Hi in a previous video linkedin down below and up here and at the end. If you haven't seen it, uh, this is part of the series of designing a better microcurrent And the last video, we took a look at the Opa-189 and how it looked like. in most respects, it was better than the Max 42 38, 42 39 that we use in the micro Current, so I ordered some. Here you go: Opa 189 Id Bvr This is the Sot 23 5 package.

It is available in different packages, but I got the Sop 23 5 because it is actually directly pin out compatible with my existing micro Current. So what we're going to do in this video is we're going to actually compare the original design microcurrent to one that uses the Uh drop-in replacement Opa-189 and measure the Uh, noise and other performance characteristics of it. Let's go, but it's actually not completely pin out uh compatible because here's the existing Max 42 39. It's actually a six pin, so you can see It's got three pins on the top here, but the Opa 189 only has two pins on the top.

It's a five pin slot 23 package. The only thing that's missing is the enable disable pin, which we don't actually use on this. We just tie it high so we can, simply, um, simply remove the existing chips and solder in the new one. Doesn't matter Beauty.

So to make this comparison as fair as possible, what I've got here is two identical boards from Uh, actually the same batch, just a serial number apart. so they use all the same parts that you know come from the same reels and everything. So they're basically as identical as you can get. but one uses the 189, one uses the original max 4239.

Now the problem is uh, though, I can't use the coin cell battery to do the comparison test because the Opa-189 Uh has 4.5 volts minimum Uh supply voltage, whereas the Uh Max 4239 can go down much lower. like 2.3 volts or something. if I recall correctly, is the dropout voltage something like that? anyway? Um, but it only goes to 5.5 volts maximum. So I'm going to run this from a four and a half volt battery.

So just enough. I've got uh, three triple A's here, which will give us if we use brand new cells. it'll give us just over uh, four and a half volts. So yeah, that'll be good enough for Australia, right? So we're going to start out by measuring the noise.

Now, as I said, a previous video linked in down below where I've actually, uh, really, quite a while ago. This video is where I use my Uh, Hp 35060a Dynamic Signal Analyzer to actually measure uh, noise floor of Op amps and that's a really great video. Remember, I think it's from about like 20 15, 20 minutes onwards. In the video is where I show you extensively how to set up a dynamic civilian analyzer to measure noise to match the data sheet noise density actually, which is nano volts per uh, Root Hertz.

Uh, now to do this, of course. Um, you can't just have the boards flapping around in the breeze on your bench. You've got to have a nice die cast shielded box like this. So I've got that with a B and C output going into our input.

So we'll put the batteries and everything else inside the diecast box just so we eliminate any external noise. any uh, you know, external 50 Hertz? All right. First thing we're going to do is just measure the Uh power supply current because the Opa 189 is supposed to take more current. Yes, So let's get the old one.

the Max 4239. I've actually removed the Uh dropper resistor for the lead here. Otherwise, the little draw, like you know, seven milliamps or you know, a huge amount of current like that. So let's give it a bill.

This is the Max 4239 1.31 milliamps and the 189. Oh, two point, six, eight. So you know what? what? One point, Three milliamps more overall for uh, the two Op amps and you know there's some residual uh, you know, other current which will remain the uh, same. But yeah, you know that's it's neither here nor there really, unless you're after some ultra low power design.

So nice. Now I know that we only have a sample size of one here, so take that with a grain of salt. But let's just, uh, see, it'll at least tell us if anything's grossly wrong with the offset voltage. So I've got the Max 4239.

I've got it set to, uh, the 10 milli ohm shunt so it's effectively, uh, shorted. and we're getting about 61 microvolts. So they're about offset And that's you know. Fairly typical for a uh, Max 4239 mic based micro current and the Opa 189.

Well, there you go. Um, 73 micro volts? Meh. Neither here nor there. good enough for Australia 67.

Look at that. Oh, it's going down. It's going down, dropping, that's plummeting There you go. just turn filtering off, there, filtering back on.

So yeah, um, there you go. Like 53. You know, microvolts? like in the order of 50 micro volts. So yeah, um, that's fine.

And we didn't really expect a huge amount of difference because as we saw, like, on average, uh, based on the bin in, it's basically equivalent to the max 4239. But it does have slightly wider offset margins. but as I said, sample size of ones, but still. hey, and that's really shouldn't change with our range.

So that's our it's our next range and our nano amp range. Whoa. And then our range has jumped up. But that's because the input's not shorted.

Let me short the input. There we go. it's back down to 50 micro volts Nice. And just to check the Uh ranges, make sure the accuracy is there, it's still not.

you know, mucking up. Uh, we're generating one micro amp here with my Keithley 2400 source measure unit and we're getting our point 999 eight? Yeah, that's good enough. Um, like, I haven't like properly calibrated any of this so like. And the Max 4239.

Uh, 0.992 So there you go. That's more out than the Uh 189 is. but hey, it's to do with all the other stuff involved. all the other range resistors, everything else, and Max 4239 generating one milliamp on the milliamp range down.

or the micro ampere range down there. Um, yeah. 0.9999 And the Opa 189 1.0011 Um, no worries. That's like, you know, uh, 0.01 percent Beautiful.

Generating 1 amp here. Opa 189. We're getting 1.0023 I know that's like 0.23 but I've had an issue with, uh, the range resistor. Um, on here.

and that's what I'm yielding at the moment. So anyway, um, we should expect. uh, very similar with the with this one as well and the 4239 board 1.0014 Look, don't worry about these values. it's to do with uh, the precision of the Uh 10 milliohm, uh, current shunt down here.

As I said, I've been having like batch issues with that. So yeah, it's all fine. I just wanted to show that it's all working and I assume it's working, but just want to double check, uh, to make sure that the split rail is uh, working just fine. 2.3 and minus 2.3 No worries.

Okay, so let's measure the noise first. I'm going to just measure the baseline noise of the system here as I've done in a previous video. but I'm going to include the box and everything else so we'll whack our lid on there and we need to set it up. Now, this is not trivial.

Now, I've covered this in more depth in the previous video, but we'll just go over the basics here. we're going to. I go over the full span here. Zero Hertz to 100 kilohertz.

Here it's 102, but that's to do with the bin in and everything else Now of course, what we want here is nano volts per root Hertz or power spectral Density Which is, uh, what is the value which is included in the data sheet and uh, by default. Um, this is set up for you know, Db, Volts, Rms. That's not what we want, that's not a power spectral density. So the first thing we have to do is go into measurement data over here and you can see that we're just getting the normal spectrum, but we want the Psd or Power spectral density because if we don't choose that and we just use regular spectrum, if we go to scale over here and then we go into our vertical units, you'll notice that there is no power spectral density.

There's no nano volts per nothing per root Hertz which is noise density. So we have to go into measurement type here. Choose power spectral density and Bingo volts rms per roots Hert. But we don't want the Rms part here.

So we go back into the scale. We go into our vertical units and Bingo volts per root Hertz. We now have that or nanovolts per root Hertz, but it's it's the same thing. it's volts per root Hertz.

It just happens to be Nano. so we can do that. And now we've actually set up our things. Okay, the next thing we want to do is that we want to go into the input type here and we want to do Channel One setup.

and we actually want, uh, grounded input. We don't want it floating. I won't go into the difference between floating measurements and we also want our Dc as well. Don't worry about the units there, that's just uh, the engineering units.

So we want ground and we should see, yep that noise floor drop. Nice. Next we want to go into scale over here and we want to auto scale like that. So now we get a resemblance of a noise floor and we're looking at well, it's jumping around.

but we're in, uh, like you know, tens of E to the minus nine which is nano volts per root Hertz. And of course you know you want to clean up all this uh, waveform. So you want to go over here, turn on some average in averaging. On how many averages we got, we've got 10.

let's go to 100.. there we go. Now we can actually start that and it will give us a well. it'll average down to a nice waveform, so that's actually significantly different to what we got with just the Bnc 50 ohm input.

So let me actually show you that. let's restart that. There you go. That's much nicer and that's what we'll get in in the uh, previous video.

You know, around like 30 nano volts uh, per root Hertz floor. but unfortunately. um, you know, just our measurements set up here. This is just.

I don't know what is it? Iu 58 or something. Um, the coax and the box and everything else. Not going to be perfect, but you can see plug this back in. There is a significant difference between the noise floors there, so we've got some peaky stuff happening up here.

I'm not sure why it's tailing up here, but hey, that's what we've got to work with. That's fine, because you've got to remember, we've got times 100 gain on here. So the noise floor of our Op amps is going to be multiplied by a hundred. So you know we aren't going to be down here.

We're going to be. You know, much, significantly higher than this. So it's neither here nor there, and we probably want that on a log axis as well. So we're going to scale and log.

thank you very much. There we go, that's better. Let's do that again. And as I mentioned in the previous video, uh, we're only going to get because of our 400 lines of resolution on this thing.

Uh, we're going to get large steps like that. We're not going to be accurately able to measure over the 400 kilohertz uh, bandwidth. But hey, let's actually run with that So right? So just remember that that's our baseline uh, noise floor for our setup here. Okay, so let's put in the 4239 First, I'll hook up the output.

Okay, so I'll put it on the Uh 10 milli ohm uh shunt range so it's effectively, you know, shorted on the input and let's whack that in there. And yeah, my battery's on. And of course, it's not just going to be the noise of the Op amps because we've got the split rail thing that could introduce noise. And you know, we're just measuring the whole system here.

So I'm not measuring the actual individual Op amp. I'd have to do a, you know, like a proper jig like I did in the uh, previous video. To do that, I'm more interested in in the actual current micro current application. Current micro current application.

Yeah, I said that right. All right. So let's not change anything. Let's start that again.

And so you know, 44 nano volts per root hertz around about there. What's that cursor at? Don't know. Whoa. There we go.

There, We go. It's jumped up. Six micro volts per root hertz. But let's get it.

you know, sort of at one of its like lowest points like that. There you go, you know, three and a half micro volts per root Hertz? Let's call it like that. So you know that that's what we get over here. Now you might be wondering, what is that spike in there? Well, I can tell you what that'll be around about 15 kilohertz would be my guess.

13 13.3 Why did I know that was the Uh that it'd be around about that frequency because that is the chopper frequency Because these are auto Zero amplifiers. just call them chopper. Amster is the difference between auto Zero and chopper, but we won't go into that anyway. It's an auto Zero amplifier that has a chopping frequency, and if you read the Max 4239 data sheet, it tells you that's typically it is like a spread spectrum kind of thing.

It's not one fixed frequency it did. They add some dithering uh to it, but you know it's around about 13 kilohertz and that's exactly what we're seeing there. And if you actually sweep, uh, the microcurrent, you can actually see a little bit of low amplitude. Funny business going on.

Uh, you know, at around about that 13 to 15 Kilohertz mark or whatever that's you know that's a loosey-goosey spec. It's not going to be exactly that, but that's a roundabout. uh, the frequency that we're measuring there. So there you go.

So yeah, I'm going to call that like maybe three and a half micro volts per root Hertz. Okay, let's measure the Opa 189. Okay, the Opa189 got to make sure trap for young players that none of this shorts out. So okay, so we'll whack that in there.

And once again, I've got it on the Uh. 10 milli ohm current shunt. Let's run that again, shall we? So 3.5 micro volts? Ah, it's lower. Oh our ref.

What Dbm Per Hertz? What did I touch? What did I touch? Okay, let's try that again. Don't Here we go. Now we're talking 800 900? Yeah, we're talking. Yeah, it's it's lower noise.

So we went from three and a half micro volts uh, per Root Hertz to around about one mic. Let's call it one micro volts per root Hertz. So yeah, the max 4239 in this implementation of the microcurrent is about three and a half times worse noise than the Opa-189 So yeah, it certainly is quieter. And the switching frequency? The Opa 189 that I don't think they tell you exactly, but it's up in the like somebody said, it's something like the hundreds of Kilohertz range they measured onto something.

Unfortunately, we can only go to 100 kilohertz with this. Um, and you can see a little something in there. But I don't think that's it because somebody else said they've that. they've measured it in a project and it was in the couple it was In the hundreds of like 250 kilohertz or something.

So 25 kilohertz? that doesn't sound right. So there you go. Um, that is a uh, our scale has actually changed a little bit here, but our input, uh, range I believe is our range the same. I have to double check that.

Damn it. There we go. I just set it back to exactly the same range: 10 micro volts are per root hertz at the top there, and uh, you can see, yep, that one micro volt per root hertz? Let's call it so. one micro volt per root hertz versus 3.5 micro volts per root hertz.

Winner! Win! A chicken dinner for the Opa189? Nice. Even right at the top there, it's only one and a half so it's still twice as good. Okay, let's just measure the noise on the scope here. I've got a 20 megahertz band with limited times.

one input with the uh coax coming out of the uh shielded box and uh, peak to peak. We're not that concerned about we're two uh, millivolts per division because the sequence has a 500 micro volt uh, per division front end. But let's just leave it on uh, that scale and you know, 7.4 millivolts peak to peak. But we're more concerned about the Rms noise here.

940 odd micro volts. So that's for the Opa 189. If you want to see the noise floor when the microcurrents are disconnected, there you go. So let's just reset those stats.

and that's the noise floor with the Uh microcurrent completely disconnected inside the shielded box. so we're only down around 90 micro volts rms there. Reset the statistics: Max: 4239. Uh, see, it is.

uh, significantly more. Uh, 12 and a half millivolts peak to peak and 1.6 millivolt some Rms noise. So there you go. It's oh no.

it's like 50 worse or something like that than the Opa-189 So significant. Uh. Noise improvements Just using the Opa-189 in the standard microcurrent circuit configuration. Nice.

Okay, I'm just curious to see, um, how much noise was actually generated by the Uh 200k resistors there and the split rail uh generator. Basically, because we are, we have seen noise greater than what you'd expect for the Uh baseline noise of the Opa189. Uh, plus the time or multiplied by the times 100 gain. So we're going to get some.

I mean, the noise. Of course, you're going to get your thermal Johnson noise on your resistors, but they are not actually directly coupled into the input as such, so you're talking about through the power rails. common mode rejection ratio, all that sort of stuff. Anyway, I've left the resistors in there because it's a Nothing burger, but I've removed the Lmv 321 Op Amp because unfortunately, you can't just whacking these Opa189s in here and a higher voltage, you get it to work.

because the Lmv Uh 321 Op Amp that only has a maximum supply rail of like five and a half volts as an A version, which I think goes to six or something. But yeah, basically. um, yeah, you can't just do that. You'll have to get another Op Amp if you want to get, uh, like, operate this thing from say, a nine volt battery or something.

Anyway, Um, yeah, I've done that. and now I'm gonna, uh, power it from split rails here and we'll see how it goes. So it's slightly higher voltage. I've got six volts now, but yeah, that's neither here nor there.

If we go back to the scope, reset our stats. There you go. Uh, peak-to-peak what? we're getting Before I think it has slightly dropped 910 when we getting like 980 or something. Uh, mean Rms before.

so it has dropped somewhat. Okay, I can't remember exactly where we were before on the frequency, but it has dropped a little bit in the uh oh, it does. Its auto calibration there periodically, so we're a bit under where we were before. but uh, look, there's not much in it.

Um, it should. Maybe it'll go under one here, but yeah, it's You know, there's not a huge amount in it. So basically, uh, the Op Amp and those resistors weren't really contributing much to that. Which is as you'd expect because as I said, it goes through the power rail system and then its effect on the Op Amp itself has to do with the power supply rejection ratio, so it's not actually coupled into the input.

As such it because even though it's a float like it's introducing noise into the ground reference, the ground reference is still the reference. So even if the reference is jiggling around due to the noise, it's You know it's not introducing much. So there you go. that's doing some basic tests on the Opa-189 Op-amp and as expected I it was.

you know, relatively significantly lower noise than the Max 4239, an actual worthwhile upgrade for not much real change in power consumption, really, and it is effectively a drop in replacement for the micro current. Although as I said, unfortunately, that Lmv 321 Op Amp, that one will have to be changed to one that supports a higher Uh rail voltage. So, uh, offhand, I don't know, just a regular Lm321 um might do the job. Anyway, I need to investigate that, but if you've got a microcurrent, um, and you want to like have a drop in uh, replacement for lower noise um, and higher bandwidth, this looks like a quite a decent option.

Now, of course I haven't actually checked the bandwidth yet, or rather, uh, dynamic performance aspects of it. Um, this video is long enough so I just want to measure the noise. make sure the Op Amps work, make sure the microcurrent gold like still works, it works at Dc, and its noise is lower. And ah, everyone's happy.

So I think we have a pretty decent new candidate. And I don't expect any show stoppers in the dynamic performance aspect of this. In fact. um, it could be potentially even better because as we saw, the actual chopper frequency is up you know, past 200 kilohertz or something like that.

So in theory you could put a low pass filter on the output as well, even mod existing micro current or in a new design micro current which we're working on. Here you can put in maybe an optional switched output filter so that the performance of it is completely below the Uh chopping frequency of the zero drift amplifiers in here because unfortunately the micro current it was uh, that you know, like 13 to 15 kilohertz or so. It's pretty much you know, bang on in like the measurement range that you're uh, trying to do. So even if you didn't want the increased Uh bandwidth, you could still get the same bandwidth as the existing microcurrent goal, but actually put in a low pass filter on the output and have that chopper frequency outside the operational range.

So right there. another big benefit. So yeah, I'm I'm liking the look of this uh, Opa 189. It's pretty schmick so anyway, hope you enjoyed that.

Found it interesting. If you did, please give it a big thumbs up and let me know down below how you uh liking this new series in quote marks. um which will be just, uh, random videos going forth. It's not some official design project design series, it's just I'm doing.

You know, the occasional Rando video. So this is part three, and I'm sure there'll be more parts, especially like another one. dynamic performance or something. So let me know if you want to see that in the comments down below.

And as always, don't forget all my alternative platforms. I'm on library, I'm on bit, shoot, I'm on dailymotion, I'm on video. I'm on my own website. You can even download the 720p podcast from my own, uh, web server on my Rss feed and all that sort of stuff.

So yep, there's plenty of alternatives outside of Youtube and as always we want to discuss Eev. Blog forum is the place to do it. I, you know the comments do, but forums are better than comments anyway. Catch you next time you.