Hi just a quick little web video even though it's not really a blab. I'm not ranting on about something I just thought I'd show you something that I've got here I've you've seen this before? This is my week on 10 K Lab Resistance Transfer standard very expensive, custom handmade resistance standard and all Lincoln in down below if you haven't seen this, so thank you very much week on. They actually sent this to me to use as a lab standard here and there's some tear down photos as well. Anyway, I've just been having it run overnight.

you'll notice I've got four terminal resistance measurement here so it's got a A Sense terminal on it. and of course as all good lab resistance standards will to compensate for the test leads and so I just wanted to show you what I actually captured overnight. Here you can see the data actually go. this is a graph of R I've got 675 thousand samples.

I'm actually doing a sampling like every couple of every couple of seconds or something. there's 10 power line cycles in this thing and you'll notice that I've actually went right up there at the start, little dip and then up and then that's I Think that's when I went home last night and then I came in this morning and it started going up and then when it went down back down like that. that's when I came in this morning. So I didn't turn on the aircon right here at all.

So but I did turn on the lights of course so you know maybe there's some heat generated from the lights or something like there's something changed I Definitely came in this morning and then it started hitting, drifting back down again. But yeah, I haven't gone into all the details. This is not a, you know, it's like a really controlled scientific test. This thought I'd leave it on overnight and see what I got.

but as you can see, it's bugger all. it's 10. If you can't actually see that, take a look at the scale there. it's actually a 10 point zero, zero, zero, zero.

So looking at four decimal places there and it basically doesn't change, you know, like I can't actually see the act your values on the scale here. I'd have to download the data to actually extract the real data from it. and I'm using my Keithley seven half digit multimeter here. it's the best.

Pretty much the best meter I have in the lab here and you'll know that this is like the original work cow certificate with it. and these are the bloody Stupid tripod. I Got to get myself a new tripod. doesn't stay still.

Cameras too heavy. This is the original cow certificate. It's nominal value is nine point nine, nine nine nine. So that four nines there, two, two, eight and that's what's written on the back of it.

It's it's you know, nominal stated value and there's alpha and beta values if you know all about that sort of jazz. But yeah, they actually measured this at five point five degree temperature gradients. and as you can see, there's just four nines all the way through there, so it doesn't change much. So what we're seeing here up on the screen is clearly not a a temperature drift of the resistance standard itself, because it's better than that.

It does not drift by really, or it shouldn't drift by that amount of value. But you know, once again, when you're trying to measure this sort of this sort of precision, everything comes into play. You've got your temperature, your leads in your contacts, and things like that. You know, low noise to lure Iam copper contacts.

you know we can actually see you can see the sample noise in there. I could actually turn on the filter in this thing and actually I get that, you know and actually smooth out the signal. but I just put in the tent I didn't put a filler and just put it in the ten power line cycle. so we do get a fair amount of noise there now.

I can actually go home I won't and that won't kill the data being sampled here. and it's It's saying it's not Ten Point Double Up No. hang on. No, that was an old value.

Was it or is it jumping around? Anyway, here we go. It's Nine Point. Yeah, Yeah, Okay, there we go. So I don't know what that anomaly was.

Nine Point Nine Nine nine, nine, seven, zero And of course the the value that's supposed to be is nine Point Nine Nine Nine Nine, Two Two eight. So you know we've got discrepancies here. What that you know like is it I haven't actually checked. if that's within spec of the meter or not, it most likely is would be my guess.

Anyway, we can go back to the graph here and you'll notice that this this little dip here was caused just before I started shooting this and I I didn't touch it I didn't know I didn't physically touch it but I was shuffling stuff around on the bottom bench here and maybe some vibration actually went through shells here. they're held on with it like they're supported by these metal things. And if you touch the table, of course, if you rock the table a bit, then that can you know send vibrations up through the top table up here. So it's not not the best environment to test in.

But anyway, it's interesting I just thought I'd show you that There you go. So I'm not sure why I left last night. I did actually well this it would have Jeez no. it actually got a bit cooler last night.

but I'm in the middle of a big building here so I didn't actually log the temperature with this thing. which is you know, which is a mistake if you're trying to do this properly of course. But I'm not I just ran it overnight to just you know to see what value it would settle down to. So looks like that's what it's That's what it's settle down to it.

So there is a bit of a discrepancy from the measured well from the calibration value of this thing. And it was done by the KCB mob in Germany who's you know, some big reputable arm, you know? Cal House or something Casey BDK calibration all that sort of jazz. and I think there's the actual calibration certificate from Mark Kcb. and yep, it's all in German There we go.

Oh no, no, there we go. toughing. English This calibration seemed documents the traceability to national standards. blah-blah-blah-blah-blah And anyway, this was one of the first units off that they actually made.

It's a new line that they're making so this is one of the first. Well, they're all handmade, but this was one of the very first units that they that they actually built well. I'm pretty sure my meter is not good enough. You know you really need a you know, an eight and a half digit class meter really.

And if I left my aircon here, it my aircon can actually control the room temperature to within plus minus half a degree C or thereabout. So I've actually got some data on that somewhere so you know if I turn the aircon on, it does actually regulate fairly well, but as you can see, I didn't have it turned on I don't think I had the aircon on last night I And this is not warm up. The meter was already on for an hour or two before I started, so not actually sure. But anyway I went home so the aircon was already off.

so I'm not. You know entirely sure what happened that? yeah, it's something dropped overnight, so some characteristic of either the resistance standard or or and or the meter itself drifting. But yeah, we're really quite. You know, down into the precision, not realm of things.

And if you don't do it properly, yeah, and sort of eliminate all sorts of variables and everything else. but you can tell I really haven't used this much. They've actually got a reading table in there and you can actually that gives you the time and date stamp and the absolute values in there. so can we just scroll that? So I started out at a ten-point double-o-double 0:05 Okay, so at the minimum point there, roughly it's nine point, nine, nine, nine, nine, six, seven, or thereabout.

So yeah, yes, still not down around the point two, two, eight that we're after. So we got six eight instead of two. Wait. So now if you take a look here at the calibration data this resistance standard actually.

well, this particular one there, they might all be different because they're all handmade and hand selected resistors. Everything else. This one actually has a slight negative of temperature co-efficient What that means is that as the temperature increases here, say from 20 degrees to 25 degrees, the resistance actually drops. You can see at point three and like 99 99, 99, 3, 6, 2, 3, 6, 4, 3, 3, 1, 2, 9 so it actually drops so 4 5 degrees C Range there.

that's a change of point or negative Point Double-o-double O to 3. Ohms so or okay, sorry So yeah, it's You know it's not a huge change and that's not nearly as big as what we're seeing on the screen here, because if you look at the highest value we've got there at at that point there roughly, you know we're looking at point 10 Point Double-o-double O 3 4 K And if we go to the you know the low point point 9999 6, 9 And if we go back to our graph here, that's a total difference of point Double-o-double o66. So just the temperature coefficient of the reference standard doesn't account for this, so there must be some drift in the meter as well. and of course you'd expect that and I Think this is actually we've come back later.

Eventually, come back from having some lunch and you'll notice that that's the dip we're at before in fact discovered that you can pinch and zoom this thing which is rather a rather neat so but you can see yeah that's the that's the, that's where we were before and since then I Think this is the point. Came back from lunch and then it started ramping up again. I Turn the lights off when we went to what lunch and then we came back in once again, have not turned on the aircon. Then it's going to continue to hover around here based on the ambient temperature.

So these are not just our temperature stability, but you know the physics to do with the reference and you know everything else to do with that sort of stuff. Because you know you're talking about material science and you know how the actual reference itself actually ages and works. you know, and drifts and everything else. so you know.

Even though you might read the data sheet for this meter for example and it might say you know allow I haven't read it, but you know it might say might typically say okay, allow it to warm up for an hour before it meets specs. Well, that doesn't necessarily mean that there's not better performance beyond that one hour figure so you know it might need to stay on for a day. And that's what we might be seeing here. We might be seeing some sort of longer-term characteristic.

It'll still be within spec after an hour or whatever, but you know we're we're sort of and possibly trying to measure. Like beyond that, we're trying to see data. actually beyond the spec. Fascinating stuff, but yet to do this properly.

We need really controlled conditions and you can separate things out. We can try and determine which ones actually got the most effect having the most effect here. is it the meter itself? I Think that's most likely is that the resistance standard it's going to be. Ultimately, it's going to be a combination of both.

but how each one ways? you can actually, you know, put one in a thermal chamber and one not or you know, and you can cycle them to see which home has the greatest effect and all that sort of stuff. But yeah, I could do that. I've got a thermal chamber here I could put the reference out standard in there and you know, use my aircon as a secondary thermal chamber effectively because I can control that too with him and half of the grease and you can have fun like that. And you know, spend weeks and weeks and sometimes months and months gathering data like this Now I've come in the next day and look at this.

We're pretty much what we expected essentially from the drift of the reference. Our in with eight. Well, what we expected from a negative temperature coefficient. Here's where I left off the filming yesterday and then we stay it like a stayed in the office a bit longer.

Got really warm here I didn't deliberately didn't turn the aircon on so it actually got up to 26 degrees in there. Now it's about 24 and a half. So I got into 26 and then this is the roughly the point that I went home obviously and it's started to cool down a little bit because I switch the lights off obviously. I've got, you know, a fair few lights in here pumping out.

you know, half a kilowatt or something, right? So you know. And in computers, you know everything else that I've got going during the day and so I switch those off. and obviously it's that cooled down a bit, only by like a degree and a half. But you can see and it's actually drifted back up in value like that.

And I bet you now that I've got the lights on and I'm back in for the day. we'll probably start see it drifting back down, but we didn't reach anywhere near that point that we got before. Although I'm actually I might turn the aircon on now. What do you think I turned the aircon on I normally have it to set to 22 or something it is.

so I'll turn it on just now and we'll see if it jumps back up, but maybe it might get back up to the peak that we had before. That would be an interesting experiment. Otherwise, I think we're just going to see it go back down and that's not hugely interesting. So I'll try that aircon on and Bam look at that.

look at the huge big jump up there when I immediately turn the air con off and it's gone to actually a higher value than initially over here and you can see some oscillation in there. I'll zoom in a second just like we saw over here. and I reckon that's got to be the cycling of the air conditioner. and I believe it's like five minutes on five minutes off or something nominally to keep this room to temperature.

although that varies depending on how much heat is being generated in here and you know all that sort of jazz. But if we can, we zoom into that. There we go, we can see some oscillation. Check it out there we go.

You can see the oscillation in there, up down, up down, like that be can see it progressively ramping up slowly so it hasn't actually even though it's now 22 degrees in here by the way. and before I started it back here, it was at 26 degrees. So I just turned the aircon on and you can actually see some look some sort of I guess you call it undershoot there, but that's really it. It's absolutely amazing how it ramps up that quickly.

but then there's a slow rise like that, it's almost like there's a a thermal dampening and this is what you'd expect. Of course, in if it is the resistor which I don't think is the only thing going on here. Okay, there's you know, the resistor. Here's a photo inside the thing and you can actually see.

You know we've got a dyke. I'll start out a box here. shouldn't actually tap that, but I cast out a box here, plus an internal sponge which actually stops the stops that you know vibration. It dampens any vibration getting into the thing, but there's an air space in there and it takes time for outside temperature differences to make their way through to the resistor.

So that's you know, partially what we could actually be seeing here, or some percentage aspect of it. There's a slow ramping up. It actually takes time for that resistor inside to reach thermal equilibrium to the temperature outside. But let's actually take a quick look at the figures and see if this large change that we've got here this large change that we saw is actually just due to the resistor.

Now if you can see, we've actually jumped up to well over 10k again. TenPoint Wo3 six and if we actually go into their reading table once again, we'll be able to take a look here. actually. I think that value there was 26 like that was, that was the absolute highest temperature that we got in the lab.

and I think it's sorry it just cooled down a little bit so I won't take that one before there. I think it's a four. it's currently 22 degrees C here in the lab now and according to the same the same thermometer so not absolute. I'm just using one of these little you know crappy things but hey, you know so we don't care about the absolute, we care about the change in temperature.

so I'm going to take that point there as 26 and then go back to our current value up there and the difference, see if it matches the datasheet for the resistor. So we've got the absolute values for this resistor from the calibration lab at 0.5 degrees C temperature intervals here. and of course, they would have let it, you know, settle for like an hour or a couple of hours at each chart, temperature, etc. They'd be doing this in a proper thermal chamber using all the best available test leads and the best available test conditions everything else.

So it can be pretty confident in these unless the resistor is physically changed or physically damaged since you know that since they actually did this. But we can be pretty confident this is a hand selected, hand age, hand characterized resistor. you know, I'm pretty confident in it. And so these are the different values at 22 and 26 with a delta of 0.001 96k.

Let's round that to point Double-o-double O2 and this is what we actually measured on our meter here for the same 4 degrees C difference we're getting. You know, roughly point double-o-double Oh 800. So it is. This is actually 4 times greater change we're reading on the meter, then the absolute value in the resistor.

So assuming that the resistor is good, which is a pretty safe bet. The rest of that 4 times change must be coming from either the meter or possibly you know it could be the the interconnection. Of course we could have thermal EMF issues here. maybe with the test leads these are you know, a different metal matching, all that sort of stuff, but you know it's it's likely most likely to be the temperature characteristic of the Keithley meter.

Now if we have a look at the data sheet for this side Keithley DMM 75 10 then we can see that the the spec actually has a temperature coefficient to do with it and it also depends on whether or not you've done auto calibration on this thing which I have nots on taken the table that has not had auto calibration done and the figures are higher and it's basically on the 10k range 2.5 ppm per degree C But if you have a look at the Asterix note 4 that it actually says these are apply outside of plus -5 degree or yes, plus -5 degrees see about the temperature, the temperature it was calibrated at. so we're not outside that. So according to the datasheet in theory, we shouldn't actually be applying in this. but because we're actually doing more metrology, metrology type stuff and actually measuring within the performance.

You know the spec performance of the unit measuring, You know some inherent temperature changes? Then we've got to take this thing, this sort of stuff into account. So let's just take the nominal figure that they've given here of 2.5 ppm per degree C that's at Point Double-o-double o2 Five K per degree C Just to put it in the same sort of format as what we've got up here. And of course that's and we're talking. we're looking at a four degrees C temperature range here.

So that's actually Point Triple O1. So you can see our difference here between our what we expect from our resistor here and what we actually got on the meter was that we were getting a point Double-o-double O6 change from what we were expecting. So we're going from basically point double O double O to 2008. So that's point Double-o-double O6.

You know discrepancy that we're trying to account for here. And you know, just according to that datasheet spec, you know it's going to be actually be higher than that. So bingo, Have we found the culprit? It's the Keithley meter actually drifting? Assuming of course that our resistor is that you know, hasn't you know it hasn't drifted, or it you know isn't faulty in some way, and that it's as per the data sheet, which is a pretty safe bet. So we can easily account for the discrepancy that we're seeing here and on this meter by saying it's essentially the temperature coefficient of the reference inside this thing.

and that's what's causing all of these. That's what's causing this change. And of course, as I said, this actually didn't start out I didn't start this video out to actually be a tutorial or whatever in how now in you know in how to do this, in how to you know a measure a precision resistor like this. Otherwise I would have got out of my thermal chamber and I would have I would have done it properly in quote marks.

but oh, I can just expand it that way can't I Yeah, silly me. hey I'm not a pinch and zoom kind of guy. I guess all the young whippersnappers are with the modern shoe phones I don't know. Anyway, yeah so it looks like the discrepancy were seeing here is actually due to essentially as I say combination are both, but the dominant thing seems to be the Keithley meter.

but hey, as I said, this is not a proper way of, you know, not a controlled experiment. Didn't start out as one but kind of turned into one. I was a bit interested in in the results here. it was just a thing I did on a whim.

but I think we need to I won't do it for this video. it'll have to be the next one definitely. But we can put the resistor in the thermal chamber so it doesn't change and just have the leads coming out and then we can, you know, ramp the temperature up in the room. or I did effectively have to do that because it's not like we can put this Keithley meter inside my temperature chamber because it generates too much heat.

You really need a big beefy thermal chamber to actually for a product that puts out 510 watts or something of radiant heat, you need a big thermal chamber to heat up the ambient. My little toy thermal chamber can't do that, but it could certainly heat up the resistor box if you leave it in there out long enough to soak. So yeah, might have to do that for a second video just to see the difference to see if we can actually get a smaller change in this. But yeah, it's rather fascinating.

So there you go. This turn. This was just going to be a simple second channel video. then I thought I might do it as a blurb but then I thought no, it's kind of, you know, turned into an interesting and rather lengthy and waffle type look at just measuring a simple resistor standard like this.

With that pretty much the best meter I've got here in the lab. so there you go I Hope you enjoy that. If you did, please give it a big thumbs up. Catch you next time you.