Hi uh I was just going to have a quick play around with this ryal DP 832 power supply did just to uh Trace out the output circuit and uh probe a couple of waveforms and uh, stuff like that. maybe uh, see if I could possibly find out uh what's going on? investigate that power on. uh, spike a bit further but got the guts hanging out and one of the first things I noticed was that um suddenly like I couldn't see the display but the fan would suddenly start revving up and then uh, going crazy and I found out what was happening without even probing. The thing is that the power supply would actually reset after a little bit of time and I'm going to leave it here and uh, just see if it reset.

So I switched on all the outputs there and uh, hopefully if I won't touch it cuz I think maybe you know the reset might have something to do with. you know there might be an Emi issue. the boards all hanging out here, all the wires, the you know, the some some of the digital stuff going here is all Loosey Goosey hanging out in the breeze and uh, maybe that has something to do with it. but anyway, um I'm just going to leave it here.

there's no load on it of course and I found that the damn thing reset so I'll just leave the video running and see if we can capture that. Usually it doesn't take too long, it only takes like a minute or two. No, as Murphy would have it, it's not going to reproduce the problem. the old white Coat Syndrome strikes again I've left it for a couple of minutes and annoying.

Anyway, what I was doing is I was adjusting the output voltage in steps just as a first uh thing just to see where it. um, you know, got the uh Transformer Taps and uh, stuff like that. So maybe if I turn the voltage back up or something like that. Anyway, I'll leave it for another couple of minutes I don't remember moving it or touching I Didn't move my multimeter from here out of the shot to to here.

but I don't know better. Hold my tongue at the right angle. Hang on. Woohoo! Did I get it? did I get it? I got it.

see it reset. it took bloody what? 9 minutes on the camera here. but I got it. sorry I was just in the other cubicle and heard the fan rev up.

bastard. So the thing just resets itself. Let's turn that back on, see if we can get it to to reproduce even faster. I don't know why I have no idea.

Um, I'm not probing anything I'm not touching anything. Um, the outputs are still seem to, uh, work. but I've got it under no load and uh, the thing just reset itself. there.

We go there, we go, got it? Bingo What was that? like? A minute? there we go. It just switched off and reset. Why? I Have no idea. Um, my only guess is because it's you know it.

it's hanging out here. it's not uh, in the system system so maybe an Emi issue. Um, getting in somewhere that's resetting this sucker I don't know. um, all I can do really is, uh, fold this back in, screw it back in place and uh, run it as a bench supply for you know, an hour and see if it, uh, see if it resets.

Now there's one thing I found is that uh, this regulator in here is tiny little. it's an Lm317. it's actually delivering the uh, that isolated uh 5 Vols for the main uh logic that we uh saw in the tear down and I noticed that was getting really really hot Jen look at that. I mean we're talking.

You know this isn't going to be super accurate, but look I mean it's it's not going to read over though. so you know I was getting close to 90 or something at one point. There we go 95. It's crazy.

that thing is getting super hot and I'm not sure if that's normal or not. Surely it can't be running that hot. Unbelievable. Have no idea why now I've actually uh, measured this thing and there's the 5 volts that it's actually uh, supplying and the input voltage.

If I can actually get in there correct ground. Yeah, there we go. 12 Vol Now I'm reading well over 100 on that now. And I can should be able to get in there with the thermocouple and even confirm that.

Look at that. That is ridiculous. 110 on the heat sink and this thing's only been on for like 10 115 120. Unbelievable.

No wonder something. that regulator is just going to shut down. and no wonder the thing's going to reset. So clearly something is, uh, drastically wrong there because you know, like an Lm317 is only got an operational temperature range to like 120 125 or thereabouts.

and it's got built-in thermal overload protection. So I think that's why this thing's resetting actually as I think that even though I haven't measured it. Um, I think the regulator is just getting too hot and it's shutting down for some reason. And of course, you can't have it operating at that temperature.

It's just ridiculous, let alone the like. I was measuring what 110 Plus on the case after a minute or two, let alone the actual D temperature? uh, itself. So yeah, it's going to thermally protect itself and shut down. and I think that's what's happening there cuz this supplies the main 5vt rail which Powers all of the logic board of course so you know it'll just shut off the output and uh, it'll reset like that.

And of course, you can't run this thing at too hot a temperature Anyway, being right next to these two large Fielder caps here, even though they're 105 CR rated, they're just going to die. um, you know, in very short order. So there is something wrong here. Surely this cannot be normal.

So I'm just wondering if there's anything that I've done in the tear down to cause increased power consumption. I mean my Supply seems to work just fine. So I don't know I'm I'm at a loss here I wasn't you know? I wasn't expecting to troubleshoot this thing I thought I'd Um, you know, just be able to probe some uh, waveforms and and uh, you know, uh, do a little bit of reverse engineering and stuff like that. but no now I've got this bloody problem to contend with.

What the? And there's also a software bug in this thing which uh, people on The Forum have actually uh, reported and confirmed and I'll just, uh, confirm it here. What a is is. if you set the current limit here below 10 milliamps then it actually shows. Even with no load, nothing connected at all.

No trickery going on here whatsoever. Um, it will actually show 30 milliamps reading current. So let's try that. that's lower at 14 13, 12 10.

Here we go Bang. Look at that. It jumps up and reads 30 milliamps. Crazy Why? Unbelievable.

Of course when goes down to zero it's it reads zero. but anywhere from 1 to 10 to 9 milliamps. Look at that 30 and then it just goes back to zero. Over that a what's going on there now? what I've done is put the lid back on here and uh, it? Well, it's not actually resetting now, presumably because it's got more air flow.

Actually, you know it's sucking in the air through here and out the back. so we're getting some forced air flow there where we didn't have that, uh, before. it was just, uh, still in the air. and I can still get in here and and probe the uh heat sink in here and yeah, it it rises.

It takes some time to get the correct contact on the thing, but it's still running incredibly hot. Check this out. Look at that. we're already up to 90 and remember that heat sink is right next to those caps as well.

90 2 93 we're still Rising that's just crazy and you know I'm not making the best contact there with the heat SN Of course. so you know it it. basically, you know, round it to at least 100. You got to be kidding me.

And just to make sure I wasn't hallucinating this thing or that I there was some bizarre fault that happened in the tear down or something like that I decided just to make sure I'd put this to The Forum and uh, I've left it overnight and sure enough, there's a whole bunch of responses from other people I asked if they could test their units and they have and they have absolutely 100% confirmed exactly what I'm getting this heat sink. Some people are even getting 110 in the case like this. Poking their thermac couble through, somebody showed a Flur uh IR thermal image of the heat sink up to 130 C Of course to do that, they have to take the case off and then there's no air flow and all that sort of stuff. so you know much higher temperatures than what you get in the case here.

but it's absolutely confirmed it the this thing. The design of this Ryo DP 832a is totally flawed I Have no idea how it even made it past the first design review meeting with a bloody 5V uh sorry, yeah. 5V regulator for the digital logic operating at a nominal around about 100 C You got to be me. It's it's one of the worst design oversights I've ever seen.

It is absolutely bre. you know, awful. This is bread and butter stuff for a power supply. One of the first things you're going to look at is the bloody thermal design of this thing.

Verify the thermal design when you're designing this product. Unbelievable. Anyway, definitely confirmed. but what I'm going to do is I'm just going to um I've left this overnight I haven't powered it up today so I'm going to switch it on and see if I can get the heat sink through here so it's got the proper air flow.

Uh, coming out the back of course the Fan's on a minimum when you first, uh, turn it on cuz it's no, it's not loaded so potentially the heat sink could actually cool down even further when I Ironically, when uh, the outputs are loaded because then it will turn the fan on uh, greater. at a greater speed, you'll get greater air flow over that heat sink and potentially could actually cool that heat sink down. ironically. But anyway, I'll switch it on see what I can get.

As you can see, my ambient temperatur is around about 23 C which is, you know, a typical office ambient temperature now. I've only had it on for couple of minutes and it's difficult to probe I've got it going straight through there I've actually uh, took out some of the sastic between the uh, the two filter capacitors there. so I can actually get through and I am probing the heat sink, but you know it's it's not ideal contact at all. But as you can see, I mean you can never actually read too high on this thing so it's not like you know I can get bad contact and accidentally read high.

so I'm getting 75 C on that heat sink. I've got to put a bit of pressure on the thermo couple just to get you know I'm I'm right on the side of the heat sink trust me it is, that's not making good contact at all. So I think the true temperature of the heat sink. is greater than that.

But we're up to 77 and at the moment and still climbing 78. This is not looking good folks. Here we go look at this 91. That is just insane Really.

I Mean you know a lot of people might think okay, what's the problem right? but it's all about design margin and you saw what happened when we'll verify this again when we open it back up, but I'm sure that this is causing that 5V reset problem. Now the to run something at 90 to 100? Uh to run a you know, an A 317 regulator or any heat sink at that sort of temperature for something dumbass like a nominal, you know, 5volt rail to power. Uh, you know, the digital circuitry in this thing is absolutely insane because there is no design margin in there. Um, yeah.

okay, maybe this thing might work and might continue to work for most people for a couple of years or something until those caps dry out, of course. And not even me in the Caps Yet, okay, without resetting at all, right? there might just be adequate air flowing there. But what happens if you stick this damn thing in a rack? You know they. I think they even sell a rack mount kit for it.

If you got. you know, a rack can easily get 40 or 50 ambient in there. Something like that raises, even if you raise the ambient temperature here in the lab by an extra 5. that could be enough to actually reset.

Uh, to trip the thermal overload in the D in that LM 317. And who knows what? We'll check which, try and check which LM 317 they're using, but that'll vary based on batch. It'll vary. they might have, um, you know, they might declare that they can source that from any manufacturer, so it's not going to be consistent across your entire production run of these units.

all sorts of stuff. It's just. it's just crazy. There is no way in hell that anyone could sensibly make a decision that says to run that heat sink at 90 would be a good idea for anything, let alone a production unit like this.

It's disgusting. A and I've managed to get almost 110. Look at that. 108 depends where I wiggle it.

Disgusting. Who the hell designs a 5V regulator that runs quiescent at 109? C What A Now I'm going to see if I can actually verify that the reset problem I'm seeing on the on the unit is actually the 5volt regulator. uh, going into thermal overload and, uh, dropping out so to speak. So I've set up the Uh scope.

Nominally, it's uh, there it is. We've got our 5vt output there. it's all pretty clean, triggering at about uh, 4 1/2 volts on, uh, negative going, so we'll just leave it there. and I've got my outputs switched on.

so I'll definitely be able to, uh, tell when it's uh, done that and well, let's see if it correlat see if we trigger anything on the scope or drop out in that 5V rail when this thing resets could take a while though and look at the temperature that thing's running at 141 and climbing when there's no air flow over it. Unbelievable. It's even got some additional little heat sinking on it due to the oscilloscope probe. Oh, and I only had to wait a minute or so and it's uh, it has reset here.

but uh, I didn't get anything triggering on on the scope over here so it looks like it didn't drop below. Uh, that? 4.5 Vol value I mean I You know, maybe I've got to uh, tweak that up a bit. Maybe there's a voltage supervisor or something on the 5vt rail inside the main Uh chip that even could be a function inside the main processor or something. a uh that, um, you know is actually detecting a smaller Dropout than what's there H Let me tweak it.

No. still couldn't get a to do it I'm triggering at 4.8 Vols and uh, the thing just reset itself but I still couldn't detect any. uh, drop in that 5V rail. so maybe my Theory's wrong there.

but jeez, I don't see what else it could be I mean now, just as a rough indication is what temperature those caps are running that just by sitting near that uh, heat sink at uh, you know, 130 or whatever it is. there you go, they're almost up to 70. just the cans on there. Of course, there's no air flow.

It's going to help when there's air flow over these things, of course, but that's how you can get with just you know, just the coupling over to those capacitors. Well, I'm at a loss now as to explain exactly what the mechanism is for resetting this unit. I could have sworn it must have been the Dropout of that regulator, but I cannot seem to capture any Dropout uh, AC or DC coupled um, at any time base of this regulator. So I I don't know.

But anyway, what I've done is this board when it was sitting here before, was resetting every 1 minute or 2 minutes absolute tops with monotonous regularity. Then I just put this: uh Pace uh, fume Exhauster I've got air blowing over it like this. It's probably more than uh I'm pretty sure it's more than what uh, was getting inside the case and all of a sudden Bam. I've left for uh, 10 minutes and it's not resetting at all.

so it definitely looks like it has something to do with that the heat of that regulator. but the exact mechanism ah, still eludes me now. I've even got to the trouble to set up a window trigger here. Oh oh, hang on yeah! I just saw that I Just saw it switch off I don't know if you saw there was some dip in the waveform there.

there was something. There was something there I hadn't damn it. I hadn't had the trigger on I was too busy shooting this video. but anyway, looks like we may have finally got it.

Let me switch this back on anyway. I've set up a uh window trigger here so that it can trigger anything outside of that or those two windows there. So I'm going to put it into single mode again and see if I can capture that. I've got it back on.

but let's uh, let's see what happens. Let's just leave it running. well. No, it switched off and reset.

but ah, we didn't capture anything. even with that tiny trigger. we window there. no I still couldn't get anything there.

So I've gone back to Um AC mode and there it is. There's our AC coupled mode 50 millisecond, uh, time base there and I set up a window trigger mode just around that. so let's oh, there we go. Look at that.

You see that. Look at that. There we go. So we've got some s Yep, Yep, all.

I think we're We're getting very close to getting this anyway. you can see that right there. Looks like that's normal. You know, in quote marks, right? that's the normal 5V AC coupled output We're only at 100 molts per division there now.

let's trigger off that and see and watch it and see what happens. See if we can trigger anything when these switch off and it resets? Bingo There we go. We finally got it. Yes, it triggered.

You saw it. There you go. This waveform started. Um I Don't know.

Maybe a second before this thing, uh, reset itself. So there you go. My theory was right. Well, you know it could not be wrong.

Really, it had to be that voltage. uh, regulator. in this case, just doing. you know, just some subtle Dropout There, it's as you saw, we couldn't trigger on the 5vt any.

V significant variation on the 5volt rail. uh itself. But when you get down, you know we're only talking like 100 molts per division. So just that sort of uh, noise or Ripple uh, that regulator is doing something and it's not regulating as it should.

um, anymore. It's yeah, it's still regulating at 5 volts. but the AC uh component of it has actually changed. So that is causing something on the digital board.

I Don't know what? I'm not going to go into the digital board and try and you know, dissect why and you know it. It doesn't matter. The fact is that regulator's overheating and is causing reset on the digital board somehow. So that's what we captured there and that is normally so you can see that it was.

you know, double or triple in amplitude before, uh when it actually, uh fails and we can no doubt capture that again. it'll be fully repeatable. You watch Bingo Too easy. There it goes.

Switched Off Barely even had to turn around for you know, 10 seconds and we captured it. Woo! So actually, this is a really good example of a little bit of a tricky world real world troubleshooting. uh scenario where I had a theory. Okay, this regulator was overheating.

it was dropping out, causing resets on the Um on the processor inside this thing in some manner. uh, but you know I I my theory was almost blowing, blowing out of the the water. I expected the 5V rail to just you know Plum it down to zero or drop down to 3 Vols or I don't know, do something stupid, allow you know a couple of volts or Ripple to come through or something horrible like that and I couldn't capture it even with a tight couple hundred molt window triggering around the 5volt rail that I had there before I couldn't do it on that 5vt scale so I had to SK uh switch to AC and I originally couldn't even find it on AC as well I'd done that before but it turn out I wasn't setting my uh trigger point. uh, narrow enough and in this case I switched to the window uh trigger and you can see it's probably just you know gone over that top one.

so I had to set that I had to use uh window triggering to go outside of, uh, the normal operational window to capture a really what is quite a small variation and most circuits would tolerate that quite well. You know, if you've got an additional 3.3 volt local regulator on your rail, or you're just powering some 5V logic, it's going to tolerate this sort of. Ripple Generally, no problem whatsoever, but there is something subtle on the particular processor inside uh, processing circuitry inside this Ryo that is causing that thing to reset. So if I wasn't absolutely confident that that regulator was dropping out, then I you know I could have, uh, thought Okay, well I've checked that and that's not the problem, you know.

Not an issue at all. it must be something else. You go away, go away, you chase red herrings until the cows come home. But no, we nailed it because I finally got down to a point where I could trigger off something that was causing this to drop out.

Now, a couple of people have already been a little bit confused by this issue. So I'll make it very clear. this 5vol regulator that we're looking at here: it has absolutely nothing to do at all with pairing these outputs or what load you put on the output. Whether you have a load or not, it's a completely isolated circuit with its own tap on the Transformer and its sole purpose is to power all the digital circuitry in the front here.

so that Applications processor the LCD um, some of the Io stuff here at the back. That's all there is to it. You can load down this all the outputs to the full 195 Wats, and the dissipation on that heat sink is going to remain exactly the same. Although, as I mentioned before, when you do load down the outputs uh, the firm wind knows that, or it's measuring the uh temperature of the main heat sink.

but I can't see any thermometer on thermister on there at all measuring, uh, the temperature of that. It does increase the flow rate of the fan, and as I said, the increased flow rate of the fan could actually have the effect of actually cooling down a little bit that 5V um, heat. SN But it's got nothing to do at all with loading the outputs. Now, the question is, how much load does this thing actually take? How much does all this digital stuff in here take? Well, let's have a look at it.

Uh, let's measure our power. Let's switch it on, shall we? Let me put my probe in here and we'll switch that on. Here we go. and it's powering up.

It's powering up. It's not much at all. Aha silly me. I Figured out what's going on.

Well, not silly me. silly. Riyo. This connector here, which, uh, you know, obviously carrying all of that 5 volts over to the board.

they've actually got the of the wires back to front I Assumed silly me that the positive wire would actually be positive. it's not. It's actually negative reg relative to that regulator. So now I've pulled out the uh black wire there because, uh, the reason why we're only measuring like, you know, 20 30 Milli before is because this ribbon cable here was taking.

um, that, uh, return Uh current. So here we go. now. we should be able to get it measure the actual current Here we go.

if we break into the positive wire here which is actually black Bingo There we go. We're getting in 300. it's boot in now. probably can't see the screen there.

we go, it's just booted up. There we go. and now after it's booted, we're getting. you know, let's call it, say, 700 milliamps.

Something like that. Let's switch the outputs on. Doesn't make any difference of course, but yeah, you know it's jumping around as you'd expect. But let's call that 700 milliamps and the input voltage about 11.8 volts.

which will of course vary with the uh line voltage because it is uh Bridge rectified with just some uh Fielder caps coming from the Transformer So you know that could vary. But let's just call it 12. V Now if we're getting 12 volt in and 5 volts out, well, we've got a delta or a voltage drop across this regulator of 7 Vols And because it's a linear regulator, it's got to drop seven. Uh, the power is going to be 7 Vols time the current flowing through it which is of course the output current which is 7 amp 7 Vols * 7 amp 4.9 Let's round it to 5 Watts This thing is dissipating five Freaking.

Watts Now anyone with any Electronics design experience knows that no way in hell you're going to use a heat Sinker that size for 5 Watts Even if you've got a fairly high air flow. uh, you know, going through your design and good thermal management, it's just ridiculous. 5 Watts Did no one even stick their bloody finger on this? or even think about it? Oh, I'm flipping the finger, that's for sure. Let's just go to a representative heat sink here.

I haven't Wasn't able to find the exact one, but this is going to be fairly close. It's an aid uh, Thurm alloy one to220 free? uh free stand? Well, it's actually a PCB Mount one. It's got a a PCB Mount tab. This one looks to have two PCB Mount tabs.

so um I'm not sure if there's any heat sink on the copper or the bottom side of this board. I Haven't taken it out, but anyway, we're going to be easily able to get some ballpark stuff here and if we have a look at it, uh, let's go in here. It's talk. We're talking 24.4 de uh C Thermal resistance there? uh, per watt.

But this is what we're interested in down here. Let's have a look at the graph, shall we? And what we've got here is, well, we don't need to really worry about the thermal resistance what we're talking about here because this is um Power dissipator. We know we're dissipating 5 Watts Look, it's for a heat sink of this size. It's off the graph already.

That should be ringing alarm Bells right? And this is the Uh Moun surface. uh, temperature rise from 0 to 100 C above ambient. And that's the key of course. And what were we measuring on this thing? Well, with uh, no air flow? you know we were getting basically um, you know, you know, well over 100 130 or something like that.

And this is the ballpark that we're operating up here at. 5 Watts with this size heat sink I mean it's going to be very similar. We're just talking Ball Park Calculations Here we're looking at 100 CR rise above ambient and that's exactly what we're getting. It's ridiculous.

I Could go into the you know, draw the thermal uh graph of all the things in there and the heat, syn compound, and the oh bloody, you know, everything. the junction case and all that. and imagine what the if this is what the heat sing temperature is at. imagine what The Junction's at.

Well, we actually don't even need to guess what the Junctions are getting at because let's look at the data sheet for the LM 317. Just take a typical one from Fairchild for example. Let's go down here. Let's get the thermal characteristics here.

it is. Um, there we go. thermal resistance Junction To case we already know that, let's assume that there's no loss between the case and the heat sink, right? Let's assume that uh, it? That's just fine. Well, the case here it is 5 C per watt.

We're trying to dissipate 5 wats in this thing. the junction is going to be 25 C at least above that's above the already measured and Quantified uh temperature on that heat sink. which even in the case with the proper airflow and everything else is over 100 C You got to be kidding me. Ah, Face palm.

Hang on, Double face palm. Well enough of that. Fiasco I might as well, um, do a little bit of uh, poking around of what I originally do before I Bloody discovered this ridiculous issue. Anyway, I was just going to have a look at the a little just a little bit at the output uh circuitry here and uh, see exactly what we got and it is very easy.

You've probably already guessed it, but um I've drawn a simple Dave card here and uh, this is basically what we got on the output. We've got the uh th000 mic output uh filter cap right on the front panel, uh, terminals as you saw I haven't shown the uh sense wires going back out there, obviously going back to a Sense amp. but yeah, there's nothing in this at all really. There's no output relay switching of course, or uh, any sort of electronic.

Well, there's electron switching, but it's done by the series pass uh, mosfet that we've got in here. but we basically got a big uh, shocky diode in there. There we go. We got some shocky diode protection across the output as you typically find.

Then we've got a couple of Ms here going to Main's Earth here and remember, this output is not Main's Earth reference. it's actually floating. So these things are going to uh shazzy Earth ground. And then we've got a another couple of Ms on our high side current sensor resistor here.

So there's our highs side current sense resistor. We saw that close up the tracers go off there to. there's our highs side current amp. they've rubbed the number off that the bastards.

But and of course the output of uh, that will be tied into the constant current uh circuitry which then controls the gate so that'll all be uh, analog Loop stuff going on in there. And then we've got a bleeder resistor across here. That's that one down in there. There we go.

it's a fairly large one, and then we've got another bleeder resistor across the Um Fielder caps there. our main three filter caps up here. there. they all are boom boom boom uh.

they're 2200 mic each. uh, 63 volt and we'll measure some. Uh, we'll get the scope out and we'll actually, uh, measure some things on here and have a look at the Gateway form there. But uh, and basically the input here.

Um, as I mentioned in the tear down that they, actually, well, I mentioned that there were Trix in there. but I didn't mention that they're actually switching the secondary uh Taps on the Transformer Here are the Transformer Taps coming in here and there. Our two triacs in there. There are our two triac uh, drivers down in there as we saw in the tear down, but that's what they're using instead of your more traditional relays.

Uh, to select the secondary Taps because uh, you? because this is a linear Supply You want to? You need, uh, some sort of tap on there. Imagine you're delivering only 3.3 volts out of here and you're getting you know, 40 volts out of your uh Transformer That's a lot of power to dissipate in your uh, linear regulator Like that in your part series pass transistor. There, it's called. So really, you want to choose that? They've got a couple of selections on those uh Taps there.

So that's basically, um, what they're doing. There's nothing in here at all. There's no output relay switching the S when you press that onoff button on the front. All it's doing is just effectively grounding that gate and pulling the output down to zero.

So it's not actually isolating the Um output at all. It's just switching off the output series pass transistor so we'll just see where those voltage Taps actually occur. You can see I've got the full 31 volts output voltage I'm measuring the voltage and also looking at the waveform. Uh, so this is identical to that.

You can see the voltage down there. 54. Um, we're getting. yeah, basically uh.

54 volts out of those Uh filter caps with 31 volts output voltage. Now lower my Uh voltage in 1vt our output voltage in 1vt increments and see where the tap drops. And of course in that, you'd normally hear a relay click in a regular Uh power supply. but this one uses a triac switching and Bingo There it goes.

It looks like it when it goes from 20 uh 21 from 22. Once we get down to 21, it drops down to 35 Vols So we're getting you know, a 15v um uh Delta there. so it's got to dissipate 15 Vols Uh, best case there. 14, 13, 12, 11, 10.

There we go. We got to drop down to 10 volts and then we drop down to our next tap which is 7 about 17 1/2 volts. So there you go and they will be the two Taps We won't find any more than that because we've only got two Tria on this things. And as We saw in the tear down, there is only a single Uh To220 mosfet in there and you know a lot of people are probably going to argue.

Well, you know 90 watts is this particular. Channel This is one of the Uh 30 volt uh 3 amp channels, so 90 wats dissipation. is this enough heat sink and air flow? I Don't know you could try and get the Uh data sheet and stuff like that, but what we should probably do is uh, measure the temperature on that heat sink at a full 90 W load and see what it gets to. yeah, you know, um, some designs would actually, uh, parallel up the mosfets there just to uh so you're not actually stressing just the one individual mosfit.

you're spreading the power against a couple of them. But Ry go well. now. I'm doubting their design decisions after that.

Ridiculous. LM 317 Fiasco But anyway, they've assuming they've done it right, they've determined that well. no, we can get away with a single Uh mosfet on there. So anyway, let's let let load it down and see if we can, uh, measure temperature on this thing somehow.

But yeah, there's no real easy way for me to stick my Thermo couple on that and get a really good connection. I'm afraid and just as we saw in the review, it can't actually deliver the full 90 WS on that. uh Channel but I was able it just, you know, shuts the voltage down I was able to get you know 85. it can do a couple of Watts more than that, but let's just that'll do 855 watts output.

So I'm drawing 85 Watts I've got the lid kind of sort of closed so we're getting extra air flow over the top instead of uh through here. but it's going to be. You know it's going to be reasonably close and I am probing the Uh main filter cap as well and as you can see there, we got 10 Vols per Division, 10, 20, 30 40. So you know 46 Vols minimum.

Plenty of margin in there for the Ripple and some initial probing of the heat sink there. it's at least to 45. So you know, as I said, it is quite difficult I'm not going to get in there and actually uh, you know, can't really uh probe the mosfet itself. it's getting up there.

but I wouldn't uh, call that particularly hot and of course the you probably can't hear it but the fan actually has turned on louder and we have been able to hit 70 there. So there you go. I don't know, you know I'm not going to go into the full uh thermal calculations but that you know that's not too bad. That's what you'd expect.

Really, You know that's a ballpark of what you'd expect for a full load on this thing, so you know, Not a problem. and I just took that back out and with no air flow there guys jumped up to 84. It's a bit hot, but you know there's no air flow so you'd expect it. Okay, I'll have a quick probe of the Uh gate of the series pass mosfet there.

and as I said that, actually control switches the output off and on and we'll actually see that here. What? I've got uh, dual channels here. Channel one, which is the yellow waveform is the gate voltage There, we're at 5 Vols per division on both channels. So 5 10, 15, 20, 25, 30, Uh, 30 volts output.

Uh, by the way, set 30 volts there. Um, the output is actually switched on at the moment and as you can see, if you switch it off, oh, run it There we go. We can see our output and Channel 2 is our output. uh, waveform.

so our output voltage. So we smack on 30 volts there. and if we just single shot, capture that bang. There We go.

We can see the rise there. There's got some little something happening down there. I'm not sure what, but anyway, that's still. There's no overshoot on that, uh, output at all.

It's ramping up. The output ramps up to 30 volts and of course, directly controlled from the gate voltage. So if you actually bring that up, you'll find that those two waveforms are virtually perfectly superimposed there. Because there's you.

Know there's nothing else, uh, switching the output. It's just that actual uh gate uh, series pass transistor via the gate there and I'm afraid I'm no real closer to that uh, turn on uh, glitch when you actually power the thing on. So I don't know that was the aim of this when I started this uh, yesterday was maybe to get down into that detail, but uh, who cares Now I mean we've got that show stopper which is the bloody Lm317. And here's another thing.

the Lm317 they're using. Check out the ridiculously thin tab on that. That is just one of those really pissant thin tab ones. Absolutely horrible, and you can see the difference on these two devices over here.

Whereas this has got like a normal thickness tab on it, that's what a you know, a proper, well-designed to220. These ones little pissant thin tabs on them look hopeless. So that LM 317 I mean you've just got less thermal Mass right there. And for the record, that looks to be an St Brand 31 lm317t and I check the data sheet for the St brand lm317 and just like all the others, you know maximum Junction operating temperature is you know around about that 125 C Mark and we've found we've you know, proven by measurement and based on the data sheet values The Junction to uh case.

and that sort of stuff that they're operating above that they're operating above the recommended Junction temperature. That's just it's just a complete fail right there. So there's something I wasn't expecting when I started probing this thing yesterday yesterday I Noticed an issue where it would reset I tracked it down to a bloody, overheating, piss poor designed uh 5V regulator for the main logic. Can you believe it? Bloody Ridiculous.

I'm pissed off. This is a huge, serious design oversight. No one in their right mind would deliberately design a little Lm317 with that piss an heat sing to run at Five Bloody Watts and think that you can get away with it. Right next to the put Fielder caps, these power supplies are going to fail in the field, no doubt about it.

Probably even see the resetting feature I had or in a couple of years time, those caps are going to dry out. It's just absolutely shocking how this thing even made it past the bloody Designer a view meeting, let alone into production, let alone to people's hands. How many months has this thing been on, uh, sale for now? and well, yeah, okay, nobody's found it. Okay, maybe a few units I don't know, might have, uh experienced issues that that we haven't heard heard of.

but ah, this is a huge, very serious design oversight and Ryon need to explain what the hell happened here and I don't think they can explain this away. How like you know, apart from we missed it or we swept it under, somebody found it and they'll told to no, that's not a problem. Shut up. Go back to your bench you know and it's ridiculous.

100 plus degrees C on a on a you know on a regulator and just quiescent static current driving the thing as by multiple people through the case with the proper air flow. Yeah, they'll probably do like a firmware upgrade. Oh, we can fix that firmware upgrade, make the fan run all the time God Man, unbelievable And there's probably going to be some people who will say well, what's the problem right nobody's hadn't AR look it just reset again. Oh bloody hell people are going to a few people say oh you know R might even say oh, had no problems in the field, no returned units.

That's beside the point. The the point is is that this is has no design margin in it whatsoever. You mount this thing in a bloody rack, right, sits in a rack. your ambient goes up by 20 or 30.

Uh Cel you're going to be screwed. Those Regulators are going to shut down the you know, who knows what regulator what the thermal cutout is in that particular type of regulator. You got variations in your Junction to case I don't even think oh yeah, they put some heat sink in there. There's going to be variations in that there's going to be iations in the air flow on the fan.

Huge variations. going to be variations in the life of those caps and all sorts of stuff. It's just it. It's just bad engineering.

It's not going to work. Needs to be fixed anyway. Riyo, please explain because this power supply now gets a huge thumbs down until this problem is fixed. I Think it needs to be fixed before you sell any more of these bloody things.

Unbelievable. What can you do? Well, you know it's tempting to sort of move that 5V rig over to, um, you know, an existing heat sink on here. but then you break your isolation stuff between your uh, isolated 5 volt po of digital and your output so that's not really going to work. Might you've got? There's a couple of mounting holes in there.

They probably could solve it by manufacturing some sort of custom, uh, heat sink or something which goes in there. And yeah, that'll probably be a fix. Maybe you know that? Well, it probably would be a half reasonable fix if they upped the size of that heat sink by, you know, four or five times or something like that, but it definitely involves something custom. Other than that, you'd probably have to relay out the board or something like that.

This is clearly unacceptable. Anyway, let's see I'll definitely let Ryon know about this and let's see if we hear back from him cuz this is just complete n of catch you next time.