Hi check out what just turned up on my bench. Whoa. look at it. Is't it beautiful? It's a um electronic uh Development Corporation uh MV 106j DC voltage standard look at all the knobs I love it Precision City it's awesome.

um it's now uh done I think it's it's owned by a company now called Kronite. They still sell it, it's still a current uh model and uh I thought we'd uh turn it on. check it out, see if it still works, calibrate it, look at how it works, internal construction, tear down bit of theory maybe and uh should be fun. Let's go and here it is.

It comes in a rather old style aluminium uh rack size uh cabinet. It's got a uh, it's got a little aluminium tilting uh baale like this but it's also got another um flip out tilting bail like that which is really quite nice I like it and uh it's in Fairly good uh Nick I'm not sure um how old it is I think it's um the manual um which you can download says it's 19 or last update in 1991 no 2001 sorry so it's like a a 10-year-old model but it could be based on older than that cuz it seems quite old school with the Uh with the old style uh, rotary encoded NOS but maybe it's been updated or something like that I'm not sure uh full scale of Uh 10 Vols or 100 MTS or 10 M volts three ranges there are. There is another type which has a current output as well, but uh, this one uh doesn't Now the uh, cow sticker here. if you take a look at that, it's uh, last cowed in uh 200.

Well, 2007 so due 2008 and made in Boston Massachusetts Do they still make anything in 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 um, the uh sense outputs as well cuz when you're talking about a Precision uh DC instrument like this, it's got to uh, sense the outputs.

But it does have uh, little short in bridging bars in there. If you're just hooking it up to a voltmeter then that's uh, fine. You don't need the sense output so can just be tapped directly off the output. and you can just use a two wire instead of four wire.

Um, you can invert the output. uh, nothing unusual there. and I love it's got six, uh, decades here and I love these style uh, rotary selection knobs and you'll notice I'm quite a fan of those. I've got uh, quite a few of the uh Keithly instruments there I've actually got four of those and I just love those uh style knobs.

So I love this thing. It's brilliant and of course, uh, you can set. um let's say we got 10vt full scale. here.

you can set 10 0000001 volts. That's the resolution of it. Beautiful. It's uh 10 uh, microvolts resolution on the Um 10vt Range but if you go down to the 10 molt range, we're talking uh, 10 n volt resolution.

Wow. Nothing exciting on the back I'm afraid. just the model number and a 240 volt selection switch. So um, it was uh 110.

switched it over to 240 and uh, we'll see if this sucker works all right. We'll just try it with my fluke 87 here. First, as I said uh, we've got the uh current sense input shorted with the current bars there. Ideally, if you had a uh, a big load, you would actually take the sense directly onto the load uh 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 Vols and hey, it lights up overload and uh, y, the overload lead is still on. Oh no, there we go.

It's gone out. Looks like it requires some settling in time perhaps. Um, and bingo There we go. 10.0 volts and we're pretty darn spot on.

Almost to the least significant digit there, so it looks like it. uh Works let's change it to 100 m volts there. and ah, beautiful, look at that 100. spot on and uh, let's we've got 100.

well let's whack the this up one notch. That should give us one least significant d it and it does. There we go, changing one at a time. Wow, it's spot on and of course we' got two decades below that, which won't do a damn thing, but if we turn that up by 10, it should jump to eight.

And it does. There we go, so it looks like it's um, it's going to still be well, I'm not going to say within spec, it's within um, you know the 05% of my uh fluke 87 here, but uh, this thing has specs of uh, what are they um, uh, 0.003 uh or 30 uh 30, 30 PPM over the temperature range. Now this thing actually has Um specs of uh plusus .3% or 30 uh PPM over um, well this the span of a year um and uh Point uhp 5% or five uh PPM uh per degree C temperature drift So it's not super high-end as far as Uh DC uh standards go, but it's not too bad at all. It be a nice addition to the lab here.

I think just for uh A A Precision voltage source for uh, testing, meters and things like that, but it seems to work just fine. That's the 100 molt range and we go down 10 molt range. Yep, not a problem at all. I like it.

It's nice and uh, there doesn't seem to be any noise on the pots either. With these things, you've got to 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 no, 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, uh, still within um spec. Well, I'm not going to say within spec because it's you know, uh, only within.

um, pretty. You know it's pretty close. um to to meet to actually check. its uh, performance? uh spec.

Um, 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 uh, it basically works. So I'm pretty Dar happy with that now. I could check the uh linearity of it and uh, things like that.

but generally with these sort of things, you don't have to because the linearity is set by by Precision uh, voltage divider resistors. um, set up with these pots either in a kelvin, or more, uh, more, uh, popular format as the Kelvin Valley um configuration. So basically, these things shouldn't drift unless those actual resistors. uh, drift.

which, uh, generally doesn't happen too often. It's more likely that the uh, voltage reference uh, diode in there would drift or something like that, but even those are pretty darn stable. So I'm not surprised that this thing, um, still works cuz there's not much in them to go wrong. Uh, we'll find that there's only a power supply in there, a reference diode, an amplifier, and a bunch of a bunch of precision resistors on these knobs.

Here, let's go one step further than the fluke 87, Shall we? and uh, get my HP 3478a uh bench meter which is actually got going to have a similar, um, short-term uh drift capability to this uh DC voltage standard. Uh, and we've got it set to 10 o there and it's not too far off at all. But I Haven't checked the Cal in this in quite some time and of course you know they 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, maybe even 27 see in the lab here.

So um, you know it's not exactly ideal calibration conditions, but we'll fix that later. now. let's uh, go down to 100. Ms There you go, That's not.

Uh, that's not too shabby at all. Let's go down to 10 ms 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 Uh 3478a is out or this is out or a combination of both or whatever. Um, I'll have to take it to a U Standards Cow Lab to find that out. but if we can calibrate this baby, um, then we can use that as a transfer standard to then calibrate.

Um, my 3478 I Hey, that would be neat. All right, let's open this sucker up and see what we got inside. Now as I said what you're going to 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 an uh like an old school uh PCB with all through hole components.

Obviously, you'll have a uh power supply um Main's power supply. you'll have a uh Precision um uh voltage reference which will be like a Zena or uh, you know, a buried reference uh Zena um uh, temperature temperature controlled uh and it'll have a Kelvin Valley divider um on the front and an amplifier and that's pretty much all it is. It says it's got a chopper amp I don't have the full manual yet but um I've asked for it so that should, uh, turn up shortly I hope and let's pull it off and see what we get. Power's off? Yep, hey, there we go.

Tada and that's not too too far off at all. It's pretty much what I expected now. Uh, one of the first things I uh noticed though, is that um, quite possibly. Uh, this doesn't look like a Kelvin Valley uh voltage divider Arrangement it just looks like a standard uh Kelvin voltage divider.

And the giveaway there is that there appears to only be a single link between each Uh decade. Bank There not a jewel link as you'd uh, expect and a more complicated uh switch Arrangement as you'd expect on a Kelvin Valley which is, uh, what you'll find in my uh, keithly, um instruments. Over here they use Kelvin Valley uh dividers in them, but this one looks like it just uses standard Kelvin divider. We'll have to get the schematic to verify that, but there you go anyway.

It's a whole bunch of uh, very precise resistors on there. We'll have to check out the value. Some uh, trim pots over here on the main decade? that's your, um, that's your first decade there and it looks like it's got um trim pots for all the various ranges. You got a main board uh down here and you've got a Main's Transformer Now one of the interesting things to note about the Main's Transformer Here is, look at all the exposed wireing here.

It's um, there's no heat shrinking, uh on that at all. It's all totally exposed as is the Uh switch on the main switch on the front panel and that pretty that you wouldn't get that past these days. But um yeah, it's uh P past the safety standards anyway. and you'll see down here this is the Main's cable input and there's like an insulated uh standoff there which is just used to join the two wires and once again, fully exposed.

So um, there's not really much uh put into the um, you know, internal safety of this instrument, but uh, that's you know, typical old school stuff and your output Jacks over there. Uh, they've got gold plated. You know they'd be really good quality. uh, goldplated, uh contacts and things like that, crimps and uh, going off.

the Uh wirings are recently neat and tidy. I guess As for holding the Uh PCB in place here, this is, uh, very old school construction technique. They've got the Um aluminium Uh plate with the cut out for the board and the board uh, screwed on the underside there and they've just cut it out for the Uh components. Um, quite.

Uh, you'll see this construction method used quite a lot in these sort of Uh rack mount. Um, early early designed instruments. Now one of the first things I I'm looking for in something like this is where is the Uh Voltage Reference standard now? I've got this curious looking Uh package here and um, but I don't think that is the voltage reference I Get a feeling that that's not it I'm not actually sure what that is as well and at all. It's got two, uh, common terminals.

here. it's just a four terminal device, two pins on either side. There's an in, there's an out um, but that's not what you'd expect in a Uh temperature compensated uh, buried Zena reference like um in the Um HP 3478a multimeter. we just uh, had had a look at.

It's got an LM Uh 299 temperature comp ated voltage reference in it and uh, that's basically just a bed buried Zena um 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 if 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 um, you know, just uh, 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 4593 at 61730 Vols at 6.5 milliamps Uh current. So um, obviously they've and because that that diode is in, uh, those like solded after um, 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.

A fully expected a temperature you know, with uh, what is it5 um, or five uh parts per million uh temperature coefficient? uh without getting that without a uh temperature, um, compensation on it, that must be one hell of a good uh diode. So we're going to have to wait until we get the Uh manual for that and uh, see what part number it is now I've turned the power on and let's see if we can actually measure that Dio down in there and if I'm correct, it should be 61730 Vol So let's give it a go. Bingo 6173 Vols it is. now.

As for the rest of the board here, there's no uh, silk screen at all. somebody's handwritten you know? R4 R3 r2a R2 R1 there. no silk screen. really old through hole, uh, kind of stuff.

and uh, these voltage Regulators Here look at the check out the pins. the spacing is not even on them. They've had to spay those pins out. that's you know that's a bit bodgy.

I Don't like that at all. and we've got a date code on this. Um LM 741 here of uh, the fourth week 91. So there you go that dates this unit to uh, at least uh, 1991.

Now of course one of the things you notice here is this, uh, little daughter board here. um and I am I'm going to assume that that's the uh Chopper uh amplifier cuz they talk about that in the manual So, but it almost looks like it's a bodged 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 taada.

There we go. Look at that that that is. It looks like they've purposed designed that. um, it was actually designed I Thought it looked like a bodge board, but it's not.

It looks like they actually uh, manufacture that module. um, as a separate thing. they probably manufacture and uh, test it. I mean there's no, uh, trim pots on there, but maybe they, um, you know, they test things.

They actually select the components and uh, put them on there or something. But yeah, it's all. it's all very, uh, it's all very old school school and uh, I'm quite surprised at the whole thing. No.

Kelvin Valley at the impreciseness. really? um, of the whole thing, you know they're not using a temperature compensated uh reference Dio down here, they're not using a Kelvin Valley divider. but hey, I you know they've they've got away with it it. I'm sure it uh, will meet its uh claim specs and you can see the Uh Precision resistors on the Uh decade switches there.

that's a 2K uh, plus- 0.005% Precision resistor and uh, as you go, they won't all be that precise, but all the ones over there will be like the ones, um, further down. Uh, you know they don't have to be as precise. uh, not necessarily as you go down, but at least some of them are at least 0.005% and the ones down here on the first decade here, these are at 0.02% Uh, they don't have to be as precise because each one of of those is uh trimmed to its actual value actually I just looked up that part number the Uh TSC 7652 There that is actually a chopper stabilized amp. So there you go and I've talked about Chopper stabilized amps before.

The reason you're going to use one of those in an instrument like this is that it has essentially it zeros out. it nulls out any uh 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 got to use a chopper stabilized amp for that. And thanks to Joe Englas from Kronite, we've got the schematic for it in record time and what can best be described as the manual.

It's a bit of a hotch poch actually the manual, but uh, 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 MV 106j. this just says MV 106 so I'm not sure what's uh actually going on there, but uh, anyway, it's um, pretty much uh what.

I expected apar from the Kelvin Valley thing in the temperature controlled 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 um, we've got our supply rails up here and we've got Bingo a constant current uh generator which you have to. Here's the Uh obviously the Xena reference diode here and um, it because to get a constant voltage out of a a high Precision High Stability Zena not high Precision but High Stability.

Go into that later Zena Diode. Um, then you need a constant knowing current through it. So that's what this is designed. It's got an adjustment pot there and you tweak.

Uh, the voltage. You tweak the current through the diode until you get the exact Uh voltage you want. and then in theory, it should just, uh, stay like that. Um, based on the Uh, temperature, stability of the diode will be the main contributing Uh factor to the drift of this absolute reference.

So there's our voltage reference here. so a constant no and current flow which Um is inside. It's actually labeled as 6.5 milliamps so that you adjust that pot there which is pot number one and they've labeled it uh, pot number one on the board too, so that's handy. Um.

so we got pot number one 6.5 milliamps through that Zena which gives that uh, measured voltage drop which you saw written handwritten on the label inside the unit. So that's where all the stability comes from. bit of bypassing across it there. And then we've got our Um, we've got our decade resistor networks.

There's the first Uh decade there. Once again, it shows all those trim pots we saw on that board if you remember and then uh, once again, it's um, this is the next decade, then the next decade. And as you can see, they do go down in steps 20K 2K 200 Ohms, 20 ohms, 2 Ohms and 2 Ohms. So there you got your six decade.

And as you can see, um, because it's not a Precision a real Precision Kelvin Valley divider, It's just a standard Kelvin divider. Then, uh, you need these little trimmer. We've got little uh trimmer resistors here for the various Uh ranges. Um, to just allow you to uh, tweak the values there.

and it's not the best drawn. All these rain switches are quite complicated and convoluted, but um, pretty much the operation is pretty simple. There's our overload indicator down here. it's just a comparator, an LM 741 which then drives the overload lamp here.

And here's our Chopper amp. So it's basically just a negative uh feedback. Um, zero offset. um Chopper amp based the gain of which Um is based on the Uh based on the feedback here.

So it's It's pretty darn simple. And of course, there's a driver. They've got an an emitter follow a driver here which drives the output. but A Apart from that, it's um, it's fairly simplistic really.

and uh, it. 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.

uh down here and uh, as in the manual it uh, match. Basically, it's got a 3K um output, uh impedance on the Uh 10 molt uh range and then uh, the 100 Molt range has got 300 ohm output impedance and then 3 ohm output impedance on the Uh 10vt range there. And if you grossly simplify this circuit, you end up with nothing more than a simple uh inverting amplifier here with a uh, very low drift Precision uh buried Xena reference with a constant current source and uh, as you can see, the gain doesn't need to be very high because you've already got six volts here for the Zenner and our maximum output is only 10 volts. So you only you know you don't even need a gain of two here 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 uh, the uh, emitter follower driver and stuff like that and the scent circuits.

This isn't actually grounded here, but basically that is pretty much what we've got in here. It's very, very simplistic in its basic operation, but that's all it needs to be Exactly is 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 its real name is actually a Kelvin uh divider named after Lord Kelvin Obviously who you're probably familiar with now. Uh, this is what the what is used in this instrument.

Here you have. this is one decade. Okay, you have a string of 10 resistors like this and you tap off the various uh voltage that you need and that's fine. If it's just one decade, it's F 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 uh, 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, uh, 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 10 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 10 times in order: 1/10th in order of magnitude. Okay, so we got the AL First decade up here is 20K. Then it's 2K. Then it's 200 Ohms 20 Ohms 2 Ohms No.2.

Ohms. Now you think. Okay, what's the big deal, but aha, think about it. These are all mechanical switches in here.

They all have uh, you know, uh, 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, uh, you know you don't want to start off with too high a value cuz then your thermal noise is too high, and all sorts of stuff. So uh, but when you progress down like that down to 0.2 Ohms, you're right 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, uh, work as you can see in this instrument. yeah, you know, but it's a lot of fiddling around, a lot of other adjustment pots, you got to have lot of mucking around, and uh, if you've got dodgy, uh, contacts or contacts that, uh, 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 uh Precision Instruments will use a Kelvin Valley divider.

So let's take a look at that one. So to get around this problem, a clever dude named Valley 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 uh, it's once again, it's a multiple decade uh system as we'll see, but it doesn't use progressively lower value resistors. Or it does, but nowhere near the order of uh magnitude drop for each decade that we saw on the Kelvin divider.

Now, Um, this is actually the internal schematic of an um, a very high Precision IET brand um, Kbd 700 Kelvin Valley voltage divider And you can, Actually, it's in a box. You can actually go buy it and they're really, uh, very high Precision Laboratory uh grade bits of kit for voltage division. Now as you can see, this is a seven, uh, decade one. So it has Uh1 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 resistance, it doesn't have 10 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 a single wiper contact coming off. here. It will actually have two coming off.

There's two contacts like that that actually move in, uh, parallel. like that, up and down. Now, the key to this is that, uh, because you've got 11 11 resistors here. Okay, and and you've got the contacts on on.

Uh, 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 cascad through seven, paralleled up through seven decades like that, is actually uh, equal to Um 20K. Which then when it's put in parallel with in this case, the 21 K resistors here forms a 10K resistor.

So those 11 resistors actually drop down to 10. So you're tapping off 1/10th of the voltage at that point. And that's the 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. or right down the bottom, it would be there and there.

and you can tap off Uh, the same. Exactly the same as the Kelvin divider we saw. tap off one or dial up 1/10th uh for each decade. a value of one.

Right down to, in this case, seven decades 0.1 PPM resolution. But look at this. the lowest value resistor in here is only 1K. So it's nowhere near.

uh, the switch contact resistance that's going to cause an issue so you can get and you don't don't have to trim anything. so you can, uh, actually manufacture this thing and not really worry too much about your contact resistance uh, 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 40K there and there's various Uh configurations of it as well as as well as on the Um output as well.

It might have another divider on the output here which will then uh tap off. There's very slightly little different configurations of it, but that's basically how a Kelvin Valley voltage 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 Uh voltage accuracy on the 10vt range of 0.003% That's actually if you convert it that's actually 30 PPM uh of the setting of the actual setting you got.

So if you've got if you got 10 volts uh dialed in, it's going to be 30 PPM of that absolute accuracy. but you've got to add on another five PPM here. Basically the difference between percentages and Ppms is when once you get below about um you know 0.01% you just it's bit of an industry standard to start talking in terms of uh PPM parts per million instead of percentages, they're exactly equivalent. but you know it's just the industry speak.

Really, you're going to once you get into small stuff. ah you know it's it's a bit more uh, bit bit more professional to talk in terms of PPM cuz 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 two microv volts 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 Uh range and the range.

Of course, down here is the full scale range of 11111 Vol So if you work those out uh, 30 PPM of the set in, in this case, we might use a 10 volts and then we would, uh, have 300 microvolts. Absolute error there. Plus we have to add on 55.5 microvolts because the Uh percentage error of the range plus 2 microvolts a total of 357 odd microvolts. But one of the keys to something like this is the stability.

And it's going to, um, good quality instruments will give you different stability over different periods. In this case, it's given you over a one day8 hour period of plus - 0.01% or 10 PPM stability so it won't drift any more than that over a period of 8 hours and over the period of a year, it's still fairly tight at 25 PPM once again, plus 2 microvolts. and even for the 10 molt range down here, it's actually identical 03 % uh, or .35% basically. plus the 2 microvolts.

Now that for a 10 molt range, that's going to be 10 000037 molts. 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 10.0000 37 or plusus uh, 37, that is, its absolute uh value. So when you, uh, 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, uh, into the noise effectively.

You can almost say it's not quite useless because, um, from an absolute accuracy point of view it is. But because you can actually dial it in there, its 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 plusus: 0.005% per degree celsius. And if you translate that over to our dials here, it's 0.005% per degree celsius change. So uh, if if you're sitting in the lab and your temperature changes by 1, you've effectively dialed an extra five onto 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 work, you got to take that worst case spec.

Now, if you're curious to know what uh reference diode is used in this thing I Looked up the parts list for it and it's actually a 1n 821 Aru uh device. Now it's available from different manufacturers I Don't know which manufacturer is the one uh used in here. It's most likely not a micro semi, but uh, this was the best data sheet I could find for an equivalent uh, second sourced part. Now as you'll see down here, this is, uh, not.

point there. it is 05% per degre C and you probably recognize that because that is the temperature coefficient uh, spec of this actual unit. It's based purely upon the reference diode, as you'd expect. Now, one of the curious things that might puzzle a lot of people with something like this is that, uh, look at the reference value.

it's uh, plusus 5% It's got a huge tolerance and that 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 Plusus 5% reference diode because you because you're not relying on the factory uh 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 and and age as well is 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 Tempco What? What's called a Tempco or temperature coefficient? uh reference device in there. In this case, it's a Zena um, uh, temperature compensated? uh, 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, a some sort of uh, you know, secondary or cow lab uh, 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 3478a mold meter we we used before um, industry standard bit of Kit it uses and off-the-shelf uh reference. Um, Diode. In this, it's the which you can't get anymore.

It's been obsolet, but it's the LM 299 and basically all it is is once again, as I mentioned, it's a Zena Diode in there. it's a buried reference. very Uh, High Stability Zena Diode Just like the Micro 71 used in here, except it's got a little internal heater inside the package which uh, 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 M at 2% 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 5-digit multimeter. It's hopeless. Aha, but look at the temperature coefficient: 0.00001% or 1 PPM uh per degree Celsius.

Fantastic. So as long as you calibrate and tweak this thing, it'll stay stable. So getting back to our 1 1821 Uh diode here, it's available in various grades. In this case, we've got the top grade which is the one in 8 29 Aru with the lowest uh Tempco but this is from uh micro semi They would have uh, you know it's a Second Source Um, Well, you know it's a multiple Source part they would have actually used because this doesn't quite meet the uh Tempco spec of the unit.

They've obviously got an even better device like this now. I Searched uh, Fine Chips.com and one manufacturer had this for $8 another one had it for $6 So you know prices are all over the shop and uh really Um, they would have sourced a really uh highs speec uh unit probably uh Bend and individually selected from a specific uh manufacturer. a very reputable manufacturer. not just, you know, left it up to their purchasing people to purchase it from one hung low in China or anything like that would have been very specifically uh specked and probably even uh bined for them.

And they might have even done their own in-house binning as well. Now here's some interesting and stuff we love: curves, don't we curves characteristic curves are brilliant. They tell you a lot. Now in this case it is the uh change in temperature coefficient um in in um, percent per degree celsius based on the operating current or the constant current through that diode.

and uh, they specify it at 7.5 milliamps here, which it has the best. It's not actually zero, it's actually 0. uh, 005 % per degree C if you read the notes down there. but if you operated at any other current, uh, then you're going to be Elsewhere on that curve and you've got to take that into account now.

Um, this unit, uh, here uses 6.5 milliamps, but as I said, it probably it likely uses a died 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 um, actually change uh with temperature. your operating current must stay stable. And here's why.

Look here is your uh change in in your Xena voltage in hundreds of Mill volts, 100 molts, 200 MTS 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 by just a smidgen, you know .1 you can be out by well, in this case, 100 MTS and 6.5 Vol So what? 1 12% It's massive, right? It's Huge. That's not Precision that's you know.

$2 One H 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. It'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, uh, put a 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, uh, tested that circuit to the hilt and uh, designed it so that it has uh, as good a Tempco as the reference diode itself or possibly even better. Well, it's either that or they've got some really tricky Dick Speec uh, special, uh, buried Xena diode in here, which uh, doesn't actually change much on that curve, which maybe has a flatter, uh characteristic curve like that? Who knows? You'd have to know the exact one and where they sourced 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 N Cow lab and we're going to check it out against a HP meter and I've got Charles Holm here who you've seen last time? hey Charles What we've got here is a H Packard Stroke adelent 3458a which is basically the world standard in Long scale multimeters and what we're going to do is take your box which is actually uh, a 30 PPM per year. Yep, has got about a test on certainly ratio of about 3.5 to the standard version of this meter. so that's close enough to give you, um, an idea of just how good sounds, near enough your eBay purchases. So the uncertainty ratio is only what 3.5 3.5 to one ideally would' like.

you know, as close to 10 to one 10 magnitude. but uh, you know you haven't got that amount of money. No, exactly neither of we, but this is good enough. Let's give it a go.

Turn on What we let's have a we'll have a look at the rating at Turn On and then we'll let it warm up and uh, we'll come back and say 20 minutes or so and see how well it's stabilized. Excellent, Well let's have a look. What are we getting now? let's Uh, so we're just hooked up. Simple two wire I feeding system so we'll turn it on.

You've got an overload light? Come on there. Yeah, it comes on temporarily until it there we go. So uh, at switch on you've got uh, are you better than uh, are you better than 0.001% ready? So that's pretty good there. Let's come back.

let's come back in about 20 minutes and have a look. All right, what's the temperature in the room here? Um, that is uh, 20 Plus or - 1 20. They actually specify this. its calibration temperature at 23.

Is that like an older? That's an older standard, older standard. The new standard is 20. Yeah, we we. Well, we operate this one at about 20.

Okay, excellent, Yeah, Well, we'll come back later and see what happens. All right, we're back half an hour later and what do we got? Charles We're back. Okay, well what we've got is uh, we're on the 10vt Range and uh, as you can see, we're uh, on the fifth digit. So the fifth digit on the 10 Volt range would indicate the parts per million.

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 absolute error Assuming that this is absolutely assuming that this is right, and this is right. This is actually quite amazing given the fact that uh, you bought this thing on eBay and it's and it's within a couple.

and it was last calibrated um July 2007. So I think you got yourself a bargain? A bargain? Yeah. Fantastic. Don't go setting up a Raval C Lab with this now, will you? no? Let's uh, turn it down to a vault, shall we? No, you will have to allow for stabiliz.

Do take time to stabilize, but even so, this is pretty good. Yeah, you're uh, about 7 PPM at the moment? Yep, 78 PPM So given that the Spec's 30, it's uh, pretty good. I like it. Yeah.

take it home and never turn it off right? I'll leave it powered up 247. But 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 the calibrated they burn in and uh, we just get better as we get older. All right.

Excellent. Looks like it's spot. I Was thought that we'd have to tweak a few pots, but obviously not. No tweaking necessary, no tweaking necessary.

That's good enough. So I can now take this home. use it as a transfer standard to trans to calibrate my HP bench meter. My 3478? Yes, Absolutely.

Yeah. Yep. You could certainly do it for, uh, your voltage anyway. 1.8 PPM out on uh 10 Vols range.

Let's uh, well, you, you've got a little bit of drift. There's a little bit of drift. but that's nine. It's just same.

I'll I'll sell it to you Charles Yeah I Give you what did you pay for it. We'll give you I'll give you double what you PID double what I paid. All right. Oh, look at this.

This is obscenely pornographic. It's a good box. It is a good box I Like it I was I thought I bought the instructions on how to calibrate and everything. but we don't I don't think we need to I don't think we need to like this.

will eventually. Uh, settle down. You know, you know, that's just. that's ridiculous.

Well, you've also got the settling time and the meter have take into as well. Exactly. It depends if these switches here are actually just operating on a passive divider. They are just a passive div.

put as a passive divider. it's going to you're going to see because it's a um, it's a chopper amp. There's a chopper amp in there and it's used in the feedback loop of the chopper. but it is a passive divider.

Um, so let's go to 99 999 996 Ah, too good. Not even going to wait for it to settle. It's already too good. It's too good for what you paid.

Not not that I'm bragging. it's ah, this is silly. It's linear too I Like it. Linearity looks excellent.

And if we check out the uh, 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 uh. zero offset error I Like it. Yeah, so we're really down in the uh down in the noise here because uh, these digits actually, uh, match the digits here on the Uh on the voltmeter. So um, it's you know it is fluctuating down in the noise there.

We could dick around and try and actually get its performance a bit better in various uh ways. but uh, this um is actually a shielded uh test cable we're actually uh, using here. Um, but yeah, as you can see, it's you know you've got to not play around with it. Um, and so we are down in the noise pretty much.

but if we bump that up by .1 we don't see it yet. There we go. Okay, now we're starting to jump up. So 2.3 and it's starting to not quite follow that.

But you know we are right down in the uh noise region where you know, uh, where, uh, thermal, uh, noise and stuff like that comes into play. based on your connections and your typ of metals and stuff like that it it can actually become an issue. Um, so it's a bit of an art of measuring that low stuff. but if we go up in the to the 100 molt range then we can play around with that.

.