Hi Welcome to the Eev blog an Electronics Engineering Video blog of interest to anyone involved in electronics design. I'm your host Dave Jones Hi I've had a bit of a problem with my micro current project and I'm a bit pissed off. Quite frankly. what's happened is I uh kitted up for a new batch of these and I uh tested them and shipped away a bunch of them and then I found.

oops there's a slight problem with them. Now the thing with the microcurrent project is that it's a little Precision current adapter and it basically relies upon the uh, precise nature of the .1% resistors in here to give the accuracy that this unit actually has. And you know those resistors are usually very, very reliable. You either get the right part and it's spot on within spec .1% or it's not or it's you know, the wrong part or something else.

So I don't fully test every parameter of this thing before I send it out I just do a couple of SP I just do a spot check on each current range to make sure it's working and I check the first build to make sure it's okay and then you know I just do some spot checks and ship them out well after I shipped out a batch of them uh, the uh second lot I started to test it and it looked a bit looked a bit funny. The readings on the um on the microamp Range were slight off and I thought I'd investigate. Now here's what typically happens: Okay, I've got the microcurrent adapter here I've got the output hooked up to a Uh to a Precision meter. Here the Metr hit extra I switch it on and I've got my uh Keithly current Source here which um I generate a constant current into the input.

Now this is on the nanoamp Range and I've set it to 99.9 nanoamps on my current Source here and as you can see, it's 99 7. It's pretty darn good because it's not. This isn't Uh, this isn't like 0.1% or lower accurate, but that shows me that I've got the right value resistor installed and everything's hunky doy with that and I switch the current range up to the milliamp range. Here we go.

so it's 99.9 milliamps and check it out. There we go: 9989 Pretty darn close to 99.9 Not a problem, but let's Swit swi it down to the microamp range and switch it over. and here we go. It's 101.2 M volt.

So it's saying it's 101.2 milliamps Now what we need to do. Okay, that seems to be a bit out way out, um of spec, but what we need to do is whack a current meter in series and actually see what current's going into here to confirm that. Okay, so what we've got now is we've got a second Metr hit meter. The Metr hit energy.

Nice precise meter I Love it. Haven't done a review of that yet, but uh, it's me now measuring the input current to my microamp range and the um other metr hit. um, extra meter is still measuring the uh output voltage or what the microcurrent is measuring. Okay, so I'm actually feeding in 99.76% So you might think that that's you know it's pretty darn close.

but if you punch that into your calculator, that's like a 1.4 1.5% error. it's hopeless. I We expect around 0.12% or better accuracy. So it's you know, it's almost an order of magnitude.

uh, worse. Really. So something is wrong with the microamp range on this meter. Now we know if you look at the circuit, we know that it's not the two uh gain set setting resistors or the three gain setting resistors.

because uh, the other rangers which share those resistors, the Nano amp and the Milli ranges are spot on. So there's nothing wrong with those resistors. So it must be the Uh 10 Ohm Uh current shunt resistor used on the microamp Range something sus with it and I found that it was like this right across the board with most of the units I measure they were above. Um, they were over spec.

This one's saying it's 3.28 microamp with 99.7 microamp input. It's crazy. So let's investigate a a really bad case board. This one's about 3.5% out now based on these Uh two readings here.

if you punch in the numbers, we instead of that 10 Ohm resistor. Uh, being 10 Ohms, we calculate we should get about 10.35 ohms. So let's actually measure this resistor and see if we get around about 10.35 where we're just ignoring the other resistors in the circuit. We're assuming this spot on.

so there's going to going to be a bit of play there, but let's actually measure it and see what we get. Now because we're talking about a .1% resistor and it's a low value of 10 Ohms. .1% Uh represents 10 milliohms or .01 Ohms. So uh, the Tesla When you short out the test leads on your multimeter like this, we'll use the Metr hit energy for this.

Okay, then you're talking. You got to zero out the residual resistance in these test leads when you're doing something like this. So it's. 29 Ohms 29 Ohms.

Let's zero that out. Okay, it's pretty close to zero and we're after 1035 or thereabouts. let's have a look. 10.33 34 There you go.

Bingo The resistors about 3.3% out of Tolerance Unbelievable. Now of course we can't just leave it at that. Let's try another meter, just in case you know something might be happening with the meter. I've got the fluke 87 here.

Let's zero out the test lead shall We 0.15 There We go and let's probe this resistor here. You've got to really have sharp when you're talking. um, this lower resistance you got to have really sharp probes. So the ones that come with, um, say the uh, the the Metr hit energy aren't very sharp at all.

They're You know, really thick ones. So you've got to actually get really sharp probes to get in there to, um, get through the oxide coating on the solder joint and all that sort of stuff. so you really need to make a good connection. But there it is.

10.35 34 Ohms. It's exactly what we got with the other one confirmed. Hey, but I'm still not convinced. So let's get the Agilant meter and give that a go.

Shall we null that out? and there it is. 10.34 Ohms gotcha now. Thankfully I Just so happened to have some resistors left over from the build. and there's the part number.

Okay, you can clearly see it's the correct part number. I can go through that on Digi key. but um, I've got some leftto over resistors because who knows, they may have been damaged during soldering or something like that. They could have drifted.

who knows. but these are brand speaking new from the packet. I've got one here on the bench and let's check it out. It's really hard to probe a little 0805 resistor, so bear with me when it's not on the board.

it uh, can be a little difficult. 10.23 Ohms, there you go, you saw it. and I've tried a few others and similar thing so nothing to do with the solder in at all the ones out of the packet that are supposed to be 0.1% It's hard to see there you you think it says 1% but there's actually a decimal point in front of that .1% Are you know, a couple of percent? It's crazy. Just as a sanity check.

I've got one of my original units from my very first batch here and let's measure that and see what we get. Let's measure the 10 ohm resistor and what do we get there? It is 9.9 10. exactly what you expect within .1% Not a problem. So there's something seriously wrong with this new batch of resistors.

And here's my original unit. hooked up measuring as you can see input current 99.74 microamps and the measurement out of the micro current is 99.76% .1% spec. So these new units are rooted okay I Know what you're thinking I've only done what's called a two terminal measurement with a regular multimeter. What about a four terminal measurement? I'm glad you asked.

Now, if you don't know what a four terminal measurement is, let me explain or a four wire measurement, it's a special resistance measurement. You got to have a special meter that supports four wire. I Happen to have my Hiller Packard 3478a bench multimeter here that support regular two wire and what's called four wire measurement and this is how it works in. if you have a regular, you try to measure a resistance like this.

what it does is it has a current generator just like it has in a normal multimeter but instead of reading back the value right in here, What it does is it actually has a second set of wires. uh, second set of inputs that is a Voltmeter. So you're measuring the current and the voltage. but you're measuring the voltage right at the point of the resistor.

So you're not measuring the extra voltage drop along the resistance of the wires of of the wire that's actually carrying the current. So what you do is you probe it directly on the resistor itself and you get a very, very accurate measurement, effectively nulling or zeroing out any effect due to the test leads whatsoever. Now this shows it. two separate Uh leads like this, but often, well, usually they.

they're actually Um. They're actually the same set of test leads, but they'll have instead of just one wire, they'll have two wires running in them and they'll be connected right at the tip of the actual Uh probe. and that goes into the multimeter like that. But you've got to have those special four- wire resistance Uh probes now.

I Don't have one of those. So I Decided to make my own and it's real easy. It's it's pretty trick trivial now because it's pretty hard to hold four probes at the same time. You've pretty much got to solder it.

So what I got is Uh is one of my microcurrent boards and I solded just the 10 Ohm resistor on there. As you can see, there it is there. the the 10 Ohm resistor and I've got two wires coming in which is comes from the Uh from the current generator basically. and then what's called the sense wires are connected directly onto the resistor like that, then solded directly on like they should be.

because the microcurrent layout. the board layout already has the wires actually going directly onto the resistor, but you normally sold of these sensor resistors directly onto the contacts of the resistor and let's see what we get now. Hopefully you can see that it's not that great in the light, but what I've got it is on is I've got it on two wire measurement and as you can see, it's 10.22 7 Ohms basically. but let's switch it to four wire measurement because I've got the Uh two current Uh ones here what's called the input and then you've got the Uh Sense wires here.

So let's switch to four wire measurement and that effectively takes out any effect due to the Uh test leads. and as you can see, it's 10.08 Ohms. So it's actually 0.8% out. That's no good at all.

It's supposed to be. 1% or better. So we're expecting 10.01 or better on this display and we're not getting it so clearly. Um, it shows that there's something wrong with these resistors.

This was just a random one picked out of the brand new batch. Now, if you actually want to do your own full terminal resistance measurement at home and you don't have a meter which supports four terminal resistance measurement, that's fine. You can actually do it with two multimeters, one to measure voltage and one to measure current. and you use Ohms law.

Simple. And here's an example of using two multimeters. I've got my full four terminal board again. I've got my constant current generator.

If you don't have a constant current generator, that's okay. You can just use the voltage source generally with a series protection resistor. Um, because you don't want to blow a low value like this, you don't want to put 10 volts across 10. OHS Okay, so um, you need a series dropper.

So we're measuring the current we're putting through 9963 milliamps. There it is. and we're getting. We're measuring across directly across the resistor, directly probing it.

100 .43 M volts. And if you whack that into the calculator, you get 10.08 Uh Ohms. Which is exactly what we got with our H packo up here. We got 10.81 actually, so near enough.

Okay, and that's how you can do a simple four terminal resistance measurement with two multimeters. And one of the main benefits of this is that you can measure actually really very low values of resistance. You know, down to milli Ohms or there whereabouts which your regular multimeter just can't measure on its Ohms range no matter how good it is. Um, you know you can have a real super Precision bench meter like this If you try and measure the resistance, your probes are just going to swamp anything under.

Generally about you know an OHM or 10 Ohms or Under. you really should be using a four termo resistance measurement to get accurate results. Okay, let's use our little Xtec uh microscope here to take a look at the one from that we actually got from the batch. Now it looks to me it looks pretty close to my original ones, but we'll have a look at that in a minute and there's the identifying mark on it.

There you go: 10 r0 which indicates that it's a Precision resistor so that looks pretty good. Now let's uh, do a comparison with the one on my original board. Now as you can see there it is, it's a bit harder to see because it's mounted on the board I can't quite uh zoom in as as well as I could on the other one, but it looks like it's an identical resistor. So I think it's actually the correct uh type, but why? it's um, several per out? I've got no idea.

So there you have it, What's going on with these resistors from Digi Key? Now these are actually from a company called Burns Uh, Who manufactur some pretty good resistors. They're world-renown. They're great manufactur great resistors now. So either dig are at fault here, they've shipped me the wrong part or burns are at fault because they've uh, sent them to digi key incorrect I don't know who's actually at fault here, but if you look at the part number, the part number is here it is.

Let's look at the Digi Key uh website. it's the B Ct805 B-10 r0f and the Digi Key website says that's 0.1% uh 0805. And if you look at the data sheet here from Burns you can see that the uh, the B in the after the Um 0805 there the B stands for 0.1% and the burns. This particular CRT series of resists is available um, up to anywhere from 1% tolerance down to 0.01% tolerance.

So, but clearly I've had some that are greater than 1% out. so it's it's almost as if they aren't actually this CRT series. They? you know, they've actually screwed it up Beyond just giving me the wrong particular uh type. they uh, the wrong, the wrong tolerance they've actually gone for I think it's a totally different type of resistor Anyway, I'm not happy.

not happy with Digi key or Burns whoever damn well fault it is I don't care I've been screwed over and yeah, it's my own fault I should have tested them a bit more thoroughly before I send a couple of units out. but jeez, you trust these things to be. You know when you buy 0.1% you expect .1% Now if they if I got the wrong part, odds on I would have got the wrong value or something like that and it would have been obvious. but in this case, no I got the wrong tolerance and that can be a real pain in the ass when you're rely on that tolerance in your design because sometimes unless you do exhaustive testing, you may not notice it.

It's a real pain in the ass. Not happy at all. Digy: KY What's going on.