laser rebranded IC. Did you buy it at AliExpress? :D I have suffered this multiple times and unfortunately I have some experience with it: rebranded MOS 6502 CPUs, MOS 6581 SIDs and even Rockwell 65C02 CPUs... Check it out: it has been sanded... Look on the back of the chip and check if there are any markings, but to me it looks like another "chinese knockoff" :(
You're probably totally right. I was lucky enough to buy it through Ebay (although it came from China), so I was instantly refunded my money. Let's see what happens to the next one when I get it.
Wikipedia, Laser engraving :"Before the year 2000, lasers only produced lower-quality results in rubber-like materials due to their rough structure. In the 2000s, fiber lasers were introduced, giving a much increased engraving quality directly into black polymeric materials." So the laser engraved marking chip is most likely fake (9918,9928/29 were developed in early 80's )
Just watched all 3 videos of this repair and I must say you have much more patience than I would have had. Especially at the end of the 2nd where you had it working and decided to replace those messy bodge wires, only to have the VDP completely fail. At that point I may have just given up or at least pushed to the back of my projects to work on list. Great repair!
Haha, it helped that it was a friend's computer and I need to return it. Otherwise, I might have done just that. Also, the fact that it was a new computer to me made it more interesting and me more willing to keep digging. Had that been a plain Speccy, it would have totally gone at the bottom of the drawer 😃
So good you got the machine working in the end. There seem to be a lot of fake TI-99xx chips on eBay and other markets so it can be very much hit and miss.
Thanks! Yes, fake VPS are a big problem. That's what prompted me to make the "Fake Chips" video actually (which is currently by far my most viewed video).
My cousin had one of these in -84. His family bought it for their 'family business'. Fair to say that machine never saw any 'business use' as my cousin had it at his room from the start.
Nice video series. At 5.34min I have done similar but using a sewing kit pin. The ones with the coloured tops. Solder doesn't stick to them either. Replaced many caps using them.
Oh nice tip! I didn't know about that (although I suspect my wife wouldn't be happy me taking her pins and doing that with them 😀). Still, good to know in a pinch.
Yeah, finaly, it works. Great fix. It looked like those annoying keyboard wires were soldered at the K/B side as well as the main board. Looking forward to seeing what it can do in part 4.
Oh I didn't notice that. I was assuming it was the keyboard membrane and forgot to look closely even when I opened it up. I'll have to watch my own video closely :-)
Just binge watched this series. Geez! How painful! (For you). And yeah, those hard ribbon cables are a nightmare. The TRS80 model 1 used one for the KB and is so annoying to work around.
Haha, yes, it was painful. Especially having to wait 2 months for parts to arrive and then find out they were fake. At least that sparked the fake chips video which brought in a lot of views! Right now I'm finally working on the conclusion of this one where I get to use it and see what it can do. Should be fun!
I do a lot with this JST connectors! Older projects and also new things. I also have a cimping toool for those (and also other connectors as dupont and so on), so I can do my own cables with special pinouts and cables. If you have trouble with to many pins as you have on the 13 pin cable, you can make a little addon board with a piece of strip board and some pin headers. In your case I would make a little piece of strip board with 13 lanes of 3 solder points each. So I can solder the pin headers right into the middle holes, and the JST sockets on the outer sides. So I don't have to cut the ends of the headers. A little board like that will float above the main board perfectly just soldered with the pin headers. Just take a bit of care not using to much force while plugging in and out the plugs. For the keyboard cable (Did I count 12 lanes two times correctly?) you could unsolder all the lanes and make a little adapter board on the under side of the keyboard. So you can adapt from the ribbon cable to normal wiring and use some kind of plugable connector on the mainboard again. Maybe you find some screws you can screw a small board onto? So you don't have to deal with the ribbon cable any more, it will stay on the under side of the keyboard.
I am in awe of your skill and ingenuiety at finding and fixing these faults. Even when you occasionally bugger it up, it's impressive how quickly you realize what the problem is and what do do about it. I wish you could have devised a way to replace the other hateful ribbon cable (to the keyboard), since I have a few of these machines which I hesitate to touch because of these shoddy builds.
Glad to see it finally working! That's been a long ride that one. In the keyboard part, it seems like the ribbon is soldered to a PCB, maybe you can put connectors there too after all.
@@jjermar1 The problem is that I'm constrained by the spacing on the boards, which seems to be 2.54mm. I think floppy cables are tighter than that and have two rows of pins. What I need is something like the C64 keyboard connector, but I don't think those are easily available.
@@NoelsRetroLab Both IDE (the 40 wire ones) and floppy cables are half the spacing cable-wise because of, like you said, the dual row connector. I mean, you could just use a floppy cable and use only one row. I'm not that familiar with the C64 connector, I never seen one in real life, but you might want to have a look at Molex' KK series (and various clones of it) if you have enough clearance (the kind used on PC CPU fan headers and the like.)
The broken tms9929 might even work if the crystal is replaced with crystal oscillator of same frequency. The xtal oscillator driver or input might have an ESD damage which lowers the crystal drive level or oscillator input resistance ... Odd quirk with the external oscillator is that externally generated clock signal must be connected to the both XTAL inputs (pins 40 and 39); This is a footnote in Texas Instruments VDP databook 9918 and 9928/29 terminal assignments (page 3-9 to 3-11)
Great and informative video as usual Noel. I would suggest a way to get rid of that black wire from the video board. You could connect it to the unused pin of the top connector and run a bodge wire from that pin to GND directly on the main board. Keep up the good work!
So ends the saga. Excellent video and series Noel. I thoroughly enjoyed it! Your content is top-notch. One thing that I noticed, at 25:28 or so when you disassembled the keyboard, is that it appears the ribbon you thought was part of the keyboard membrane is soldered in place. You could replace that with another home connector job in Part 4! :D Your solution to the problem with the video board connection was brilliant. One more time perhaps?
Thanks so much! Yes, a few people pointed that out (I didn't even notice until after the video was done, I was so focused on just getting the plunger in without tipping the keyboard 😀). The main problem is finding a good connector for that. I counted 25 pins, I'm assuming 2.54mm spacing... not trivial. I'll look into it though. Cheers!
@@NoelsRetroLab I became suspicious when I noticed that the keyboard and video ribbons appear identical. I would not be surprised if they came from the same roll of material. No doubt the ribbons were added in the final assembly of the computer.
True. On the other hand (and this is totally a guess on my part), maybe 80s components weren't as susceptible to static discharges since they were larger components. But I have no idea, really :-) I'll try to get in the habit of using it.
@@NoelsRetroLab Climate seems to play a role, too. Here in the damp north of England static build-up is rare, but I notice that when I'm working in the south lightning bolts fly from my fingertips like the Emperor in Star Wars! =)
Good work on getting the SVI back online :D It was a good idea to go with a switching regulator module instead of the linear 7805. Without a heatsink, it could maybe do 1A max before it overheats and goes into thermal shutdown. With a heatsink, 1.5A is pushing it, unless using a beefier TO-3 with heatsink is used, then could get 2A. The older CMOS tech in that circuit is power hungry and likely that board is pulling more than 1A at 5V... On broken or missing pads and vias, be careful to check the circuit ("net") as it may be relying on the (plated) through-hole connection to connect top and bottom layer traces (that's what the primary purpose of vias are) as some board layouts may use a pin pad as a via as well. If the hole is no longer "plated through" then be sure to solder both top and bottom side of that pad/pin. Not saying you are not doing that already, just something to watch out for. 2.54mm is the old 0.100" spacing standard. There are several options for the cabling (eg. good ol' ribbon cable) and your idea seemed to work fine, but could the wire that was removed been re-purposed to bring the ground over and then the additional ground strap removed? Maybe some traces would have to be cut... but would clean up the inter-board connection... just a thought. Thanks for the video! Cheers,
Thanks! That's a good idea about the ground cable! I didn't even think about it, but I think it would be quite possible with just a couple extra wires in the boards themselves.
Great work Noel. You really need a wiring pen or some polyurethane insulated transformer wire for making fine repairs, saves me lots of time. One comment about probing crystal oscillators is to be know that the scope probe capacitance on its own may stop a working crystal oscillator because the 10pF adds to the small caps already there. It may also start a non working oscillator! Usually it just shifts the frequency slightly lower but the rule is to buffer the crystal. This is usually done for you in a working circuit so measure the output pin, not the input pin.
Interesting. I didn't know about wiring pens. But the wire in those pens is not insulated right? Or it is and you need to scrape it off or something? As for the oscilloscope, I always run my leads at x10, so I think the capacitance added is negligible. I've never had trouble with crystal oscillators, at least at these frequencies.
Noel's Retro Lab : The wire is insulated but the trick is to get the right insulation. You just solder right through the insulation at the tip so you don’t have to scrape it off. Instructables has an example. www.instructables.com/id/Wiring-Pen/ You can buy the wire from various places like Amazon. As for the scope probe, very astute to stick to x10 but it’s still at least 10pF, add that to the 22pF or so usually specified for the crystal and it may be a problem. I’ll admit it usually makes a oscillator that has a broken cap work but can also stop it if you probe at the input side of the crystal. It will definitely shift the frequency slightly and you may see some designs use a trimming cap to allow fine tuning.
I know this is from last year, but I'm wondering if you had the composite cables plugged in when you changed the wires underneath the board, as those can cause sparks from a soldering iron through to ground via the composite. I killed two famicoms that way sadly, as I was new to electronics and didn't realise. Now I just unplug everything before I work on anything, as any sensible person would do.
Good point. I didn't because it was already uncomfortable enough working on this board and that would have made it even worse. But it's a good tip to keep in mind. Thanks!
Well done Noel.! You kept at it, and got there in the end. The SpectraVideo 328 is a tight looking machine. Cant wait to see it running. I guess they have a modern SD Card Drive solution for it as well?
Thanks! Not exactly, and that's part of the problem. Even the cassette connector isn't a standard audio jack!! So it's going to take some creative problem solving to actually feed it some software next time :-)
The dropping voltage at the output of the 7805 is not just a random symptom of overheating, it's a feature of the 78xx series! In your system, the input is quite high (8-9 volts would be standard design, but 11 is quite high), so the chip generates extra heat from the extra voltage it has to dissipate. If the chip starts overheating, it starts passing less current than required to keep the output voltage at nominal value to protect itself from breaking. As passing less current means dissipating more voltage over the chip, it is common for the voltage to plummet quickly until the current gets so low that dissipation is low again.
Right. I was just surprised it happened so quickly. But you're right, I wasn't thinking of the high input voltage, and then the fact that I replaced the VRAM with 4164, which probably uses more current from the 5V rail than the 4116s.
This computer could be properly named a MSX 0. The MSX standard was based on it. If I'm not mistaken, what it doesn't have that the MSX has is the bank-switching and multiple slot slelection architecture.
That's right. I'll get into the details of what makes it different from the eventual MSX standard in the next video. I'll hopefully load a few games and see what it can do too.
Well you get my admiration for never 🤬 🤬 🤬 through this nightmare of a project. With the projects you come up with I think a well placed punching bag somewhere in the shop would be a must have item. Amazing video series. Thank you so much.
Hahaha, this was by far the most frustrating project, made worse by how uncomfortable it was to do ANYTHING with this board with the other parts dangling around with stiff ribbon cables. But in the end it all worked out 😃
As always you have done a perfect job here! And you gave me an idea how to go on testing what is wrong with my Philips VG 8010 MSX computer. In fact I have 3 TI 99/4A to borrow parts from... ;.)
I'm glad that was helpful! Looking forward to watching your MSX repair. I actually have a couple of them (different models) on the way, so I'm sure I'll end up making some MSX repairs as well.
I'm starting to get used to your volume of your voice and it's just perfect nice tone, and your videos are great very inspiring, certainly a man who loves his work, great stuff Noel
Thank you! I'm glad you're enjoying them. Was the volume of my voice too loud or too quiet? Just curious. I try to put some effort in editing to make sure the dialog is fairly even and understandable, so any feedback is appreciated.
Thanks! I'd like that too! This is just a friend's machine that I'm borrowing, but after getting this intimately close to them, I always end up wanting to get my own :-)
At the 25:27 mark you can clearly see the keyboard connector in the upper right hand corner of the screen... you can do the same mod as you did for the RF connector...
Very true! I do get up and move around a fair amount to reach over to the computer or grab some components, so that might get even more in the way though.
Yes, good point! I had an instance that the top track didn't make contact so part of the board had no +5V. With the Traco power is much harder to do that, so I just made sure I let the solder flow thoroughly from the bottom and it was fine. But yes, definitely something to watch out for.
Noel's Retro Lab I install it on a msx machine and used a basic program by dvik, ident.bas, that identifies vdp, psg, type of machine , among other things.
Great job Noel! I remember a friend of mine, back in 1984, who had a Spectravideo 328 with disks and CP/M, also with the MSX converter cartridge. By the way... Have you noticed the empty space for "ROM 01"? It seems that there may be possible to add some additional functionality to the SVI 328. Perhaps some embedded apps?
Thanks! Interesting, I didn't know about CP/M on the SVI. It must have needed extra RAM in the cartridge I would think. As for the ROM slot, I saw that. Maybe they had some apps planned like you said, although at least with cartridge systems like these it's really easy to provide them externally, which is great.
Hi Noel !! Nice job .. I've been waiting for this part.. so the ribbon cable was not part of the keyboard membrane after all ya... I'll be waiting for part 4 of SVI 328 hope to see it soon .. have nice day Noel!
Right! And to be honest I didn't notice that until just now. Haha... So I may have to change that one too, although that may be too many cables to come up with a good connector.
It's one of those things that financially it would make no sense for anyone to be spending that much time in a repair. But as a hobby, I find it very satisfying to get to the bottom of an issue, understand it, and fix it completely 😃
It does an amazing job in particular types of cases: light color and rough texture, like this one, Commodore 64, Apple IIe, etc. Check out The 8 Bit Guy for a lot more use of baking soda 😀
Even I would remove those horrible connectors that had hot glue all over them. It's not difficult to add connectors that you can easily unplug as needed if possible. It's also one of the first things I would do given the chance.
Wondering if you tested the original so called dead VDP in the Ti-99 4/A. I would have adjusted the trimmer cap to see if the oscillator circuit loading was just a bit off. I'm sure you leave out lots of details in these videos for time constraint reasons.
It was a TRACO TSR 1-2450 (eu.mouser.com/ProductDetail/TRACO-Power/TSR-1-2450). It's rated at 1A, which in retrospect might be a bit tight, so it's probably better to get one rated for 1.5A.
Hi Noel! Greetings from Argentina! I love how detailed your troubleshooting videos are. I learnt so much watching them! Can I make a small suggestion? When editing your videos, try to remove the high pitch noise that comes from CRT's screens. It's kind of uncomfortable for headphone users with good hearing (like me :P) You can remove it by applying an EQ to the audio track and lowering the 16K Hz band. Keep up the good work! Saludos!
@@NoelsRetroLab Yeah, it is just when a CRT is on. It is not so noticeable in this video. On others (I can't quite remember which one) it can be clearly heard.
I always edit the videos with headphones, so I'll keep an ear out for that (literally). I have some CRT shots in a video I'm working on, so I'll try to do that. Cheers.
Why would the input voltages to the regulators be double the output voltages? The unregulated input voltage should be close to the output voltage so the linear regulators don't get overheated as they waste 50% of the electricity they are being powered with.
You know, that makes a lot of sense. I didn't question it somehow, but I agree it seems really high. I know the input power supply is fine, and the path for the 7805 and 7812 are completely different, yet they both run high. I wonder if it's a badly-designed circuit or there's something going on in there.
@@NoelsRetroLab The input DC voltages make sense considering the AC voltages. 8VAC (nominal so higher with no load) through a full bridge gets you (8*1.414) 11VDC. So the 12.x Noel had was within what could be expected. No way were the 7805 ever going to survive without substantial heatsinking, which is why the previous person probably stuck those 2A 5v and 12v LDO on there except they would have needed to dissipate the same amount of heat just from a larger package. Very nice result in the end, great perseverance Noel!!
If 7805 had dodgy ground and dodgy heat stink - there is your vdp killa. Cheap chinese switching modules with lm2596 clones + 5.6v 5 watt zener or a maybe a crowbar will make it bulletproof...
The people at SpectraVideo had a typical Hong Kong mentality, it only had to work at the moment it was sold. It didn't came to their minds to make something better then others, they followed to loosely the MSX standard. Programmes had to be made in two versions, one for SpectraVideo and one for MSX.
That's really too bad. I didn't know much about their quality standards form manufacturing, but I'm definitely not impressed. As for the MSX, the SVI 328 came out before the MSX standard, so at least they have that going for them. Or are you saying that the later SVI 728 didn't follow the MSX standard completely either?
I've been working on it for the last two weeks and I keep running into problems! 😳 The good news is that it's mostly done and needs editing. The bad news is that there will be a part 5.
Does anyone have any suggestions on where to find a working TMS9929ANL? I'm also on a hunt for one and would like to avoid the Chinese ones if possible :) Preferable from a retro store or somewhere similar.
I've bought a lot of Retro computer chips from this seller on Ebay www.ebay.co.uk/usr/ande_8351 , nothing that hasn't worked yet and he is very flexible.
@@jameshearne891 Thanks, I'll save it for future use. He doesn't seem to have any of these VDPs, but I see lots of other interesting goodies :-) Cheers.
Really? They run THAT hot? Someone was just telling me on a FB group that there was even another computer released with this same VDP and they didn't install a heat sink from the factory. I doubt that running it for a couple minutes would do any harm.
@@NoelsRetroLab They run hot but not that hot, you can run them without the heatsink, there are computers that did. A heatsink just reduces the risk of them dying eventually. to be honest, i think the slab of metal on the top doesn't really do much as a heatsink, it needs fins to work properly.
I actually never saw any email from you! Weird. Do you know where I can order one of those cartridges, or at least a kit to build it myself? I'll send you another email and see if this time they go through.
Hi Noel. That VDP TMS9929 seems to be a laser rebranded IC. Did you buy it at AliExpress? :D I have suffered this multiple times and unfortunately I have some experience with it: rebranded MOS 6502 CPUs, MOS 6581 SIDs and even Rockwell 65C02 CPUs...
Check it out: it has been sanded... Look on the back of the chip and check if there are any markings, but to me it looks like another "chinese knockoff" :(
You're probably totally right. I was lucky enough to buy it through Ebay (although it came from China), so I was instantly refunded my money. Let's see what happens to the next one when I get it.
Wikipedia, Laser engraving :"Before the year 2000, lasers only produced lower-quality results in rubber-like materials due to their rough structure. In the 2000s, fiber lasers were introduced, giving a much increased engraving quality directly into black polymeric materials."
So the laser engraved marking chip is most likely fake (9918,9928/29 were developed in early 80's )
That is so bad... so the chip isn't even a TMS9929? It could be anything?
I bow to your dedication on repairing this non-so-MSX computer, good sir. Nice job that is!
Just watched all 3 videos of this repair and I must say you have much more patience than I would have had. Especially at the end of the 2nd where you had it working and decided to replace those messy bodge wires, only to have the VDP completely fail. At that point I may have just given up or at least pushed to the back of my projects to work on list. Great repair!
Haha, it helped that it was a friend's computer and I need to return it. Otherwise, I might have done just that. Also, the fact that it was a new computer to me made it more interesting and me more willing to keep digging. Had that been a plain Speccy, it would have totally gone at the bottom of the drawer 😃
Your persistence is very admirable, Noel! Great job as always.
Thanks! Much appreciated.
So good you got the machine working in the end. There seem to be a lot of fake TI-99xx chips on eBay and other markets so it can be very much hit and miss.
Thanks! Yes, fake VPS are a big problem. That's what prompted me to make the "Fake Chips" video actually (which is currently by far my most viewed video).
My cousin had one of these in -84. His family bought it for their 'family business'. Fair to say that machine never saw any 'business use' as my cousin had it at his room from the start.
Nice video series. At 5.34min I have done similar but using a sewing kit pin. The ones with the coloured tops. Solder doesn't stick to them either. Replaced many caps using them.
Oh nice tip! I didn't know about that (although I suspect my wife wouldn't be happy me taking her pins and doing that with them 😀). Still, good to know in a pinch.
You must have felt some frustration at times after being hit with challenge after challenge but great result, I enjoyed all 3 parts. Great work.
Thanks so much! Yeah, it was a bit frustrating, but I was so relieved to get it to work at the end! 😃
Noel when removing hot glue try using some isopropyl alcohol, it loosens up the glue and it just comes right off
Cool. I didn't know that. Thanks!
Yeah, finaly, it works. Great fix. It looked like those annoying keyboard wires were soldered at the K/B side as well as the main board. Looking forward to seeing what it can do in part 4.
Oh I didn't notice that. I was assuming it was the keyboard membrane and forgot to look closely even when I opened it up. I'll have to watch my own video closely :-)
Just binge watched this series. Geez! How painful! (For you). And yeah, those hard ribbon cables are a nightmare. The TRS80 model 1 used one for the KB and is so annoying to work around.
Haha, yes, it was painful. Especially having to wait 2 months for parts to arrive and then find out they were fake. At least that sparked the fake chips video which brought in a lot of views! Right now I'm finally working on the conclusion of this one where I get to use it and see what it can do. Should be fun!
I do a lot with this JST connectors! Older projects and also new things. I also have a cimping toool for those (and also other connectors as dupont and so on), so I can do my own cables with special pinouts and cables.
If you have trouble with to many pins as you have on the 13 pin cable, you can make a little addon board with a piece of strip board and some pin headers. In your case I would make a little piece of strip board with 13 lanes of 3 solder points each. So I can solder the pin headers right into the middle holes, and the JST sockets on the outer sides. So I don't have to cut the ends of the headers. A little board like that will float above the main board perfectly just soldered with the pin headers. Just take a bit of care not using to much force while plugging in and out the plugs.
For the keyboard cable (Did I count 12 lanes two times correctly?) you could unsolder all the lanes and make a little adapter board on the under side of the keyboard. So you can adapt from the ribbon cable to normal wiring and use some kind of plugable connector on the mainboard again. Maybe you find some screws you can screw a small board onto? So you don't have to deal with the ribbon cable any more, it will stay on the under side of the keyboard.
Noel, you're definitively the best ! All these videos are awesome.
Wow, thanks so much!
Wow, soooo many issues, but finally you were able to fix it!! Congrats and respect!!!
Thank you! Yes, it felt good to finally get that one working in the end.
:-) Thank you for showing us, there is a lot I can learn from it!!
I am in awe of your skill and ingenuiety at finding and fixing these faults. Even when you occasionally bugger it up, it's impressive how quickly you realize what the problem is and what do do about it. I wish you could have devised a way to replace the other hateful ribbon cable (to the keyboard), since I have a few of these machines which I hesitate to touch because of these shoddy builds.
Glad to see it finally working! That's been a long ride that one.
In the keyboard part, it seems like the ribbon is soldered to a PCB, maybe you can put connectors there too after all.
Yes, I just noticed that. It may have too many cables to be able to use the same trick. I'll look into it.
@@NoelsRetroLab How about a floppy cable and connector?
@@jjermar1 The problem is that I'm constrained by the spacing on the boards, which seems to be 2.54mm. I think floppy cables are tighter than that and have two rows of pins. What I need is something like the C64 keyboard connector, but I don't think those are easily available.
@@NoelsRetroLab
Both IDE (the 40 wire ones) and floppy cables are half the spacing cable-wise because of, like you said, the dual row connector. I mean, you could just use a floppy cable and use only one row.
I'm not that familiar with the C64 connector, I never seen one in real life, but you might want to have a look at Molex' KK series (and various clones of it) if you have enough clearance (the kind used on PC CPU fan headers and the like.)
The broken tms9929 might even work if the crystal is replaced with crystal oscillator of same frequency. The xtal oscillator driver or input might have an ESD damage which lowers the crystal drive level or oscillator input resistance ... Odd quirk with the external oscillator is that externally generated clock signal must be connected to the both XTAL inputs (pins 40 and 39); This is a footnote in Texas Instruments VDP databook 9918 and 9928/29 terminal assignments (page 3-9 to 3-11)
Great and informative video as usual Noel.
I would suggest a way to get rid of that black wire from the video board. You could connect it to the unused pin of the top connector and run a bodge wire from that pin to GND directly on the main board.
Keep up the good work!
So ends the saga. Excellent video and series Noel. I thoroughly enjoyed it! Your content is top-notch.
One thing that I noticed, at 25:28 or so when you disassembled the keyboard, is that it appears the ribbon you thought was part of the keyboard membrane is soldered in place. You could replace that with another home connector job in Part 4! :D Your solution to the problem with the video board connection was brilliant. One more time perhaps?
Thanks so much! Yes, a few people pointed that out (I didn't even notice until after the video was done, I was so focused on just getting the plunger in without tipping the keyboard 😀). The main problem is finding a good connector for that. I counted 25 pins, I'm assuming 2.54mm spacing... not trivial. I'll look into it though. Cheers!
@@NoelsRetroLab I became suspicious when I noticed that the keyboard and video ribbons appear identical. I would not be surprised if they came from the same roll of material. No doubt the ribbons were added in the final assembly of the computer.
Nice to see an anti-static wrist strap being used! =)
Modern ICs have a certain amount of ESD protection built in. 80s CMOS ICs really didn't!
True. On the other hand (and this is totally a guess on my part), maybe 80s components weren't as susceptible to static discharges since they were larger components. But I have no idea, really :-) I'll try to get in the habit of using it.
@@NoelsRetroLab Climate seems to play a role, too. Here in the damp north of England static build-up is rare, but I notice that when I'm working in the south lightning bolts fly from my fingertips like the Emperor in Star Wars! =)
@@Xoferif Very true. It's also very humid here in northern Spain, so it's probably a similar situation.
Yep, he has successfully taken part of the 1st step to ESD safety, he owns a strap, not using it correctly yet but he will get there ;-)
It took a while, but it ended well! I’d love to see what this computer can do since I know nothing about it. Great video, Noel!
Thanks!
So much repairs, congrats!
Your patience Is amazing!!!!
Good work on getting the SVI back online :D
It was a good idea to go with a switching regulator module instead of the linear 7805. Without a heatsink, it could maybe do 1A max before it overheats and goes into thermal shutdown. With a heatsink, 1.5A is pushing it, unless using a beefier TO-3 with heatsink is used, then could get 2A. The older CMOS tech in that circuit is power hungry and likely that board is pulling more than 1A at 5V...
On broken or missing pads and vias, be careful to check the circuit ("net") as it may be relying on the (plated) through-hole connection to connect top and bottom layer traces (that's what the primary purpose of vias are) as some board layouts may use a pin pad as a via as well. If the hole is no longer "plated through" then be sure to solder both top and bottom side of that pad/pin. Not saying you are not doing that already, just something to watch out for.
2.54mm is the old 0.100" spacing standard. There are several options for the cabling (eg. good ol' ribbon cable) and your idea seemed to work fine, but could the wire that was removed been re-purposed to bring the ground over and then the additional ground strap removed? Maybe some traces would have to be cut... but would clean up the inter-board connection... just a thought.
Thanks for the video!
Cheers,
Thanks! That's a good idea about the ground cable! I didn't even think about it, but I think it would be quite possible with just a couple extra wires in the boards themselves.
Great work Noel. You really need a wiring pen or some polyurethane insulated transformer wire for making fine repairs, saves me lots of time. One comment about probing crystal oscillators is to be know that the scope probe capacitance on its own may stop a working crystal oscillator because the 10pF adds to the small caps already there. It may also start a non working oscillator! Usually it just shifts the frequency slightly lower but the rule is to buffer the crystal. This is usually done for you in a working circuit so measure the output pin, not the input pin.
Interesting. I didn't know about wiring pens. But the wire in those pens is not insulated right? Or it is and you need to scrape it off or something?
As for the oscilloscope, I always run my leads at x10, so I think the capacitance added is negligible. I've never had trouble with crystal oscillators, at least at these frequencies.
Noel's Retro Lab : The wire is insulated but the trick is to get the right insulation. You just solder right through the insulation at the tip so you don’t have to scrape it off. Instructables has an example. www.instructables.com/id/Wiring-Pen/ You can buy the wire from various places like Amazon. As for the scope probe, very astute to stick to x10 but it’s still at least 10pF, add that to the 22pF or so usually specified for the crystal and it may be a problem. I’ll admit it usually makes a oscillator that has a broken cap work but can also stop it if you probe at the input side of the crystal. It will definitely shift the frequency slightly and you may see some designs use a trimming cap to allow fine tuning.
18:55 you can see the sanding marks on the plastic injection molding dots, thar blows either a fake, or a cleaned up and relabeled busted chip
I know this is from last year, but I'm wondering if you had the composite cables plugged in when you changed the wires underneath the board, as those can cause sparks from a soldering iron through to ground via the composite. I killed two famicoms that way sadly, as I was new to electronics and didn't realise. Now I just unplug everything before I work on anything, as any sensible person would do.
Good point. I didn't because it was already uncomfortable enough working on this board and that would have made it even worse. But it's a good tip to keep in mind. Thanks!
Well done Noel.! You kept at it, and got there in the end. The SpectraVideo 328 is a tight looking machine. Cant wait to see it running. I guess they have a modern SD Card Drive solution for it as well?
Thanks! Not exactly, and that's part of the problem. Even the cassette connector isn't a standard audio jack!! So it's going to take some creative problem solving to actually feed it some software next time :-)
The dropping voltage at the output of the 7805 is not just a random symptom of overheating, it's a feature of the 78xx series! In your system, the input is quite high (8-9 volts would be standard design, but 11 is quite high), so the chip generates extra heat from the extra voltage it has to dissipate. If the chip starts overheating, it starts passing less current than required to keep the output voltage at nominal value to protect itself from breaking. As passing less current means dissipating more voltage over the chip, it is common for the voltage to plummet quickly until the current gets so low that dissipation is low again.
Right. I was just surprised it happened so quickly. But you're right, I wasn't thinking of the high input voltage, and then the fact that I replaced the VRAM with 4164, which probably uses more current from the 5V rail than the 4116s.
This computer could be properly named a MSX 0. The MSX standard was based on it. If I'm not mistaken, what it doesn't have that the MSX has is the bank-switching and multiple slot slelection architecture.
That's right. I'll get into the details of what makes it different from the eventual MSX standard in the next video. I'll hopefully load a few games and see what it can do too.
Phew, what a lot of hard work, well done. That is how I ended up owning 3 of the little blighters.
Now to see if you can load any software!
It should be a fun change of pace.
Ooo another Video to watch after work tonight. Great. Thanks Noel .
My pleasure!
Love the videos. I really enjoy learning and troubleshooting that you show in every video. Keep up the great work.
Thanks! Glad you like them!
Well you get my admiration for never 🤬 🤬 🤬 through this nightmare of a project. With the projects you come up with I think a well placed punching bag somewhere in the shop would be a must have item. Amazing video series. Thank you so much.
Hahaha, this was by far the most frustrating project, made worse by how uncomfortable it was to do ANYTHING with this board with the other parts dangling around with stiff ribbon cables. But in the end it all worked out 😃
Well done. Great series.
Thank you!
As always you have done a perfect job here! And you gave me an idea how to go on testing what is wrong with my Philips VG 8010 MSX computer. In fact I have 3 TI 99/4A to borrow parts from... ;.)
I'm glad that was helpful! Looking forward to watching your MSX repair. I actually have a couple of them (different models) on the way, so I'm sure I'll end up making some MSX repairs as well.
I'm starting to get used to your volume of your voice and it's just perfect nice tone, and your videos are great very inspiring, certainly a man who loves his work, great stuff Noel
Thank you! I'm glad you're enjoying them. Was the volume of my voice too loud or too quiet? Just curious. I try to put some effort in editing to make sure the dialog is fairly even and understandable, so any feedback is appreciated.
Hi Noel!!! Nice vid as usual! Keep them coming! I would like to have an SVI 328 also, nice little machine :)
Thanks! I'd like that too! This is just a friend's machine that I'm borrowing, but after getting this intimately close to them, I always end up wanting to get my own :-)
Congratulations ... that was a bit of an odyssey.
Thanks! Yes it was!
At the 25:27 mark you can clearly see the keyboard connector in the upper right hand corner of the screen... you can do the same mod as you did for the RF connector...
Just a note that you can wear the static strap on your ankle, if you don't like dragging the cord around.
Very true! I do get up and move around a fair amount to reach over to the computer or grab some components, so that might get even more in the way though.
You should really solder the top sides of those two regulator pins as well... otherwise they might not be connected to every circuit that needs it.
Yes, good point! I had an instance that the top track didn't make contact so part of the board had no +5V. With the Traco power is much harder to do that, so I just made sure I let the solder flow thoroughly from the bottom and it was fine. But yes, definitely something to watch out for.
Nice video! Lots of “rebranded” chips lately. I recently got a ay-3-8910 that was a ym2149 🤷🏻♂️. Really glad things turn out ok at the end. :)
Thanks! Yeah, apparently I've been told the market for TMS9929A is mostly all fake chips! How did you find out what that rebranded chip was?
Noel's Retro Lab I install it on a msx machine and used a basic program by dvik, ident.bas, that identifies vdp, psg, type of machine , among other things.
Looking forward to part4, I must admit I don't know anything about the electrical computer range.
Great job Noel! I remember a friend of mine, back in 1984, who had a Spectravideo 328 with disks and CP/M, also with the MSX converter cartridge.
By the way... Have you noticed the empty space for "ROM 01"?
It seems that there may be possible to add some additional functionality to the SVI 328. Perhaps some embedded apps?
Thanks! Interesting, I didn't know about CP/M on the SVI. It must have needed extra RAM in the cartridge I would think. As for the ROM slot, I saw that. Maybe they had some apps planned like you said, although at least with cartridge systems like these it's really easy to provide them externally, which is great.
@@NoelsRetroLab th-cam.com/video/xe_K0AzyEio/w-d-xo.html
Hi Noel !! Nice job .. I've been waiting for this part.. so the ribbon cable was not part of the keyboard membrane after all ya... I'll be waiting for part 4 of SVI 328 hope to see it soon .. have nice day Noel!
Right! And to be honest I didn't notice that until just now. Haha... So I may have to change that one too, although that may be too many cables to come up with a good connector.
Such perseverance!
It's one of those things that financially it would make no sense for anyone to be spending that much time in a repair. But as a hobby, I find it very satisfying to get to the bottom of an issue, understand it, and fix it completely 😃
Baking soda to clean, interesting.
Thanks for that.
It does an amazing job in particular types of cases: light color and rough texture, like this one, Commodore 64, Apple IIe, etc. Check out The 8 Bit Guy for a lot more use of baking soda 😀
Even I would remove those horrible connectors that had hot glue all over them. It's not difficult to add connectors that you can easily unplug as needed if possible. It's also one of the first things I would do given the chance.
Wondering if you tested the original so called dead VDP in the Ti-99 4/A. I would have adjusted the trimmer cap to see if the oscillator circuit loading was just a bit off. I'm sure you leave out lots of details in these videos for time constraint reasons.
Congratulations 👍 nice one!
Thank you! 👍
I wonder if you had an ammeter on the 5V regulator to see what current is passing and compare with spec sheet.
Oops, commented before you replaced with a modern part 😄
as soon as the voltage was dropping i was like its getting HOT shut i down its going in to thermal shut down lol :p
Yup, exactly. You could probably even picture the smell of the regulator getting hot 😃
Was that keyboard a PCB with carbon and rubber cups? If so you could replace the cable with a new one and connectors.
It's alive, it's alive!
thank you for posting. what was the device w/ which you replaced the 7805 ? source ? take care & stay safe.
It was a TRACO TSR 1-2450 (eu.mouser.com/ProductDetail/TRACO-Power/TSR-1-2450). It's rated at 1A, which in retrospect might be a bit tight, so it's probably better to get one rated for 1.5A.
See my comment in the previous video about that static strap!
Hi Noel! Greetings from Argentina! I love how detailed your troubleshooting videos are. I learnt so much watching them!
Can I make a small suggestion? When editing your videos, try to remove the high pitch noise that comes from CRT's screens. It's kind of uncomfortable for headphone users with good hearing (like me :P) You can remove it by applying an EQ to the audio track and lowering the 16K Hz band.
Keep up the good work! Saludos!
Thanks! Ah, good point. I can easily do that since I already do some audio processing. Is it just in the CRT shots? Thanks for letting me know.
@@NoelsRetroLab Yeah, it is just when a CRT is on. It is not so noticeable in this video. On others (I can't quite remember which one) it can be clearly heard.
I always edit the videos with headphones, so I'll keep an ear out for that (literally). I have some CRT shots in a video I'm working on, so I'll try to do that. Cheers.
@@NoelsRetroLab You're too old to hear it 15khz, that's the issue :-)
Well, the other ribbon cable is soldered to the keyboard PCB, so if you want to walk the extra mile... 😁
Yeah, I saw that afterwards. Getting the right connectors could be a chagellenge because it has so many pins though.
@@NoelsRetroLab what about a simple pin header strip and dupont style connectors?
Why would the input voltages to the regulators be double the output voltages? The unregulated input voltage should be close to the output voltage so the linear regulators don't get overheated as they waste 50% of the electricity they are being powered with.
You know, that makes a lot of sense. I didn't question it somehow, but I agree it seems really high. I know the input power supply is fine, and the path for the 7805 and 7812 are completely different, yet they both run high. I wonder if it's a badly-designed circuit or there's something going on in there.
@@NoelsRetroLab The linear regulators are terribly inefficient.
@@NoelsRetroLab The input DC voltages make sense considering the AC voltages. 8VAC (nominal so higher with no load) through a full bridge gets you (8*1.414) 11VDC. So the 12.x Noel had was within what could be expected. No way were the 7805 ever going to survive without substantial heatsinking, which is why the previous person probably stuck those 2A 5v and 12v LDO on there except they would have needed to dissipate the same amount of heat just from a larger package. Very nice result in the end, great perseverance Noel!!
If 7805 had dodgy ground and dodgy heat stink - there is your vdp killa. Cheap chinese switching modules with lm2596 clones + 5.6v 5 watt zener or a maybe a crowbar will make it bulletproof...
Yes, that's probably very true. Do you have a link to one of those clones? I'd love to check them out.
@@NoelsRetroLab ebay search lm2596 dc-dc 3a. Make sure module is trimmed to exactly 5v before connecting load
At 6:17 you soldered the 5v to Ground on the 5v Regulator
The people at SpectraVideo had a typical Hong Kong mentality, it only had to work at the moment it was sold. It didn't came to their minds to make something better then others, they followed to loosely the MSX standard. Programmes had to be made in two versions, one for SpectraVideo and one for MSX.
That's really too bad. I didn't know much about their quality standards form manufacturing, but I'm definitely not impressed. As for the MSX, the SVI 328 came out before the MSX standard, so at least they have that going for them. Or are you saying that the later SVI 728 didn't follow the MSX standard completely either?
Did you try that fake 9929 in the TI-99/4A just to be sure it's bad?
Not yet. The TI-99 is still dead. Working on it this week though.
@@NoelsRetroLab Ok...thought just for giggles to give it a try before declaring it dead.
Part 4 When :) ???
I've been working on it for the last two weeks and I keep running into problems! 😳 The good news is that it's mostly done and needs editing. The bad news is that there will be a part 5.
Looks like you could have easily replaced the keyboard ribbon cables as well.
Yeah, I saw that afterwards. The problem is getting a connector long enough to have that many pins. I need something like a C64 keyboard connector.
Does anyone have any suggestions on where to find a working TMS9929ANL? I'm also on a hunt for one and would like to avoid the Chinese ones if possible :) Preferable from a retro store or somewhere similar.
I'd love to hear about good sources too. I have another one on order from Ebay from China and I don't have my hopes up after reading all the comments.
I've bought a lot of Retro computer chips from this seller on Ebay www.ebay.co.uk/usr/ande_8351 , nothing that hasn't worked yet and he is very flexible.
@@jameshearne891 Thanks, I'll save it for future use. He doesn't seem to have any of these VDPs, but I see lots of other interesting goodies :-) Cheers.
@@NoelsRetroLab It's worth asking him, he can get all sorts.
I wish he would have used a 14pin JST connector :D
Me too! I just couldn't find any of that size ☹️
I wonder why would anyone choose TMS VDPs for anything. Monochrome-only 6*8 40 character text is useless.
13 is about as ODD as it gets.
You should not have powered up that new chip without a sink.
Really? They run THAT hot? Someone was just telling me on a FB group that there was even another computer released with this same VDP and they didn't install a heat sink from the factory. I doubt that running it for a couple minutes would do any harm.
@@NoelsRetroLab They run hot but not that hot, you can run them without the heatsink, there are computers that did. A heatsink just reduces the risk of them dying eventually. to be honest, i think the slab of metal on the top doesn't really do much as a heatsink, it needs fins to work properly.
Wrote you an e-mail, but didn't get a reply. So I am posting the link to Caro's SVI-328 cartridges here: github.com/Wierzbowsky/SVI-328_Cartridge
I actually never saw any email from you! Weird. Do you know where I can order one of those cartridges, or at least a kit to build it myself? I'll send you another email and see if this time they go through.
@@NoelsRetroLab I didn't get a bounce. So check your SPAM folder please.