Hi welcome to Te Down! Tuesday Got a real interesting bit of kit today CU It's not something that you'll find in your average Lab They're pretty high priced and they're pretty obscure. What is it? It's a source measure unit or SMU or better known as a Schoo and this is an Agilant one which we'll take a look at. There's not too many companies on the market which are make these. uh Keith Le are one of the main players.

Agilant do a really good one. Uh, yoka, go do another one and well, you know there's not too many other players out there. What an SMU or Schmo Uh does is it's basically a pretty accurate uh voltage and current Source Four quadrant source which I might get into in another video and basically this one is a 6 and2 digit voltage and current meter and voltage and Source meter built into the one unit. What are they useful for? Well, uh, things like a semiconductor characterization.

If you want to, uh, characterize, say a a lead for example, then you can feed it with very accurate currents. From this, you can measure the voltage and then you can couple that into Optical measurement units and measure the light output and characterize it. All those sort of semiconductor characteristic curves which you find in the data sheets fets, you know, transistors, other types of Uh devices. How they get those characteristic curves? Well, it's using one of these news generally, so they're really handy bits of kit, but not something you'd find in your ordinary lab.

So thanks to Agilant, they've loaned me, uh, two of these actually? um, one that they're going to actually let me take apart and well, they think they might have to recalibrate it after I take apart I don't I suspect not I Don't think it's that critical, but hey, that's just their policy. Um I take this thing back I send it back. they'll get it re-calibrated so hey, maybe we can go as far as we want. So so there you go and this is a dual Channel unit.

This model is the Uh B291 12a comes in several models. This is the highest end unit. It's a dual channel one, so it's got the second channel on the back. There dual channels is handy because you can do stuff that you can't do on a single Channel unit.

It's got um, four terminal measurement, all sorts of stuff, fancy graphical LCD display in it and this one can go down to 10 fto amps resolution. So anywhere from 10 fto UPS amps up to you know, a couple of amps sourcing and measuring as I said 6 and 1 12 digit resolution on voltage and current so you can sort of do the same job as one of these with you some current meters and with with some uh, constant current power supplies and some 6 and 1/2 digit regular multimeters and network them all together and stuff like that. but these Sho are just all built into the one unit designed to cover a large range uh for and up to hundreds of Volts For example, I think this one goes up to a couple hundred volts as well for characterizing high voltage transistors. uh, for example stuff like that.

So really, really interesting and really Precision bits of Kit 6 and 1 half digits. 10 fto amps. Oh beautiful. You can just smell the Precision in this thing.

So you know what we say here on the E blog. Don't turn it on, Take it apart. So basically what we've got here is four Precision Instruments in one. First of all, it's a Precision current source as I said from 10 10 fto amps up to several amps uh and also it's a Precision voltage source or a Precision Power Supply Basically, anywhere from 100 NTS resolution up to uh, a couple hundred volts as I said depending on the particular model.

and it's also o Got a built-in Uh 6 and 1/2 digigit volt meter and current meter as well to measure all of those ranges. So you know, really precise we're talking. you know, 0.005% all that sort of stuff. Really high precision as you'd expect out of a 6 and A2 digit thing.

So four instruments in one and it can do uh, two terminal or four terminal measurements as well. And because it goes down to really low currents. then we've got things like uh and well guard so that you can get uh, triaxial test cables and things like that which we might go into in another video actually using uh, this thing In a real world scenario, if you want to take a look at it. in fact, using these is a little bit unlike.

It's a bit unusual. It's unlike a regular uh, power supply and multimeter which you typically use. so um, thumbs up if you want to see a separate video on that. Anyway, this is a dual Channel unit.

Uh, it's got the one channel on the front force and sense so two wire or four wire selectable and then on the back is just the secondary Channel exactly the same. It's got its own guard terminal plus Main's Earth terminal as well. Gpib interface little fan um, some digital IO stuff because these are designed uh, well, not this particular one, but Schmo Um. in general, often designed and implemented in production test systems where they manufacture semiconductors.

You know when they say in the data sheet that uh, this device is 100% characterized for its you know, bias, current or something, you know for its base current or something like that, then input current. Then they'll be using something like these Smoo. And you can actually get rack mount versions of these multiple channels. So you put them in a big 9in rack and you might have you know 10 or 20 channels of these things.

Testing a whole bunch of semiconductors tied into a big automated test system. So this one is more like a bench R&D use kind of one. but it does have those digital iOS if you want to integrate it with sort of like a desktop kind of you know, semi production test or pre-production test. or you know, R&D testing uh uh, scenario.

something like that. You usually wouldn't get one of these bench top ones if you're using them on the production test 4 in a silicon Fab for example. fully networkable of course. and as they said, it's got a big graphical display on it which can, uh, data log and plot graphs and do all the usual jazz.

Anyway, let's take it apart and in case you were wondering about the price, well if you have to ask, you can't afford it. this one's about $135,000 Of course this one is the Dual Channel fully optioned up one. So these things are not cheap, which is why you'll never find them in your typical you know lab. They just you know you really got to have a very specific need for this thing.

And as I said uh, before, four quadrant as well which I uh, explain in a separate video so it can source and sink Uh Current Excellent. Very useful bits of kit kit and this should just, uh, slide out I think cuz there's only two screws on this I'm liking this new agant gear that just slides open anyway. all right I've taken off the handle here and that should just Tada slide out. Oh, look at that.

Oh, look at that. Oh beautiful. Well, we can just smell the engineering that goes into this. Ah, this beautiful board on the top here.

Beautiful on the bottom, as well as we'll take a look at, looks like it's duplicated there. So uh, of course you'd expect uh jeel boards in this thing. Um, one for each? Uh Channel Beautiful shielding on the top here. As you'd expect, lots of beautiful shielding.

Lots of lovely cable routing going through here. That's for for the main switch. No, no, that's a soft, uh, soft button main switch on the top, but that's uh. I don't know.

they got something running from the back right to the front there. Oh, it's just lovely, lovely, lovely, lovely. Might not be easy to take this sucker apart though. so uh.

anyway, this could take a bit of work. Check out some of the interesting stuff we've got happening around here. Obviously some high voltage, uh, cam caps here. Three paralleled up like that and then four of them.

four Banks of those in series for a total of 12. uh, ceramic caps there high. They obviously doing this high voltage. We've got the bleeder resistors across here.

or balance in resistors across here. I've done uh uh, videos on these before B Balancing resistors across the individ. These are high value. These are.

yeah, these are 10. Meg Obviously, this is part of the high voltage power supply as I said couple hundred volts output. so clearly they're doing really buil and braces there. We got ourselves some uh Power diodes so they could be some protection.

uh, something like that. Don't know what's going on under this bit here. This is interesting that it's got its own, uh, double sided taped uh film on there to actually protect it. That might be interesting to have a look under there and then of course.

uh. testing to these things is a big deal. So this is for. you know, this is a built-in test connector for pre- testing.

Uh, for example. That's why you pay a lot for these. They're not made in high volume, they're engineered incredibly well, and uh, you know it takes a lot to actually. uh, test and characterize these.

And then we've got some other interesting stuff. couple of big Power resistors here by the looks of it. and look at the huge amount of huge number of Vas we've got stitching going there. so obviously they're trying to get.

um, that's most likely not for low inductance that's trying to get the heat down to the other side. so they're using those as heat sinking. And then we got more power devices over here. more power resistors.

They could be sense. oh, maybe not sense resistors I don't see, uh, any sense wires actually coming off them or anything like that. so we'll have a look. uh OB probably part of looks like power supply.

uh, 3.3 volts there happening. So maybe some part of the linear power supply so they're relatively uh, you know High values they're 10 Ohms a pop and power resistors. so maybe they could be doing some uh, current sense happening on those. For example, for the inbuilt, uh Power Ciz I think that's part of the output, uh, circuitry at all.

And there you go I missed the labels there. plus 245 Volus 2455 volts across that entire network. So we're looking at 500 volts there. That's why they've uh, you know, parallel their series uh, and parallel stacked these capacitors they needed x amount of capacitance on there for whatever reason and well, the only way they could do it.

They needed ceramic for the stability and the only way they could get that was to parallel them up and then series them up for the high voltage. And if you're wondering about that circuit there, that that's just a quad comparator LM 339 So obviously they're just doing, uh, some testing stuff and maybe getting some past fail results out of here or something like that. And you can see on the bottom side board here, exact duplicate of the top board. So obviously one of these large fullsize boards for each.

Channel That's why you pay a lot extra for the two channel units. Not like it's just building. They've got to build in this whole extra board for that. So that's why I pay.

you know, a couple of thousand. At least a couple of thousand more for that compl complete duplicate. We've got another board here. It looks like, you know.

processor big BGA device on there that's probably running the OS. That's that would be the embedded uh computer would be my guess. We'll take a closer look at that and then we've got some, uh, more processing on the back here that looks like the, yeah, that's the Gpib board dead giveaway. There you go.

it's all in the one board Gpib there. That' be our Gpib chipset or go curiously. uh, going down into the main Channel board down in here. So that could be just using that as a convenience and then bypassing that sort of stuff.

maybe through to the front and coming out these ribbons here and into the main processor down here. and there you go. That's just as I said plug-in processor module. And there we go.

It's a Uh spear, um, Arm processor of course we've seen before. Very common. We got a Zyink Spartan Fpga on there, so that's presumably doing all the main processing running all the main OS and whatever. Windows C embedded.

uh thing. Very smart to put that on a card. They can just upgrade that later cuz all this analog uh stuff doesn't go out of style. You know they want to manufacture this for the next 15 years so this stuff doesn't go out of style.

but this stuff might. They may have to upgrade this at a later time and I figured out why that other uh, metal work on the back had that plastic on here. It's just a duplicate of that, so obvious. L They're using it on that side, plugging into that uh card Edge connector on the front panel.

but they're not doing it over here. but they're just duplicating the metal work on the other side and that's why it's there. They've just got the same part and it looks like we might have a relatively easy time getting this board out because look it, It obviously undo a couple of screws and then it slides forward like this and then pops out. We've seen this, uh, construction before and that's at a couple of points three points along there, so hopefully undo a couple of screws in here.

This board just slides forward like this and then lifts out once you undo the ribbon cables. Curiously though, I Do see a bodge and it looks to be a ground bodge and it's solded directly down. They scraped off some of the uh, solder mask down on that screw terminal there and they've taken that snaked it obviously all the way through over to this point over here which has a diode which will probably be able to get in and have a look at. but yeah, bit of an afterthought there and I gave that a little tug forward and it looks like it is going to pop out exactly as expected, hinged out.

Yeah, we got the wireing that's connected in there like that. We should be able to get that. oh you know, probably best just to undo the connector there and there's a couple more under here, but we should be able to get those out and just lift out the board. Br Brant Engineering look at that and that is just beautiful.

Look at this heat sink here. Beautiful. RFI Gaskets on the bottom of this. so when this flips up, goes in there makes excellent uh RFI contact with the shazzy down in the center there.

We should be able to do likewise with the bottom board. that'll just sort of like pop out and it'll just have this Center shazzy held in there. Oh, it's just gorgeous. It really is so really the only Iio on this board.

this wire coming along the uh top here. that's obviously the power connector going to the board, so that'll be coming out of the switch mode power supply which we'll eventually take a look at that's in its own well shielded enclosure. Here it has to be. and then all we've got is the output connectors here.

This is for the second channel on the back here. so this, uh top board. If you bought the single Channel unit, of course none of this would be populated inside here. You wouldn't get this top board and you wouldn't get this.

um all the connectors on the back back panel which pops out in its own little uh, own little back panel there. so they just put a blanking plate over that for the single Channel model. but there's nothing basically. And then as we said before, all the all the stuff goes from the ribbon cable up here over to the uh main processor board.

over well the front panel board which is not only used for the front panel, the display and the connectors and stuff like that. got battery back up in there as well but then that obviously go. There's a um, then the processor orders on the bottom of the unit which uh then plugs into that front panel. So really very simplistic interface and that's how you want to design it.

You don't want to have cables running every which way over the shop. DC Input: You got your IO connectors and you got your connection to your processor very well thought out. It shows that they were thinking about the system design of this thing when they actually laid out this board. It's not like you know some Guru designed all of the analog stuff on here and when I'm just going to lay out this board and I know what I'm doing and then just whack it in there and then have to run cables all over the shop.

No, they really thought about how it was going to be integrated with the front panel and everything else. Fantastic lot of engineering hours went into this. Let me tell you and our Channel board here. Pretty easy to understand just looking at the physical construction as you can usually do with tear.

Downs Like this look, we've got our DC input one over here. Coming from the DC to DC converter, it's not marked, but it might be. you know, 12 volts DC or something like that looks like it's fused. got uh yeah? big nice uh inductor on the input.

Everybody's happy. uh big high voltage Step Up DC to DC converter here cuz as I said, this thing has a maximum output voltage of like a couple of hundred volts like 200 odd BTS that that's why they needed to generate that high voltage we saw before. here we go. We need to generate the couple of the plus minus uh, 245 volts on here for example, that ultimately then goes out.

So here's our power supply under here. This will all have the local voltage regulation as well. You know, 3.3 volts or whatever to power all the digital and the other analog stuff under here as well. We probably have plus- 15 volt Supply There will be very low noise linear Regulators of course.

local regulation. Uh, when you're talking about currents like you know, really 15 F amps or whatever. it is absolutely incredibly low, so you really have to look at low noise. uh, local linear Regulators But anyway, high Voltage Regulation: Step Up DC Todc converter.

that's the output filter in that we're generating now. plusus 245 volts. There probably got some sort of load or something there to keep it, uh, to keep it all. Uh, nice.

And then we've got some other control or Lo or regulation around here. No, it's saying plus 20 Vols here. So yeah, plus - 20 Vols So there you go. there's our local regulation I Don't know why it's sort of all the way up this end because I would have expected the high voltage to flow from over here and then power the Um output, um, side of it.

but anyway, uh, they tapped on. They got plusus 20 volt rails there, most likely for the Uh analog, all the analog stuff, and then because they don't need any of that, uh, shielded at all. Which is why all this stuff out here is just basically the power supply to get the thing ready to generate all the high uh, all generate all the precise voltages and currents which are of course under this completely shielded set in both top and bottom. And obviously because it can generate a significant I Think it's 10 amps at a couple hundred volts, you know, so it's significantly.

wattage is maybe not that high. it's not a couple hundred Watts but I think it's like 30 watts output capability, source and sink. so you need? That's what this heat sink uh channel is here for doesn't look like a heat sink, but there it is. it is a heat sink.

They've got some uh seal pad underneath that so that big aluminium Extrusion there is connected to your power devices which look like they're on the top part of the board here. So these are the screws likely that hold down those uh Power transistors in there, coupling them down into the heat sink on the other side. And of course this is where your thermal management comes in. Here's your fan at the back here? Okay, which got some shielding tape over that? look at that.

That's attention to detail for you don't want to muck up your um Emi by you know, having that big fan, that big path in there where things can get in and out of? No, they've shielded that puppy. Anyway, we've got ourselves a fan guide here top and bottom. so this one and there's obviously a mating one on the other side cuz as I said, there's another channel in here for the other channel board. But here you go.

there are probably I'm not sure. yeah they would be uh, blowing out so air would be coming in the uh side? Yeah, air would be coming in the side of the units here I suspect and then being blown out the back I think is the way it's uh uh. set up yes, air on the side of here. the case, as you may have notice right at the start, had the vent holes on the side and then sucking that through the heat sink and coming here.

So that's why the heat sink is channeled is open like that and is channeled like that because then all of the air will be sucked right through the heat sink like that. Very efficient, beautiful thermal design. You got to think about that in something like this that needs you know, 30 50 watts of sort of power management inside. Uh, you know, a reasonably, uh, you know, smallish closed case like this.

But as for the heat sink though, it's interesting that it is on this side. the power devices are on the other side. It's not like they mounted them on the back side of this board with the tabs directly connected onto here. so they're relying on the screw going through there to get thermal cup.

they're relying on the seal pad. They've obviously got some exposed sections of the board, exposed copper on the board, and they're probably using hundreds of little Vas in there. Um, as we here, of course, hundreds of little Vs to get uh, power through to the other side like that. Not terribly efficient.

not the most efficient way to do it, but good enough. And when you constrained in in this sort of you know, case size and things like that. a lot of thought would have gone into that lot of measurements. lot of testing to make sure that that was adequate and now we'll lift the skirt on the hey, there we go the uh source and measure part of the unit and look at that.

There's a lot of nice stuff on there. o I like those, look at those TDK Lambda things. Oh, they're very nice that looks absolutely beautiful and some lots of guard traces happening around here. As you'd expect, let's go take a close look at those cuz you need those down when you're talking 15 fto amps and we' got some really nice Precision lowle measurement stuff happening around here.

Take a look at that that looks for all the world like a coaxial re Reay there. Fantastic. We'll take a better look at the part number little tiny Qfn down in there. What a pain in the ass.

But look at all the gold tracers. These are all guard tracers and I've gone into guard tracers on previous videos, but the reason that they're gold is because the solder mask has been left off and that when you put those around a component or around a node in this case, we have a node here. For example, they don't want any surface leakage into that node so that guard Trace that gold guard Trace will be kept at the same voltage potential as the point that they're actually uh measuring so that any leakage going into here won't affect the critical node right in there. And this is one of the reasons why they think they might have to recalibrate this cuz if I uh, put my greasy fingers on there and well, you know guard tracers can only work so much I can actually probably affect the calibration of this thing.

if I started, you know I'm spitting as I'm talking and it's all going all over this board, all over the shop. But uh, yeah, probably shouldn't have to. uh, recal this thing. but this is where all the critical measurement stuff is happening.

And check out that is that a capacitor there? Look at that. It's got its own little isolation slot under there, so they really didn't want any leakage from one side to the other there. It's not a high voltage isolation slot. that would be a leakage isolation slot slot for sure.

And there is our Re relay I believe and you can see. look the shield around that. I mean there's our input and our output contacts there and that sucker is shielded brilliantly. and uh, then we've got our coils coming in on the bottom so that'll probably be magnetically and electrically shielded as well.

and just at a quick uh Google I can't find any info on that. That part number does turn up as a relay, but it's not one of these sexy read relays so I can't immediately find the data on that, if I can I'll link it in. And then we've got our Kelvin connection relays here. and here's an internal diagram showing the physical connection of how the voltage and current sources and uh, the four terminal measurement is connected on the output and likely high voltage Precision 10 Meg ceramic resistors in here, that probably for the high range current measurement shunts.

And if you're wondering what that little Qfn is, well try and decode that sucker. And just around that, we've got a couple of Fairchild Fds 8978 en Chanel mosfets down in there so they'd be doing some switching and it looks like we got some protection happening down in there as well. We got ourselves our four D they could be you know Zena type things in old style m uh cylindrical package which I love. They would be incredibly low leakage of course.

bet your bottom dollar probably cost a few bucks a pot. And clearly we got some Precision current shunt resistors down in here. There we go that looks like 50 Milli ohms down in there R5 there 50 milliom current shunt. Probably another one there, but it's hey, it's upside down.

Woohoo! Back to front. So there we go. Um, that would be are they tapping that off? Is this the for that? I don't know. Are they tapping it off there? Looks like they might be tapping it off there and going off, but you can certainly see these are obviously mosfet switches.

You can tell by the Uh pin out and also the you know fact that big tracers are running into the multiple pins. So they got some high current switching and this is another current shunt resistor of Sun description with the two Taps coming off there and then you've got your differential signal heading over there to some sort of precision instrumentation amp. and for all you vault nuts out there. Yes, classic linear technology LM 339 They're still making these things.

They're still using everything. that's our Precision voltage reference for this thing. So yeah, it does have that weird ass part number 1826 D 1249 but that is basically a uh, rebadged LM 399 manufactured by LT And you also notice yeah, of course this is going to have extensive build-in self test. So yeah, various tests signals.

What? Those test signals? uh, actually are you know Precision uh DC voltage sources or something like that? I don't know. but yeah anyway, all of this is built in and that's what adds to the expense of some really high Precision bit of Kit Like this and needless to say they haven't mucked around on the capacitors. Quality: Nichon brand 105 C-rated yep, Spared no expense and these beautiful looking TDK Lambda Top quality uh bricks Here these are DC to DC converters in this case uh 9vs DC in Uh 5 Vols DC out isolated, so made in Japan They really haven't mucked around here. You can buy these on Digi key.

These are about $16 each in uh 10 of quantity. And really, you know they might only make you know 10 or 50 of these units at a time. In fact, they might even make these units to order. Uh so to speak.

You order one of these and they might make I don't know 10 or something and and fulfill the order. But yeah, very low volumes for these type of products. And really, they didn't want to roll their own so they sped expense doesn't matter in something like this, you just want something that works. a designer's Time shouldn't be spent around on an isolated DC to DC converter like this.

If you need it, go to Someone Like You Know Top quality manufacturer like TDK Lambda get top price off the shelf uh unit. It's fully characterized. You got the full data on it. You know exactly how it's going to perform.

Guaranteed whack it in there. Job done. You're better off spending your time on all this fancy whizbang stuff over here. And yes, that is a xylin Spartan six on here.

What is something like that doing on something like this? Uh Precision device like this current and voltage source and measure unit? Well, it's got arbitrary waveform capability as well, so I suspect that could have uh, something to do with it. Perhaps another little xylin part over here, complete with that separate J tag headers for each one. So curiously, you can see the you can see the parallel traces running out of there through little are they little RS and C's there I can't quite see on the screen into that little beasty over there. Is that like a parallel? uh, Dack would be my guess.

So maybe this is generating the arbitrary waveform? a parallel data out straight over to a Dack Let's look at that number, see if my hunch is correct. I could embarrass my myself here. and no, I wasn't uh, wrong. These things were pretty obvious.

There you go. It's a TX Dack in there. Pretty decent one. going over to This Ti Opa 1611 opamp.

No surprise for gessing, that's pretty darn. Schmick In fact, it's designed as one of those Audio Level Ultra ultra low Distortion ones .2% or thereabouts. I mean ah jeez. I mean you fart halfway across the solar system and you're going to generate more Distortion than this thing has.

Unbelievable and then no surprises for finding a digital isolator sort of coupled onto that Fpga to get uh Digital Data in and out while being completely isolated ISO 724. And then there's the other. So we basically got that coupling between the two Fpgas over here. If Fpga doing the arbitrary waveform stuff, maybe some other, uh, housekeeping stuff, who knows, but that is digitally isolated from this Fpga which presumably dries all of the control side over here.

And this is what this isolated power supply is. powering. this section down in here. notice the white line there.

The PCB design has been very kind to us and showing how they've separated the two halves. What are they? Well, they've even labeled them V sense is up there. so there's our voltage sense path or voltage sense Channel up here and here's our current sense Channel all the way down the bottom and these have to be the Adcs. Let's go in and take a look at those and sure enough, both of those are a real high-end ADC A in this case a Texas Instruments Ads 16751 Let's go to the data sheet I Think you'll be impressed and look at this: Beast 4 Meg sample per second 24bit Delta Sigma analog to digital converter.

Here we go: 103d dynamic range at 4 Meg samples per second and down to 125k where they may be operating I don't know if they would be sampling at the full full Meg sample here. they may be using it at a lower rate to get H better performance, but 111 DB dynamic range - 107 THD fantastic p 3 PPM DC Accuracy Oh beautiful. Not only 4 PPM per degree SE gain drift It's got programmable digital Fi filters. Uh, the bandwidth is 1.7 MHz so it's uh, right up there at the uh Nest limit.

Really? Pretty much so. That's uh, quite impressive. I Like it and it draws half a watt. And of course there it is.

test and measurement. That's what we're using it for. Got ourselves an interesting little resistor Network there by the looks of it, probably a Precision resistant Network and by the color of the green I'd say probably the same man facturer as these Precision resistors over here. And once again, we got another one of those Qfn packages the same as we had uh down in this channel over here and of course more guard tracers going around there.

don't want any leakage at all into those nodes that they're measuring and down in there. No surprises for finding a Precision opamp. It's more than a Precision opamp. It's an ultra Precision opamp uh 30 36 volt Supply 2.8 NT per root Hertz uh.

dual opab Analog Devices ad 8676 Nice. Then we've got a couple of more Analog Devices Parts there don't know what they are and of course, some uh switching you'll find in here for uh, range changing and stuff like that. And back over to the Fpga here. But going back to the digital side of things.

how it gets all that information back to that main processor board. Well, no surprises for guessing that's galvanically isolated too. We've got another one of these digital Opdo isolators here. couple of dip switches I'm not sure what they're setting there, but you can see that the silk screen that all of this stuff right up here is all electrically isolated all outside here.

That's what that silk screen is designating. and if you have a look on the bottom where it's going there, we go. There's some more what looks like probably some more opto isolators there bringing data over to those ribbon cables which then go over to the main main processor board and all of that designated by the silk screen. Electrically isolated.

And let's flip this metal plate off here and see what we get underneath here. If we can get it off. There we go. Oh, that's all very nice and integrated.

And today there's all our power devices. all uh, surface mount ones. Then I Thought there might have been maybe even some uh to 220s or something on the bottom going through those um, screws actually getting that heat through. But no, they're all surface mount and they don't have a huge number of Vas around them, just a a small number I'll get in there and show you that closer up in a second.

But anyway, we've got the various supplies coming in here, so if we start at this one here, you can see that we got plus 245 volts here plus 35 volts here plus 20 Vols and then the other side of this. then we' got the negative supplies going in - 245 - 35 - 20. So various supplies for the various Um output ranges and they'd be switching those through uh, depending on which range you select. So there you go.

That's rather interesting and very symmetrical for the positive and negative supplies, but that's what you expect cuz as I said, this is a four quadrant device, meaning it can source as well as syn current. so they've got to have those devices symmetrical on the power rails like that. And this thing has current sensing coming out the Wazu as we saw on the top side there, but we're basically being fed in here. Look, we got four 10 ohm power resistors there paralleled up uh, giving you 2.5 ohm current shunt resistance there and then that's being tapped off by I Believe that little tiny uh so 23 uh, that would be like a highs side current sense uh amplifier and there's probably more current sensing going on in all parts of this thing.

uh, plenty on the top side and plenty on the bottom. that's not for the direct output Uh current. that's just for, you know, self testing and all that sort of stuff just to make sure everything's working fine and no surprises for finding a real beefy quality high voltage end. Channel Mosfet on the positive Supply rail in this case, the Fqb 860c from Fairchild and it's a 600 Vol 7 half amp 1.2 ohm on Resistance mosfet.

Very nice and as you can see, they got four of these paralleled up the drain connection which is the tab there. They're all connected in parallel, but they've of course got their own Uh current shunt resistor down here. So they're using these as a uh, basically a constant current uh, load. That's pretty much what it is.

and there's the current sense resistor side of that uh, they've got some extra rain switching in here here, which is rather interesting .1 Ohm resistor here and then two separate 1.0 Ohms, this is a duel uh n Channel mosfet that's a Fds 8978 and I presume haven't decoded that number. but I Presume: that's a little uh mosfet as well. So they can switch in any one of these current range resistors and probably might have some uh, current sense amp or something here cuz they're tapping these off so you can see the uh, um, the value coming being tapped off that thing. And that's how they're able to uh, sense and maintain the precise current.

So we've got our Big N channel uh, Power device up here and then we've got our uh, lowside uh, current range setting down here and then that's duplicated all across there four times, all with the same values and same components. And here's a quick little Dave cat of how that all works. We got our Big N Channel power mosfet up here of course that dissipates all of the Power all, well, a quarter of it anyway, because we've got four of these whole sections paralleled up like this. So This Is Our Big uh Power device which uh SNS and sources our current here.

And then we've got our uh range switching down here. So our current uh, sense resistor as you've seen in my, uh, constant current load. Similar, basically similar sort of thing. except we' got some extra mosfets for switching there.

But of course, all of the current does actually flow down through here. It must flow through here. It can't flow out in the gate, so it's got to come down here. So these mosfets down here.

while they're lower power and lower voltage, uh, they still have to, uh, handle the maximum current. We saw this was a a 7 amp rated mosfet up here. These and no surprises. This Fds 8978 is also 7 amp rated as well.

So of these three range resistors here, you could get a reasonable combination of current range values and of course then there I didn't show that, but they would be tapping off the value from these current sense resistors. And then that's how how they're able to uh regulate and maintain the current. And they're doing that paralleled up across all four sections. like that.

Really quite nice belt and braces. they've gone to town and then on top here. flipped upside down so all the electrons are going to run down to the bottom. Get better performance like that? You know, higher power dissipation.

Anyway, Um, this is a P Channel Mosfet. This is a Uh 25p O6. So it's a 60 volt 272 amp mosfet. so it's a real beefy one.

So obviously they're switching on the main Supply there for all of these. Um, these are your Uh load resistors down here, but this is just effectively acting as a switch for all of your current flowing into all of these fets. Now the product spec for this thing is only Uh 3 amps Max continuous DC current. but it can handle Uh 10 amps uh, constant current uh pulse current sorry Um, so really you know this thing is overrated in in terms of Uh and and so are these overrated in terms of Uh current handling capability.

But that's really not the Uh main factor here. The main factor would be the on resistance of this thing and especially these the power dissipated. and yeah, they probably could have handled this with just one Uh device or maybe two, but hey you spread that heat across multiple channels becomes much more reliable and then over here we' got ourselves another P Channel mosfet. but in this case it's a Uh 500 volt unit so another high voltage one which can easily handle uh, the full Supply capability of this thing.

1.5 amps. Uh, that is an Fqb 1 P50. So it' be interesting to sit down and figure out the exact Arrangement that they've got here for the Uh positive half with all the Uh switching and load arrangements and current sensing and uh, all sorts of stuff cuz it is a little bit involved I'm not going to sit down and do it now, but if anyone wants to go through the exercise by all means, there is where the split is going to happen. Everything on this side is going to be for the negative Supply rail.

so it'll be an exact mirror image of what's happening with all these devices over here. And of course you know we've figured out our main load resistors and we've got some. We've got a low voltage and a high voltage switch on this side by the looks of it. Not sure what's going on around here, but yeah, it's very interesting.

Arrangement So I' would love to see a uh a schematic or theory of operation for this uh thing because as I said, it is a four quadrant device and you can tell that these ones up here are pretty much only Uh switches because of the limited amount of heat sinking on there. So of course they're over specified for this purpose so they're not going to dissipate much power if they're just you know, slamming these things on hard and then uh, just switching the current through to these devices. These are the ones. You can tell that there's not much heat sinking because there's no Vas around these that are dissipating the heat down into the other layers and the heat sink.

They could have done that, but obviously they don't need to. So it's only these devices here, which are, um, taking all of the current of course and you don't know what your load's going to be. So these things have to dissipate a lot of power. That's why they've paralleled up four of them and they've got that via stitching.

I Expected more via stitching than that to get the Uh to get low, lower thermal impedance through to the other side so that they can get to the heat sink. But these are, you know, very over uh specified. So um, you know it. It does the job.

obviously. Maybe they got some Vas there underneath the actual Uh devices themselves. that's usually not recommended cuz that leeches, uh, solder away and they often don't solder all that well. So, but I'm surprised there's not.

You know, hundreds of the damn things like there was up here. See those there? We go there. We go bloody 100 each side or something like that. Crazy.

So I'm just a little bit, uh, surprised at that. they didn't Uh, Guild the Lily there. We got a high voltage isolation slot down in here just to protect it from this other lowside stuff. And as I said on the negative, Supply over here, it's going to be exactly the same configuration as before, except they're going to have everything reversed.

so the N channel mosfets are going to become P Channel Mosfets and Tada. This is a bit surprising though. not the same brand uh Fairchild So they've got an Exis um 10p 50p in here and once again, but it's a similar spec component to the positive side. it's uh, 500 volt 10 amp.

It's a little bit higher current rated 10 on Resistance so very similar style mosfet and of course those ones up there they were P channel before they will be n Channel now the 30 and 60s I don't even have to look those up. Check this out! I'm glad they've got the guard rings around here because look at the residue left on the board there. H Pretty horrible. You definitely need the guard rings when you're manufacturing process leaves behind that sort of stuff.

Not a huge deal. I Mean the guard rings are going to do their job, but you know you know it just looks untidy and you might remember way back in the video how we said as part of the uh, really nice systems engineering, they were having the ribbon cables comeing over this side and here they are over here and they're just routing those right across the board all the way all the way over to this side and you can see the isolation. Uh well, the ground is still the same. You can see the silk screen Mark right around there so all of that is at the same.

All this stuff around here is still at the same potential. So this is all the digital side potential of the board so they're just running those right across instead of having daggy cables running throughout the thing. but that would have been easier from a system design point of view. But no, they've integrated them, integrated them onto this board.

Very nice. All right. If you're anything like me, you want to see underneath that heat sink. Not this.

The only thing as I said before is going to be they're just going to have some gold, uh, exposed. uh B pads. A big thermal pad on the other side. so I've undone the heat sink.

but H yeah, it's not going to. It's not not going to come off without a fight. Aha, there's nothing you can't fix with the application of more. Force There we go.

Tada each one of course electrically, uh, isolated, but nice and sticky. and aha, there you go. It was yeah too few viers under there. They have completely popular ated the bottom side of the bottom side of that dpack there with V's look at them.

Ah, probably 100 vs in there or something. And as I said, the disadvantage with that is that the solder does flow through to this side. but that actually can be, uh, beneficial because then you can get a lower thermal resistance coming through. So eh.

Pros and cons both ways. Sold a leeching money shot. There you go. So that's the there's the thing I mean you look, we've got bubbles under there like that, so not the greatest thing because that means that you're not getting.

There's going to be air gaps around there where this, um, thermal. this seal pad is okay. it does conform around these, but not entirely. It is going to be better if you're going to have a completely flat surface on there.

This one's not bad. it's You know it's reasonably flat like that of course. Pros and cons both ways. Um, if there's soldar leeching coming through of course means less controlled solder on the top side of the device because you got to apply the paste.

and if, well, you know a lot of it's leeching in through here and you don't take it into account, then you might get a poor, or you know, insufficient solder on the top side to hold down your device. But on a power device like this, there's usually enough Pace to go around. But anyway, um, some people like to do that, others don't Others would have uh I probably wouldn't have done that I would have uh, had all the Vs around the outside uh, part of it like that and not had the solder Leach through to the other side. but eh, each to their own.

And of course, if you're wondering why they've got these plastic spaces here, uh, holding down the heat sink? well, it's electrical isolation. Of course you can't have that heat sink. Well, well, you could have it at some live potential, but generally, uh, you wouldn't. That's why they got the seal pad on the bottom side.

It transfers the heat, but it leaves it electrically isolated. And that's why. Uh, because all these devices need to get their heat out separately. And apart from these four devices.

Um, they're not paralleled up at all. so they all have to get their heat out separately. So that's why the heat sink is at no potential whatsoever. It's just floating.

In fact, they've only got the one screw down here actually holding it down and making electrical contact through to the PCB. But in this case, they don't seem to have anything going to it unless there is, uh, a trace on the top side. They might be holding it at some nominal potential, perhaps. But anyway, you don't have to.

All right. Well, that's enough for the main board. Let's take a look at the rest of the shazzy here, as, uh, the front and the rear terminal connections. This happens to be the rear one there.

Practically identical. Looks like they've got uh, shielded, uh, coax coming down to the Uh coming down to the rear connector here nicely. all nicely heat shrunk. There's one of the sense wires going off there.

We've got a nice big Ferite bead in there, a Ferite clamp sorry in there, and uh, they're tapping those off very nicely so that so you can see the guard terminal here that's actually connected through to the force line. Yeah, that is the force line there. So the center of the force line. So the guard terminal is actually connected to the shield of the coax.

and then that goes through that Ferite clamp up into here like that. So that's what the guard terminal does. and then you can extend that guard terminal further out on the other side with your uh, triaxial uh leads? Your you know those test leads that cost you know, a th000 bucks or whatever. You in fact, I might have some in the box somewhere.

Let me have a look and yes, what do you know in the other box with the other unit that I'm supposed to use for uh, testing? Yes, we got the triaxial leads and some component test fixures. These component test fixures probably cost more than an oscilloscope and all they are is a metal box with, uh, you know, a nicely shielded metal box, got silk screen on the front and um, basically a shielded component test box. So you put your component that you want to test inside here and you hook them up with these shielded triaxial leads. And here's a triaxial coax connector if you haven't seen one in the while cuz they're not used too often.

they're used in these, um, low leakage, low current measurement systems. but apart from that, you know not a huge amount else. Any anything that needs real critical shielding, you're going to use these triaxial coaxes. It looks very similar or regular coax.

in fact, from the outside. Without looking at the end, here, you really, uh, have a hard time telling that it would be a triaxial type. So we've got the outer shield around here, which is, of course, connected through to your regular uh, coax body here. and that's going to be connected to Main's Earth So that one, uh, you know, ground, uh, physical.

Uh, Main's Earth ground. so that one keeps out the 50 Hertz hum. But that's not good enough for these Um measurement systems. You also need another secondary Shield inside there.

So it's physically going to have its own braid inside the coax and that's connected through to that inner ring there. and that one's connected through to your guard terminal. so that's connected to your um your measurement system common inside the instrument and that's very important. And then finally, you've got your test signal in the center conductor.

And of course, this thing doesn't actually have triaxial coax connectors on it. It's got those banana plugs, but that's what this is for. You Just plug it directly in like that, push it in, locks firmly in place, and Bingo! You got your triaxial coaxes in here so you can connect those through. And of course, you can see that the outer uh box.

Of course, this whole box here is connect and the outer part of the coax is connected through the your Ms Earth So That stops any Mains hum getting in and then inside that inside your triaxial. the guard inside. well, the coax inside there. The Shield Inside there is connected through to your guard terminal which is your circuit common inside your instrument.

So therefore, you get a double shielding on this thing 50 HZ plus your protection plus your shielding through to your ground inside the instrument. And that's how you do proper measurements on these sorts of very low current bits of gear. So you'll find these triaxial connectors and these sort of uh things. Usually, um, Fe really precise stuff.

They'll have a triaxial connector on the front panel so all of your Uh High Resistance measurement units like say, the Uh HP 439b or something like that, or your really low current measurement units like some of the Keithly gear and stuff like that. And for this one, which goes down to 10 uh, Fto amps, this is what you really need if you're measuring down at those sort of currents and those sort of leakage levels Really important. And these are these aren't just regular coaxes, these are. There's a reason these are expensive.

They you know you can say that they're not worth the money, but when you're testing a measurement system, you really need these. You can't just use ordinary coaxes. they're just too crap, too lossy. You need the proper measurement leads, so that is the proper way that you would, uh, test your component and your component is going to sit inside this test feature.

Hey, hey, we have a We we already have a lead in there. There we go. So I'll be able to do a video uh, playing around with components here so you feed in. It's just got a little, uh, internal wiring diagram.

You can connect your lead or your transistor or whatever it is you want inside this thing and Bob's your uncle. It's all shielded. It's all connected through the outer boxes. You can see that the wires coming up there that's all connected through to your Main's Earth keeps out the 50 HZ Mains hum.

And then inside that, you've got your guard terminal going right up to your test connector on there. Fantastic! So that is the proper way to characterize a lead or a transistor or something other semiconductor device. And this test setup, even though it might cost thousands of dollars, let alone the $133,000 instrument, is how you get those characteristic Curves in the data sheet that you see all the time now. I don't know about you, but I have no compulsive need really to, um, get out and have a look at that main front panel board.

There's nothing much on there, it's just driving the display and the and the connectors. and uh, the front panel keypad and stuff like that. Not very interesting. It's got a battery back up there for the real time clock and it just basically connects uh, the ribbon cables through to the main Uh boards and then through to the main processor board as we saw before.

so I won't bother taking that out. And likewise, I'm not going to bother taking out the Um second Board second Channel board on the bottom because it's going to be exactly the same as the one we just extensively looked at. But we do want to look inside the power supply so that Shield just came off very nicely. It all clipped under there.

Ah, screwed down Beautiful. beautiful, beautiful. Um single a job. very very simplistic cuz as I said, I think it's just got you know, one or two voltages coming out here.

Incredibly simple, so just a a universal uh switch mode power supply. Um, single-sided board looks very well laid out. no problems whatsoever. Once again, Agilant likely didn't do this.

it would have been a third party. There we go. we can just see inside there, they've got a label on there. third party? Um, it could be TDK Lambda Again, awesome.

Anyway, the output is just a single 24 4 Vol that looks like 6.3 amps I think so you really don't get much better in terms of power supplies. and TDK Lambda And you know, first class Azured expense, they' they've spared no expense. This thing would not have been cheap at all. but hey, in a $13,000 instrument, eh, it doesn't matter.

They're not pinching pennies. They got proper nichicon uh caps of course. The uh Mains wiring over here of course is uh, properly crimped and then Shake proof washed down to the side and they've got all the requisite input protection and everything happening. Beautiful, spaced out that it be super duper reliable and would last well, not forever, but a long time and easily repairable too.

So there you go I hope you enjoyed a look inside that high-end agilant B 2912 a Precision Source measure unit or Shmo as they're called cuz you don't get to look inside these things every day. And yes, it does look like it works. Oh 10 10 fto amps and 100 Nan volts resolution up to a couple hundred volts. Unbelievable.

Thank you very much Agilant! And uh, yes, I do have two of these units for a little while longer. so if you want me to do a video, uh, showing you how to use one of these things and playing around and maybe characterizing that uh lead they gave us or something else, please let me know. leave it in. the comments: it's boot in.

it's boot in. Will it work? Will it work? I'm sure it does. I'm sure they don't have to recalibrate this thing I don't think I spat on it too much as I was talking or didn't get my grubby hands on those uh, critical nodes in there. Anyway, they got those beautiful guard terminals.

should take care of it anyway. Hope you enjoyed it. If you want to discuss it, jump on to the Eev blog. Forum The link is down below and also if you want to see the highres, uh, tear down photos for this thing, the link is down below as well.

Um, but it's on the Evev Blog.com website so hope you enjoyed it if you did, give it a big thumbs up and thank you very much Agilent Catch you next time. It's pretty sad these things have to go back. This sex on a stick really?.