You are perfectly correct in your logic deduction that the amplifier was shorted and overloaded the regulator. Let me add some info about the regulator, though: The 78xx type regulators have over-current protection and over-temperature protection, and you have seen both of these protections in action. Over-current protection is limiting the output current to a value some amount above the nominal maximal current. As the standard 7809 is specified to be good up to 1 amp, it will likely clamp the output current to around 1.3 to 1.7 amps if shorted. If the output voltage is around 3V at 1.5A, the regulator dissipates 9V * 1.5A = 13.5W. This is why that thing got hot so fast. The 125°C you quote is the maximum die temperature for simple logic chips to operate as specified. This specification limit is due to rising leakage currents disturbing the operation, not to silicon breakdown. The maximum junction temperatures for simple bipolar transistors that does not damage the device is usually specified as 175°C. There is some margin to destruction, but not that much, expect the device to break down at around 200°C. At that temperature, the P/N barrier will break down, and the transistor will *turn into a short circuit*. This is a *very* bad thing to happen on a voltage regulator, because that would mean it will pass the input voltage to the output. On that sound card, passing 12V to the TDA2025 wouldn't be the end of the world, but think about a logic board getting 8.5V instead of 5V because the transistor integrated in a 7805 shorted out. That's why there is a thermal protection that kicks in way before the die is at a crititcal temperature of 175°C, but it might be well above 125°C. If the die gets too hot, the output current is further reduced. It seems the noise from the card is the overtemperature protection kicking in and disengaging periodically: The overtemperature protection oscillates.
Seems like a common defect on SoundBlaster cards with sound output chips for unpowered speakers The amp drags down the VR with excessive load and causes the VR drops the voltage to reduce heat
The truth is that I know nothing about electronics, I know pc´s because I worked all my life on them, but is incredibly easy to understand your explanations so I guess I thank you for teaching me about electronics the way you do. keep the good work you´re doing.
BTW, TEA2025 amplifier datasheet says in Stability sections: "No Sockets". GND pins 4,5,12,13 has also second function - its thermal pads for IC cooling. So this pins must be solder to big GND polygon on the PCB for same reason - for dispantion 2Watts power
@@tony359 The datasheet PCB design recommendation is totally irrelevant here. It's a Sound Blaster board, that can't be changed. A datasheet is only the manufacturer's recommendation. It's configured in stereo mode (not the more power-hungry bridge mode) and a socket won't make any difference in this application. It is not using much current and your PC speakers are powered. I doubt the IC even draws 100mA and I doubt it even gets warm. But test it and update here if it does....
@@tony359 Sponsor pleasing idea: make a PCB that sits in place of the amplifier IC and carries it that allows you to bypass the power amplifier and have line out. The PCB can also be equipped with a pour with plated through holes that would make a heatsink for the amplifier IC.
I would try to solder some decoupling caps to the ISA riser. Something like 220uF + 100nF to each power line right at the bottom side of the ISA socket. Adding some shielding (grounded alu or copper tape) around the cable could also help.
@@tony359 on our extenders at work (for debugging boards that are plugged into small chassis), they put test-point loops on common lines. Really handy for clipping a probe in place while debugging. Just a thought for version two.
Best video yet Tony!! Not only the technical data in on point but the editing with schematics and you showing the test points on the card at the same time was perfect. Thank you!
Thank you so much, it takes me quite some time to do that so it's particularly rewarding to read a comment like yours. I appreciate your kind words! Cheers!
That first sound out of the speaker i could only describe as a "deflating digital fart". That's the sound of a poorly sound card pensioner. 😁 Yet another excellent video, sir.
What a great video this was!!! The level of details, B-roll, schematics, close views on the board, etc... I followed perfectly your reasoning and that was truly entertaining too. Your content is improving.
Thanks so much - I'm learning! :) Really thanks for your kind words, it takes quite some time to get to that final editing so your comment is particularly appreciated!
My all time favourite card, Very useful info, Unfortunately my non functioning card is 'undetected' and nothing is getting hot on my card. Using the extender is probably the best way to really compare it with a working card and figure out what should do what. Congrats on your repair, this gives me hope.
undetected is bad. I see a 245 IC there, that should interface with the ISA bus. You might want to swap that. And check the usual: voltages, clock, reset. Good luck!
Careful with the hot parts :) I just burnt my finger on an Athlon CPU. It's a big blister. I didn't get the memo that these reach 370 degrees C, 23 years ago, and this week, while testing, I paid the price.
Thanks! I'm so happy it's working-ish! It would have been almost impossible to troubleshoot this SB while plugged on the motherboard. Particularly with a camera in the way.
I really enjoyed this video today, i remember yesteryear when you had to have a sound card to have sound in a game unless you loved the BEEP from the pc speaker lol
Thanks, it was a cool repair which I enjoyed a lot! I could not afford a sound card back then so I did play DOOM with the PC speaker. I remember turning on the radio to be my soundtrack :)
Great, to the point faultfinding. No parts cannon was used in fixing this Soundblaster pro :D Great job and entertaining and engaging to watch. Your positive attitude makes me come back, every time theres a new video. Thank You! :D
For the test jig, I think I would've dropped the cables and put the new ISA-socket straight on the male part as a regular riser. Then you'd just need something to prop up the card being tested. As the mounting bracket would always be in the same place in relation to it, that's a natural place to put the supports for it.
Yes that’s also an option. I just thought some flexible cables would give me more… flexibility, also to turn the card around. But yes that’s an option for sure. Thank you!
I agree that making an ISA extender card, just so it can rise above all other cards (so you can get to both sides) is the way to do it. We use that extender card technique a lot at my work to make our cards be accessible (they are normally slotted into a unit).
Shielded cables is a good idea to try, but it may not be the problem. Some thoughts: 1) Termination impedance which will cause ringing and if the ringing exceeds the logic voltage threshold, you are flipping bits. 2) Timing. The speed of light (actually 68-72% through wire) is not that fast at all. Note on 386 and higher motherboards you have those sqiggly traces to match buss length signals.There may be too much delay added by your extension card for the required handshake time of the video card with the CPU. 3) Power and ground. Too much resistance with either one due to the ribbon cables.
Thank you, your points are excellent. Someone has already mentioned points 1 and 3. I'll try 3 soon. For point 1, I might need to turn the design into an active one where the signals are received at the plug and transmitted to the socket. I'm hoping this is not necessary and what you say is only valid for 100% reliability, not for a 10 minutes test on a bench! :) But it's a good point. For point 2, I'll bear it in mind though there are around very long ISA backplanes, I'd be surprised if my extension could be a problem. But I need to make sure there is not much difference between the length of the traces indeed. Thank you!
The Roland MT-32 also does some similar tricks so they can use a mono 16-bit DAC for stereo which in the day was a very expensive part. It outputs through a analogue switch to a pair of R/C networks. The MCU would switch L/R samples and the switch at the same time to allow the single DAC to be used for two channels. Note too, you stated that you were adding leaded solder first, cards of this era were before lead-free solder and are all using leaded solder. I'd have done the same simply to assist in removing the part, but just a small note :)
Oh yes, I always add leaded solder to "refresh" the existing one and also as I read that low melt + leaded solder = super low melting temperature. Besides that I feel that the extra fresh leaded solder + flux helps with flowing and saves (expensive) low melt solder :) interesting trick the one you mentioned! thanks!
Hey Tony, great Video, Always enjoy watching, forget about the people complaining about advertising, You clearly explained in other videos WHY PCB way is Important for this channel , If they don't like it, they don't have to watch, pretty easy. You are doing a BIG service for the retro Community on testing and fixing and putting it in a way us hobbyist can clearly understand. You also have some really knowledgeable people commenting and i have also learned quite a bit from what they commented on. I like how you also say "we will fix this together" Looking foreword to the next video.
Thanks for your kind words! In an ideal world I could make my videos without having to worry about the bills at the end of the month but hey :) At the same time though I want to make sure that the videos remain enjoyable, so negative feedback is also welcome! Thanks for your kind words!
Super.....I have heard that sound cards are notoriously difficult to fix and troubleshoot....glad you got it working......Another great video...thank you....😊😊😊
I agree, mostly because when I tried in the past there was not much available in terms of schematics and datasheet. Thankfully the fault was in the Yamaha section!
I might suggest next time you design the extender is use a few more sockets and wire grounds between most of the logic signals and possibly double up or use multiple pins for the pins used for power and add capacitors to the relevant power rails (particularly 5 v) electrolytic bulk cap (careful with the negative rails) at the plug on the extended point plus some ceramics close to the ISA slot. You may have been getting ground bouncing or ringing signals with the video card which likely has rather high frequency signalling
Absolutely - next design will have a separate loom for power and the caps for sure. I'll do some tests on the current setup in case that's all it takes! Thanks for that!
For your extender card, most bios's have setting for bus speed, wait states etc. Using these may get you over the line with your extender. Also some supply rail bypass caps would be advisable.
it didn't make it in the video but I did try on the 386 board which has some settings - nothing worked unfortunately. But thanks for that. I will definitely add bypass and bulk caps as a test ASAP!
I'm surprised the riser didn't work, because PCIe risers work. I think someone was saying the wire/traces need to be the same length to keep timings, ie, maybe put the ribbons side by side rather than over each other.
I can trim the ribbon more accurately but I think it's ending up being the same length. Side by Side might be a good idea - I ended up with this design to keep the footprint as small as possible. I am also told PCI-ex is "designed" to be extended, that's why you can buy a riser for £5! :) Anyways, thanks for your input, I'll consider the side by side option for sure!
Maybe shield the ribbon cables with aluminium foil? Good old TEA2025. On one of my opti soundcards, it has a hole in it (was like that when i bought it). No problem, just don't use amp, you also don't amplyfy noise!
mega repair, Tony, it can be seen how much you enjoy doing this sort of repairs. also noticed that you're not something of a dancing man, but you managed :))
Regarding the extender I suspect signal integrity could be the problem. I would do an extender using a better flat cable like the ones used on scsi and pass each isa signal interlaced with ground wires (signal1, gnd, signal 2, gnd, ...) to make it like an rf transmission line. Regards.
ISA bus has very poor distribution of ground among all of the pins. So I would use bigger ribbon cables and put more grounds in there. Also double/triple/quadruple up the 5V etc power pins on the ribbon to counter voltage drop. You could test the voltage drop theory by measuring it on the video card, it would actually be quite interesting to see what is like. 😅 As others have suggested it would probably pay to add some bulk capacitors on the remote card to help provide a lower impedance source of current at the end of the ribbons.
Absolutely, I look forward to scope those lines and see what's happening there! For the ground, I'm considering using 80 conductors ribbons - we'll see how it goes after I upgrade the power lines! Thanks for that!
There is a thread on vogons about extending the ISA bus is a similar way to add more ISA slots using a backplane. You might want to try adding capacitors across your power rails to ground on the back of the ISA slot and using 80 conductor cables instead of the standard 40 conductor IDE cables.
Absolutely. In fact, I started modifying the design for the 80 conductor cable and then realised I had to re-do everything (ground has to be on specific pins etc) and... I'm still very slow with PCB CAD software so I crossed my fingers for this design. Next one will be better :) Thank you!
Signals with fast rise times, like digital bus signals, do not like going long distances. What is defined as "long" changes with the rise time of the signal. Essentially, the faster the rise time, the higher the frequency content of the signal, and the more you need to make RF-type considerations in your design. Impedance mismatches between the signal source, the transmission line (your ribbon cable/pcb), and the signal receiver can create signal reflections jeapordizing signal integrity. As others have mentioned as well, crosstalk can be an issue here. That also gets worse with higher frequency content. If you add a GND wire in between all the data signal wires in your ribbon cable, crosstalk should reduce significantly if present. It doubles the conductor count of your cable but is the ideal connector configuration. The way I think about it is, if you imagine an electric field surrounding your signal wire, you want that field to "hit" a voltage reference (either your positive rail, negative rail, or GND) before it hits another signal's field. I dont think adding capacitance on the power rail is necessary, as the ISA card itself will have any decoupling caps on it. Adding some wouldn't necessarily hurt, but I don't think that's your problem. Another issue is parallel busses have strict timing relationship requirements between all the data and clock lines. Again considering those pesky RF transmission line effects, length mismatches between the bus signals can cause timing mismatches that violate those timing requirements, causing undefined behavior. Like the other issues, this gets more intricate the higher the frequency content. Hope I gave you a couple things to look at for troubleshooting! If your scope has the bandwidth for it, you can check for these issues by checking the shape of the ISA bus signals at the card end of the extender. Crosstalk and reflections and whatnot will turn clean square waves into ones full of ringing and random level changes.
Yes, these are very good pointers indeed! I will start with the basics (thicker power lines, that ribbon is CHEAP!) and then inspect the data lines. My scope can do 200MHz so I should be able (if I am capable of!) of spotting issues there. I will also consider going with 80 conductors cables (or just simply adding another ribbon and make it signal-ground-signal-ground as you suggest, this is great as I have plenty of those connectors!) and consider the length of everything in the design. It's going to be fun! Thanks for taking the time to share your opinion, it really helps!
Fantastic video. Smiling from ear to ear hearing those classic sounds coming out of that board. That ISA riser is really interesting too. Good up the good work.
One idea that would work for some MB's would be a card that plugs into the ISA slot, then an ISA board can plug in at 90 degrees, so horizontally. No ribbon cable capacitance. You could have connectors on either side of the pcb, to support plugging the test board in which ever way works best. Not as flexible as the ribbon cable one, but better signal integrity on the ISA bus.
Try to check the negative voltages on your riser. As far as I know VGA and serial ports need -12V which could possibly be a source of some problem. (?)
Usually on ISA extensions you have a 10uF capacitor on the 12v line after each ISA slot. Even motherboards have a cap like this. Your extender doesn't have that so it's possible that there's voltage drop or too much ripple on the card that's on the extender. Some cards might be tolerant of that but not others.
I suspect the problem with the extender is primarily due to power - the ribbon cable cannot provide the power needed for the videop bord. Try adding heavy duty power cables and some capacitance on each power rail at the test board end - 100uF electrolytic + 100nF ceramic.
Yes, I plan to try that for sure. Rev 2 will have a separate connection for power and decoupling caps. But I'm curious to see if this works with the extra power cables - this ribbon cable is VERY cheap, it's got 5 strands inside each wire....... Thanks for that!
@@BogTheWombat My multimeter didn't see anything wrong but it's just a multimeter. Maybe the video card is causing a momentary voltage drop when it tries to initialise.
@@tony359 A meter won't do it - you will need a scope with a decent bandwidth, I suspect you will see large spikes on the supply rails in the microsecond or sub-microsecond level that is throwing the logic into all sorts of invalid states
Great repair! I wonder if the voltage regulator running at that high of a temperature over a period of time may have "cooked" the op amp chip. No telling if the card was left installed in a PC for any length of time after the amplifier IC shorted. I could easily see someone simply unplug speakers from the output of the sound card after it initially failed and continue to use the PC as with no speakers connected there was no annoying audio present.
Nice fix! I made an ISA extension on my channel. It has some weird behavior too. Some cards work fine, others have issues. I haven’t done too much diagnosis, but it feels like it could probably use some caps to smooth out the voltages.
I don't have one and they're not super-cheap online for just a test. And there are only straight risers, I would like a 90 degrees riser - which would not allow me to probe the card at the back. I think I'll play with the scope a bit, hopefully I'll see what's wrong.
You're more than welcome - I didn't mention in the video that of course the project will be free for everybody to make. I'll share it... when it works :)
It will be good idea to add separate PSU connector for 12V and 5V on adapter output slot board, maybe voltage drop on those cables is that high that some circuits stop working on video card ☺
Yes absolutely. I am going to try on this revision but next one will have power on a separate cable. That said, my multimeter doesn't measure anything nasty but who knows, maybe the voltage drops momentarily when the board tries to initialise. Thanks for that!
To avoid signal cross talk via Ribbon cables you should use Double the leads, "instead of 40 have 80" Double as you will need an earth on either side of every signal lead, to act as a shield from each other and this is especially important if the signals are pretty low or worse really fast switching ergo the leads are "Close to behaving as if an AC signal" is being fed through, Speaking of which.. it may also be critical that the leads are all the same length "Due to Video Timing Requisites" and they also must be as short as possible for numerous reasons Yeah.. "Too many theoretical reasons to detail here".
Yes, next revision will use 80 conductors "IDE" cables I guess. I went to modify the design before manufacturing this one but I realised I would need to re-do everything as ground would need to be on specific pins... I'm still slow with PCB CAD software so I gave up and crossed my fingers :) Next one I'll definitely do that! Thank you!
A much smaller 90º angle adapter, can only be used in the lowest ISA slot and can't probe the back of the card, but could it be a better solution? Super nice job on fixing the card.
Yes, maybe I am overthinking that. Though I couldn't find anything online and I ended up with this which gives you a little more flexibility - but doesn't work! :D I might make a simple 90 degrees one as you say, as a "backup riser". Thank you!
Molto bravo, per quanto riguarda l' extender i moderni pcie riser's hanno condensatori e connettori per fornire voltaggi stabili alla periferica che viene connessa, varrebbe la pena provare ad aggiungere anche solo i condensatori per esempio 220nF sulle power rails 5v e 12v
Ti ringrazio - Si', e' il mio prossimo passo. Saldare dei cavi per l'alimentazione e aggiungere condensatori dove c'e' il connettore ISA. Spero sia tutto quello che serva anche se un mio follower ha trovato le specifiche ISA per me e ha scoperto che consentono l'estensione del Bus al massimo per 2.5 inches quindi il mio adattatore e' fuori specifiche. Ma magari per testare la roba funziona lo stesso. Vedremo! Grazie!
use 64 and 40 way ribbon cables splitting your signalling to as it is laid out in ISA slot, not unpairing them as you did - yes, there is no diffrential pair in ISA, but still that bus pinout is not there by chance, but througly thinked over. your pinout on 64 way ribbon should be like pin 1 - A1, pin 2 - B1, not like you did right now with one ribbon with all A and C pins, and the other with B and D pins
Nice repair. For the ISA riser, I would prefer a video on how to get it to work with the video card (if possible). Could be a useful tool for the retro computing community. Greetings from Germany.
Thank you! I'll experiment and might document that in a video for sure! I didn't mention in the end but it would have been incredibly difficult to troubleshoot that issue ON the motherboard... not to mention with a camera in the way. So yes, I want that to work with video cards too! Thank you!
Think if you design a isa card that was the normal height of a isa card so the card your testing would be above the rest of the cards would work better because you could space out the rails to reduce interference and have places to test the bus eaisier.
I thought of that but 1. it would need the motherboard in the case so the riser can be screwed on something 2. it would still be wobbly and difficult to probe. I wanted something that made the card rest on a rigid surface - I'll eventually get to a working one! :) Thanks for watching!
I had an issue with a long USB cable and it turns out the problem was delivering power rather than signal degradation. Maybe explains why the VGA card doesn't work because is more power hungry than the other cards.
I checked with my multimeter, there was a little drop but nothing terrible - though I should check with the scope. Maybe there is a drop when the board is trying to initialise. That ribbon cable is awful, the gauge is miniscule. I shall try jumping 5V and 12V via extra wires to see if that is the issue - if that is the problem I can either replace the ribbon with something better or make a Rev2 with a dedicated connector for power. Thanks for that!
Thanks! I went to buy one first but I don't like the "card" risers because you end up with an even more unstable solution. Also, I think they're only available in 8 bit format. The 90 degree ones are ok - but I can't find them for sale. Hence my partially working design! Rev 2 will be cool, I promise :)
A VGA card needs to talk much faster with the CPU and memory via DMA. The higher bandwidth needed is just too high to reliably travel through your test rig. Nevertheless great video. Like it.
Thank you! I thought the speed was dictated by the ISA bus. I guess there might be many ‘modes’ to use that but, I don’t know. I can’t wait to troubleshoot that! Thank you!
@@tony359 I'm interested what your findings are. Max speed is indeed dictated by ISA bus. But cards not always use the max speed. For example a sound card needs to do some I/O read/writes in the tempo of the music to the FM chip. Digital audio is 44,1 kHz. But VGA has to update video memory via DMA transfers for 640x480x18bits.
I wonder if the rig not working with VGA cards has something to do with the fact that the VGA card has BIOS extentions and the IDE controller and sound card do not. So the PC has to run code from the VGA card bios but not so with the other cards...
@@tony359AFAIK the the card only needs to send extra pulses down the clock line when no other device uses the bus. OR, so I have been told by someone who toyed more with the ISA stuff... my history with PC is short enough that my first daily driver PC had a PCI video card already, and the next one AGP. The ISA gear I have toyed with later on was ancient already at that point and I wasn't studying them very deeply.
It might be some special mode the card uses - the term "overclock" is probably what puzzled me as the ISA clock cannot be changed by a card. But whether the card is doing some wizardry to speed up, that's very possible indeed!
Yeah, opamps, dacs and power amp chips are all common failures on any board that makes sound. I've replaced hundreds of those types of chips when repairing arcade games over the last 30 years. If you use a logic probe with a built-in piezo speaker you can actually 'hear' the analog audio when probing the chips.
@@tony359 well I've repaired 1000's of arcade games over ~30 years and I actually only ever used a logic probe for that. I only use the scope on analog audio because a logic probe can't detect it, usually just shows as a high. In 30 years I think I used the scope maybe 10 times and half of those it didn't help heh!
Ran into someone extending cables via ribbon who swears that tin-foil( covered afterward with something non-conductive) works as a shielding. IDK, try it?
TEA2025B was a very popular amplifier IC in the late 80' and early 90' i think. It had several equivalents ( like KA2206 for example ), and it was widely used in those "boombox" players back then ( either this one or TDA810 and it's equivalents, which was a 3W amplifier, thus a bit more powerful ). Anyway, it's still your basic class AB amplifier with a very basic schematic that used only one power rail, with bootstrapping VAS and push-pull output stage with an output capacitor to remove DC bias ( which was half of the power supply ). Now, when it fails it usually does so in the output, can't remember of the top of my head if it had interna SOA protection, but usually it happens either from too much load ( IE paralleled speakers ), or accidental shorts on the outputs, that shorts the output transistors in the IC and you get a shorted power line. There is a very basic test you can do ( very fast too ) to determine if an amplifier IC is shorted or not. You just use the multimeter in the diode function, and you test the outputs with respect to the power rail and ground. In both ways you should either find diodes or nothing at all ( open circuit ), if the IC is fine. If the IC is shorted then you will measure shorts between one or both the outputs and one or both power pins. Sounds complicated, i know, but really is just basic stuff, and it works the same not only with audio power amplifier IC, but with any type of IC that has outputs ( gate driver IC's, opamps, switch mode power supply controllers...).
Again, glad you show proper diagnostics instead of just swapping ic's and crossing your fingers you dont fry an expensive good IC in the process. Really enjoy the deep dive diagnostics videos like this. What pcb designer software did you use? Personally I exclusively use Kicad as I find it easiest to work with.
I used EasyEDA but I think I'll switch to KiCAD now. Thanks for your kind words. Yes, I could have swapped parts at random but it's so cool to learn something in the process! Happy you liked the troubleshooting!
@@samuraidriver4x4 I was recommended EasyEDA but every time I ask for advice online everybody seems to be using KiCAD! :) Since I have to learn more or less from Scratch, maybe better stick with the most popular. I'll take a look thanks!
Nice repair Tony, I like the way you say FM, sounds like effing LOL and CHIPS sounds like sh*ts hahaha I love the diversity of smarter people (dual language) accents :)
I was told that PCI Express is somehow designed to be extended. It makes sense to take a look - though I'm not sure the solution there will apply to ISA. But it won't hurt to check. Thank you!
There is a small drop - I definitely need to route power via a dedicated cable and that ribbon I got from Aliexpress is very cheap so the gauge is too small. That said, I don't think that's the problem, with a good ATX PSU and some load to balance the rails I get something like 4.9V if memory serves. That video card works with 4.8V when plugged directly in the slot. Thank you!
You mean the title saying "the wheel has no effect"? Well, sometimes I add voiceovers, in this case it was a simple thing and I only have one microphone which is normally in my workshop :D (aka I felt lazy!)
@@tony359 4 days ago. But i can re-post it here: Maybe try to add som capacitors to the powlines on your risercard with the ISA socket on. Something like this backplane card has: i.ebayimg.com/images/g/JIYAAOSw0vdeQXDV/s-l1600.jpg There is also something called ISA termination resistors, that should be installed and matched according to your cable impedance. For at cable that long, I belive you should only be able to get a stable result with proper matched termination resistors
I don't remember all the comments and I don't seem to be able to find your earlier one - anyways why would I feel offended? :) Your suggestion is sound about capacitors - what kind of resistors would you add to terminate the line?
Ah gotcha! Well if you start replacing components around the regulator you never stop. I guess it should be used for a while to decide. It might be less reliable than the average, who knows!
@@tony359 it's marketing speak or they were just told to say it's a dsp by engineering. A real dsp is something totally different much more expensive and powerful.
@@g4z-kb7ct If audio is being processed digitally, then surely the device doing it will be a DSP, even if it isn't specifically designed for the purpose.
@@SpeccyMan You clearly don't know what a DSP is. That was just marketing spiele. That chip is not processing audio digitally. It's a CPU with internal ROM and RAM and built in timers and other peripherals. It's the brains of the system controlling the board. Look up the datasheet for the actual chip being used... intel 8051.
You are perfectly correct in your logic deduction that the amplifier was shorted and overloaded the regulator. Let me add some info about the regulator, though: The 78xx type regulators have over-current protection and over-temperature protection, and you have seen both of these protections in action. Over-current protection is limiting the output current to a value some amount above the nominal maximal current. As the standard 7809 is specified to be good up to 1 amp, it will likely clamp the output current to around 1.3 to 1.7 amps if shorted. If the output voltage is around 3V at 1.5A, the regulator dissipates 9V * 1.5A = 13.5W. This is why that thing got hot so fast.
The 125°C you quote is the maximum die temperature for simple logic chips to operate as specified. This specification limit is due to rising leakage currents disturbing the operation, not to silicon breakdown. The maximum junction temperatures for simple bipolar transistors that does not damage the device is usually specified as 175°C. There is some margin to destruction, but not that much, expect the device to break down at around 200°C. At that temperature, the P/N barrier will break down, and the transistor will *turn into a short circuit*. This is a *very* bad thing to happen on a voltage regulator, because that would mean it will pass the input voltage to the output. On that sound card, passing 12V to the TDA2025 wouldn't be the end of the world, but think about a logic board getting 8.5V instead of 5V because the transistor integrated in a 7805 shorted out. That's why there is a thermal protection that kicks in way before the die is at a crititcal temperature of 175°C, but it might be well above 125°C. If the die gets too hot, the output current is further reduced. It seems the noise from the card is the overtemperature protection kicking in and disengaging periodically: The overtemperature protection oscillates.
Thanks for the explanation, it's very interesting. I knew I should have researched that 125C number :)
Seems like a common defect on SoundBlaster cards with sound output chips for unpowered speakers
The amp drags down the VR with excessive load and causes the VR drops the voltage to reduce heat
Adrian's party dance! 😂 Nice fix!
There you go! I shall come up with my own version at some point! :) Thanks for watching!
The truth is that I know nothing about electronics, I know pc´s because I worked all my life on them, but is incredibly easy to understand your explanations so I guess I thank you for teaching me about electronics the way you do. keep the good work you´re doing.
Thank you so much for your kind words!
I like it you leave the mistakes on the video instead of editing them out.
These are important for those people less experienced.
Of course, the fun is in the journey :)
Thanks for your comment!
@@tony359 Also, the ISA extender is a brilliant idea!
BTW, TEA2025 amplifier datasheet says in Stability sections: "No Sockets". GND pins 4,5,12,13 has also second function - its thermal pads for IC cooling. So this pins must be solder to big GND polygon on the PCB for same reason - for dispantion 2Watts power
Oh thanks for that! I was a bit unsure about the socket to be honest. I'll remove it and solder directly on the PCB - cheers for that!
@@tony359 The datasheet PCB design recommendation is totally irrelevant here. It's a Sound Blaster board, that can't be changed. A datasheet is only the manufacturer's recommendation. It's configured in stereo mode (not the more power-hungry bridge mode) and a socket won't make any difference in this application. It is not using much current and your PC speakers are powered. I doubt the IC even draws 100mA and I doubt it even gets warm. But test it and update here if it does....
Well, I feel it's prudent to solder it back on the PCB in case someone plugs some low impedance speakers and decide to run DOOM at full blast :)
@@tony359 Sponsor pleasing idea: make a PCB that sits in place of the amplifier IC and carries it that allows you to bypass the power amplifier and have line out. The PCB can also be equipped with a pour with plated through holes that would make a heatsink for the amplifier IC.
Not a bad idea for sure, thanks!
I would try to solder some decoupling caps to the ISA riser. Something like 220uF + 100nF to each power line right at the bottom side of the ISA socket. Adding some shielding (grounded alu or copper tape) around the cable could also help.
I have tried with kitchen foil around those cables. But decoupling caps are definitely a good idea for rev 2, thanks!
In addition, you might consider attaching extra 5V and 12V cables. You might be getting some voltage drop under load.
indeed, that's definitely on my list. Actually something I'd like to try with my current setup in case it's just that. Thanks!
@@tony359 on our extenders at work (for debugging boards that are plugged into small chassis), they put test-point loops on common lines. Really handy for clipping a probe in place while debugging. Just a thought for version two.
Uhhh, very nice! Thanks for that!
I have a SB16 without OPL-sound, so now you gave me some ideas as to how to fix it!
Good luck, I hope you can sort it out!
3:04 - Never heard a sound card farting so much like this one! 😝. You really *amped* up this repair Tony! 😁
ahah yes, the card was farting at me :)
Best video yet Tony!! Not only the technical data in on point but the editing with schematics and you showing the test points on the card at the same time was perfect. Thank you!
Thank you so much, it takes me quite some time to do that so it's particularly rewarding to read a comment like yours. I appreciate your kind words! Cheers!
That first sound out of the speaker i could only describe as a "deflating digital fart". That's the sound of a poorly sound card pensioner. 😁 Yet another excellent video, sir.
eheh I'm becoming an old fart after all :)
@@tony359 I can't in good conscience up vote this. 😬😁
What a great video this was!!!
The level of details, B-roll, schematics, close views on the board, etc... I followed perfectly your reasoning and that was truly entertaining too. Your content is improving.
Thanks so much - I'm learning! :)
Really thanks for your kind words, it takes quite some time to get to that final editing so your comment is particularly appreciated!
My all time favourite card, Very useful info, Unfortunately my non functioning card is 'undetected' and nothing is getting hot on my card. Using the extender is probably the best way to really compare it with a working card and figure out what should do what. Congrats on your repair, this gives me hope.
undetected is bad. I see a 245 IC there, that should interface with the ISA bus. You might want to swap that. And check the usual: voltages, clock, reset. Good luck!
@@tony359 Thank you. I will start there.
You could re-work the board to have every other wire of the ribbon cable being common ground. Crosstalk is an issue at high speeds.
It's a good idea, I think next revision is using 3 x 80 conductors IDE ribbons and I'll have ground where it's supposed to be. Thank you!
Hello Tony
Clever troubleshooting, cool repair, nicely done video, thanks!
Best regards
Alex
Thank you Alex! I enjoyed this one for sure! I'm glad you liked it too!
Careful with the hot parts :)
I just burnt my finger on an Athlon CPU. It's a big blister. I didn't get the memo that these reach 370 degrees C, 23 years ago, and this week, while testing, I paid the price.
I'll have another memo sent to you :D
Sorry to hear that, I can imagine the pain!
Cool ISA riser.
Thanks! I'm so happy it's working-ish! It would have been almost impossible to troubleshoot this SB while plugged on the motherboard. Particularly with a camera in the way.
Canyou I get the Gerberand file a and Parts list for yourhands iSA test riser?
Have you thoughtit about a PCI and PCI-E test riser card?
@@highforresttechnologies4325 Sure, I'll share it somewhere. No, haven't thought of a PCI extender. PCI-E are readily available though.
I really enjoyed this video today, i remember yesteryear when you had to have a sound card to have sound in a game unless you loved the BEEP from the pc speaker lol
Thanks, it was a cool repair which I enjoyed a lot! I could not afford a sound card back then so I did play DOOM with the PC speaker. I remember turning on the radio to be my soundtrack :)
Great, to the point faultfinding. No parts cannon was used in fixing this Soundblaster pro :D Great job and entertaining and engaging to watch. Your positive attitude makes me come back, every time theres a new video. Thank You! :D
Thank you for your kind words!
For the test jig, I think I would've dropped the cables and put the new ISA-socket straight on the male part as a regular riser. Then you'd just need something to prop up the card being tested. As the mounting bracket would always be in the same place in relation to it, that's a natural place to put the supports for it.
Yes that’s also an option. I just thought some flexible cables would give me more… flexibility, also to turn the card around. But yes that’s an option for sure. Thank you!
I agree that making an ISA extender card, just so it can rise above all other cards (so you can get to both sides) is the way to do it. We use that extender card technique a lot at my work to make our cards be accessible (they are normally slotted into a unit).
Shielded cables is a good idea to try, but it may not be the problem. Some thoughts:
1) Termination impedance which will cause ringing and if the ringing exceeds the logic voltage threshold, you are flipping bits.
2) Timing. The speed of light (actually 68-72% through wire) is not that fast at all. Note on 386 and higher motherboards you have those sqiggly traces to match buss length signals.There may be too much delay added by your extension card for the required handshake time of the video card with the CPU.
3) Power and ground. Too much resistance with either one due to the ribbon cables.
Thank you, your points are excellent. Someone has already mentioned points 1 and 3. I'll try 3 soon. For point 1, I might need to turn the design into an active one where the signals are received at the plug and transmitted to the socket.
I'm hoping this is not necessary and what you say is only valid for 100% reliability, not for a 10 minutes test on a bench! :) But it's a good point.
For point 2, I'll bear it in mind though there are around very long ISA backplanes, I'd be surprised if my extension could be a problem. But I need to make sure there is not much difference between the length of the traces indeed. Thank you!
The Roland MT-32 also does some similar tricks so they can use a mono 16-bit DAC for stereo which in the day was a very expensive part. It outputs through a analogue switch to a pair of R/C networks. The MCU would switch L/R samples and the switch at the same time to allow the single DAC to be used for two channels.
Note too, you stated that you were adding leaded solder first, cards of this era were before lead-free solder and are all using leaded solder. I'd have done the same simply to assist in removing the part, but just a small note :)
Oh yes, I always add leaded solder to "refresh" the existing one and also as I read that low melt + leaded solder = super low melting temperature. Besides that I feel that the extra fresh leaded solder + flux helps with flowing and saves (expensive) low melt solder :)
interesting trick the one you mentioned! thanks!
Very good my friend !
Thank you !
Thank you for watching!
Excellent repair! I like your Probing-Board for the ISA slot and I want to see a video with some info's or a Rev2 of it :)
Thanks! I'll document my troubleshooting on rev1!
Hey Tony, great Video,
Always enjoy watching, forget about the people complaining about advertising, You clearly explained in other videos WHY PCB way is Important for this channel , If they don't like it, they don't have to watch, pretty easy. You are doing a BIG service for the retro Community on testing and fixing and putting it in a way us hobbyist can clearly understand. You also have some really knowledgeable people commenting and i have also learned quite a bit from what they commented on. I like how you also say "we will fix this together" Looking foreword to the next video.
Thanks for your kind words! In an ideal world I could make my videos without having to worry about the bills at the end of the month but hey :) At the same time though I want to make sure that the videos remain enjoyable, so negative feedback is also welcome! Thanks for your kind words!
Super.....I have heard that sound cards are notoriously difficult to fix and troubleshoot....glad you got it working......Another great video...thank you....😊😊😊
I agree, mostly because when I tried in the past there was not much available in terms of schematics and datasheet. Thankfully the fault was in the Yamaha section!
I might suggest next time you design the extender is use a few more sockets and wire grounds between most of the logic signals and possibly double up or use multiple pins for the pins used for power and add capacitors to the relevant power rails (particularly 5 v) electrolytic bulk cap (careful with the negative rails) at the plug on the extended point plus some ceramics close to the ISA slot. You may have been getting ground bouncing or ringing signals with the video card which likely has rather high frequency signalling
Absolutely - next design will have a separate loom for power and the caps for sure. I'll do some tests on the current setup in case that's all it takes! Thanks for that!
For your extender card, most bios's have setting for bus speed, wait states etc. Using these may get you over the line with your extender. Also some supply rail bypass caps would be advisable.
it didn't make it in the video but I did try on the 386 board which has some settings - nothing worked unfortunately. But thanks for that. I will definitely add bypass and bulk caps as a test ASAP!
Great work, Tony!
Thank you!
I'm surprised the riser didn't work, because PCIe risers work. I think someone was saying the wire/traces need to be the same length to keep timings, ie, maybe put the ribbons side by side rather than over each other.
I can trim the ribbon more accurately but I think it's ending up being the same length. Side by Side might be a good idea - I ended up with this design to keep the footprint as small as possible. I am also told PCI-ex is "designed" to be extended, that's why you can buy a riser for £5! :) Anyways, thanks for your input, I'll consider the side by side option for sure!
Your videos are really improving and becoming more helpful. Keep up the good work!
Thank you!
Maybe shield the ribbon cables with aluminium foil?
Good old TEA2025. On one of my opti soundcards, it has a hole in it (was like that when i bought it).
No problem, just don't use amp, you also don't amplyfy noise!
I tried that. It didn't work. I still need to troubleshoot that :)
mega repair, Tony, it can be seen how much you enjoy doing this sort of repairs. also noticed that you're not something of a dancing man, but you managed :))
ahahahaha - Zero at dancing for sure! :D
Regarding the extender I suspect signal integrity could be the problem. I would do an extender using a better flat cable like the ones used on scsi and pass each isa signal interlaced with ground wires (signal1, gnd, signal 2, gnd, ...) to make it like an rf transmission line. Regards.
Yes, that is an option for sure. Can you tell me more about the SCSI cables?
Awesome repair. Great work! Great to know in case I run across one in the future. Thanks!
Thank you!
Nice work and even better presentation.
Hey thank you so much!!
ISA bus has very poor distribution of ground among all of the pins. So I would use bigger ribbon cables and put more grounds in there. Also double/triple/quadruple up the 5V etc power pins on the ribbon to counter voltage drop.
You could test the voltage drop theory by measuring it on the video card, it would actually be quite interesting to see what is like. 😅
As others have suggested it would probably pay to add some bulk capacitors on the remote card to help provide a lower impedance source of current at the end of the ribbons.
Absolutely, I look forward to scope those lines and see what's happening there! For the ground, I'm considering using 80 conductors ribbons - we'll see how it goes after I upgrade the power lines! Thanks for that!
Cool Vibe, Thanks
Thank you!
There is a thread on vogons about extending the ISA bus is a similar way to add more ISA slots using a backplane. You might want to try adding capacitors across your power rails to ground on the back of the ISA slot and using 80 conductor cables instead of the standard 40 conductor IDE cables.
Absolutely. In fact, I started modifying the design for the 80 conductor cable and then realised I had to re-do everything (ground has to be on specific pins etc) and... I'm still very slow with PCB CAD software so I crossed my fingers for this design. Next one will be better :) Thank you!
Excellent :) 👌
Thank you! Cheers!
Thank you Tony! What a great video - I learned a lot! I am sure this video will be helpful for others in the future!
Amazing thanks! Glad you enjoyed it!
Signals with fast rise times, like digital bus signals, do not like going long distances. What is defined as "long" changes with the rise time of the signal. Essentially, the faster the rise time, the higher the frequency content of the signal, and the more you need to make RF-type considerations in your design. Impedance mismatches between the signal source, the transmission line (your ribbon cable/pcb), and the signal receiver can create signal reflections jeapordizing signal integrity.
As others have mentioned as well, crosstalk can be an issue here. That also gets worse with higher frequency content. If you add a GND wire in between all the data signal wires in your ribbon cable, crosstalk should reduce significantly if present. It doubles the conductor count of your cable but is the ideal connector configuration. The way I think about it is, if you imagine an electric field surrounding your signal wire, you want that field to "hit" a voltage reference (either your positive rail, negative rail, or GND) before it hits another signal's field.
I dont think adding capacitance on the power rail is necessary, as the ISA card itself will have any decoupling caps on it. Adding some wouldn't necessarily hurt, but I don't think that's your problem.
Another issue is parallel busses have strict timing relationship requirements between all the data and clock lines. Again considering those pesky RF transmission line effects, length mismatches between the bus signals can cause timing mismatches that violate those timing requirements, causing undefined behavior. Like the other issues, this gets more intricate the higher the frequency content.
Hope I gave you a couple things to look at for troubleshooting! If your scope has the bandwidth for it, you can check for these issues by checking the shape of the ISA bus signals at the card end of the extender. Crosstalk and reflections and whatnot will turn clean square waves into ones full of ringing and random level changes.
Yes, these are very good pointers indeed! I will start with the basics (thicker power lines, that ribbon is CHEAP!) and then inspect the data lines. My scope can do 200MHz so I should be able (if I am capable of!) of spotting issues there. I will also consider going with 80 conductors cables (or just simply adding another ribbon and make it signal-ground-signal-ground as you suggest, this is great as I have plenty of those connectors!) and consider the length of everything in the design.
It's going to be fun! Thanks for taking the time to share your opinion, it really helps!
Fantastic video. Smiling from ear to ear hearing those classic sounds coming out of that board.
That ISA riser is really interesting too.
Good up the good work.
You're welcome - and thank you!
Nice one tony, well done mate. Bob
Cheers Bob!
Nice repair! I'm glad you rescued the card. Regarding your riser I think I have an idea but I need to take some measurements first.
Look forward to your input! I haven't measured anything on that setup yet. I'd be curious to see too.
One idea that would work for some MB's would be a card that plugs into the ISA slot, then an ISA board can plug in at 90 degrees, so horizontally. No ribbon cable capacitance. You could have connectors on either side of the pcb, to support plugging the test board in which ever way works best. Not as flexible as the ribbon cable one, but better signal integrity on the ISA bus.
Yes, that is an option too of course! Thank you!
Try to check the negative voltages on your riser. As far as I know VGA and serial ports need -12V which could possibly be a source of some problem. (?)
Sure, I'll check all the main voltages, clocks etc as soon as I have a moment. Thanks for the idea!
Usually on ISA extensions you have a 10uF capacitor on the 12v line after each ISA slot. Even motherboards have a cap like this. Your extender doesn't have that so it's possible that there's voltage drop or too much ripple on the card that's on the extender. Some cards might be tolerant of that but not others.
I'll poke around with the scope, I am very curious. I can add caps and also extra wires for 5/12V to see if it helps. Thank you!
I suspect the problem with the extender is primarily due to power - the ribbon cable cannot provide the power needed for the videop bord. Try adding heavy duty power cables and some capacitance on each power rail at the test board end - 100uF electrolytic + 100nF ceramic.
Yes, I plan to try that for sure. Rev 2 will have a separate connection for power and decoupling caps. But I'm curious to see if this works with the extra power cables - this ribbon cable is VERY cheap, it's got 5 strands inside each wire....... Thanks for that!
Yes I don't think that will cope with the high frequency spikes from a video board.
@@BogTheWombat My multimeter didn't see anything wrong but it's just a multimeter. Maybe the video card is causing a momentary voltage drop when it tries to initialise.
@@tony359 A meter won't do it - you will need a scope with a decent bandwidth, I suspect you will see large spikes on the supply rails in the microsecond or sub-microsecond level that is throwing the logic into all sorts of invalid states
Very nice vid Tony, perfect example of using logic to get to the root problem. Made it look easy, 10/10
Thank you so much!
Great repair!
I wonder if the voltage regulator running at that high of a temperature over a period of time may have "cooked" the op amp chip. No telling if the card was left installed in a PC for any length of time after the amplifier IC shorted. I could easily see someone simply unplug speakers from the output of the sound card after it initially failed and continue to use the PC as with no speakers connected there was no annoying audio present.
Absolutely - though there was no sign of overheating under the regulator so maybe the card was lucky this time. Who knows. Thanks for watching!
Nice fix! I made an ISA extension on my channel. It has some weird behavior too. Some cards work fine, others have issues. I haven’t done too much diagnosis, but it feels like it could probably use some caps to smooth out the voltages.
Ah cool, can you send me the video link?
Well done on making your first board, defo like to see a follow up trouble shooting video, good luck!
I'm very curious myself, thank you!
Interessantissimo come sempre! Ho anche apprezzato molto la citazione dell'8 bit dance party di Adrian's digital basement.. 😁
Ecco come si chiamava! :) Grazie!
About ISA Riser. it has to do with 16 bit ISA signal integrity!!
Yes, it's a strong possibility! I'll do some tests as soon as I have a bit of time! Thanks!
If in doubt, why not try an ISA riser card you could borrow from another system in the meantime?
I don't have one and they're not super-cheap online for just a test. And there are only straight risers, I would like a 90 degrees riser - which would not allow me to probe the card at the back.
I think I'll play with the scope a bit, hopefully I'll see what's wrong.
Loved to hear your version if the 8-bit dance party!!
eheh! :)
Very pedagogic content Tony! Thank you and keep up the great videos!
Thanks for your kind words!
Amazing ISA tester ! Will wait for your rev. 2 to get one. Excellent idea Tony. And thanks for sharing your project ;-)
You're more than welcome - I didn't mention in the video that of course the project will be free for everybody to make. I'll share it... when it works :)
It will be good idea to add separate PSU connector for 12V and 5V on adapter output slot board, maybe voltage drop on those cables is that high that some circuits stop working on video card ☺
Yes absolutely. I am going to try on this revision but next one will have power on a separate cable. That said, my multimeter doesn't measure anything nasty but who knows, maybe the voltage drops momentarily when the board tries to initialise. Thanks for that!
To avoid signal cross talk via Ribbon cables you should use Double the leads, "instead of 40 have 80" Double as you will need an earth on either
side of every signal lead, to act as a shield from each other and this is especially important if the signals are pretty low or worse really fast
switching ergo the leads are "Close to behaving as if an AC signal" is being fed through, Speaking of which.. it may also be critical that the leads
are all the same length "Due to Video Timing Requisites" and they also must be as short as possible for numerous reasons
Yeah.. "Too many theoretical reasons to detail here".
Yes, next revision will use 80 conductors "IDE" cables I guess. I went to modify the design before manufacturing this one but I realised I would need to re-do everything as ground would need to be on specific pins... I'm still slow with PCB CAD software so I gave up and crossed my fingers :) Next one I'll definitely do that! Thank you!
A much smaller 90º angle adapter, can only be used in the lowest ISA slot and can't probe the back of the card, but could it be a better solution? Super nice job on fixing the card.
Yes, maybe I am overthinking that. Though I couldn't find anything online and I ended up with this which gives you a little more flexibility - but doesn't work! :D
I might make a simple 90 degrees one as you say, as a "backup riser". Thank you!
Molto bravo, per quanto riguarda l' extender i moderni pcie riser's hanno condensatori e connettori per fornire voltaggi stabili alla periferica che viene connessa, varrebbe la pena provare ad aggiungere anche solo i condensatori per esempio 220nF sulle power rails 5v e 12v
Ti ringrazio - Si', e' il mio prossimo passo. Saldare dei cavi per l'alimentazione e aggiungere condensatori dove c'e' il connettore ISA. Spero sia tutto quello che serva anche se un mio follower ha trovato le specifiche ISA per me e ha scoperto che consentono l'estensione del Bus al massimo per 2.5 inches quindi il mio adattatore e' fuori specifiche. Ma magari per testare la roba funziona lo stesso. Vedremo! Grazie!
use 64 and 40 way ribbon cables splitting your signalling to as it is laid out in ISA slot, not unpairing them as you did - yes, there is no diffrential pair in ISA, but still that bus pinout is not there by chance, but througly thinked over. your pinout on 64 way ribbon should be like pin 1 - A1, pin 2 - B1, not like you did right now with one ribbon with all A and C pins, and the other with B and D pins
Probably a good idea, thanks!
You could try lowering the ISA Bus speed. It might help make cards work better. This is an option in some BIOSes.
I tried that on the 386 board but that didn’t work unfortunately/ thanks for the idea!
Nice repair.
For the ISA riser, I would prefer a video on how to get it to work with the video card (if possible). Could be a useful tool for the retro computing community.
Greetings from Germany.
Thank you! I'll experiment and might document that in a video for sure! I didn't mention in the end but it would have been incredibly difficult to troubleshoot that issue ON the motherboard... not to mention with a camera in the way. So yes, I want that to work with video cards too!
Thank you!
IDE is slow than videocard.
Your problen is the cable.. Signal integrity.
Probably. But isn't the bus running at the same speed?
Dude, you can't use a socket with DIP audio amp chips. They use the PC board as a heat sink, and the socket blocks the heat flow.
Yes thank you so much, others have also mentioned. I still have the SB, I’ll remove it!
Think if you design a isa card that was the normal height of a isa card so the card your testing would be above the rest of the cards would work better because you could space out the rails to reduce interference and have places to test the bus eaisier.
I thought of that but 1. it would need the motherboard in the case so the riser can be screwed on something 2. it would still be wobbly and difficult to probe. I wanted something that made the card rest on a rigid surface - I'll eventually get to a working one! :)
Thanks for watching!
I had an issue with a long USB cable and it turns out the problem was delivering power rather than signal degradation. Maybe explains why the VGA card doesn't work because is more power hungry than the other cards.
I checked with my multimeter, there was a little drop but nothing terrible - though I should check with the scope. Maybe there is a drop when the board is trying to initialise. That ribbon cable is awful, the gauge is miniscule. I shall try jumping 5V and 12V via extra wires to see if that is the issue - if that is the problem I can either replace the ribbon with something better or make a Rev2 with a dedicated connector for power. Thanks for that!
Very nice trouble shooting and repairing video.
Thank you!
Hi Tony I quite like your board extender solution, I do have a riser card but I think your idea has it's place.
Thanks! I went to buy one first but I don't like the "card" risers because you end up with an even more unstable solution. Also, I think they're only available in 8 bit format. The 90 degree ones are ok - but I can't find them for sale.
Hence my partially working design! Rev 2 will be cool, I promise :)
20:04
You don't need to ground yourself because you're not working with CMOS circuits from the 70s.
It doesn't make sense for you to use that.
Not everything has to make sense :)
A VGA card needs to talk much faster with the CPU and memory via DMA. The higher bandwidth needed is just too high to reliably travel through your test rig. Nevertheless great video. Like it.
Thank you! I thought the speed was dictated by the ISA bus. I guess there might be many ‘modes’ to use that but, I don’t know. I can’t wait to troubleshoot that! Thank you!
@@tony359 I'm interested what your findings are. Max speed is indeed dictated by ISA bus. But cards not always use the max speed. For example a sound card needs to do some I/O read/writes in the tempo of the music to the FM chip. Digital audio is 44,1 kHz. But VGA has to update video memory via DMA transfers for 640x480x18bits.
I wonder if the rig not working with VGA cards has something to do with the fact that the VGA card has BIOS extentions and the IDE controller and sound card do not. So the PC has to run code from the VGA card bios but not so with the other cards...
I need to look into that, I hope I can "see" the problem with the scope!
Hi Tony, Nice repair. Keep up the good job. Greetings from Steven from the Netherlands
Hello Steven - thank you!
Молорик мужик. Люблю такое видео.
Some video cards overclock the ISA bus when transfering data, and may also draw enough power that the cable causes too much voltage drop.
The ISA clock is set by the BIOS, I'm not sure that a video card can alter that frequency? On power, I totally agree and I hope it's my problem!
@@tony359AFAIK the the card only needs to send extra pulses down the clock line when no other device uses the bus. OR, so I have been told by someone who toyed more with the ISA stuff... my history with PC is short enough that my first daily driver PC had a PCI video card already, and the next one AGP. The ISA gear I have toyed with later on was ancient already at that point and I wasn't studying them very deeply.
It might be some special mode the card uses - the term "overclock" is probably what puzzled me as the ISA clock cannot be changed by a card. But whether the card is doing some wizardry to speed up, that's very possible indeed!
Yeah, opamps, dacs and power amp chips are all common failures on any board that makes sound. I've replaced hundreds of those types of chips when repairing arcade games over the last 30 years. If you use a logic probe with a built-in piezo speaker you can actually 'hear' the analog audio when probing the chips.
Interesting! It could be a cool tool to have!
@@tony359 well I've repaired 1000's of arcade games over ~30 years and I actually only ever used a logic probe for that. I only use the scope on analog audio because a logic probe can't detect it, usually just shows as a high. In 30 years I think I used the scope maybe 10 times and half of those it didn't help heh!
I'd imagine a logic probe is a very handy tool for such machines!
Great analysis!
Thank you!
👏👏👏 another excellent repair
Thank you! Cheers!
AWEsome!
ahahah not quite AWE yet :)
Thanks!
Thank you!
Thank you for watching!
Ran into someone extending cables via ribbon who swears that tin-foil( covered afterward with something non-conductive) works as a shielding. IDK, try it?
Thanks - I tried that already unfortunately. But thanks for the idea!
@@tony359 several billion heads are better than only one :)
TEA2025B was a very popular amplifier IC in the late 80' and early 90' i think.
It had several equivalents ( like KA2206 for example ), and it was widely used in those "boombox" players back then ( either this one or TDA810 and it's equivalents, which was a 3W amplifier, thus a bit more powerful ).
Anyway, it's still your basic class AB amplifier with a very basic schematic that used only one power rail, with bootstrapping VAS and push-pull output stage with an output capacitor to remove DC bias ( which was half of the power supply ).
Now, when it fails it usually does so in the output, can't remember of the top of my head if it had interna SOA protection, but usually it happens either from too much load ( IE paralleled speakers ), or accidental shorts on the outputs, that shorts the output transistors in the IC and you get a shorted power line.
There is a very basic test you can do ( very fast too ) to determine if an amplifier IC is shorted or not.
You just use the multimeter in the diode function, and you test the outputs with respect to the power rail and ground.
In both ways you should either find diodes or nothing at all ( open circuit ), if the IC is fine.
If the IC is shorted then you will measure shorts between one or both the outputs and one or both power pins.
Sounds complicated, i know, but really is just basic stuff, and it works the same not only with audio power amplifier IC, but with any type of IC that has outputs ( gate driver IC's, opamps, switch mode power supply controllers...).
No it's not complicated at all!
Another way is to power it up and burn your finger on it! :D
@@tony359 Indeed 😁
Good call on the fault finding by the way!
Both on the amplifier IC and the opamp.
... its equivalents ... (the possessive pronoun has no apostrophe!)
@@SpeccyMan yes, indeed.
Sorry for that, i am not a native English speaker so i do mess up sometimes.
Sorry and thank you for correcting me.
ONE SUGGESTION: add one SMD LED (micro size)
YES, so I don't plug anything WHILE it's powered. I have that green LED on my AT PSU plugs for that very reason! Thanks for that!
@@tony359 yes
Again, glad you show proper diagnostics instead of just swapping ic's and crossing your fingers you dont fry an expensive good IC in the process.
Really enjoy the deep dive diagnostics videos like this.
What pcb designer software did you use?
Personally I exclusively use Kicad as I find it easiest to work with.
I used EasyEDA but I think I'll switch to KiCAD now. Thanks for your kind words. Yes, I could have swapped parts at random but it's so cool to learn something in the process! Happy you liked the troubleshooting!
@@tony359 just try a couple of different types of software and stick to the one that feels the most intuitive to work with.
@@samuraidriver4x4 I was recommended EasyEDA but every time I ask for advice online everybody seems to be using KiCAD! :) Since I have to learn more or less from Scratch, maybe better stick with the most popular. I'll take a look thanks!
enjoyable. good work.
Many thanks!
Quality video! Subscribed.
Thank you and welcome!
Nice video one of your best! Well done :)
Thank you for your kind words!
Nice repair Tony, I like the way you say FM, sounds like effing LOL and CHIPS sounds like sh*ts hahaha I love the diversity of smarter people (dual language) accents :)
Awesome video bro. Congrats. Learning a lot from you.
Awesome! Thank you!
I am but a kid in mind! hahaha it sounded like a fart and I couldn't help it but laugh!
Laughter is good! :)
Maybe look at how PCIe riser cables are constructed, and perhaps there's some insights in their design that could help with your cables.5
I was told that PCI Express is somehow designed to be extended. It makes sense to take a look - though I'm not sure the solution there will apply to ISA. But it won't hurt to check. Thank you!
@@tony359 I remember that being the marketing speak for PCIe; remember all the claims about hot-swappable GPU modules and all that?
Hot Swappable GPUs :)
Great video!
Thanks!
Is it receiving proper voltage through the riser? Maybe an external connector for power? Great Video!!
There is a small drop - I definitely need to route power via a dedicated cable and that ribbon I got from Aliexpress is very cheap so the gauge is too small.
That said, I don't think that's the problem, with a good ATX PSU and some load to balance the rails I get something like 4.9V if memory serves. That video card works with 4.8V when plugged directly in the slot. Thank you!
3:44 Look, just because you don't like dubstep...😂
You mean the title saying "the wheel has no effect"? Well, sometimes I add voiceovers, in this case it was a simple thing and I only have one microphone which is normally in my workshop :D
(aka I felt lazy!)
Sorry if my comment offended you 🙏
I really like your videos :)
Which comment?
@@tony359 4 days ago. But i can re-post it here:
Maybe try to add som capacitors to the powlines on your risercard with the ISA socket on.
Something like this backplane card has:
i.ebayimg.com/images/g/JIYAAOSw0vdeQXDV/s-l1600.jpg
There is also something called ISA termination resistors, that should be installed and matched according to your cable impedance. For at cable that long, I belive you should only be able to get a stable result with proper matched termination resistors
I don't remember all the comments and I don't seem to be able to find your earlier one - anyways why would I feel offended? :)
Your suggestion is sound about capacitors - what kind of resistors would you add to terminate the line?
Replace the OPL3?
I hope you watched the video till the end :)
@@tony359 I did! And I meant to say the DAC... Because you mentioned that it's one of the components that got scorched.
Ah gotcha! Well if you start replacing components around the regulator you never stop. I guess it should be used for a while to decide. It might be less reliable than the average, who knows!
1:02 that's not a real DSP. It's a simple 8051 microcontroller with internal ROM.... note the "(C) intel" on the chip....
Indeed - but I think it's called DSP on the documentation. Someone told me those are rebadged ICs, I had no idea! :) Thanks!
@@tony359 it's marketing speak or they were just told to say it's a dsp by engineering. A real dsp is something totally different much more expensive and powerful.
@@g4z-kb7ct If audio is being processed digitally, then surely the device doing it will be a DSP, even if it isn't specifically designed for the purpose.
@@SpeccyMan You clearly don't know what a DSP is. That was just marketing spiele. That chip is not processing audio digitally. It's a CPU with internal ROM and RAM and built in timers and other peripherals. It's the brains of the system controlling the board. Look up the datasheet for the actual chip being used... intel 8051.
your 'riser board' is smearing the signals and rounding off the nice square pulses, the vga card you have doesnt like it, another brand may be ok.
I'll check with the scope - any ideas how to fix that? Thank you!