Hi Just a quick follow-up video to my previous one where we saw this: Rd Tech: DPS 5020 module have the magic smoke escape from it and I'll link that one in here Oh at the end of the video and also down below if you haven't checked that out. So I Thought we'd do a quick follow-up actually getting this thing up and running again because I didn't have time last time. Let's get to it Now, as you saw in the previous video, the problem was that there was a multi-layer ceramic capacitor 100 nano farad one connected directly across the pins on this, while the two are screw terminals on the backside of the board here and Glenn who is the designer at Artie tech said yes, this it was the problem. Actually, screwing the terminals on the top of the board actually caused mechanical stress and cracking inside the multi-layer ceramic capacitor when you turn those screws too tightly because that force of those, even though they're rigidly sold it to the PCB.

It's just enough that a tiny bit of you know, like torsional type force is applied to the capacitor, which is directly, of course, rigidly soldered between the two pins and then a key you can get possibly also a micro crack inside. and that can lead to the failure mode of these things, which is typically a short circuit. And of course, all the energy which was delivered into that capacitor via this grunty little power supply caused it to catch a light and burn all the board. Oops.

So yes, the follow up to that is that Glenn has admitted that's a problem and I has stopped selling these until he can fix. They're actually redesigning the board to have less stress on there, so if you do have one of these, he says just remove the capacitor. It's not hugely needed, it's just some like little extra filtering on the output. So that was a huge oopsie.

So as I mentioned in the previous video, just be careful of mechanical torsion force on multi-layer ceramic capacitors and also thermal as well. having the capacitor directly across and sold it to these two pins that have to be hand soldered. not a good idea at all. Thermal stress can also lead to similar sort of problems causing a short failure.

Now, when I repowered it in the previous video, you saw that a da hiccup then turned off and that was a combination of two things. One is that I was using an external current limited power supply which was going into limiting and then causing the thing to hiccup. And the reason that it was overloading with no output on it was that the the board that's being carbonized in here actually had formed a low impedance so it's actually conducting across the output and that was what was drawing there of the quiescent current and causing the thing to trip. No, I did actually measure this before I powered it up in the video, you didn't see that and it was the in the order of you know, tens of K or something like that.

but it's now 47 ions for example and this will most likely change with the applied voltage as well. So yeah, it's like it's physically changed. but anyway, um, you can. If we swap the leads around there, we go.

It's the same either way. So we do have a low impedance directly across this and what we have to do is get in there with a dremel and or a knife and just gouge out all of that carbonized fiberglass in there. So why it actually changed from when I originally measured it before I powered it up to now where it's like forty seven ohms or whatnot. It could be different if I pair it up again.

I don't know. There's some sort of, you know, some sort of chemistry thing happening in there with there with the carbon-based you know, burnt fiberglass or whatever. I won't pretend to try to explain it all. I know is that we have to get rid of that, grind it out and hopefully this thing will repair up now.

Unfortunately, it's not as simple as just like grinding out a track like cutting a slot into that because if you have a look at the top, there's actually two fuses down in there and also let the trace going over to there like I could do it, but I'd rather try and keep those intact. And those fuses there. 228 fuses by the way. And yes, they are intact.

They didn't blow at all enough to go gene out a decent part of that. Let's measure it again and bingo you can see it's changed so you can see that the more material that we're going to remove from there the wall, the more resistance that we low resistance. I'm sure that we're actually taking out, so if you physically remove all of it, we should remove the upward short. Unfortunately, we're gonna have to go join a few things.

All right, let's take a look at this. I've gone out some of it in here, but as you can see, it goes all the way practically all the way through the board. and if we flip it over under the top side here, you can see down in there that under this twenty eight fuse, yet the burn has gone all the way through the board. The top 20 amp fuse there is fine, but yeah, that's just a complete fail so we can't just scrape out get out like half the depth of the board.

We really have to route out the whole blinkin lot between here, which is a real shame. I Can probably maybe get the grinding wheel in there and if you have a look, you can see that there's these two two two 20 amp fuses aren't in series. Ones actually come in from some Vias down in there over to here over to the positive pin and the other ones also come in from the positive pin from the output of the inductor and if you flip it over here, you can actually see what's happening there. So this one goes straight in down those vias, through that 20 amp fuse and then into the positive Jack here and then the other one is of course, on the top side coming from the exact same pins of the inductor.

So they're effectively two separate 20 amp fuses going from the output inductor in parallel to the output pin. So why they've put two 20 amps in parallel in there? My guess would be that maybe 20 amp is the most that they could get in an SMD size and they didn't want all the current going through one because this is their a 20 amp output rated power supply. So they did want to use a 20 aerated fuse in there in SMD size so that when I put two in and yeah, it's overrated. So yeah, I wouldn't have I would have put maybe two 15 amps in parallel or something like that in that particular case because it's going to be much harder to actually blow both of those at 20 amps as opposed to like you know, two 15 amps in parallel or something like that.

I'm not sure if that's the best choice, but of course you can't use a single 20 amp one for 20 out rate of supply because eventually, in theory a 20 amp fuse will blow of course there that they don't just magically blow at 20 amps. There is a characteristic curve for all fuses like this and it's all to do with temperature rise inside the fuse, melting the internal wire, and how long that takes the fusing time and all that sort of jazz which I won't go into. But anyway, there's two 20 amp fuses there so either I can get down the side and Dremel it maybe like right down here or I just grind out the whole blink a lot. but then I have to wire the fuses in separately and that's a real pain in the butt so don't know if I like doing that.

Maybe the grinding wheel is the best option. and there we go. That seems to have worked at read where like the you probably can't see it very well but the internal charring on that board basically stops there so we scored away most of that. I'm gonna have to repair that track is still intact, the cop is still there, but yeah, I'll just wick away that sort on there.

Of course we have to repair these sense lines here. There's one on the top and let's have a look. well, one on the top here and one on the bottom on the other side. but we've basically still got that fuse intact, all the veers, everything coming through and that I think should work.

a treat. Let's go measure it anyway. that's saved around with having to re manually rewire those fuses. That would have been a pain so that should do it.

Alright, so let's try it again. You saw that we'll get in 47 ohms and then up to 70 or whatever. So I expect to see like some sort of capacity of output charging or something. Hello, there we go.

Started out low of course and it's yep, it's gone up. That's what I expect. Put it around the other way and woo There we go. Start out low cuz there's some output capacitor that's somewhere that's charged and Bob's your uncle? Look at that.

Alright, that's where I'm not sure what it's supposed to be, but hey, I deem that to be as normal as it possibly could be. So let's put it back together and pair it up. Oh and by the way for those who were spotted it and pointed out thank you very much the short across the pins on the microcontroller in there that was and just a solder dag that got in there from part of my soldering or whatnot I fell off. Not sure how but that did not cause the issue at all.

It's gone now and because that's only on the keyboard display side of thing had nothing to do with the rest of the circuitry. Alright, so if we actually power this thing on and got it powered from a 40 volt source with a 5 amp current limiter this time, so it's capable of 200 watts and you can see that it's drawing two watts qu some current and you'll see that everything is just hunky-dory Now, let's switch that's got set to four Point Nine volts. Whatever, let's switch the output on. It's not drawing, it's not measuring anything at the moment.

Let's there you go. Bingo, Four Point Nine Seven. It's reading precisely what it's set to and there's no current draw on the output because this thing, you've got to assume that's still working because the fuses were intact. It's got all sorts of protections built in, so it no worries whatsoever.

It just flamed in that capacitor on the output and the magic smoke escaped. So it should still be a fully working power supply, and it looks like it is. So let's switch on the load now and see what we get. Alright, so let's set it to a 10 watt constant current load I Set it to wire 10 volts precisely on the output.

Um, in fact, let's uh, switch that on. Now there we go. No worries at all. It matches and - we have current limit so that's 20 watts.

No worries whatsoever. So let's switch, snare, put on and see. Bingo 9.8 What says and it's a little bit out, but it's so I Think it's like half a percent or something is the spec on this thing, so no worries whatsoever. 10 watts? fine and dandy.

One thing I don't like and you can probably hear is the fan in this thing. I Don't have the top on it, but it is just whirring at constant speed even with no load. It's just really annoying. and it's loud enough to be quite annoying, but you know you can retrofit any fan you wanted to.

Okay, now we're drawing a hundred watts. There we go and everything's hunky-dory Don't worry about the voltage loss in here. I haven't said that by remote sensing, so we're just getting some loss across the cables here. even though there are a decent size when you're talking 10 amps, yeah, you're gonna get some drop.

Let's just have a look at the output noise here. like 10 volts. There's no load on there whatsoever and that's actually quite substantial. I mean, and that helpful cap shouldn't make it like the the nor point missing, not point one mic on there shouldn't make a huge difference.

If you're wondering where these spikes are coming from, you can see those there. If I capture it, look at those um they like are still there. even if I switch the output off I've got no load at the moment so that's a that's the low note, low no load noise. and if I switch the output off as in like soft button on the front of it, we still get that high frequency stuff so that is coming from somewhere in the system and that's actually coming from my environment here.

So yeah, I've actually switched off the input the power supply. You can probably hear there's no fan noise now and it's completely switched off. So yeah, we're just picking up crap because I've just got the leads flapping around in the breeze. Hear me? ignore the man behind the curtain.

And if you're wondering if that no load ripple isn't caused by the lack of the capacitor on the output, the answer is no I can whack in a 22 mic 200 volt cap here and make sure I got around the right way and I can put that directly on. Again, we've got no load and bingo, it makes no difference whatsoever. So that measly I think it's a naught point. 1 micro Farad cap across the output.

It's just for, you know EMC Like, you know Cee compliance. So and I believe this thing is a, you know C II compliance. So not having that cap on there actually technically changes the compliance of this thing. But yeah, even a a 22 mic cap on the output doesn't fix that.

but you might see it change if you put a load on so we can actually do that. I've got 200 watt load on the airport. Let's put the cap on it. 22 micro Farad's Bingo.

You can see it, that's off, that's on. So you can see that change. Just a turn. so there you have it.

That's the repair of the DPS 50 20 power supply at that Completely smoked and you know, Thumbs up to our D Tech They did actually admit it was a problem and they'd actually seen it before as well in one or two were cases before, so they should have fixed it at the time, but they didn't They sent me one and a classic mechanical and all thermal could have had some that could have contributed as well. But and the most likely scenario is that when you screw in those screw terminals on the top, it just got some micro cracks inside the multi-layer ceramic capacitor and in most of those circumstances they will fail short and then this thing is capable of I mean we're only 100 watts at the moment, but even that, like you dumped 10 watts into that capacitor or something and it's going to start smoking and flames. The magic smoke escapes and you get the flames and it starts burning. But anyway, it's robust enough to handle all that.

No worries whatsoever. So sorry. but I'm not going to fully characterize this thing in the video and do a full review of it and everything like that. I'm Glen I Believe it has actually stopped selling these and until such time as they can change the PCB to do the layout properly and then they'll be reselling these with the corrected Ml/cc but I hope that's an interesting lesson to not only I detect, but to everyone out there that is, multi-layer ceramic capacitors are susceptible.

As I said in the previous video, it's not only mechanical stresses, which is a big thing for them, thermal stresses they're susceptible to are just a failure in manufacturing. You know you've got an infant mortality thing on these components, so some will eventually fail and a good way to actually reduce or limit basically eliminate. The problem is to put two of them in series, so if one happens to short Judah whatever issue it is, then then the other one will still be a capacitor. And yeah, you will effectively double your own capacitance, but it's generally not a huge issue.

At least it won't smoke and catch on fire. especially for high power supplies like this that are capable of delivery. not not just on the output of a supply, but as you saw on my Ness alarmist sister which had the same as smoking ceramic capacitor that caught a light and burnt the entire board that was on on the input side. So I just like the AC plug pack, the rectified AC plug pack if the capacitor was on there just to you know a filter is a little bit on the input side of the regulator after the bridge rectifier.

and of course the plug pack can deliver watts and we know 5 or 10 watts and the capacitor shorted Judah where in this case it wasn't a mechanical file is just like an infant mortality failure, a component manufacturing tolerance thing. It happens and it caught a light as well. So not only just the output, but input side of things as well. So just be careful with multi-layer ceramic capacitors.

When you lay in our boards, taking your count, your location of the ceramic capacitors next to any stress components. In this particular case, here we've got our screw terminals on the PCB so when you put your note, screw them up real tight. The torsional force on there can couple through to the PCB. Or if you've got mounting holes or something like that, you're putting screws into those.

You have multi-layer ceramic caps next to those can ruin your day If you get a generally they'll fail short. If you're lucky, they might fail open and well, the product might fail. You might get more noise or whatever, but when they fail short and you've got a supply across it that can deliver a certain amount of power that can make these things catch a light. So anyway, it's a real interesting real case of how component fire can in this case lead to something quite dangerous in a component catching on fire anyway, hope you enjoyed it.

If you did, please give it a big thumbs up because I've always helps a lot and as always, discuss it down below catch you next time. Oh by the way, if you're wondering why this video is maybe a little bit different I Actually I'm shooting this thing on my Sony NEX V G 30 which I normally only use for my talking head mailbag and I'm using the internal mic on that main air which is a shotgun so it may not be audio, it may not be as well. It's probably going to be a little bit different or what you expected. Anyway, catch you next time you.

.