TSP #191 - Teardown, Repair & Experiments with a GW-Instek APS-1102 Programmable AC/DC Power Source

[Music] hi welcome to the signal pad in this episode we're going to try another repair this is a gw instax aps-1102 this is a programmable ac dc power source and as you can see from the front it actually has a multi-purpose ac plug essentially where you can plug anything into it and control it and power it from this so why did i even bother buying this one well i do have a sencor pr 570 which is an isolation transformer adjustable isolation transformer and it allows me to bring up equipment that are broken or if i'm repairing let's say a power supply it allows me to work with it safely and bring it up slowly and you know that has some good advanced functions it has chronic limiting and leakage detection and so on but this is a far more advanced instrument and it can do sequencing and some other fancy things including of course providing a dc power source i don't know if you can change the frequency or not but we can take a look at it so i picked this up from ebay for about 400 which is not cheap it's still fairly expensive and again i can do that thanks to my patreon supporters which enable all of these things of course and even if you can't fix this i'm still interested in taking it apart and see how it works what kind of architecture it uses this is almost certainly a solid state architecture so i'm curious to see what they've done to enable you know the advanced functions it has it's actually quite large also i'll show you the back and see what kind of interfaces it has but i'm eager to see if we can figure out what's wrong with it and here's the back of the instrument you can see it also has the output brought into the terminals in the back one kva maximum and plus one minus 400 volt peak maximum that's where the power goes in to of course power the instrument and it's made in japan straight up so that's nice to see and we have usb and io control and some external signal in for synchronization so it's you know a fairly advanced unit and this was actually calibrated based on a sticker on the top in 2018 valid until 2019 so it must have been working not too long ago and it's in fantastic shape it looks almost brand new so i think at this point we should turn it on and see what happens all right power on sequence it's plugged in here we go nice back light is on loud beep we see it has a version 1.5 and a usb id is there and there's a problem right there so system locked communication failure 2. so obviously something inside some processor is trying to communicate with something else and it can't and it just basically disables the whole thing there's absolutely nothing you can do there's nothing else you can do to get into any of the sub menus or systems and i looked at the manual there's no schematic or even a block diagram but they say that if you get this send it back to gw instead which of course we're not going to do we're going to try and take it apart and see what's going on inside and check this out look how long it takes after i turn it off for it to actually turn off it's going to take a while that's that's a long time so i wonder if that is some kind of shutdown sequence or it's just a big capacitor that's holding the charge all right i think i got all the screws off we can take a look at this for the first time together let me see that this should slide right off and what do we have here wow that's a there's a lot of stuff in there that's a pretty complicated instrument but i do think it is indeed solid state based as there is two other pcbs down there big heat sinks at the top the main processing unit seems to be the one on the side interesting way of getting the usb port out i wonder where that goes i don't think it goes to the front because i don't see a usb port in the front but yeah pretty cool so we're just not going to have to figure out what is talking to what and what could have possibly gone wrong so the communication failure is probably coming from something over here not working because it looks like this part in the here that's actually connected to the front so this is what's driving the display so this controller is figuring out something is wrong somewhere else so let me take a little bit closer look at it and scan over it a little bit and then we can go over it piece by piece so i spend a little bit of time going over the different boards and figuring out what's connected to what and now i have a little bit of a better idea of what to look for so this board at the top is called the pfc inverter maybe power factor correction power factor controller i'm not sure yet and the board in the middle and the board at the bottom have are labeled the two different kinds of inverters obviously all of these boards are responsible for generating the ac and the dc signals and potentially putting some imperfections on the ac line so you can stress test your device that you connect to the front so that's whatever is going on from the ac now over here this main board is quite interesting because this also has multiple different domains on it i assume that this instrument probably has many power domains galvanic be isolated in strategic places to make sure that you know whatever it needs to talk to whatever it doesn't necessarily have to be in the same potential and that's important for power supplies of course now on the right side over here this area is connected to the usb and the digital ios in the back so this thing is galvanically connected to the user and therefore it has to be isolated very carefully if you look on this left side there's a dotted line around this section that's connected to this connector with all these cables going to the top and this appears to be almost entirely isolated and if i look this at this dotted line that's probably where the isolator section is by the way this instrument has really really good silk screen there's a lot of voltages and test points labeled you can actually see what you're supposed to measure so it's super handy even though there is no schematic now this usb here is nothing more than just the front panel uh controls the lcd screen nothing unusual there so if i look at this a little bit this obviously has to get power from somewhere now if i look at this portion this portion is coming and connected to this one which means that this connector here is bringing power into this board in some way and getting potentially some digital data out controlling this whatever is in the back side now this one over here seems to be fairly well isolated from this connector which means that the power from to this one is either coming from here or there may be some inner layers that connect these connected to this one and i think in order to answer that we need to x-ray this board so we can at least understand what is connected to what so when we trace out the power supplies we can actually figure out if everything even makes sense there are a lot of labels here which i would like to measure so i propose before we do anything else get an x-ray of this which would be helpful during the debugging and then we can put it back and do some actual measurements on it here we go the board's inside the x-ray machine and now we're going to get some nice images all right so here's the photograph of that board you can see this is the area we're interested in and that's the area where the one connector comes in and the other connector leaves the board now i already took some x-rays and here it is so in this x-ray we can see that there is definitely a connection from this bottom connector to this connector at the top so there is and these are thick lines so definitely some kind of power is being fed to the other board so we have to watch out for that on the other carrier the big board at the top of the instrument we have to see where this thing ends up now i took an x-ray with a different exposure and you can see this line actually disappears because it doesn't have the correct power but now you can clearly see that there is a section in the middle of this board that has a completely different power supply it's fully isolated it's also pretty cool to see the opto isolators how they're fully galvanically isolated as well as these ios that are also isolated as well so this is a this is how these two boards actually talk to each other there's nothing else between them there's no metal between them that's kind of nice to see there's a big ic here in the middle you can see the capacitors for example so this is almost definitely a dc-dc converter and you can see a couple of other power supplies that enter this port so looking at this i'm fairly convinced that this board is powered through this connector at the left so we go over here it means that this section of the board is powered somehow through this connector but the connection going from this following this mouse all the way to this corner should not be ignored we have to follow that as well you can see there's 1.8 volt let's written here 3.3 volts written here so we would definitely be able to check those and here's the up to isolators that you can see in this one that's what they look like under the x-ray so we have a lot of information now we can put this board back into the instrument and make some measurements check out those power supplies and even you know find out if there's a way to separate this board and maybe run it just by itself and i'm suspecting that that digital problem may actually be from this side because if you look this is a some kind of a serial link or some kind of an io interface between these two boards so if this processor board at the very top here which you cannot see in the x-ray it's right above this side if this board is not able to talk to this one it would generate some kind of an error so that's kind of my intuition so let's go ahead back to the instrument make some measurements and figure out if that's correct okay so here we're now looking at the top of the instrument this is the big ribbon cable that's coming from the board we just saw the x-ray off and if i look at this it's beginning to make more sense so some power enters this and i think that if you look there's a little bit of a dotted line around here as well and that dc power whatever that is coming into here goes into this chip and this is actually a switching power supply i see and i believe this then generates a minus five volt and a plus five volt and those two then are fed back onto this and on to the other board so we should be able to measure these plus and minus five volt power supplies and make sure they're okay if they're okay that means power is reaching the other board and that's entire link is working so that's a pretty good place to start so i'm going to turn this on okay so i've connected the multimeter on the ground side to the input of the switching power supply remember this transformer here isolates the switching power supply from the other side so we have to measure differently on each side the ground is connected so this should be the input power v in there you go 24 volts that's good so that's what's coming from the other board so we already saw that in the x-ray those were the two lines i'm going to take this ground i'm going to move it to the other side connect it to the other side of this i'm going to make sure don't touch anything else there we go looks good i already tested those uh these heat sinks that are not live of course i took the precautions and let's try and see if we can find the minus 5 volt here's the minus 5 volt and what do we have oh minus two and a half that does not look good it should certainly be minus five so that doesn't work what about the plus five volt nope 1.15 there we go so we already see a problem there's also a plus 12 volts and eleven and a half okay that one's not bad acceptable so the plus and minus five volts don't work that's not a good sign i'm surprised because there is a and 7805 here this is a 7005 this is a 70 905 so that's plus or minus five volts regulators so these regulators should work i mean there's no reason for them to to be doing this unless they're heavily loaded let me touch one oh yeah that's pretty hot okay so that means that something is loading them heavily let's see if we can verify that so it's fairly clear that this bottom board whatever is happening over here is loading that plus a minus five volt power supply now if i remove this unfortunately i won't be able to measure it anyway because there is power coming from here and then coming back but i can remove this cable and just bypass everything i know where i need to connect my 20 forward power supply onto this board and then we can see if the plus and minus 5v power supplies come back when they're not connected to this circuit this is an easy way to verify if they're really being loaded heavily by this board that's a good debugging step all right so i removed that ribbon cable you can see it's completely gone and i brought in the actual 24 volts directly from the bottom to the top so basically we only have that one connection these two cables are the only only interface between the bottom board and the top board now let's measure those voltages again so we had we should have minus 5 here there you go minus 4.9 that's much much better i wonder if this regular has aged a little bit and it's not regulating as well as it used to and let's see the plus 5 volt one there you go plus five volts okay so that already pretty much confirms that the bottom board is loading the top board somehow through that ribbon cable the power supply section in here indeed is working at least good enough for our purposes now it could be that these are ghost voltages and you have to load them to see the effect but it's very unlikely for both of them to be like that so i'm you know reasonably convinced that those regulators are working if i touch them now yep they are now completely cold there you go that's good so it means that that that part of the circuit is indeed back to life so now we can turn our attention to the bottom board again all right now let's do some thermal imaging this should be pretty straightforward i just have the thermal imager pointed directly at the area underneath the connector zoom you can see my fingers there and this is the connector coming in and this is the area that i'm interested in looking at see if we find anything of issue i'm going to turn it on and let's see let's see oh my god look at that that's the first beep so it's going to do the second beep and the communication failure there you go wow those components are hot they're already reaching almost 90 degrees celsius and there's three of them and you can find out where they are yeah these are super hot so these three components all look like op-amps actually i have to take a closer look interestingly the dc-dc converter is right here and it's totally cold and if you look all over the board i don't see anything else light up not naturally because this is at 90 degrees celsius so everything else is kind of swamped out but yeah three components and they appear to be identical so why have they died that's a good question and uh we can also find that measure around and make sure that the power supply is indeed the same power supply we just measured i think it is these things are probably clamping it down look at that 94 degrees celsius i'm gonna turn this off now okay so now we have something concrete we understand the power supply flow and this also explains why we're not getting any communication because if those power supplies that clamp down the processor which is down there is not going to produce anything it's not even turned on and therefore there is no communication to it so now we finally can explain why we have the communication failure so the three suspect op amps are this one this one and this one and they're indeed all identical there's no difference between them this one is sharing the same power supply but it doesn't seem to have an issue now three identical op-amps to die at the same time why would that happen i mean i can see if let's say the power supply has some weird behavior but those voltage regulators look reasonably good so i'm not so sure what's going on we're going to remove them because there's something's definitely wrong with them anyway and if i remove them and let's say the whole power supply comes back and there's nothing else wrong that's already a pretty step big step forward and we can then replace them is it possible that all three of them have something loading their output that's a bit skeptical each of these has two operational amplifiers on the inside of it so i think removing them and running it without the op-amp is probably not a bad idea all right i removed the op-amp so you can see that the three footprints are cleaned up so without it we're going to run the instrument and measure the power supplies again okay so let's start from the plus and minus 5v power supply everything is not connected except obviously the op-amps are not there let's see this should be five volts ah look at that so the five volt power supply has returned this should be the minus five volt which was always a little bit low there you go minus four point eight that's also working that's great so the power supplies are now back so they're not being loaded by those damaged op-amps there's a dc-dc converter here which is supposed to generate 1.8 volts and 3.3 volts we can go ahead and measure those two that's but powering the entire processor down here so measuring this one there you go 1.8 that's a very good sign so that's that supply is working where is there 3.3 here's a 3.3 and that looks good too okay so that entire station has now come back to life that's a really good sign it's probably time to order some parts now well check this out the error change it no longer says internal communication error it now says dcps failure which is some kind of a maybe dc power supply failure and this is not a reliable error of course because it's missing those op amps and those op amps were probably part of some kind of detection circuit or some kind of drive circuit and that might be why it's not working so we've indeed changed the nature of the error and i think this is a step forward all right so i received the parts from digi-key and i decided to change all the four op-amps even the one that wasn't heating up the one that wasn't broken just so that they are all matched from the same batches from the same manufacturer at the same time i replaced the 707805 and the 7e 905 the plus and minus 5 volt regulators because i think they had aged and as you saw the negative one for example wasn't very good at all so now we should have hopefully a nice symmetric plus or minus five volt supply i have done the same thing i did before i bypassed the ribbon cable only powering the dc-dc converter up here from the voltage coming from the bottom just so we can make sure that the plus and minus fibo supply is actually working and it's accurate so i think everything is wired up correctly so i'm going to put this on the 5v supply so we can catch it as it powers on here we go there you go that's nice 5 volt that's very good so the 7805 the replacement is working and the 7905 there you go minus 5 volt so we've got a very nice symmetric plus or minus 5 volts as expected with the brand new parts now the exciting part we connect the ribbon cable back and try again okay all back together here's the moment of truth let's see so far no explosions that's always a pot plus okay let's see if it gets past this point come on you can do it there we go look at that we booted no more errors okay so i think we should take a closer look at this gui and see what's going on with it well of course as soon as i finish repairing the circuit i go back and there is a block diagram which describes the operation of the system at least in a very simplified block diagram there is really no actual description there is no schematic but yeah i found this only after i repaired it but nonetheless we can look at it and see if we kind of reverse engineered correctly so the board at the very top of the instrument is this this the pfc inverter it actually does the power factor correction because it's what's connected to the ac line that's where the ac line comes in that sub dcps circuit that's the circuit that generates the power supply isolated that powers that entire system controller board which is the one over here and that power goes over here and it powers another dc-dc converter and that dc-dc converter was the part we were measuring with ldos right after it all these isolators these are the what separates the two bores so there's some up to isolators here that completely isolate this portion the bottom of the circuit from the top so all of that kind of now makes sense so the second error that we saw the dcps failure is actually the failure of this which means those op amps that i thought were somehow to do with monitoring of the voltages of the power supplies and conditioning them were indeed doing exactly that they were monitoring this so when those op amps you know reported something some problematic it thought that the tcps power supply was broken so indeed all of that makes sense and that is consistent with what we thought it is it's also interesting to see how they have designed the actual amplifier so this is indeed fully isolated as you can see there is transformers here so this is not connected galvanically to the line coming in which is great so it's similar to what an isolation transformer would do but they have replicated this amplifier twice it seems and there's a range changer here which both of them come to i think what happens is when you put it into the higher range in order for it to be able to achieve voltages higher than the ac line you put into it it puts them in series it looks like so they essentially run them in parallel and if you need to go to 250 volts from a 100 and you know 20 volt ac line just put them in series and i think that's what this is and this range changer is almost certainly all done with relays i don't there's no need for this to be a solid state of course and these power amplifiers all have their own protection current limits and all the fancy things to make this work so as you can see it's quite complicated and this block diagram probably doesn't do it justice but at least we got the rough idea of how it works so looks good okay so the very first thing i did i reset the instrument to a factory setting so we can take a look and see what's going on in terms of the modes there's quite a few things you can do you can apply a lot of different types of waveforms not just applying a sinusoid to the load you can apply even an arbitrary waveform so it's really quite advanced there's two ranges 100 volts and 200 volts i set it to 100 here in north america where i live is 100 volts of course the frequency default is 50 hertz which is not what we want so let's make that 60 hertz which means that we can of course change the frequency you can change the phase here is when you can apply any kind of arbitrary waveform you want to your load very cool stuff and then there's a maximum peak current limit that it applies as well as an average current it applies under miscellaneous you have memories you can sequence your waveforms very cool different limit stuff you can tell it what it measures and uh yeah looks good let's see i don't want to do that anymore so output is turned off okay i think i haven't actually turned the output on yet you can see the buttons at the bottom actually you can't see there's buttons at the bottom that allow you to run a particular sequence and on the left side you can do the measurement it measures the voltage the power you know the different kinds of parameters it measures the phase and power factor all that stuff that you would expect from power analysis so first is we should hook it up to an oscilloscope and see if it's producing the correct waveform and then maybe we can stress that something that i have in the lab okay so let's talk about our measurement setup here so here's the gw instead in the middle here and at the top i have an high voltage differential probe these are fully isolated the output of this goes into an oscilloscope on the right which i will show you in just a second i have a tiny little 130 volt light bulb here which i'm using as a pure resistive load it's a nice load that it doesn't change the shape of the waveforms allows us to monitor the quality of the way from directly from the gwen stick without having it floating so there's actually some load at least then i have a current probe here which is monitoring the current going into the light bulb of course on the right side those two things are being measured by the setup right over here sorry about the camera we have here a keysight dsox3104 and this is going to have channel one voltage channel three is our current probe and i have a power supply here that powers the probe so we should be able to now see how it works and here we go so we're going to do this measurement with this instrument set to ac plus dc so we can get both of them at the same time so we can see the dc performance too these two scales that you see in the measurements on the right side have already been appropriately scaled for the differential probe and the current probe so right now the output is turned off and the voltage is zero let me turn it on nothing changes of course because we're not applying anything so let's apply some ac voltage there's 10 volts oh look at that we do indeed see a sinusoid it's a beautiful sinusoid okay here's 80 volts as an example and let's see if this makes any sense so it's measuring 79.9 volt rms look at that 79.8 volt rms on the oscilloscope very good it's measuring 0.06 amp rms roughly this is measuring about a little bit less than that the measurement and the current is probably not very accurate because this is a 30 amp probe and we're not using it in a very good part we're losing at this lowest kind of range it has a very accurate probably this also might need calibration of course and uh let's see we can change the frequency which is very cool here we go very nice look at that yeah we can change the frequency it's very good for some motors if you use this on motors of course you can control their speed in some situations now let's try a different sinusoidal that's square wave look at that beautiful square wave coming out and of course here you can see that the current and the voltage are perfectly in phase because we're dealing with a purely resistive load and i can actually hear it now because it's a square wave and the harmonics of it are now in the audible range let's turn that back to sinusoid and the power factor doesn't show up here because it's probably below the measurement it can make this is a very very tiny load so it doesn't show anything it's unfortunate i would have liked to see it show but from the range you can see that it is it's getting much larger loads heavier current flow through the device let's try some dc let me turn this uh ac completely off let's see how the dc works oh look at that very nice of course the light bulb doesn't care of course about whether it's dc or ac it's a 100 volt dc let's see how accurate is 99.9 volt rms and oh this is not going to show anything but you can see it's right on the 100 volt range it's because i'm measuring ac rms only but it's right on 100 very nice and accurate same current this again is not going to show up but i think it's in the right place very good it's amazing so we should be able to combine these so let's say a 30 volt dc you can put ac on top of it yep absolutely it works no problem very good i have to say i'm very happy that i managed to fix this because as you know i was trying to replace my other isolation transformer and now we now that we saw the block diagram we see that this is indeed a fully isolated instrument as well so it's going to have a lot of use cases but anyway i think we can stop here i'm fairly satisfied with it there's a pc control software too but i think that's going to be no issue at all yeah it's good let me know what you think in the comment section i have a whole bunch of other instruments still to repair waiting in the queue but this turned out to be a good one i'll see you in the comment section

Channel: The Signal Path

Views: 27,555

Keywords:

Id: TYFt-vlXcew

Length: 25min 0sec (1500 seconds)

Published: Wed May 26 2021