Hi welcome to Teardown! Tuesday I've had this one sitting around for quite some time. This is the Keithley 2400, SMU or Shmoo Source Measure Unit. and you've seen source measure units on the blog before, so I won't go through all the details. but this is pretty much the industry standard that source measure unit or it was because they've been superseded by the new whiz-bang fancy Touchscreen 2500 model.

So anyway, this is still a really great bit of kit if you can pick one up at a reasonable price. But they do hold their value very very well. and it's a point. Oh, one percent class instrument.

Five-and-a-half digits in this particular model, the 2400. There are others in the 2400 series, but this is like the general-purpose one so it goes up to 200 volts. output voltage are there abouts and one amp. so it sort of has that general-purpose capabilities.

Other models which have higher current capability, higher voltage, all that sort of jazz and these things do exactly what their name implies. They source/measure unit. They can source both current and voltage and they can measure current and voltage as well. They're generally use for our semiconductor characterization and things like that.

So if you want to get the like like a modern replacement for a traditional true curve tracer for example, you would have got back in the old days which can draw your our semiconductor characteristic curve A Via Transistor, your Dire Junctions all yeah, you know your T Vs devices Emma Diodes. Whatever it is, these things will pretty much test it. They're designed basically to be automated with our PCs and things like that, but they're really precision bits. a kit.

they really asked. So I Expect some really nice precision part in this I Expect it to be very, very well made. It is a very expensive instrument. They sell for many, many thousands of dollars.

Like the brand new, our 2,401 is like four thousand dollars or something like that's. got front and rear terminals. GPIB Rs-232 interface. Now that new fangled Ethernet rubbish.

this model light dates back quite some times. Some nice heat sinking on the side of course because this is I think it's a hundred watts our output capability on it. So yeah, we need some heat sinking on the side. So you know we say here on the Eevblog don't turn it on, take it apart and these are four quadrant devices and I probably been meaning to do a video on that and that basically means that it can source and sink current as well.

so you can. Actually yeah, so you could use them as a load effectively. So let's try and get this case off here. We probably need some more stuff on the bottom, but I reckon that's going to slide off.

There should slide over the heat sink like that. The warranty void if removes sticker is already broken on this one. I Got this at our auction quite some time back. I'll link in the auction video down below I Cut it with a whole ton of other stuff.

so oops, oops we're coming apart. we're coming apart. Look at this. it's looks very very nice.

Let's come on, you can do it. haha. we're in like Flynn Check it out. doesn't that look nice? And surprisingly I think I'm quite spread out.

The reason I say that is because look here's the front panel wire terminals over here. These will no doubt be low thermal tellurium copper contacts same with the ones at the back as well. You know it's spared no expense on the terminals because this thing does have full scale current range of one micro amps so you know and five-and-a-half digits resolution on that. So I think it goes down to you know, 10 nano amps or something like that resolution.

So it's you know. So we're talking very low current in here. not as low currents as the new model by the way. the 2450 goes and order or to lower than that I think.

But anyway, check this out though. this is the interesting part. We've got our wiring coming over here. We've got some relays down in there as you can see, we'll take a look at those in a minute.

But basically here is our output terminal right and here our output driver transistors right all the way on the other side here. I can show you the the basic general was schematic of this in a minute. but yeah, it's just like these are our these are our. output driver transistors.

Looks like down here is our output currents shut by the looks of it, where it's doing some measurement. That's by the looks of things. Anyway, we've got a main control over here which is not the main digital processor. I Don't think because there's a second board right down under this which has all the digital processing and stuff like that I Believe.

But yes, so that's some sort of maybe you know custom ASIC or something like that, some sort of, you know, the the high-end converter and DAC for example. So it's I Just find it interesting that's Holton completely spread out like this. We've got our high voltage driver here driving the transistors. I'll show you that in a minute with the different taps.

We've got some isolation slots along here. just some short ones actually cut into there just to differentiate the output driver around here with the rest of it. But yes, somehow has to get from here over to here and vice versa right across the other side of the board. Interesting.

Now of course this is a top shelf unit. top specs. You won't be find any corners cut in this thing in terms of devices use. So we'll get you know, analog, top-end analog devices, parts, linear tech, parts, things like that.

So yeah, it's really precision. Our Op amps and there'll be precision. current shunt resistors, precision range set resistors, all that sort of jazz, you know, and the capacitors will be top class. everything else.

So that's what you expect from a top-shelf Keithley model. And there's the rear panel jacks. and as I said, if they're not low thermal tellurium copper, I'll eat my hat I think and we've got that going through a choke here and they're headed all the way down to a couple of relays down in here. We've got some any C's and a Koto exactly the same for the front panel terminals here as well, and no surprises for finding a super high quality Koto 8200 Series reed relay in there.

couple of NEC jobs. They make terrific relays as well, but it looks like that's a there probably the high current read relay, so they're using that for the main outputs, which it USA USA USA Now here's a simplified schematic from this service menu which I'll link in down below of the output stage here so we don't actually have to take off these little arts sill pads against these are transistors to, you know, find out exactly what they are. we've got MOSFETs Basically, there's that going to be for MOSFETs in there and to power transistors as well. So that's why we've got two, four, six devices on the upper half and six devices on the lower half here.

And of course, as I said before, this is a four quadrant device so it's able to source current as well as sync. So we've got our main 225 volts supply rail up here. and then of course, it's able to source current out like this so. but it's also able to sync through these transistors down to negative 225 volts as well.

It's supposed to be a negative down there. and you'll notice that these are plus 36 volt plus or minus 36 volts. HAP's what are they for? Well, on lower ranges, you don't want to dissipate all the power up here if you only had your 225 volt rail. If you're down on the lower ranges generating, you know, 10 volts or 5 volts or 1 volts or whatever, you're going to be dissipated a huge amount of power across your power transistor output stage here.

So you want to tap in these lower voltages here. So I think they've got a 20 volt range or something like that. So that's what those two taps will be. Therefore, we've got ourselves a drive-in amplifier here.

Whether or not that's a discrete amplifier I Don't know, you'd have to have a look in. There could be a transistor amplifier? who knows. But anyway, interesting thing is is that they've got a cascode arrangement here. So these two transistors Q five one six and Q five one eight here are a casco configuration.

You can tell because we've got our sellers are basically a common emitter and then a common base joined together configuration and they do that for the extra output stability. the cascaded transistor arrangement with they're sort of I Guess you can call a hybrid arrangement because they've got the bipolar transistor and the MOSFET as well. but hey, you can do that. You know, a transistors a transistor, so that really gains them output stability.

The cascade arrangement is really known for its stability and of course, what you need. This sort of device? Really high-end. Really? No low noise. Really.

ultra stable. You really need a stable output driver arrangement. And of course you go into a ton of detail about exactly how this works until the cows come home. But I'll spare you the details.

But yes, suffice it to say that pretty much matches the physical arrangement here of the output stage. And you know, as a PCB layout person, when you're actually laying out this product, of course you'd have the designer designed this thing that's usually on a high-end product like this. it's not going to be the same person typically. So yeah, they've designed the output stage and you go right.

I'm going to do an output stage and I need to hook it up to a heatsink? It's got to be on the side here, of course. So, and I'm going to physically arrange all of my transistors like that because it, that's how it rounds out. I Don't know whether or not that's positive here and negative here or vice-versa I'm assuming that the positive is on this side, but hey, you get the gist of it. And here's our high voltage transformer up here generator presumably to generate the high voltages required for the 220 plus minus 225 volts and the capacitors.

Here, we've got a combination of a hundred volt niche econ here. Of course you get top quality praying caps in this so they're hundred volt and nichy cons for the lower rails and then we've got four I think four hundred volts or 350 volt caps up here. these are IC brand Illinois capacitor and with they've been acquired by Cornell DubLi R So you know once again, super high quality our caps in there. So therefore, the higher voltage rails, you notice the big high voltage ceramic ones in there as well.

So yes, looks like we've got a couple of those for each rail. so you'd have likely the positive rail, the ground, and then the output and then the negative rail as well. And there you go. If you've never seen the Illinois capacitor symbol IC in there, I Mean this is a 13 year old, a 12 year old unit.

This was manufactured in November 2003. so since then they've been acquired. But yep, world class caps. And there we go.

That's likely our buffer driver. Analog devices ad 847 as always old linking the data sheets down below. Now this is an interesting they've got it in a dip eight arrangement and there's a couple of debates on this board and you'll notice that Primarily like you know, almost all of this is our or surface mount construction, but why have they chosen a debate there? Well, it was the right device for the job. It had the right specs they needed, and it was almost certainly only available in a debate package while at the time anyway.

so they used it. They didn't worry about that sort of thing, no compromises. it's the device the designer wanted. So whatever package came in, that's what they used and you can see that a few times around here as well.

We've got ourselves some amp. Oh three, there's a couple of those and of course you'll get them on the opto couplers as well. But yeah, there's a couple of amplifiers in there that just buck that SMD trend there. And of course that shows really good design intent because the designer goes.

Notes: screw you manufacturing I Don't care if that dip parts going to be a problem for you, That's the part we need. This is a high-end market leading precision product. Use that damn part. No substitutes, Thank you very much.

and it's quite a few of these spread throughout the board. There's Nao is that top brand of course our solar state photo Masri Laser Iqv two one fours. So what is this puppy under here? My guess is some sort of custom part. hence you know they have got the model on there and they've got some sort of code after that.

so I might just lift the skirt on that and have a look run at 12 megabits. Whatever it is. Well, that's pretty surprising out. Aeramax: I'm old school EP M 7000 series PLD Not even FPGA Yeah, we've got a digital beast under there I Expected some sort of analog magic beans from Keithley but no, obviously they're doing that to drive the nearby looks like we have a Ladakh C is so yeah, and maybe something else down here.

Anyway, that could be yeah, just some generic glue logic to our drive the various dachshund ABCs required. And yeah, on the upper side of this, we have a couple of Arts serial input Dax These are our analog devices Seventy Eight, Forty Nine, 16-bit multiplying DAC Not a bad DAC at all. Typically a typical of what you'd find in such a bit of kit. Next to that, we have a couple of our precision Op Amp C Linear Tech one triple one to up here.

that's a that's a really nice up Pico amp input level op amp the one the 109 seven is a bit more jellybean, but I don't see a voltage reference next to those off hands, so there's probably going to be one common up voltage reference for this whole whole unit. I think for the various ADCs and DACs I suspect. And of course, on something like this high voltage, high current precision unit, you're gonna have everything optically isolated. All this entire analog board optically isolated from the digital board down in here.

So that's why we've got our Opto isolator is here. This cable goes down as I said the lower digital processing board down there and the yut CP order would handle some of that comms as well. between that board and then just Fanning that out to the various ADCs and DACs and everything else. So and not a huge amount going on in there.

But yeah, it's you've got to have some sort of smarts on the other end of this because there's too much going on on this analog board. so that's just it. Takes care of all the housekeeping here. Alright, here we go.

Let's have a pan around the ball without Sugano microscope shall we? and we can get reasonably good detail. see all the chips and everything. So everything's hunky-dory So let's have a look around here. and what I'm specifically looking for is the voltage reference now.

I Found a V ref test point here, but look, it's a linear Technology 109 seven. When linear technology one double-oh-seven these are just all our parents. We've got a TI 7705 that's a voltage supervisor, some seven 4hc logic indeed for 400 series marks. another max.

three to six is nothing special happening here at all. Got some more awesome for HC discrete stuff all around there. We've got another amp oh three up there measuring across a current shunt and most likely maybe that's the input there is that I don't know coming from somewhere I Don't think that our precision current shunt resistors are around there. more.

DG series maxes. they got those everywhere. We've got some precision resistors there I don't know who they're from offhand, but Op2 Eight Twos and tons of different types of Op amps look eighty Seven, Nine, seven S, Double one, two fours. What? I've got classic Ad71, 2's they're just, you know, fairly jellybean.

o P2 A2 is eighty Seven One One which is the single version of these seven One Two there. Oh goodness, that looks weird on the camera. It looks like a like this. two holes burning the chip, but there's not, That's just the shadow coming from the light.

You can see it down here as well on the on the Takano microscope. So some like Lm339 S and but so a whole bunch of Op amps everywhere. How much just stopped? 23 s everywhere. But these are all Op amps and I please scream at me if you can see the voltage reference on here.

that puppy looks unusual, but it's an R So that is a resistor network by the looks of it, so nothing fancy going on there. There's the 16 bit Dax we were talking about before Marks, we've got some Op amps near there. Curiously, we've got a seven for F-series fast series our TTL Ah, really old school. So I don't know when this design dates from.

but yeah, to put a 7 4 F in there. Wow, that's that's really something that it's got our next to it. So I presume that's another resistor network more Op1 77071 ones. I'm double five, three, four Op amp g so they're mixing it up LM three one one comparator and well I don't know.

Scream at me if you can see the voltage reference because I haven't damn wolf seen it yet? And there's lots of test points all around of course. and that V ref is got to be doing something. There's an ad presumably ad converter test point, but once again, I haven't actually found any analog to digital converters or a reference. Unbelievable.

If we go down here to the relays down here, there's nothing much doing down there as well. Anyway, here is our oops. Just hit my microphone. We got ourselves and Dale Precision Shunt Resistors down in here.

They got those in a series parallel combination. there got some back to back diodes happening and I've got an Opto coupler there and but there's a couple more of those NAIS Moss relays in there. So this is this would be our main output current shunt resistor because it's quite near the output. Here it's near all the there's a output filter inductor and there's our looks like our main output switching relay.

The code: I1 here's all our wires going off, so presumably that's got to be the main output current sense resistor so that somehow has to presumably get back into the amp o3 over here because that's our that's all this one over here cuz there are two closest differential amps for measuring that current shunt. but geez, I mean I've got some MOSFETs down here, but ah man, where is the voltage reference? I don't know. I must be blind. Looks like we got ourselves some my precision polystyrene caps there.

they'd be using those for a rust ability. I've got a couple of more precision resistors tucked away down in there. By the looks of it, looks like we've got ourselves a big-ass precision resistor to Meg up there. so that's doing some high range shunt stuff presumably.

and I think it's just there for protection. But hey, you know when it's to Meg and you use one that size I You know, is it a high voltage one? But there's our output stage for those who wanted to see the output stage. The light isn't the best on this Takano microscope when you're just using it for. you know, when using it for high components, it's only got the one light at an angle.

Anyway, I'm still not coming close to an ADC or a voltage reference. Unbelievable. And if you're wondering what that Cal SD 5400 is there? That same high speed must quad fit and log switch arrays. So yeah, that makes sense.

That's all fine and dandy now based on the fact that I can't find a dedicated ADC chip in here, and of course, the precision high-end performance of this thing adds it's a five and a half digit converter in the thing. It must be by deduction, absolutely must be I have no doubt implemented in the RTO max, P or D They're doing a multi slope conversion in there, similar to our other high end up multimeters we've seen before. So yep, some sort of you know, charge balancing variation of a dual slope converter or a multi slope art convert. It won't be just a dual slope.

Sorry so and that is. yep, that's what what they must be doing there. but I still don't know where the voltage reference is coming from. I might be able to power it up and measure that perhaps.

but I still don't see the voltage a reference. But yeah, that's what. they're implementing the P or D in there and talking to the ADC over there. so obviously it must do some.

There must be some other interfaces coming from microcontrollers would do all the range switching and other miscellaneous our housekeeping stuff but that one that so that could be done somewhere else but that that could actually be dedicated just to the ADC itself. So sorry do or you are volt nuts out there for the life of me. I can't find the actual voltage reference I used for the multi sloper converter in here Could be one of the stock 23 packages. Could be a precision buried Zener reference for example.

Something like that, perhaps mm-hmm Well, there very well may not even be a specific voltage reference in here. It may be that the that the rails are. you know, ultra-low noise, ultra low drift voltage rails for example, and everything is referenced from that. Perhaps because of course you don't need an absolute reference in there like an absolute precise value reference like two point five, zero, zero, zero, zero volts.

I mean as long as you've got a power supply that is stable ie. you know however many ppm drift this thing is, or ppm reference yo. 5 ppm, 10 ppm or whatever it is and then as long as your power supply doesn't drift, then well, you can use those as your voltage reference. So maybe that's what's going on here.

but I don't know I'd Need the specific schematic to figure that out. And sure enough, it pays to read the service manual. It turns out that the PLD here is the ADC As I suspected, it's the control element format. and yes, it is a charge balancing multi slope converter.

That's the only way to get the performance you expect out of an instrument of this class. And no, there are parts on the bottom of the board of how to. looking there down the bottom. Here, we've got ourselves the digital control board.

It's not worth taking apart to look out. it's just a 68 double 3 2 micro controller with driving some GPIB Nothing fancy whatsoever. Deep down inside there is the power supply. so there's switching power supply down in there, there's a metal shield that's separating them all it would be I think quite a pain in the arse to take this sucker apart I'd have to take out all the heatsink tower components and lift everything out.

Now it's looking really messy. so sorry. I don't think I'm kind of go any further with that today. Nothing interesting, this switching power supply and just a dumbass micro controller board.

All the interesting stuff we want to see is of course on the top, which we've taken a look at and there really is a lot of analog goodness going on in here. It is actually quite complex if you've got the schematic. yeah and I know you may not be able to make heads or tails. Oh, it is a quite a complex art measurement system and our train switching of course.

it goes from anywhere from you know 500 volts down to millivolts and 1 amp down to one micro amp so it scales. You know, quite a significant range there. and the outputs are drivers. That's an art in itself.

All the output that stays there. the Moldy slope conversion which Keathley no doubt have a pattern on that Agilent. I've got their own system as we saw in the three Double One 680 multimeter and that's so forth. and yet they all have their particular art flavor of implementation.

but that is really quite a nice design. It's you know, impossible to fault that there's no budges on it, but it is a very well proven design over decades, really. So I mean this one was manufactured in 2003, but I think it even predates that. I think it came out before that it and it really is quite a nice example.

As I said of the design driven aspect to this, it's not driven by anyone else is driven by the design side of things. The engineers who actually designed it to meet the really awesome art performance specs this thing and they said yeah, we must use that device, we must use that, must use that available and it doesn't matter. All the different types of Op amps and that switches on this. It's not like they've gone in there.

Oh yeah, we're just using an Op O 7 for everything you know? No, they've chosen very specific parts for very specific purposes on both, like the high voltage side of things. Over here, you're going to use high voltage maxes, the to precision that current shunt differential amplifiers, and I'm sure if you looked in the type of precision resistors they've used in here the polystyrene caps for example, and they've chosen these parts very specifically to get the best performance possible or the performance that they wanted. Any way out of this thing. it's not.

you know, absolute world best start performance. But yeah, these are a really nice, accurate, better kit. And I mentioned right back at the start. Here's our I'm IO terminals over here.

Our wiring to those and our output stage is right at the back here. How are they getting it across? Well, I think they're I'm It's different. It's differentiated by the cutouts, those little slots in there. There's our high voltage driver over here.

Here's our output stage. it comes around. We've got some high voltage maxes over here. We've got the Big O the big two, mega precision and resistor over there for high voltage and that's coming down here I think that's coming through this stage which is where our output current shunt resistor is and then it's going directly to our switched output terminals there.

So I think that's the that's the general flow of things. And here's our block diagram. We've got our voltage Duquette in our current DAC over here these are the two ducts that we would have seen over here, no doubt. And then we've got our voltage and current clamping where that's actually happening in here I'm not sure but I've got some error amplification and then our high-voltage output stage which were say well switched, high N volt and a low voltage output stage.

And then we've got our sense resistors that as I said I think flows around here. So we've got our current limited sound, our limiting stuffs probably over here somewhere, and then flows around here to our output sense our output sense resistors there, which is down there and Bob's your uncle. We've got some feedback and stuff like that, all sorts of MUX in got some overload protection and stuff like that, but that's about all she wrote. It's love to see a more detailed block diagram of that and the schematic.

So if anyone does have the schematics for this thing, if an is available online, not, please leave it in the comments. Yeah, Anyway, it is quite a nice design. I Hope you enjoyed that look inside a Keithley 2400 source/measure unit. Very high-end bitty kit, and if you can pick up one of these I highly recommend it.

There's nothing like having an SMU for doing a characterization of basically almost any type of that component you can think of. They're absolutely fantastic devices. and yeah, I've done that. Separate videos on that I'll have to try them.

link them in if possible. As always, high res teardown photos of this are mainly just the board. really because I didn't tear down the rest of it. Mmm sorry.

But yeah, high res Ted Our photos available on Eevblog comm. The link to that is down below somewhere or up above depending on where you're reading this. And as always, if you liked it, please give it a big thumbs up. And if you want to discuss it, jump on over to the Eevblog Forum Link is down below as well.

Catch you next time you.