Hi. Unfortunately, there's been a tragedy in the eevblog lab. the arcade machines on the blink all that's coming out of it is just like this low level mains hum I Don't think that's ever been there before, but usually it was going. you know due to the 8-bit sound and just silence.

Huxley was playing that today and no she's gone on the blink so tear down time now. It actually does still work, but it appears to be intermittent because it's initializing at the moment. The monitor is obviously working fine and something to do with the well, the Gemma board, the power supply something like that and that time that how am I can here before I think it's gone now. maybe it's only when it goes on the fritz.

Anyway, there we go. we're in like Flynn Um, so let me play it for a few minutes and well there we go. There we go. It just went.

You saw it. check video cable. there it is just went. kaput.

I can hear the high-frequency hammer. so yeah, I was playing that for like five minutes tops. Turns out the hum is actually coming from the speaker I Thought it was coming from the power supply but you can clearly hear that. Yeah, I reckon the power supply is gone.

Kaput on that. It just pulls out like that. Bob's your uncle. we're in like Flynn and that's the I've done a separate video on that.

This is where I got the video feed out of this thing. the VGA video out, convert it to HDMI and power it from an external 5 volts here. So VGA to HDMI our converter which then I then capture. so I can capture game footage and stuff.

Alright, so take you through a guided tour of this. We got ourselves a mains power supply over here. that's just 5 volts, 15 amps as you can see I don't think it needs the 15 years, it's just powering a gem aboard. Then we've got our jammer board over there I'll show you that in detail in a minute and then that pipe there is just for the coins.

If you do want to use coins, this one actually I've got it actually bypass but there we go. You can see that for those playing along at home. I'm sure the arcade aficionados have tell me what one that is I don't know but I have used coins in it before of it. Anyway, it is disabled and up here and there.

the joysticks and the arcade buttons. Once again, the arcade aficionados might be able to tell us the brand and the model there. I don't know I didn't specify this I just had someone make it for me down in Wollongong soda and don't know if they're around anymore. but see the micro switches on the bottom big and beefy? Look at those and then there's just a monitor and that's just a HP LCD monitor of course, compatible with the well, there's a 15 kilohertz.

All these 30 kilohertz? isn't there two different frequencies? I Think this one might be set to the 30 kilo. Hertz Mobis a lot of LCDs I Believe can't handle the old-school 15 kilohertz. It's not much in it. so I think the monitors fine.

So something's gone kaput in here and my money would be on the power supply there. First thing we're gonna do of course is measure that 5 volt rail there and nope. Well, there's your problem. 1.4 7 volts.

Ok I won't even bother looking at the jammer board. Alright, so one of these one Hungary and power supplies before we do that, I just want to measure that again. You can see that our lights come on now. so L5 out really good and sure enough, there we go.

5.18 They've adjusted it a bit high so there's a bit of drop and the jammer boards working I can see it flash and it's LEDs and stuff like that. So yeah, everything's hunky-dory so it looks like we have like a dropout after about five minutes. That's interesting. Hmm.

Confidence is not high I Repeat Confidence is not high. By the way, there was an extra hour 12 volts supply on that thing as well. Yeah, One hundo? No. NEMA Yeah, this case doesn't instill a lot of confidence in me.

All right. Is it just gonna be something lame like a cap? Or is it gonna be something more interesting? Okay, there we go. We're in like Flynn dodgy looking label on that cap. For starters, it was right that this didn't instill a lot of confidence.

This looks really old school. Just the switch on the the right angle switch there me wrong there. Rubicon's there all right, No workers. it's just something else.

the output cap so there's no bulges in them. Yeah, they're uh. I don't know what brand that is Yulie or something? I don't know. No, that Loius odd.

you'll be there. You can usually tell where the green and the yellow. Yeah, certainly the design and layout leaves a bit to be desired. At night it looks like something that was slept together in the 80s or something.

I Wonder if we could actually get a date code on this thing, but we got our two primary side driver trainees up there. Look at that. they're all the way over there. That's interesting, just totally old-school design and layout and build quality.

Well anyway, given that it's five minutes use, that means you know something could be heating up. First choice wouldn't necessarily be the Cape's because I they don't look like they're bulging at all and these are Rubicon so they should be okay. not sure what one that is actually. I can't read that gold writing.

It's really a girl. gold on green. It's really quite knowing you've got to get that in the right light. But I wouldn't rule out a dodgy joint somewhere.

We've had those before. like a dry joint. You know, when the thing starts heating up after five minutes and or something else. You know we've had like diode bridges or repairs I've done in the past fail after I did an oscilloscope repair I have to leak that one in if I think of it, that one was really interested or spoiler alert.

Yeah, it's the best part about their video was tracking that down. Um, anyway, it's an interesting video I'll link that one in. but yeah, I wouldn't rule it out myself. tap is going into them into the metal.

That's terrible Muriel Good thing is it's all single sided so you can have a good look at other joints. Pass it to the left. Oh Yep, there's a definitely a good chance I've called that one. Look at the output connections over here.

They look dry as a dead dingoes. Donna These are the five volt and ground outputs Wow dry as a dead dingoes Donna Unbelievable and 12 volts go over. Oh, this is just shocking Wow Who soldered that? Stevie Wonder And check out the main driver chip over here. It's almost as this.

Somebody's had to go with that. like everything else has waves soldered and that has been Harry Hacked. That is terrible Muriel Wave soldered so it looks like this is like a repaired board or something like that. Even though I don't know, the driver trainee up there looks alright.

But yeah, get down to there. you can see the flux residue there. okay ish. but I'd go in there and resold of those as a matter of course.

Is that actually a track that's been cut? Nope. Nope. it's just some insert, just a dead bug. Clean all that up at.

Oh geez, you know yeah I would suspect that as the first port of call because what happens is you know these are carrying over I don't know what the drawer of that board is that have to look. but but let's say it's drawing like five watts or something like that. You know it's drawing an amp or something like that. so these can.

These joints can actually heat up and just go open as yeah after like five minutes or something like that. But why it's only happened now? I You know it's one of these marginal things. so first thing I do is simply re solder all those and then put it back in. Pair it up and play it for like an hour or something like that.

And if we don't get any dropouts then yeah. I think we're solved it okay. I'm just curious to see what happens if we just heat up that joint there. so let's just give her a ball, see if she says she crumbles or not.

I Go on this side here so we can hopefully see it on camera. Oh look at that. Yeah, that's that's not good. The fumes.

oh you have to suck it all out. Don't reuse this salt up. suck it all out. Put on fresh stuff.

Now we're talking. Look at that. Those little notches in there. They're actually are taken out of the copper PCB pad.

So yeah, that's not me. It's physically taken a notch out on each one of those. There you go like a boy. These self tapping screws into the metal work.

Awful. really. I could just rip that extra quarter-turn she's Gonski Yeah, it's got to be one of the most how you doing power support like brick power supplies. I've seen just the lack of quality in the construction.

I Was surprised to find Rubicon caps in there actually. Well, it turns out when you make assumptions like that, you can come a gutter um and plug it in and it got a brief flush there, but not nothing. So it's it could be a combination. I mean those tweets what dries the dead dingos dong and say I had to redo those.

They were dodgy as something else. Mike I didn't screw up back in. That was the whole idea, right? So if we go back in here and have a squiz I've actually you know, found a few more our suspect joints and things like that and I've had a little lob ash at them cleaning them up and still no I get exactly the same result. So I've gone around with my LCR meter in impedance mode I've measured some of the cat like I've mentioned, well, most of the caps, almost all of them.

and if they're all okayed like they're down in the you know, the expected hundred kilohertz type value should expect for these caps. So if it was, say like a cap on the secondary side for example, then like these ones here might only be for the five volts, whereas this one here is probably for like the twelve volt output here. So one of these cats went on the five volt rail, You'd expect the 12 volt one to still be okay. Alright, powers on.

Actually, the lid is kind of on its measure. Note 2.3 No, that's no good. This is the this is the twelve yeah for honey and five volts on the 12 volt rail and this is minus twelve 1.3 So it's most likely to be something primary side related rather than an individual thing on the secondary side. So let's have a look at the bottom here.

and by the way, it's always good. Like just that with one hand, just hold on to the plug. Uh-huh just so that you don't, because if you are in this unparent at all the time trying to troubleshoot something like this, you don't accidentally forget plugged in and then you start touching the thing. That could really ruin your day anyway.

Yes, it does have our bleeder resistors across the two main caps here, so that actually that's for sharing as we'll see and also will help our bleed those capacitors as well so it's safe to touch out fairly quickly after power down. Of course you can measure that Anyway, we've got mains input over here. Yeah, input cap across here we've got and there's a varistor as well. and we've got a common mode choke.

and this is our bridge rectifier here. so it comes in to our bridge rectifier. Here's our output of our bridge of full wave mains. That bridge rectifier here goes down to one of the caps.

As you can see, they're actually split. Here's the other cap here. so they're actually in series like this and it goes back. and each one has its own current sharing and bleed resistor as well.

And this here is your 240 volt, 110 volt switch. So it's currently these two terminals are shorted at the moment, so this is just isolated. so both of these caps are in series across the 240 volt mains there. so we'll just measure that.

Okay, so let's probe across there and there and we should get 330. Yep, 330. There you go. So our full wave bridge rectifier.

It's working just fine. All right. At this point, it'd be really handy to have a schematic. Unfortunately, this being the one hung low that it is it.

Wow Hang on. I Just noticed the day code 84. What? Yeah, that explains a lot. I I'm debating whether or not I should - repair this, but I'm just going through it as an academic exercise.

I'm thinking about yeah. I I Think the guy who built this cabinet has just like found any old power supply yard lying around it in so this thing is ancient. Couldn't get a date code from the chip, but we did get a part number and it's a ki a four, Nine Four killed in action. That tells you a lot.

Anyway, it's obviously a clone of the Texas system as T or Four Nine Four. Absolutely classic PWM Controller which doesn't necessarily have to be used with a main switch mode like this. It's just a pretty much our generic up here. WM Controller: you can use it for, you know, power supplies like mains supplies.

like there's other DC to DC converters or other applications. It's like the 494 S, just it's been around forever. It's used in tons of different things anyway, so it's a clone of that. So I couldn't get a schematic.

So what I found is this one from Eric Taylor Thank you very much. There's the website address there, so if you don't have the exact schematic the next best thing? Rather than stabbing around in darkness to at least know what the topology is going on here, mains comes in here full wave our bridge rectified into these caps here. We've got 2 switching down here as we'll take a look at. We've got some isolation transformers here and here and of course this is our main power transformer that's going from the primary over to the secondary side.

This is very similar to the topology I believe to what's going on in here. Basically they're 240 volts in full wave bridge rectifier. We get our you know 230 volts DC across here. Then it goes into two switching power like this and if you actually flip it over and have a look you'll see it is exactly the same.

Follow the money. Okay call like this is. just say it's the topside here I Don't know whether or not is it goes into one side of the power training. the other side of the power that comes out goes to the other switching transistor down here so that there it is in the in there like that and then the other side of that goes back right around to today, the negative or the other side of the mains DC input.

So those two power transistors I swear simply alternate switching. And of course they've got to be isolated of course from the drive side here. So this is what this isolation or this gate drive transformer as it's called does here. Now why they've got the two there? There might be a little like a little last secondary thing that's powering that's tapping off some voltages, but I reckoned that one there puppy.

If you have a look, that configuration seems to have a dual winding in there that seems to be employing one whine in two windings like that. So that's equivalent to one gate drive and the second gate drive. and I'm sure you actually traced all that out and follow the money. Then of course it's not just a direct connection like that.

There's other you know stuff in there. there's diodes and resistors and various whatnots. but basically all that's doing is driving the two switching transistors there. of course in the middle here goes to this is our main power transformer over here.

it's going to call it a power, but people have complained that I call both transistors and transformers and well, that's true. They both have the you know going back since before I was born called. so yep, it's just going to be the primary there and the secondary over here. They'll have multiple taps of course for different voltages and that's about all she wrote.

Yeah, there's no reason why we can't get in there. not with our regular. Crowe Pro I've done a video of how not to blow up your acro probe Crowe Catherine Oscilloscope here in Australia you CRO probe. As always, I'll link that in so we'll use our high voltage probe and we'll be able to see if these little puppies are switching.

They're obviously doing something because we did actually measure the outputs at you know, a volt, 2 volts things like that. so something sort of sneaking through. so there you go. I use my high voltage differential probe I'll probe directly across the primary side of the transformer there see if we get a switching waveform.

Alright, power on we are. Switch in. There you go. That's a hundred volts per division getting something there? Yeah, switching frequency there about 27 kilohertz? That sounds about right.

But the problem is we're only 20 volts per division. I Do have my probes set up correctly to one hundred to one. So yeah, that seems a bit low 20 volts per division. Although having said that, if you'll notice that this particular schematic here has a low ratio transformer 1 to 2.4 with this AC series cap in here.

And of course we have the AC series cap in here like this: It's one side of a transformer. Here, there's only a there's only one coil on the primary, and there's the AC cap going back over and it jumps right back over to the center point of those two caps. Exactly. So this is practically an identical circuit to what we've got here traced out.

except in this particular case. We've got a 1 micro farad film cap in here and yes, I have actually measured it. and yes, it is bang on one mic so nothing wrong with that cap. So with a small turns ratio transformer, they're kind of the levels that would expect for like 5 volts and 12 and minus 12 volts output.

So I don't know about the waveform though. Mm-hmm Okay, just for kicks. What we'll do is we'll our probe at the secondary side, some on one side of the output diode. Ease there that is, that's not a transistor, that's actually a dual diode package.

So I'm probing one of those. We can actually go back to using our scope, probe and the common ground because if you have a look, earth is connected to this, which is connected around to your ground because it's mains earth referenced output because we're measuring the output. Nothing on the primary side though, because you'll blow your scope up and yourself. probably, but on the secondary side.

Completely safe to do so. Peel that up and prove it, and well, there's your problem. Yeah, that's one sick puppy. We're getting some volts on the output, but we get in like just little lousy bursts.

High frequency burst. not what we actually want to have any decent current output. That's why we get in like some voltage on the secondary. But yeah, obviously as a primary side drive problem.

and I'm probing the diode output, it's actually on the choke there of the five volt output. And what one? white wire? Well, yeah, there's your problem. That's so without in getting that on both the five volt output and the twelve volt output, that so that shows it's not the secondary. It is definitely a primary side drive problem.

Okay, so what I'm doing now is just probe in pin Eleven, which is one of the drive outputs. There's two pin 8 and Pin 11 and there you go. There's the 27 Kilohertz drive signal coming from the PWM controllers capture that and there you go that seems to be doing its well. it's doing something anyway.

and there's the other drive signal there. So well it's yeah. it's playing up yeah, like that should be consistent and it's not. It's really something wrong there that should be like a consistent output.

Yes, it is the 27 Kilohertz if you actually measure it. But yeah, there's all sorts of weirdness happen in there, so you know it could be anything like could be a sense, You know, the feedback, the voltage reference, or anything like causing an issue. But it does seem at least to be the PWM chip seems to at least be working because it's outputting something right. So what I want to do is I Just want to go back to why: measuring that primary side transformer.

now that we've got more information on the secondary. Well, the actual drive side of that, the PWM control side of that. and there's our signal again. But if we actually run that, you can see it's just going.

Silly buggers. I mean it's just all over the shop. So yes, if you go, Oh yeah, okay, that's a bit repetitive there, but then you've got big periods in here where it skips and everything and I thought that this chippy it didn't skip like that. It's supposed to just be consistent and change the duty cycles.

so it's just nuts. It's just oh, look at that. look at that. what's going on there? All right, Because that output is just complete.

Silly buggers. we're gonna have a look at the supply for the driving chips. So I determined that this link over here goes over to PN at 12 and that's the VCC supply for the chip. I Think it's like oh, and this schematic here at 16 volts.

I'm not sure what the maximum is, so hello, unless that settles, he's not a great power supply is it? That could explain why we've got serious problems such 15 volts with a ton of ripple on. There may be one of the bypass caps for that rails Gonski That's pretty terrible, but there is a supply there. but yeah, you can't have that much ripple. that's oh yeah.

Look, that's gone silly. Yeah, it's gone. Silly buggers. So whoa, whoa.

Now we're getting huge transients on there. Wow that's your power supply for your control chip. There's your problem. What's causing that though? It's obviously getting and initial feedback from the primary side to to power that show up there we go.

Whoa. Look, it's just jumping. This is just oh. it's got the heebie-jeebies and even replacing the cap down there was four point seven Mike fifty volts with a hundred Mike and it tested reasonably okay.

And now it's still there. So not even like a 20-fold increase in the capacitance there can smooth that puppy out though. it's a bit better, but yeah, not magically better. And it's still doing that weird ass thing.

Seems okay at the moment, but you saw it before. Yeah, there there it goes. It's just it's it's playing silly buggers again. This thing's just going.

Ah no no, no, no, no, no. Okay, now from what I can actually I Gather here is that this little secondary transformer in here is actually in series with the primary side power switching transformer. So that's just tapping that off. and that's just all that sort of stuff.

You can actually see it coming over. This is actually a center tapped. This is the main primary side. one centre tap.

it jumps over there. There's the primary side for the secondary transformer. They were auxilary transform or caller. There's the secondary side tap and that's got all those all those components in though.

all those resistors and die. it's there. So that looks like is that like a full bridge? That could be a bridge configuration there and that's going in, but that's not actually powering the rest of the circuit. What? that's that's just sensing.

The power is actually tapped off from up here on the gate driver transformer. Here, there's actually on the center tap on the secondary side that actually goes directly to VCC here. if we have a look here, it is here. Here's the center tap for that.

There's that's a jumper link that goes over. follow the money. there's a resistor and that goes to pin 12 and that's our VCC there. so that's powering that.

So there's no like auxilary transformer auxiliary oscillator actually powering this rail here. So how it's actually and there's no other oscillator on the primary side here which is interesting which you know there's no like extra auxilary winding on here and an oscillator on this side which then drives which then you know is rectified and gives our power supply over here. So I'm not exactly sure how it's actually getting our main supply here and why it's was so upset and you know, just going ballistic there. So these two.

there's two primary side electrolytic here. I've actually taken those out and measured those. They were fine, but I replaced them anyway and everything's hunky-dory I've measured the diodes in here, they seemed fine and upon some further tracing, I've determined that the Phys 16 / R 15 volts I believe it is. Is that coming through a series resistor? A series diode actually coming from the 12 volt tap? which is actually you know one of the outputs down here basically.

But yeah, once again, it's there is no primary side separate oscillator to actually power this thing so it looks like it's bootstrapped in. you know, some way it's like, how much further do I have to trace this thing? It's just it's just ridiculous. I could waste hours and hours going down this rabbit hole and given that that is the waveform on the output of this cap here, even when I put in the much larger cap there on the rail, it's almost as if there's something wrong, possibly with the driver transistors in here. and it's just drawing excess current or something like that because to get that much ripple on that rail is just yeah.

there's something loading it down so something to do with the chip. perhaps even like I don't know, it could be like yeah, quite a few things. and if you go in there and actually measure the basis of the transistors in there, they all seem hunky-dory I've measured them all and they're all seem fine. So yeah, and well, I was about to abandon this, but today it now magically works.

They slam why in a second I've got it hooked up to my electronic load. They're sure on a constant current of one app and it works. An absolute treat. So what was wrong with this thing? Well, unfortunately I can't tell you the exact reason because what I did is that I went back to my original hunch that look it I think it's a solder joint at fault here.

Little micro, a crack in his solder joint because over time, especially like 35 years or something like this. even though this pathway hasn't really been stressed a lot, the thermal changes in the components every time it powers up and powers down, they metal things, they expand and contract with temperature changes and that can cause micro cracks that are almost sometimes impossible to see. So what I did is just like I was going to completely abandon this as I'll explain in a minute. there's just there was no point I wasn't going to reuse this thing because sometimes you just have to give up and say look, it's not worth it's beyond economical repair, not in terms of parts cost to fix it, but in term of you in terms of your time investment.

and I was never gonna reuse this thing anyway. so it was pretty much just an academic make exercise to actually get this thing working again. Yes, I switched it off now. So what I actually did was just went in there and I just resold it.

Any joint that so much has looked at me the wrong way and I sold it like I don't know a dozen or two joints and yeah, it's it's magically work again. So unfortunately unless I can systematically track down where that that actual fault because we did a reasonable amount of like a reverse engineering detective work there and we I did pretty much all the basic stuff you should be doing to track down the fault and still it was kind of like a weird issue that something to do with like the power supply bootstrap in when it powers up and things like that and keeping it running and and stuff like that. So yeah, not great. but anyway.

um, it's fixed itself. So anyway, it was obviously a solder joint somewhere I can't tell you if it was on the secondary side or somewhere over in the primary side because I saw I resold of the chip yet again I had already done that before by the way and another couple of other suspect joints and there were stuff in there that I resold at individual joints and then repowered it, but I didn't. In the end, I didn't want to just keep doing one joint repair it, see if that caused the problem. So yes, you know, did a few dozen.

and yeah, it's come good. So it was a micro crack it seems to be operating. you know, more than the five minutes under a decent load. and yeah, Bob's your uncle.

Um, so that's fixed I guess in quote marks, but I have absolutely no plans to reuse that thing whatsoever, so that was a real hope. You enjoyed that. That was a real interesting night. academic exercise just going through things even though we couldn't find the exact culprit.

Leave it in the comments down below if you want to have a yeah I guess it exactly what mechanism actually calls the stuff. We will see him, but before we call it quits on this and we go back and actually fix, the arcade machine will just probe a few things when it's working. So let's just start with the primary of the main switching transformer and there it is there. There you go.

Once again, it is the 27 Kilohertz or there abouts but that's the that's the proper way for me should be getting and it's completely consistent. No skip Ben No funny business and all. That's 50 volts per division for those playing along at homes. And if we probe our output here, then that VCC power supply.

It's actually a much more than fifteen volts actually. look at that, it's 5, 10, 15, 20, 25 volts. Thank you very much with Bugger-all Ripple. So whilst I can't tell you the exact mechanism at play there I Hope you found that in.

An interesting example of you know, having a suspicion upfront that it was a like a solder joint, a thermal type issue and then going down the rabbit hole and doing your basic checks and you know Murphy will get you every time in that it's not going to be something simple never is for me because well you know and then you follow it down. You start to you know, suspect your electrolytic sand other components and your diodes and your transistors and you start measuring things and everything's you know, looking just fine and then you eventually come back to it. Well you couldn't see any physical fault before and the symptoms that you will get in were like not quite easy to understand unless you reverse engineer the whole thing and spend a lot of hours at it and stuff like that. And in the end it was.

It looks like it was just a solder joint that's coming guts us so ya know you get him sometimes. Anyway, leave your comments down below and then to top it off I Never actually had that like five minutes thermal fault on the bench here. just by literally like taking it out, unscrewing it, physically, handling, and doing all that sort of stuff. It never actually showed those symptoms on the bench here.

it simply had those weird, intermittent, bursty, you know, failure mode kind of things. So yeah, it's like when you can't get a repeatable symptom like that when you take it out onto your bench to work on it. Ah, ruins your day, sucks your life away. And the reason I don't want to do that is because look, I don't have to pare it through this wiring harness here.

It's got a PC power supply for pin Molex connector on it for the 5 volt and 12 volt power supply is required for this thing. And I've got a million dumpster PCs were perfectly good. you know, relatively modern compared to that ancient piece of crap. No, it's just not worth it.

So I'm just gonna get a PC supply. Whack it in here and Bob's your uncle. I'll get this thing back up and running. Ah bugger it.

Even a PC past was too messy. I Found this in the junk bin of power adapters. Dozens and dozens of things. 5 volts, 2 amps, 12 volts, 2 amps that'll work a tree.

So there you go. That's completed installation. I Just put a mains cable in cable clamped. I've got a double adapter in there.

the little plug pack is clamped down the bottom there and that just started. Just splice that in line with the existing power lines there. and I just replaced the bare wire mains cord going up to the monitor with just a regular one. and Bob's your uncle.

That's it. Neat and tidy, and as you'd expect, it works. An absolute treat. Winner winner chicken dinner.

And that's better than that 30 year old power supply. a head in there that was just garbage so don't know why they installed that to begin with. That's just insane. You could have just used the PC power supply.

that's what it was designed for. So yeah. nice. Anyway, what the video? Please give it a big thumbs up.

And as always, comment down below. catch you next time.