SVI 328 VRAM Replacement and Disaster Strikes (Part 2)

[@joshpayne4015]

Your channel has quickly become one of my favorite ones. Every new video you post is really interesting. Thank you for the content!

[@NoelsRetroLab]

Thank you! Really glad to hear that!

[@stevefreegard6406]

I'd like to echo this, I'm not even into retro computing and I love watching these videos. You've got a great presentation style Noel.

[@NoelsRetroLab]

@@stevefreegard6406 Thank you! I really appreciate it!

[@tony359]

I second that too - I like the presentation style, the voice, the quality of the video and the overlays and that all the steps are documented and explained (including mistakes) but without getting boring. Thank you!

[@Xoferif]

Ah, the perils of improving somebody else's work... I'm really enjoying this multi-part repair - looking forward to the next installment!

[@BotaTamas85]

When you modified the 4164 chips and bended the leg over at 12:29 and told you need a small metal piece, I thought first you will use the previously cut leg. What a missed opportunity! 😂

[@frankowalker4662]

Well that's annoying. Looking forward to part 3.

[@tony359]

Thank you for the great video! A quick tip - and I don't mean to sound patronising: when soldering using previously-melted solder, you'd really want to add some flux to improve the flowing. Those little pointy solder spikes that came off the memory chip legs confirm that. If you add some solder flux, you'd see the solder flowing beautifully when you touch the metal with your solder tip!

[@NoelsRetroLab]

Thanks for the tip. Which solder spikes do you refer to? Can you point me to the time in the video? Just curious. I usually don't add extra flux because the solder I use have a lot of flux already. But you're right that in cases where I'm combining old solder it might be good. What kind of flux do you recommend? The pasty kind of the clear liquid one?

[@djmips]

@@NoelsRetroLabclear liquid is much easier to clean. The rosin style is annoyingly sticky.

[@mineown1861]

Loved that piece of binary backtracking , from gui to pcb , brilliant. Like the reveal moment in a vintage detective show.

[@Vampier]

It lived! It will live again!!

[@NoelsRetroLab]

That's the spirit! :-)

[@NYYR1KK1]

Aww... This is becoming my absolutely favorite thriller! When you had that DRAM upside down I felt horrible and when you had that one VRAM chip disconnected, I was on my sofa like "One bit more!!! Just one bit more!" and when you nailed it I felt a victory with my hands up like "Now what did I say?! -> Just replace that VRAM and it boots fine!"... and just two minutes later there was this plot twist where you blew it all up again and I was like "Whaaat?" Next time I'll definitely take popcorn with me... "On episode 3 we will fry PSG and PPI until on season final we blow up Z80 and get some blue smoke from PSU."? :) Naah, don't take it seriously... What is bad for the project is anyway great entertainment for us sofa potatoes... Especially since you explain your measurements and tests so well. It would be shame if there would not be next SVI-328 episode. :)

[@NoelsRetroLab]

I'm glad you're enjoying it so much. I'm hoping that by putting all my mistakes along with the advances, it makes for an interesting video and we can all learn from it :-)

[@ChrisPinnockUK]

1000 points for the chip mod. :-)

[@richardkelsch3640]

There's a great idiom to describe "fixing" the wiring, that was already working (even though it looked ugly): "If it ain't broke, then don't fix it"

[@NoelsRetroLab]

Sure, but I'd still prefer to take down what's in there to know exactly what's going on. If this happens, it's all part of the learning process.

[@herrbonk3635]

Replacing two black wires with yellow ones seemed totally redundant to me...

[@64bittz93]

Probing directly on crystal pins is not a good idea. You are adding extra load capacitance that can disrupt the oscillation and you will see nothing on the scope. You could try changing the attenuation on the probe. If it didn’t work anyway I don’t think this was the problem, but it’s just a good thing to know usually

[@thereeven]

So.... it seems I'm an old school computer aficionado. Is just that I didn't knew it until now, and also I've never touched one of those computers. Silly jokes aside, your content is very well made and I really hope people realize how underrated your channel is. I learned more about how to troubleshoot general electronics watching your videos than I've learned in my entire life. Congrats!

[@NoelsRetroLab]

Thank you so much! Really happy to hear when someone becomes more interested in retro electronics after watching one of my videos 😃

[@gyulaszabo6750]

Hi! It’s as if you didn’t put them all back in the same place when you replaced the jumper wires at the bottom of the panel.

[@AlejandroJCura]

The suspense is killing me, this looks like an episode of any series with zombies.

[@NoelsRetroLab]

Except that I'm the one trying to revive it and it keeps dying on me :-) But stay tuned because in today's Mail Day video there's a big update about this project and I should be able to wrap it up next week. So yes, it's finally coming.

[@johnsonlam]

I'm always interested in ... how to use modern RAM replace the old one, since old chips will gone one day even still a lot recycled parts. And few people teaches how to do that.

[@NoelsRetroLab]

4164s aren't exactly modern, but they're for sure much more easily available than 4116s and much easier to work with. So yeah, big win there 😀👍

[@johnsonlam]

@@NoelsRetroLab At least we can still buy it, but sadly in reality, nothing last forever.

[@prpplague]

the TMS9918 chips were notoriously sensitive to static discharge! probably needed to ground the board before soldering to it!

[@NoelsRetroLab]

Ah, interesting! I've been trying to get better about wearing the anti-static wrist band, but I didn't ground the board itself. I'll make sure I do that when I get the replacement. Thanks!

[@pentiummmx2294]

It was before any ESD protection was implemented, so wrist bands are not an option if you want to work on those things.

[@ingmarm8858]

@@NoelsRetroLab don't head down the popular misconceptions involving "ground". The ESD issue is related to the "potential difference" between you, your tools (iron etc) and whatever you are working on. The only thing "ground" has to do with it is that it is usually the common connection point between your soldering station, your ESD strap and your ESD dissipative mat. Not much point wearing a strap when the PCB is scooting around on a plastic work mat generating lots of static so you should really get a work mat as well (like Jan Beta!). I've seen various videos where people have made a song and dance about "I'm wearing my strap and it is connected to ground", yeah good for them but that does NOTHING to protect what they are working on unless it is also effectively connected to ground as the common point. The other nonsense idea people have is that due to its age this gear is not susceptible to ESD damage, that's a fantasy that will cost them in the long run. I did my first formal ESD training course with IBM in the early eighties and the most recent professional refresher last year. Even the 1960s NASA documentation lists ESD precautions relating to electronic modules (which were ignored by a popular YT group restoring lots of precious Apollo gear!!). Save your precious collectibles and invest in a mat, strap and ensure you don't use nasty plastic bins and takeaway food containers to store ESD sensitive parts (like the ebay glad wrap packing of 4164!!!!).

[@NoelsRetroLab]

@@ingmarm8858 I actually started looking into it a while back and didn't come up with a clear conclusion. Those grounded mats seem to be only conductive on the underside (as they should, otherwise it would be horrible). How exactly does that help equalizing the potential difference of the board being worked on? (It's a serious question, I'm not impliying it doesn't). And the followup question: If I'm grounded and pretty much constantly touching the board as I'm repairing it, isn't that enough to keep the potential close to equal? Thanks!

[@ingmarm8858]

@@NoelsRetroLab the proper bench mat material has a top and possibly bottom "dissipative" layer, i.e. it will not hold a static charge and any static charge will dissipate into the conductive layer and of course the internal layer or bottom layer is conductive. The equalisation can only happen if all these things are bonded together and that is where ground is useful. So in my case at home where I don't have ESD flooring (unlike work) I have matting right along my work bench and yes it is "grounded" via a special ground only plug and typically 1M resistor. Then my wrist strap is also connected to that common grounding point. This way I stay at the same potential as my work bench and any mains grounded equipment like soldering irons, bench supply and whatever I am working on is already on the mat and will not build up a static charge (relative to me or ground). You can take it as far as you want, obviously the material your clothes are made out of can make a difference, the humidity etc etc. If you maintain constant contact with what you are working on then yes just a strap will help but the risk of discharge is then between you and the device and say your plastic bench mat or real ground or your chair. That is why I'd encourage to get a mat, strap and ground connection. It only takes 50v or less to damage (ref IBM training) something that does not have integral ESD protection (like modern chips do). You won't even notice that sort of static discharge.

[@charlybrown9024]

I think you have mislinked the third bridge @15:10. Look on your RetroWiki thread.

[@NoelsRetroLab]

It does look like that from the picture because of the angle of the wire, but the link is correct. I'll post an updated picture. And most importantly, I checked continuity on the connections like 10 times in that area.

[@sarreqteryx]

15:54 your crystal is loose. did you try reflowing it?

[@NoelsRetroLab]

Yes, I noticed the pads were totally missing, so there was nothing to solder it against. It was just held with the wire on the other size, which was at least soldered correctly to the crystal. Working on a board that someone destroyed is quite frustrating.

[@Ikrananka]

Bummer about the video chip. One question regarding your mod of the 4164, if Pin 1 is NC then does that mean it is not internally connected to anything. If that's the case then do you really need to cut the leg off, as applying -5V to pin 1 would go nowhere internally?

[@NoelsRetroLab]

Probably, but I haven't tested it. It would make sense, but it could even be one of those things that is left to the implementation of the actual memory chip? Not sure, but it felt a lot safer to cut it (or bend it if you really don't want to cut it).

[@Ikrananka]

@@NoelsRetroLab Absolutely, better safe than sorry. It would be interesting though to do a test on one of the chips, applying say 5V to Pin 1 and then testing the chip afterwards to see if it still functions. I have no idea myself, but one would think that if a pin is NC then literally the physical pin would just be a dummy pin with absolutely no internal connectivity. But like you, I don't know if that is the case or not.

[@NoelsRetroLab]

@@Ikrananka Yeah. Or even better, feed it 5V from a bench power supply and see what kind of current draw it has. If it's significant at all, bad things will happen :-) But you're right, I expect it to be disconnected internally.

[@drtadjakaroghli6898]

Without words Sin palabras

[@Starchface]

Whatever happened with this? Did the replacement video chip arrive? The state of the machine is puzzling. It's as if a kid without the proper information (or soldering skills) was working on it. This could have occurred in the '90s or so. The date codes on those improper RAM chips is 1985 so obviously it was after that, but not too recent because in that case they should have had all the information they could need from the internet. The wiring to the odd Sony voltage regulator boards appears the same as the wiring to the missing crystal. Again I think this is evidence of an amateur's repair efforts. Perhaps the video output was garbled, so they replaced the video RAM. Clearly they were not aware that the 4864 RAMs would cause a short circuit. Of course, the fuse blew at once when powered up. Not knowing what they had done, they started replacing other parts. Savage!

[@NoelsRetroLab]

In a couple of days I'll be posting a video with some related updates, but don't worry, it's going forward. Whenever the video comes up you can laugh at me for... getting the wrong VDP the first time around :-b The correct one is on its way, so hopefully I'll wrap this up in a week or two.

[@KolliRail]

Those things jst happen... Don't worry! But to me it was very interesting to watch, because I have an MSX-Computer (Philips VG8010) with black screen that I can't get to work. Think I have to take a closer look at the video memory and the PAL chip...

[@10p6]

Sinclair pay for sockets, LMAO

[@AngelFernandezA]

Wearing THAT t-shirt was a very subtle "Subterfuge", if you know what I mean... and definitely, you know (...guiño, guiño 😉 😉 ). ¡Saludos!

[@NoelsRetroLab]

🤣🤣

[@fabioacrs]

I don't know if you have figured it out yet. But I think one of the new bridges you've made is conected wrong. The one that was over the large trace.

[@NoelsRetroLab]

I was told that, and I agree that it looks that way, but I checked and it's connected correctly. It looks like that because of where the housing of the cable ends, it looks like it's soldered to the wrong place. Thanks for the heads up though.

[@fabioacrs]

@@NoelsRetroLab Ok, waiting for the follow up :)

[@NoelsRetroLab]

@@fabioacrs I'm waiting on th replacement VDP. You'll see it in the video, but I ordered one and it seemed to work but... I accidentally ordered an NTSC one! Doh! So I'm waiting on the PAL one. Should arrive any day now.

[@TzOk]

There is a deep truth in the suggestion, to not attempt to repair something, that is already working ;)

[@NoelsRetroLab]

I know, right? On the other hand, I couldn't let those wires with exposed metal hanging out back there alone :-)

[@djmips]

I would've left well enough alone. ;)

[@prpplague]

great video! nice explanations!

[@BigGfromSA]

Great video. THKS

[@robertlinder6414]

Where's the link to part 1?

[@NoelsRetroLab]

It should have been linked at 00:26, but here's the link to the full MSX/SVI playlist: kzbin.info/aero/PLmbYPqkyiASxP3MVx_3yBTu30o0egZGpF

[@thomasandrews9355]

Wait, confused...why is it the chip if you're missing clock? Missing clock on video is likely dodgy solder joint or trace...Am I missing something?

[@NoelsRetroLab]

Because the chip completes the oscillating circuit. Look at the circuit I made to test it. The chip sort of implements the two inverters (or whatever circuit they actually implement). That's why the chip has connections to the crystal on both pins 39 and 40. If you remove the chip (or if it's faulty), you get no clock signal. You're right about bad solder joints and I actually did reflow them all off camera to be sure, but still nothing, that's why I'm leaning towards a bad chip. But who know, maybe I'll get a new one, replace it and it won't still work.

[@thomasandrews9355]

@@NoelsRetroLab Ah, didn't know it was implementing the divider circuit. I am going to say I doubt the chip died, unless you shorted something out its a bit unlikely. I'd make sure its getting voltages and such. Also ohm out the various pins. I'm sure it'll be something simple. I ran into this when fixing my Amiga 2000, now I can generally pin down faults by looking at them =)

[@NoelsRetroLab]

@@thomasandrews9355 Yeah, I'm not hopeful about the new VDP, but it's the only lead I have right now. I checked the connections around the crystal like 10 times (since that's the last thing I messed with when this happened) and they check out fine. I also pulled out all the VRAM in case something there was shorted, but nothing. I suppose I can take out the VDP and put in a socket. That might be enough to catch some small short or a bad solder joint. Might do that if the VDP takes too long to arrive.

[@thomasandrews9355]

@@NoelsRetroLab Okay, I'm reading the schematics, I see XTAL1 and XTAL2 are connected to a capacitor RC network? Thats actually what the yellow thing is, a variable cap, looks like 50pf? I think this is entirely a circuit issue, I'd review hansotten.file-hunter.com/uploads/files/STM-A_SVI318-328.pdf page 6, under system clock. Verify no caps are shorted. The crystal I'd imagine should still put out a signal, just not oscillate. Start there.

[@NoelsRetroLab]

@@thomasandrews9355 That's the doc I've been using. The exact schematic is in page 37, because this is a MK II version. The clock circuit is very slightly different. Those capacitors and resistors are the ones I've checked for connectivity a bunch of times and they're fine as far as I can tell. The crystal puts out a small oscillation on pin 39 (see towards the end of the video). Oh how I wish I had a working 328 side by side to compare signals and swap chips. It's so much easier when you have a working machine! :-)

[@ingmarm8858]

Sorry Noel but your "fix" of the oscillator tracks looks just as bad as the original lol. Maybe your fix has change the circuit characteristics just enough to stop the xtal from starting up. Try tweeking the trimmer capacitor to see if the oscillator starts up.

[@NoelsRetroLab]

Haha, fair enough. The oscillator tracks are horrible because it's completely missing the pads, so it's attached the best I could without getting into a bigger production. And my wire job isn't the best either :-) No tweaking of the trimmer would change it though. I did verify later that it was the VDP, so at least that's finally working. But yeah, this could definitely be improved.

[@williamsquires3070]

You probably damaged the video chip by inserting the video RAM chip in upside-down (i.e. with the notch on the wrong side.) That may have connected RAS/CAS to some other line - maybe a power bus - and the video chip wasn’t happy with this. 😢

[@NoelsRetroLab]

I thought about it, but I doubt it. All it could do is connect the address/data/control lines to other parts of the chip. The only output is data out, which flipped would be connected to A5, and I doubt it would cause any damage (and chances are it wasn't even enabled as output). The RAM chip on the other hand could have well been damaged swapping VCC and ground, but apparently it was sturdy enough.

[@Derpy1969]

Should have “fixed” something that was working!

[@KeriSqueak]

What's this about Neptunium O.o

[@NoelsRetroLab]

It's the t-shirt for a game I made a while back subterfuge-game.com

[@KeriSqueak]

@@NoelsRetroLab Oooo, you're a sneaky tricksy one. Well played, I suppose :'D

[@larixlyallii]

Noel! Subs!

[@drakepirate]

Excellent video, thanks. Just one question. Is that schematics for crystal testing correct? because those resistors are in series and not parallel. (at your vero board looks correct) Could be this one instead? www.eleccircuit.com/wp-content/uploads/2014/12/simple-crystal-oscillator-1-10mhz-using-inverter-gate-600x532.jpg. Excelente video, gracias. sólo una pregunta. ¿Está bien el esquema del comprobador de cristales? Porque ahí lo veo en serie, podrías haber puesto una resistencia de 2K... pero en tu montaje parece correcto.

[@NoelsRetroLab]

Good catch. You're right! I think I missed that extra connection in the diagram in the middle from the output of the first inverter to the point in between the two resistors. Gracias! 😃