EEVblog #210 - Krohn-Hite DC Voltage Standard Teardown & Calibration

hi check out what just turned up on my bench whoa look at this in a beautiful it's a electronic guy Development Corporation ah MV 106 j DC thought he'd stand it all the knobs I love it precision city it's awesome um it's now I done I think it's owned by a company now called arc cronkite they still sell it it's still a current model and I thought we'd turn it on check it out see if it still works calibrate it look at how it works internal construction tear down bit a theory maybe and that should be fun let's go and here it is it comes in a rather old style aluminium rack size Kevlar it's got a sort of little aluminum tilting bail like this but it's also got another flip out tilting bail like that which is really quite nice I like it and it's in fairly good Nick I'm not sure how old it is I think it's um the manual which you can download says it's 19 or last updated 1991-2001 sorry so it's like a 10 year old model but it could be based on older than that because it seems quite old-school with the with the old-style rotary encoder novice but maybe it's been updated or something like that I'm not sure full scale of 10 volts or 100 millivolts or 10 millivolts it's three ranges there are there is another type which has a current output as well but this one doesn't now the cowl sticker here if you take a look at that it's a last count in 2000 and well 2007 so June 2008 end made in Boston Massachusetts do they still make anything Boston anymore I don't know um I think they are I'm sure there's a few things made there anyway uh I quite like it it's got the sense outputs as well because when you're talking about a precision DC instrument like this it's got a sense the outputs but it does have a little short in bridging bars in there if you just hooking it up to a voltmeter then that's fine you don't need the sense outputs they can just be tapped directly off the output and you can just use a two wire instead a four wire arm you can invert the output nothing unusual there and I love it's got six decades here and I love these style rotary selection knobs and you'll notice I'm quite a fan of those I've got quite a few of the Keithley instruments there I've actually got four of those and I just love those not style knobs so I love this thing it's brilliant and of course site you can set let's say we've got ten volt full scale here you can set ten point zero zero zero zero one volts that's the resolution of it beautiful it's at 10 micro volts resolution on the 10 volt range but if you go down to the 10 millivolt range we're talking at 10 nano volt resolution Wow nothing exciting on the back I'm afraid just the model number and a 240 volt selection switch so wow it was a 110 switch to over the 240 and up we'll see if this sucker works all right we'll just try it with my yacht fluke 87 here first as I said we've got the current sense input shorted with the current bars they're ideally if you had a big load you would actually take the sense directly onto the load point itself for absolute accuracy but in this case it's going to be more than good enough so let's switch it on and see what happens let's put it to 10 volts and it lights up overload and yet the overload lab is still on right now there we go it's gone out looks like it now requires some settling time perhaps and bingo there we go 10 point zero zero zero zero volts and we're pretty darn spot on - almost to the least significant digit there so it looks like it works let's change it to 100 millivolts there and ah beautiful that 100 point zero zero one and let's we've got a hundred point zero zero well let's whack the sup one notch that should give us one least significant digit and it does there we go changing one of the time Wow it's spot-on and of course we've got two decades below that which won't do a damn thing but if we turn that up by ten should jump to eight and it does here we go so it looks like it's um it's gonna still be well I'm not gonna say within spec it's within you know the 0.05 percent of my fluke 87 here but this thing has specs of what are they not point double O 3 or 30 30 30 ppm over the temperature range now this thing actually has specs of plus/minus point double O 3 percent or 30 ppm over well this the span of a year and point triple O 5 percent or 5 ppm per degree C temperature jus so it's not super high-end as far as our DC standards go but it's not too bad at all it would be a nice addition to the lab here I think just for a precision voltage source for our testing meters and things like that but it seems to work just fine that's 100 millivolt range and we go down 10 millivolt range yep not a problem at all I like it it's nice and that doesn't seem to be any noise on the pots either with these things you've got a sort of if you jiggle the jiggle the pots a bit you can often see if there's any issues there but I can't see any problems there at all so it seems to be working just fine it and we'll try the negative on there and know it's within you know that's pretty darn close so that's not too bad at all so I determined that this sucker is pretty much still within spec well I'm not going to say within spec because it's you know only within there pretty you know it's pretty close to meet to actually check its performance spec oh it's going to need better gear than what I've got here which will which is what we'll have to do later but it basically works I'm pretty damn happy with that now I can check the linearity of it and things like that but generally with these sort of things you don't have to because the linearity is set by by precision a voltage divider resistors set up with these pots every in a Kelvin or more more popular format as the Kelvin valley configuration so basically these things shouldn't drift unless those actual resistors are drift which generally doesn't happen too often it's more likely that the voltage reference our diode in there would drift or something like that but even though is a pretty dance table so let's surprised that this thing arm still works because there's not much in them to go wrong now we'll find that there's only a power supply in there a reference diode and amplifier and a bunch of a bunch of precision resistors on these knobs is let's go one step further than the fluke 87 shall we enter get my HB 3478 a a bench meter which is actually got a going to have a similar arm short-term drift capability to this at DC voltage standard down and we've got it set to 10 pointer whoa-oh-oh-oh there and it's not too far um off at all but I haven't checked the cowl in this in that quite some time and of course you know they're they have to come up to the same temperature and well it's a bit warm today here in the lab it's probably 25 26 degrees maybe even 27 degrees C in the lab here so you know it's not exactly ideal calibration conditions but we'll fix that later now let's go down 100 millivolts here you go that's not that's not too shabby at all let's go down to 10 millivolts and once again not too shabby I don't mind it at all so um it might need well see we don't know which one's out we don't know whether my 3478 a is out or this is out or a combination of both or whatever I'll have to take it to a standard scale lab to find that out but if we can calibrate this baby and then we can use that as a transfer standard - they're not calibrate um my 3478 a that'd be neat alright let's open this sucker up and see what we got inside as I said what you're gonna find that you'll get in here this will be my guess and I think I'm going to be pretty accurate is that you'll find in like an old-school PCB with all through-hole components obviously you'll have a power supply mains power supply you'll have a precision voltage reference which will be like a Zener all right you know a buried reference Zener um temperature temperature controlled and it'll have a Kelvin value divider on the front and an amplifier and that's pretty much all it is it says it's got a chopper and I don't have the full manual yet but um I've asked for it so that should turn up shortly I hope and let's pull it off and see what we get that was off yep hey here we go tada and that's not too far off at all it's pretty much what I expected our one of the first things I out notice though is that um quite possibly ah this doesn't look like a Kelvin Valley our voltage divider arrangement it just looks like a standard Kelvin voltage divider and the giveaway there is that there appears to only be a single link between each decade bank they're not a jewel link as you'd expect and a more complicated uh switch arrangement as you'd expect on a Kelvin valley which is that what you'll find in my Keithley um instruments over here they use that Kelvin Valley dividers in them but this one looks like it just uses standard Kelvin divider we have to get a schematic to verify that but there you go anyway it's a whole bunch of very precise resistors on there we'll have to check out the value some trimpots over here on the main deck aid that's your arm that's your first decade there and it looks like it's got trimpots for all the various ranges you've got a main or down here and you've got a mains transformer now one of the interesting things to note about the mains transformer here is look at all the exposed wire in here it's um if there's no heat shrinking now on that at all it's all totally exposed as is the switch on the main switch on the front panel and that pretty that's a you wouldn't get that - up past these days but um yeah it's a path past the safety standards anyway and you'll see down here this is the mains cable input and there's like an insulator does stand off there which is just used to join the two wires and once again fully exposed so um there's not really much put into the um you know internal what safety of this instrument but that's you know typical old school stuff and your output Jack's over there are they've got gold-plated you know they'd be really good quality gold-plated contacts and things like that crimps and going off The Moorings are really neat and tidy I guess as for holding the PCB in place here this is a very old-school construction technique they've got the aluminium plate with the cutout for the board and the board are screwed on the underside there and they've just cut it out for the components um quite you'll see this construction method I use quite loading these sort of rackmount um early early designed instruments now one of the first things I'm looking for in something like this is where is the voltage reference standard now I've got this curious look in our package here and um but I don't think that is the voltage reference I get a feel in that that's not it I'm not actually sure what that is as well and and at all it's got to work common terminals here it's just a full terminal device two pins on either side there's an in there's an out but that's not what you'd expect in a temperature compensated buried xena reference like in the HP 3478 a multimeter we just had a look at it's got an LM 299 temperature compensated voltage reference in it and that's basically just a Bennett and buried our Zener diode that isn't very very precise at all and we'll go into this later but it keeps it at a constant temperature and I don't see so I've that was the voltage reference I'd expect to see a heater connection on there not just in and out I'd expect to see you know just the diode connection on one side and the heater on the other but that's not it um but looking down here they've got the reference down here and that looks like it might be the reference diode let's go check that out there you go it's got EDC reference diode at four thousand five hundred and ninety three at six point one seven three Oh volts at six point five milliamps our current so I'm obviously there and because that that diode is in those like soldered after after this thing's assembled into those little contacts there I reckon that is the reference diode for that thing so I'm very surprised that it's not temperature compensated I fully expected a temperature you know with our point double oo of what is a double O double O five or five parts per million our temperature coefficient without getting that without a temperature compensation on it that must be one hell of a good diode so we're gonna have to wait until we get the manual for that and see what part number it is now I've turned the power on and let's see if we can actually measure that diode down in there and if I'm correct it should be six point one seven three Oh volts so let's give it a go bingo six point one seven three volts it is now as for the rest of the board here there's no silk screen at all somebody's handwritten you know our for our three R to a our to our one there no silkscreen really old through-hole kind of stuff and these voltage regulators here look at the check out the pins the spacing is not even on them they've had display those pins out that's you know that's a bit boggy I don't like that at all and we've got a date code on this LM seven for one here of uh the fourth week ninety-one so there you go that dates this unit to were at least 1991 s now of course one of the things you notice here is this our little daughter board here um and I'm going to assume that that's the chopper amplifier because they talked about that in the manual so but it almost looks like it's a budged afterthought like they had to add it on but if you actually wiggle it here it will it should looks like it will pop out and tada there we go look at that that is it looks like they've purpose-designed that um it was actually design I thought it looked like a bodge board but it's not it looks like they actually manufacture that module um as a separate thing they probably manufacture and test it I mean there's no trimpots on there but maybe they um you know they test things they actually select the components and put them on there or something but yeah it's all it's all very yeah it's all very old-school and uh I'm quite surprised that the whole thing no kills and valued at the imprecise Ness really of the whole thing you know they're not using a temperature compensated reference diode down here they're not using a kelvin valley divider but hey i you know they've got away with it i'm sure it dirt will meet its at claim specs and you can see the precision resistors on the decade switches there that's a 2k plus - naught point double oh five percent precision resistor and as you go they won't all be that precise but all the ones over there will be like the ones further down you know they don't have to be as precise not necessarily as you go down but at least some of them are at least no point double-o 5% and the ones down here on the first decade here these are not 0.02 percent they don't have to be as precise because each one of those is that trimmed to its actual value actually I just looked up that part number the TSC 76 5-2 there that is actually a chopper stabilized and so there you go and I've talked about chopper stabilized dance before the reason you're going to use one of those in an instrument like this is that it has essentially it does zeroes out if it nulls out any art DC offset errors and this is a DC precision reference so what don't you want in this you don't want any DC offset errors bingo you've got to use the chopper stabilized app for that and thanks to Joe Inglis from kronite we've got the schematic for it in record time and what we can best be described as the manual it's a bit of a hodge podge actually the manual better we do have the schematic I'm not sure if it's actually this is all they had um my the model I've got the NV 106 J this just says mV 106 I'm not sure what's actually going on there but anyway it's a pretty much what I expected a partner from the Kelvin valley thing in the temperature control Zena so let's actually take a look at it here we've got um over on the left side here we've got our arm we've got our supply rails up here and we've got bingo a constant current generator which you have to here's the obviously the Zener reference diode here and it because to get a constant voltage out of a high precision high stability Zener or not high precision but high stability go into that later Zener diode then you need a constant knowing current through it so that's what this is designed it's got an adjustment pop there and you tweak the voltage you to tweak the current through the diode until you get the exact voltage you want and then in theory it should just start stay like that based on the temperature stability of the diode will be the main contributing factor to the drift of this absolute reference so these are voltage reference here so in constant knowin current flowing which arm is inside it's actually labeled as 6.5 milliamps so that you adjust that pot there which is part number one and they've labeled it hot number one on the board too so that's handy um so the pot number one six point five milliamps through that Zener which gives that measured voltage drop which is saw written up hand written on the label inside the unit so that's where all the stability comes from better by passing across it there and then we've got how um we've got our decade resistor networks there's the first decade there once again it shows all those trimpots we saw on that board if you remember and then that once again it's um this is the next decade then next decade and as you can see they do go down in steps 20 K 2 K 200 ohms 20 ohms 2 ohms and point 2 ohms so there you've got your six decades and as you can see because it's not a precision a real precision not kelvin valley divider it's just a standard kelvin divider then you need these little trimmer we've got little large trimmer resistors here for the various ranges to just allow you to work tweak the values there and it's not the best drawn all these rain switches are quite complicated and convoluted but pretty much the operation is pretty simple there's no overload indicator down here it's just a comparator and LM mum seven for one which then drives the overload lamp here and here's our chopper amp so it's basically just a negative feedback zero offset um chopper amp based the gain of which um is based on the based on the feedback here so it's pretty done simple and of course there's a driver that they've got an emitter follower driver here which drives the output but apart from that it's um it's fairly simplistic really and you know it's obviously good enough for the job if you tweak things and and you design it properly so that it's low drift well it works a treat and here's your output resistors are down here and as in the manual it don't match basically it's got a 3 km output impedance on the 10 millivolt range and then the hundred millivolt range has got a 300 ohm output impedance and then 3 ohm output impedance on the 10 volt range there and if you grossly simplify this circuit you end up with nothing more than a simple inverting amplifier here with a a very low drift precision buried Zener reference with a constant current source under as you can see the game doesn't need to be very high because you've already got six volts here for the Zener and our maximum output is only 10 volts so you only you know you don't even need a gain of 2-year total so effectively what you're doing is using this as a precision divider and then these are these are your six decade knobs on the front like that and of course I've left out all the little trimmer things and stuff like that to actually calibrate the thing I've left out the the emitter follower driver and stuff like that and the sense circuits this isn't actually grounded here but basically that is pretty much what we're got in here it's very very simplistic in its basic operation but that's all it needs to be exactly here's a kelvin valley divider i've been talking about well we have to start at this kelvin divider now you've seen this before in fact it is what you know as a voltage divider what most people call a voltage divider but it's Ed's real name is actually a Kelvin divided and named after Lord Kelvin obviously you you're probably familiar with now ah this is what the what is used in this instrument here you have and this is one decade okay you have a string of ten resistors like this and you tap off the various voltage that you need and that's fine if it's just one decade it's fine and daddy these could be nice high values like 10k or something like that now when you get to a multi-decade device though like on this one this one's got seven decades as you saw and really when you get to that point you can actually when you put them in series you would have another one here which effectively just short actually shorts it out so then you put multiple ones of these in series okay you would have one of these for each decade so you'd have ten resistors for each decade and then they short out etc etc but the problem with this as you saw in the schematic for this thing is that each decade must get progressively smaller in value by ten times an order one-tenth an order of magnitude okay so I've got the our first decade up here is 20 K then it's 2 K then it's 200 ohms 20 ohms 2 ohms point naught point 2 items now you think ok what's the big deal but aha think about it these are all mechanical switches in here they all have uh you know dirty contacts and you know and they bounce around and do all sorts of funny things and they're going to have a certain contact resistance a certain minimum contact resistance and really if you up you know you don't want to start off with two higher value because then your thermal noise is too high and all sorts of stuff so ah but when you progress down like that down to null point 2 ohms you write down in the territory of the contact resistance of all these switches and it's not just this decade but all of these other ones that you've got in series like this now you can of course make this work as you can see in this instrument yet you know but it's a lot of fiddling around a lot of other adjustment pots you got to have little mucking around and if you've got dodgy contacts or contacts that increase in resistance with time or where or whatever then you're going to end up with all sorts of problems so it's really not an ideal solution and that's why lots of our precision instruments will use eight Kelvin valley divider so let's take a look at that one so to get around this problem I clever dude named Varley I don't know his first name I don't know I've never bothered to look it up but it's known as a Kelvin Valley divider and it's once again it's a multiple decade system as we'll see but it doesn't use progressively lower value resistors or it does but nowhere near the order of magnitude drop for each decade that we saw on the Kelvin divider now this is actually the internal schematic of an a very high precision IET brand kvv 700 kelvin valley fold his divider and you can actually get in a box you can actually go by it and there really are very high precision laboratory-grade bits a kit for voltage division now as you can see this is a seven decade one so it has a point 1 ppm resolution absolutely incredible that's one decade more than what we've got on our voltage standard that we're playing with here today now as you can see each decade if you count up those resistors it doesn't have ten of them it actually has 11 and that's one of the keys to this thing now as you can see it won't also have just a single um a single wiper contact coming off here it will actually have two coming off there's two contacts like that that actually move in parallel like that up and down now the key to this is that because you've got eleven eleven resistors here okay and you've got the contacts on it on a second contact point one over like that then the rest of this all of the rest of these resistors on the next decade including this resistor here and this one and all the rest of them cascaded through seven that paralleled up through seven decades like that is actually equal to twenty K which then when it's put in parallel with in this case the two 10k resistors here forms a 10k resistor so those 11 resistors actually drop down to 10 so you're tapping off one tenth of the voltage at that point and that's a clever part about it and likewise it cascades through the system like that these two wiper arms here move up in unison like that so if they're up here you would find that the wiper arm would be there and there for example all right down the bottom it'll be there and there and you can tap off are the same exactly the same as the kelvin divider we saw tap off one we'll dial up one tenth for each decade a value of one right down in this case seven decades point 1 ppm resolution but look at this the lowest value resistor in here is only one K so it's nowhere near are the switch contact resistance that's going to cause an issue so you can get and you don't have to trim anything so you can actually manufacture this thing and not really worry too much about your contact resistance at all as you would in a traditional Kelvin voltage divider so there you go that's and you can go through and you can actually do the math of what values you need to put in in parallel here there's a 25k there and forty K there and there's various configurations of it as well as as well as on the output as well it might have another divider on the output here which will then tap off there's very slightly little different out configurations of it but that's basically our camp Kelvin Valley bonus divider works very obscure but very very useful now let's take a look at the specs for this unit shall we it has an absolute heart voltage accuracy on the 10 volt range of naught point zero zero three percent that's actually if you convert it that's actually 30 ppm of the setting of the actual setting you got so if you've got dot if you've got 10 volts dialed in it's going to be 30 ppm of that absolute accuracy but you've got to add on another 5 ppm here basically the difference between percentages and ppm is once you get below about you know naught point oh one percent you just it's bit of an industry standard to start talking in terms of our ppm parts per million instead of percentages they're exactly equivalent but you know it's too the industry speak really you gonna once you're getting the small stuff ah you know it's a bit more bit more professional to talk in terms of ppm because it it sounds more impressive and it's easier to work with than you know throwing in X number of zeros and things like that now we also have to add on to microvolts and we have to do the math to figure out what that's actually going to be and that's 5 ppm of the actual range and the range of course down here is the full scale range of 11.11 one-vote so if you work those out 30 ppm of the set in in this case we might use a 10 volts and then we would have 300 micro volts absolute error there plus we have to add on 55.5 micro volts because the percentage error of the range + 2 micro volts a total of 357 odd micro volts but one of the keys to something like this is the stability and it's got a good quality instruments will give you different stability over different periods in this case it's given you over one day eight hour period of plus/minus naught point oh one percent or 10 ppm stability so it won't drift any more than that over a period of eight hours and over the period of a year it's still fairly tight at 25 ppm once again plus two micro volts and even for the 10 millivolt range down here it's actually identical point SS 3 percent 0.035 percent basically plus the two micro volts now that 4 10 millivolt range that's going to be 10 point zero zero zero three seven millivolts so coming back to our unit here if we've got it set to our 10 volt range here then that means it's going to be 10 point zero zero zero three seven or plus minus 3 7 that is its absolute value so when you are calibrate this thing that's all it's guaranteed to be so these last two digits especially the last one you're really getting down into the noise effectively in almost it's not quite useless because from an absolute accuracy point of view it is but because you can actually dial it in there it's resolution can be handy and as we've talked about before that's the difference between accuracy and resolution just because your accuracy is not the same as your resolution doesn't mean the resolution isn't useful for various purposes and the other thing of vital importance is the temperature coefficient here and it's plus/minus an or point zero zero zero five percent per degree Celsius and if you translate that over to our dials here it's point zero zero zero zero five percent per degree Celsius change so if you're sitting in the lab and your temperature changes by one degree you've effectively dialed and extra five on to that digit there but that's going to be a maximum worst-case value it's most likely in practice going to be better than that substantially better but as always in electronics you should take especially when you're doing precision stuff and really serious worker you got it Kate and take that worst case spec now if you're curious to know what reference diode is used in this thing I looked up the parts list for it and it's actually a art one in eight to one a IU device now it's available from different manufacturers I don't know which manufacturers the one used in here it's most likely not a micro semi but this was the best data shooter I can find for an equivalent art second sourced part now as you'll see down here this is a naught point there it is no point triple oh five percent per degree C and you probably recognize that because that is the temperature coefficient a spec of this actual unit that's based purely upon the reference diet as you'd expect now one of the curious things that might puzzle a lot of people with something like this is that look at the reference value it's a plus minus 5 percent it's got a huge tolerance and is the difference between absolute accuracy and temperature coefficient over here you can design a high precision instrument like this high precision high stability using a plus minus 5% reference diode because you because you're not relying on the factory or the you know the actual manufactured absolute tolerance of this thing because you calibrate it you tweak those knobs you dial it in all you care about is the stability of it so one with temperature because in the end and age as well as another thing but let's not go there but let's just talk about temperature itself that's all you care about as long as you've got a very low temp Co but it's as much called a tepco or temperature coefficient reference device in there in this case it's a xena temperature compensated diode then as long as this figure is really really low you can and you keep that diode at the same temperature or it's so low that it doesn't matter then you can just tweak your circuit adjust it and bingo you've got a high precision high stability circuit but the problem is you've got to actually have access to a you know and some sort of you know secondary or cow lab our standard equipment to do the calibration in the first place but if you've got that as any good manufacturer does then you can do it not a problem and if you're wondering what reference is used in the legendary HP 3478 a multimeter we've we used before a industry standard bit of kit it uses and off-the-shelf reference diode in this guy that's the end which you can't get anymore it's been obsoleted but it's the lm-2 9:9 and basically all it is is once again as I mentioned it's a Zener diode in there it's a very DIF very high stability Zener diode just like the micro semi one used in here except it's got a little internal heater inside the package which are keeps it and it's got some regulation circuitry for the heater in there that keeps it at a constant temperature so if you've got it at a constant temperature and you've got a constant current flowing through your diode not a problem it's going to be very very high stability and let's look at the spec there it is plus might add two percent right it's horrible that's the initial tolerance of this thing you buy it and it can be 2% out it's useless right you can't use that in a precision five digit multimeter it's hopeless uh-huh but look at the temperature coefficient point zero zero zero one percent or one ppm per degree Celsius fantastic so as long as you calibrate and tweak this thing it'll stay stable so getting back to our one in eight to one diode here it's available in various grades in this case we've got the top grade which is the one in eight to nine AIU with the lowest art Temko but this is from micro semi they would have thought you know it's a second source well you know as a multiple source part they would have actually used because this doesn't quite meet the tempo spec of the unit they've obviously got an even better device like this now I searched find chips-- comm and one manufacturer had this for our $80.00 another one had it for six dollars so you know prices all over the shop and really they would have sourced a really high spec our unit probably our Bend and individually selected from a specific manufacturer a very reputable manufacturer not just you know left it up to their purchasing people to purchase it from one hole oh in China or anything like that would have been very specifically aspect and probably even up Bend for them and they might have even done their own in-house spinning as well now here's some interesting stuff we love curves don't we curves characteristic curves are brilliant they tell you a lot now in this case it is the change in temperature coefficient I'm in percent per degree Celsius based on the operating current or the constant current through that diode and they specify at 7.5 milliamps here which it has the best it's not actually zero it's actually naught point zero zero five percent per degree C if you read the notes down there but if you operated at any other current then you're going to be elsewhere on that curve and you've got to take that into account now this unit here uses six point five milliamps but as I said it probably it likely uses a diet from a different manufacturer might be slightly different you know who knows they've taken into account everything's hunky-dory the other thing to consider with your operating current this is even more critical that it doesn't actually change with temperature your operating current must stay stable and here's why look it here is your change in the in your Zener voltage in hundreds of millivolts 100 millivolts 200 millivolts per your operating current here so if you're operating at 7.5 milliamps here okay you smack on zero but if you change it the operating current wait just a smidgen you know nor point one you can be out by will in this case 100 millivolts in 6.5 volts of what we are one and a half percent it's massive right it's huge that's not precision that's you know two dollar one on low multimeter kind of you know accuracy so it's critical absolutely critical that your operating current remains stable and it remains stable over temperature otherwise you're not going to have a precision instrument like this that's critical so that brings us back to our constant current circuit here this is just as critical as the reference diet itself this must put a very precise current through this diet once again it doesn't matter the absolute value of it because you can you know trim it to any value you want here but it must be completely temperature stable so they would have tested that circuit to the hilt and I designed it so that it has a as good a tempo as the reference diode itself or possibly even better well it's even that or they've got some really tricky dick spec special buried Zener diode in here which doesn't actually change much on that curve which maybe has a flatter a characteristic curve like that who knows you'd have to know the exact one and where they source it from and get the exact data sheet from it but anyway that current can be just as important and I'm here at trio smart cow in the Nardo cat web and we're going to check it out against the HP meter and I've got jowls hold him here you've seen last time he chose what we've got here is a killer Packard stroke Angela 3458 a which is basically the word standard in young scale multimeters and what we want to do is take your box which is actually a 30 ppm for a year yep that's got about a test uncertainty ratio of about 3.5 to the standard version of this meter so that's close enough to give you an idea of just how good sounds near enough what's your email purchases so the uncertainty ratio is only what three point three five five two ones I really would like it was close to ten to ten or magnitude yeah but you know you haven't got that in my domain know exactly they're okay but they're this is good enough let's give it its turn it on let's alright well have a look at the reading it turn on yep and then we let it warm up and we'll come back and say twenty minutes or so and see oh well it's stabilized excellent well let's I don't what do we get when is that so we're just hooked up simple to wire I pay the system so we'll turn it on you've got an overload like on there yeah it comes on temporarily until I what's there we go there we go so that's switch on you've got are you better than there anywhere I'm going oh oh one percent already yep that's pretty good there you go that's too bad let's come back in about twenty minutes now look all right what's the temperature in the room you and that is 20 degrees plus or minus one degree 20 they actually specified this its calibration temperature at 23 is that like an older that's an older style almost standard then your standard is 20 yeah where will we operate this one at about 20 okay excellent well we'll come back later and see what happens all right way back half an hour later and what do we got Charles back okay well what we did is we're on the 10 volt range and as you can see we're on the fifth digit to the fifth digit on the 10 volt range would indicate that parts per million yeah so one there would be one part per million we've actually got about three and a half parts per million error looking at this so absolutely assuming that this is absolutely at this is right yes it's right this is actually quite amazing given the fact that you bought this thing on eBay and it's anything within a couple attends with them that was last calibrated 2007 2007 so I think you've got yourself a bargain a bargain yeah fantastic don't go setting up a rival caliber this hour you know let's turn it down to a volt Joe no you will have to allow for steady yep it will actually do take time you say we'll ask us that but even so this is pretty good yeah you're right 7 ppm of their own yeah 7 8 ppm so giving it the specs 30 it's pretty good I like it yeah take it home and never turn it off I don't leave it up to anything you know the thing is with us old things you know the older that we get the more stable we get and the same applies to the meter and the same applies to yeah the calibrated be burnin and we just get better as we get older all right excellent looks like it's what I was thought that we'd have to tweak a few pots but obviously not no drinking necessary no tweaking necessary that's good now so I can now take your time use it as a transfer standard to Trent to calibrate my HP bench me to my 3478 yes so let me yeah yep you could certainly do it for your voltage anyway 1.8 ppm out on 10 volts range let's say wow you thought you got to get the drift no idea there's a little bit of drift but that's 900 same Oh I'll sell it to your child yeah how do you pay for it we'll give you the ball I'll give you double with you double what I paid all right look at this this is obscenely pornographic it's a good box it is a good box I like I was I thought I bought the instructions on how to calibrate and everything but we don't I don't think anything I don't think we need to like at liesl eventually out settle down you know that you know that's just that's ridiculous well you've also got the settling time demeanors Edie yeah exactly it depends of these switches here are actually just operating on a passive debater so they are just a passive developer doesn't passive today here it's going to change this or so because it's a it's a chopper and there's a chopper ramp in there it's used in the feedback loop of the chopper area but it is a passive divider so let's go to 99.9999 96 ah too good not even gonna wait for it settle it's already too good too good for you but not that I'm bragging it's ah this is silly linear - I like it we already looks excellent you you and if we check out the zero error there because I've got the output actually switched off let's switch that on and it's it's increased a little bit but that's a pretty good at zero offset error I like it yeah so we're really down in the down in the noise here because these digits actually match the digits here on the on the voltmeter so it's you know where it is fluctuating down in the noise there we could dick around and try and actually get its performance a bit better in various ways but this is actually a shielded test cable we're actually using here but yeah as you can see it's you know you've got to not play around with it and so we are down in the noise pretty much but if we bump that up by 0.1 you don't see it yet there we go okay now we're starting to jump up so point 2.3 and it's starting to not quite follow that but you know we are right down in the noise region where you know where our thermal noise and stuff like that comes into play based on your connections and your type of metals and stuff like that it can actually become an issue so it's a bit of an art of showing that low stuff but if we go up in the to the 100 millivolt range then we can play around with that you

Channel: EEVblog

Views: 78,511

Keywords: Krohn-Hite, mv106, mv106j, dc, voltage, standard, teardown, calibrate, calibration, lab, burried, zener, diode, reference, lm299, lm199, lm399, fluke, hp, agilent, hp3478a, 3478, 3478a, laboratory, nata, certification, schematic, parametric, graph, curve, temperature, compensation, kelvin, varley, divider, datasheet

Id: onqsjDJq4I0

Length: 50min 17sec (3017 seconds)

Published: Thu Oct 20 2011