EEVBlog #426 - HP 3457A Multimeter Teardown

hi welcome to teardown Tuesday you've seen me mention my HP 3478 a five and a half digit classic bench multimeter for used it I'm in quite a few videos well you may have noticed behind me I have upgraded tada the HP 3457 a six and a half digit multimeter big score on ebay got it for just over one hundred and fifty Australian dollars fantastic and it's bang on beauty so I thought we'd crack it open see what makes it tick could be interesting let's go and here it is hooked up to my MV 106 transfer voltage standard you've seen in our previous videos yes I did violate my don't turn it on take it apart rule here I just wanted to tweet some photos and check its performance before I go rip the thing apart today on a teardown a Tuesday and as I said got on eBay for like a hundred and fifty bucks and it really is a very nice meter it comes with that cow stickers it's a couple of years out-of-date but still that's not bad at all we'll have a good close-up view of it but here it is it I've hooked it up to the MV 106 I've set it to one point zero zero zero zero zero volts I've only got five decimal places there this thing of course has six decimal places it's six-and-a-half digit meter in fact technically with summer averaging in high res mode it's capable of seven a half digits but you can't get that via the front panel apparently you have to do that via the GPIB remotely so I do get fairly lucky with these bits of gear on eBay absolutely bang-on of course if I dial that up one digit there you can see it basically changes by ten least significant digits there's a bit of noise there a bit of fluctuation we can maybe change the power line cycles or something like that so anyway if you want me to do a separate video on using this thing and having a play roof and play around with it let me know so that's actually not a bad upgrade to the classic HP 3478 a this one's a 35 57a and while it's not a metrology our grade multimeter like the famous 3458 a this one's the 57 a don't get confused between the two it is still a very impressive instrument and really at some hard to beat for any modern multimeter under a thousand bucks for example so you know to get this thing for a hundred and fifty bucks it's a bargain so they're certainly worth are paying a couple hundred bucks for if you're in the market for a six and a half digit multimeter its performance is still pretty impressive it's a 1980s vintage art mid to late 1980s I'm not sure when this one was actually designed we'll find out a manufactured we'll find out when we crack it open by the date codes on the chips and things like that but yeah it's still a pretty damn good meter and of course it one thing my 3478 a doesn't have it's actually got a micro amp range as opposed to milliamps so there it is bang look at that so that's a hundred Pico amps resolution very very nice bit acute indeed and in terms of you know basic DC volts accuracy on this thing like the 24 hour spec on this thing is pretty good it's doubting the order of you know 18 ppm or something like that it's quite a nice bit of art kit so if you can pick one up for a decent price highly recommend it now well it may look like a big bit of kit sitting on your bench with its big wide front like that you can see it here compared to the 34 78a which is actually quite deep and it's actually not that much physically bigger in terms of that volume at all and it's not that much heavier either which means that you can get shipping usually for quite a reasonable price on these things even international but as always some equipment dealers like to rip you off I got this thing delivered via DHL for like 70 bucks or something so for me it's a worthwhile upgrade just for the DC current range that micro amp range capability the accuracy of this thing I won't go into specs on the if you want me to do a review of it sake you know compared to it or something like that let me know but yeah the specs are a little bit better than 3478 a but it has got some additional ranges and all the math capability and stuff like that as well so a better instrument all-round basically but it has that larger are wide form factor whereas this one is you know half wit half rack width so that's probably its main disadvantage is that it can take up some additional space on your bench in terms of width you know instrument width on your arm shelving but apart from that lovely be the kit I love it and here's the front panel for those playing along at home DC volts AC volts ohms are for two and four wire DC current AC current and frequency measurement as well and it's got range you know auto or manual range selection display digits and six and a half digits as I said it is technically capable about seven and a half digits but getting that to work on the front panel is that not possible I believe you've got to do it via the GPIB as I said if we take a look over here at the configuration buttons we've got number of power line cycles so you change the integration time effectively the update time and that averages things out for you it's got auto calibration auto zero offset compensation and it's got various are triggering capabilities and it's got an input scanner an optional input scanner as well but they're usually a very expensive option to like you can have multiple input channels as we'll see there's a slot on the back to actually do that this one's not configured with that and you can store and report and recall various arc configuration settings and then there's the math capability now unfortunately this thing doesn't have a huge memory on it it's only think it's only 1k or one and a half K or something like that and I'm not even sure if that is the number of readings I think it's actually less than that in number of readings depending on the resolution mode you've got this thing in so you know you might only be able to store a couple hundred readings but you can certainly do some usable mess and get math and gets our standard deviations and like that and of course we've got that proper binding poster inputs here it's got that four wire input for ohms and that separate current fused current terminal down here now there's one minor disadvantage compared to the 3478 a this one's replacing in my lab here it's only got the front panel terminals it doesn't have them duplicated on the back for my purposes doesn't matter a rat's ass but if you've got you know Warren these into a production test rack or something like that that sort of thing could matter and check it out it does have calibration stickers on it unfortunately it didn't actually come with the calibration certificate of course but it's fairly recent you know April 2011 that's not bad at all I'm certainly aren't not going to complain about that and that's a reason why it's bang on because these things you know they don't drift much folks so really once you vote you know if you've got it calibrated a couple of years ago I was pretty confident I saw this in the ad and I went well you know that's good enough for me I'm you know I've taken a bit of a risk but I I thought this one would be bang on and it was now interestingly it's got analog devices on here and it seems to have come from and owned by and log devices themselves and they presumably have their own internal a calibration metrology department which calibrated this thing and we've got another sticker here telling you what capabilities it's been not calibrated over voltage current and resistance not a problem by the way it's a it was calibrated in that twenty ten I presume it wasn't calibrated again after that they probably just let the thing expire but still that's pretty recent it's got a sticker here that says a dpi I I'm going to guess that stands for analog devices of course production instruments so I think this was used in their production environment for art manufacturing the analog devices chips one you know those automated art test racks and something like that which is why it's in you know fairly good Nick for its age really ender it's got a sticker from test equipment depot but I didn't actually get it from them I bought it from another eBay dealer so I don't know whether or not they got it from them or what the deal is there and it's made in the good old US of A so number 27 oh three a a stands for made in America and that's the actual serial number eleven thousand eight hundred and thirty eight know our option seven hundred just the standard option and there's a slot for the scanner as you can there's no panel on that didn't come with it there's the internal connecting cable for that scanner module you can pay a couple of hundred bucks for those low EMF relay switching modules so unless you really need that capability don't worry about it one bit and it's got GPIB of course or HP ib as they call it and it's got a voltmeter complete output so a signal 5 volt TTL signal when the when it's actually completed and an external trigger input as well standard IEC mains and switchable from 110 to 240 so not a problem it doesn't matter where you buy this thing it's going to be usable and we've got ourselves a couple of our calibration void if seal broken stickers so we'll go right through those lovely the feeling you get when you crack open the energy void is steel yeah sorry calibration void if I see your broken seeker fantastic there's no screws on the back of this thing but it looks if it looks like it's all sort of done from the bottom here so don't know whether or not these screws go all the way through or not I'm not sure what the deal is there yeah I think they do the I presume that the top those things don't fall out lovely design they're captive screws so this should yep tada just lift straight off beautiful or shielded that's what you'd expect in a six and a half digit multimeter worn in this panel connected to terminal low and that's not surprising at all that internal shield there is connected to your ground it's not connected to mains earth it's just connected to the ground on the front panel and we can verify that by mains air and there we go there's no connection whatsoever through to the ground terminal because of course it's up floating that's what you expect in a multimeter a bench multimeter like this but you'll probably find that this power supply shield over here is connected yeah there we go that one is connected right through but this one is just internal ground connected for you know it's all part of the system ground design to shield all the electronics and ensure there'd be lots of low EMF our design in this lots of star point grounding and stuff like that and they're deliberately kept it from mains earth as all part of the system grounding functionality to ensure that they get the six-and-a-half digit performance in this thing so this grounded top cover here and when I say grounded I actually mean ground and not mains earth as we just saw there actually it's connected through to the bottom cover down here which is also are separated from mains earth over here so that one is just the shielding cover for the relay input switching matrix board which then slots in underneath here so we should be able to lift that out and somehow it's got instructions there should read it and if we just flip that cover open it's held on there by this ribbon cable that's attached on and won't disconnect that unless I absolutely have to but you know warning caution caution we've got some we're going to gain and flatness control here that warns you that it's input high that's connected to the input high potential and there's a V offset adjustment as well so but they're the only two adjustments that you can get through this thing so we should be able to lift this plate off and see the measurement and acquisition board on the bottom here separate from the power supply and digital display board over in this part you can see the ribbon cables go into the keypad and the LCD and stuff that's all digital control there'll be some roms under there microprocessor but this is all your interesting measurement stuff under here so it looks like this thing just unlatches yep there's some sort of there we go you just pushed these latches back like that they're not easy but unlatch this side first and then you probably got to unlatch this side over here yep there we go and that's a nice design I like that tada look at that I could see the voltage reference board straight off the bat and wore lots a relay switching oh good stuff and one of the first things I noticed of course as part of the system design keeping these grounds separate you'll notice that there's only two connecting cables on this thing one is coming from the transformer over here so it's got its own transformer tap going over there that's of course isolated due to the art transformer then it's got its own just linear rectifier and linear regulator over here and then down in the bottom is the only other connection between the digital board down here and the and the main signal processing board and a DC board over here and you'll notice that it looks like there there it is opto isolated and you can see that opto isolator down in there with the big slot you're going right through the board they've even cut this ground earthing trace over here because this would be connected you just ruin your system ground if you connected that straight across so they've actually cut that straight through with the slot and they've used these optocouplers here but they're they're like one of those optocouplers you'd see you know like in a photocopier or something like that where you can actually put something in between like that and it just cuts off the signal between the two so that's rather interesting I would have expected to see a you know a basic dip package optocoupler or something like that but they've gone with that type so they're doing that to keep the entire ground system separate this is the grounding point so if we measured say this screw over here and be on the metal shield that would be going through to the ground connector on the front panel and of course this one over here is connected through two mains earth as we saw in the measurements before now my idea for why they've chosen that type it's probably just a basic built in braces or engineering these type of you know am I going to have a larger withstanding our voltage larger isolation voltage just purely because of the huge physical gap and isolation between the phototransistor and the diode on this thing so yeah my guess is it's a they were much better than the dip package so maybe they just had a higher withstanding voltage than say a standard 4 in 25 optocoupler or something like that but yeah I don't know you'd have to actually get the data on that one but even your stead adopted a couple is a pretty darn good so I don't know maybe there's some serious belt and braces happening there and there you have it it's not a 1980s vintage folks sorry to disappoint you it's a forty second week ninety that's probably around about the earliest date code I can find on this thing the chips seemed to be around the forty or fifty ahthe week nineteen ninety so I'm not sure when they actually stopped manufacture on this thing if you actually know the start and end manufacture dates on the 3457 a do let us know in the comments or on the forum I've popped the top cover off the processor board here and you can see the lovely power switch extension bar there going all the way back I love those extension bars great engineering and there's the backup battery in this thing Lou lithium battery check it out folks made in Britain title my British viewers it's a neat earnest all three volt battery and it doesn't look like it's leaked at all you have to be very careful by the way when you're opening these things service in them prodding and poking around in these things don't short out that battery at all and don't replace it with the power off because you will lose all your calibration data in this thing so I'm going to check the voltage on this make sure it's still good now make sure I don't have my meter set two amps because that would be really embarrassing I drawed out my battery and I'd lose all the fantastic calibration and the bang on enos if that's a term I'm going to invent it of this thing so let's have a look there we go 3.0 six volts beautiful I'm not going to touch that sucker now this is interesting this processor board seems to have been manufactured at a different time four years prior to when the ATC board most of the seven 4hc chips on here all seemed to be dated eighty-six look at that 40th week 1986 44th week eighty-six all within the same you know all within the same order 23rd week eighty-six up here so almost every one of them so that is weird why that processor board was clearly manufactured you know around 81 I'm presuming around 86 otherwise they bought an absolute buttload of these chips back in 86 and they were still using them for years later I don't know um it doesn't you know it doesn't wash I mean all of them the Intel processes all these chips all around 86 whereas all this board over here is all 1990 vintage check it out even the system ROM has a date code of fifth week 1987 my guess is well this is a 1987 vintage instrument but it's had a complete upgrade of this board in 1990 or so that's the only conclusion I can come to now this is really interesting check out all of these dip parts around here including this part here on the voltage reference board here they're all got 1826 - something for five linear tech parts there plus national nti yet they've all got that same 1826 - and then a number like 1382 and zero five to one and 1265 and all sorts of stuff so it looks like HP have had these sort of like a custom-branded there of course are off-the-shelf parts aren't like custom chips they're just regular off-the-shelf parts but they've had them rebranded possibly with the HP part number now one thing you won't find in a precision instrument like this is a cooling fan why well I it doesn't generate a huge amount of heat I think these things rated for like a 30 watts maximum consumption or something it could be a lot less than that in actual power draw but because when you have a cooling fan in there and it's sucking air through the side vents for example in a typical product and then going out what you're creating thermal gradients within the product and over the parts in particular the voltage reference down here which we'll take a look at and other parts of the board and you don't want thermal gradients because if you've seen my thermocouple video you'll know that basically any dissimilar metal junction within this unit any junction at all solder and copper clad board and everything else creates dissimilar metal junctions acting as a thermocouple so you're going to get EMF voltage is generated in the unit so you don't the last thing you want in something like this you would rather have it heat up to a warm temperature and stay nice and even than try and cool it down with the airflow through the thing and we can demonstrate this alright what I've got here is my resistance standard I looked up of this thing it's close enough to what 10k here so this least significant digit here it represents 1 ppm of that 10k so one parts per million and you know this thing it's Temko is incredibly low in the order of like you know parts per million so any I you know change in temperature in here can actually generate EMF small EMF effects which can affect or thermocouple effects in the joints or a temperature gradient across components that can actually affect the reading when you're talking down in this region now I can wave my hand around in here so it's not noise pickup or anything like that okay so what I'm going to do is actually breathe on the circuit in here and you should see this significantly change let's give it a go look at that there you go that's changing by like hundreds of ppm look at that hundreds of ppm just by breathing over that circuit there so that is a really big deal for a six and a half digit precision meter like this so that's precisely why you won't find a fan in a precision bit of kit like this you do not want all that air blowing over and creating those temperature differentials across your component it's a big no-no and if you take a look at the processor board down here somebody's actually texted on the serial number for this unit 2703 811 838 and then a date code in 93 so something and some other numbers there so something was done to this thing in 1993 I have no idea what that is a couple of years after the or three years after the 90 dated chips 1990 data chips on the analog board as for the mainboard here we have more HP custom part number branded chips here's the main processor use 601 here and it's some sort of motorola part but it's got this 18 20 26 24 part number and that looks like a HP part number like here then these resistor networks look these are but these are all 1810 and 1820 you'll also find down here check this out here we go here's a National part 1820 as well here's another Harris I think that's a Harris part down here 1820 there it is again that I believe is the keyboard front panel keyboard controller because it goes down to the ribbon cables to the front panel or the LCD sorry the LCDs got its own nut control a bit yeah they got these freaking custom part numbers on them real pain in the ass and here's the schematic which I'll actually link in in the notes below so you can load up the PDF of the service manual it's got the full schematics in here they're a little bit hard to read but there's the main processor you 601 there that we looked at and then of course it's got you note the RAM and ROM it's all very traditional stuff down in here or miscellaneous control stuff we've got the speaker down there nothing really majorly important there and then yet here's an extension of it it's one of these big long scum out so it's got to be broken up and then we've got our HP IB is less GPIB control of stuff around here we've got the non-volatile Ram control there's our internal battery there so you can see how that just continually powers that that's just diode or there and there's the right calibration lock circuit as well watchdog timer it's all they're all pretty basic digital stuff but what that actual what Motorola processor is I don't know one of the motorola 8-bit processors most likely and then we have our opto coupler link between boards and that's what we looked at down below that's that and there you go it's just start serial in and serial out data connecting the two boards together and of course this analog board over here actually has its own control processor and there it is that's an until eight oh five one microcontroller and look at this national device here that's actually don't know if you hear that but that's actually not plastic or anything that feels like and sounds like ferrite so they've surrounded all of the leads on this chip here kind of like maybe isn't afterthought that they've put like a little effectively a little ferrite bead through each one of the leads on that national semiconductor device I wonder why and of course we don't know what the bloody chip is because there's one other one the stupid part numbers 1820 - three one seven four caror and here's another device down near the input down here what the hell is this doing this is an 1820 - 38 6-1 ah man but thankfully of course you can pretty much tell what that's going to do based on the proximity of these realize I don't see any discrete transistor drivers for all these relays down here so that's probably a really driver chip and if you have a look at the schematic tada there it is relay control you 121 and you can see a whole bunch of our catch diodes there which are the get the back EMF of all the switching coils but basically all of these relays they're going to be used for am not only switching the inputs but that use for the test modes as well they can short the inputs they can disable them and route them through to other test signals because this supp meter has a fairly comprehensive built in self-test so there's 805 one micro controller it basically you know it does the relays it handles the optional input switching card which sits on the top and basically handles the ADC as well and there is the custom hybrid ADC in this thing this one ain't off the shelf by the looks of it so there you go I have no idea what that ADC is it's app you know some sort of maybe HP our custom hybrid ASIC ADC or something like that to get the performance I don't know and here it is you 511 that's the hybrid looking custom ADC thing they're all with that you know that's the entire ad converter circuitry around here and you can tell you know we really start getting into our precision measurements here we've got guard nodes marked around here matched transistor matched J FET transistor pairs and stuff like that and here's our reference board over here that's got our reference xena on it and we'll take a look at that but yeah that ADC is of course that clocked and controlled via the 805 one processor and here's what a lot of people want to see what is the ref of the voltage reference used in this thing well this is the PCB assembly T this one is obviously the reference Zener diode with ability not here too of course but it's a linear technology part but once again that custom HP part number 18 26 - 12 49 - five with a date code of the 36 week 1990s and once again we've got another linear technology part over here but once again HP part number branded but that's just a precision op-amp clearly so there's you know there's nothing else going on there and it generates the plot I believe plus minus 10 volt reference fold which we should be able to measure on those two test points I'm assuming and I'll just verify that I'll just probe I assume that that nut there is ground so I'll just probe that yet minus ten point one seven five of course it doesn't have to be precise because the and of course it's exactly there you go the positive is the exact opposite of that so it generates plus minus 10 volt supplies and they don't of course have to be spot-on because this thing compensates for that with the calibration values in the software and of course the key to an instrument like this is not absolute accuracy as I've said before that voltage reference down in there it can be plus minus 5% absolute accuracy you don't care what you care about is the temperature coefficient ie the drift per degree Celsius how many ppm per degree Celsius drift is that 100 ppm is that 10 ppm is at 1 ppm you know really precision stuff like used in say the 3458 ie for example one of the world's best multimeters and the big brother to this one well it's not about its absolute accuracy you can just trim it to anything you want it's all about the drift now this reference board here as I mentioned the designator is a 25 and it is exactly the same HP part number as the one used in an earlier model watt 3456 a which we do actually have the schematic for and here it is the a 25 reference board oh three four five six double six five two five and that's exactly the same part number as what's used in this 3457 a and as you can see the same r4 pin device here with the internal heater like that and there's the internal Zener diode and of course it amplifies that and of course there's not much to it it's all about the stability of the Zener diode itself you know very carefully selected very carefully tested and characterized for the performance in this thing and you know there's nothing else on this board there's one two three resistors and a cap and that's exactly what I see on this board three resistors and a cap so it's an identical board and the service manual for this instrument gives you five hundred listed as an lms - nine 9h and that's the classic at national semiconductor reference but the one we've got here is a linear technology so and what but with still four pins and in case you're wondering it also lists you 501 here which is just an LM 301 now it's hard to actually get in and see under there but I can actually see it and this is actually a four pin device so clearly are two of the pins are for the Zener diode and the other two are for the heater you can see there that we're measuring you know roughly like 40 to 43 degrees or something on that on that chip either the lasers not directly on it because there's an offset there when it's up bar close folks so I haven't let this thing out warm up to temperature but you know if you move it to the side you can see that that is obviously got a heater in it to heat up the Zener and that would be our temperature controlled as well to keep that at a constant temperature and if you have a look at the input terminals down in here yes they are just wired directly across like that but they do go through a toroid filter they're wrapped around in there quite nicely heat shrunk down in there and then they go directly into the relays but because this thing is all like metal grounded shielded stuff like that you can get away with those effectively quite long leads go in from the input terminals and there's a one point 1 ohm high value resistor that's so most likely our input current shunt resistor and you'll notice all the star grounding branching off from that point really is quite nice and this board mounted on top here this is the AC converter boards doing all the true RMS stuff and AC conversion couple of switching relays and those adjustments which we saw in the top through the top metal lock can and yes we've got more HP part numbers look at that 1826 there 18:26 up there for that and log devices part and more of them all the way up the top HP part numbers as far as the eye can see and if we pop that AC converter board out there's only two wires levering and that thing on no silkscreen designators all on the bottom there component designators that's also got its own metal shielding can so we should be able to remove that and whoa hello and looky what we have here folks we have a whole bunch of well they at first glance you go well what are they they don't look like the other relays but of course our koto is a dead giveaway koto is a very high quality manufacturer of precision low EMF reread realizing that stuff like that so these are Oh for 900 - one triple five haven't been able to find the exact datasheet for that one on a quick glance but they would yet be some sort of maybe are shielded or right low noise low EMF relay or something like that but we've got ourselves a hybrid and we also have ourselves a genuine bodge check that out there even heat Shrunk the leads look at that beautiful and a look at our schematic folks reveals all u101 is the input hybrid board there and it's got all the switch in look at that all the FET switching required for all of the input circuitry because here's our input circuitry over here and here's the input circuitry here there's some ohms over voltage protection there here's all our input relays switching stuff there's our input fuse by the looks of it and there is switch it rain switching stuff so like it you can sort of equate that to say similar to our like a fluke custom input switching hybrid that they use in their multimeters for example so that's all on that one board and that would be very well matched and then of course the output of that goes into these matched transistor pairs here precision op amps and various other stuff so you can you know take a look at this schematic until the cows come home but there's the ohms current source for example great stuff there's the input amplifier which I said and there's the match transistor pairs and then we've got a pre charge offset circuit ah and an offset DAC as well now that certainly is a very interesting hybrid board indeed obviously are all the circus trees on the bottom they've just got this big ground plane on the top here there's only one extra trace on there apart from the ground and it's actually held in by these two screws and it looks like they just start press into those pins those gold plated pins sticking up from the board so I should be able to undo those and lift that board off without any damage well sorry folks that's not the least bit exciting it's all encapsulated it looks like it's I got epoxy in you could probably you know if you wanted to ruin this thing you could and you know take off the epoxy or break through the epoxy on the outside of that and then have a look at the hybrid under there but of course I want my unit to work so I'm not going to do that I'm just going to put it back and we've got ourselves a little metal can bird brown package down in there and the obligatory trim pot now there's one thing I don't particularly care for on this AC converter board is that it's only held down by that one screw there I mean the board does go through over here but it's not like it's really held in place with that at all so there's only you know I wouldn't like to know about the vibrational modes set up during transporting that thing I not impressed by one mount like that at all so there you have it that's about all she wrote for the HP 3457 a sorry I just can't go into more detail on some of the chips I could try and reverse engineer some of them figure out what they are but really it's it's tough when they brand the things with just generic HP part numbers real pain in the ass anyway if you want to take a look at this thing the service manual for this which has the full schematics and everything else in it will LinkedIn down below and if you know what that voltage reference is in there from linear technology please let us know I'm sure there's a lot of people into you know precision references on the forum and stuff like that and they would love to know that sort of thing maybe it's obvious I don't know I just haven't looked hard enough but anyway if you want to discuss it jump on over to the EEV blog forum that's the best place to do it and if you like teardown Tuesday please give it a big thumbs up catch you next time you you

Channel: EEVblog

Views: 71,280

Keywords: multimeter, hp3457a, hp3478a, 3478a, hp 3457a multimeter, bench multimeter, agilent, 6.5 digit, 7.5 digit, gpib, hpib, interface, math, calibration, volatge reference, mv106, edc, adc, schematic, pcb, dac, hybrid adc, temperature, temptco, temperature coefficient, thermocouple, hp 3457a, system multimeter, ebay, ebay bargain, how to, what is calibration, multimeter calibration

Id: wUcsxjvoV1s

Length: 38min 28sec (2308 seconds)

Published: Tue Feb 19 2013