Part 1 in designing a better uCurrent.
Is there a better chopper/auto-zero amplifier than the MAX4239 used in the uCurrent?
A look at the exciting world of parametric searching for components. Follow along as Dave looks for a new state-of-the-art opamp. Success not guaranteed!
Cascading opamps for increased gain: https://www.youtube.com/watch?v=ZvT9hHG17tQ
Forum: https://www.eevblog.com/forum/blog/eevblog-1318-whats-state-of-the-art-in-current-opamps/
#opamp #design #tutorial
EEVblog Main Web Site: http://www.eevblog.com
The 2nd EEVblog Channel: http://www.youtube.com/EEVblog2
Support the EEVblog through Patreon!
http://www.patreon.com/eevblog
AliExpress Affiliate: http://s.click.aliexpress.com/e/c2LRpe8g
Buy anything through that link and Dave gets a commission at no cost to you.
Donate With Bitcoin & Other Crypto Currencies!
https://www.eevblog.com/crypto-currency/
T-Shirts: http://teespring.com/stores/eevblog

Hi, I thought I'd show you one of the more interesting aspects of electronics design, and I'm sure I've done this many times before, but I'm just going to go through it again. So why not? Hey, turn on the screen recorder and go through the process of choosing a particular chip for a design. In this particular case, a particular Op-amp for a hypothetical, um, improved microcurrent design. Ta-da There it is.

The Micro Current: This is actually a really old design, dates back from I believe it was 2008. I had it published in Silicon Chip in April 2009. It dates back to 2008, so it's more than 12 years old now, and this new gold version I did in a kickstarter in 2014. so that's over six years old.

So anyway, I thought it'd be interesting to have a look at uh, the current state of Op amps, and in particular, low offset chopper Op amps. What is the best Op Amp on the market? because in the Micro current? Uh, you should know. I've done quite a few videos on this over the years. I used the Maxim Max 4239 or the 4238.

They just got different Uh bandwidth variants and I actually cascade two of them in series to get increased bandwidth. and I've done a video on that, and here it is. I'll link it in somewhere up there and at the end and down below if you haven't seen it, because it is rather fascinating how you can actually, uh, cascade or put in series Op amps like this to increase the bandwidth. So just a brief overview of that.

All Op amps have what's called a gain bandwidth product where the bandwidth of the Op-amp will actually decrease when you increase the gain. So in the case of the micro current, we need a times 100 gain. and if it's got a bandwidth gain bandwidth product of let's say one megahertz, then you have to actually divide that one. If you have times 100, you have to divide that one megahertz bandwidth by a hundred.

And that's the effective bandwidth of the Op-amp So it's not much fun when your whiz-bang One megahertz Op-amp actually suddenly has a bandwidth of 10 megahertz when you need a times 100k, which is what we need in the microcurrent. Of course, it's designed for measuring, uh, small voltage drops across a sense resistor and overcoming burden voltage on a multimeter. and but most people these days use it for like, low current measurement that was never the original uh intended purpose. really.

I just included the uh, the nano amp range just because I could. Really So so. Anyway, the more you cascade in series like this, the more you're actually going to get an improvement. But you can't just put infinite number in series and get infinite power and get your one megahertz bandwidth.

There are diminishing improvements here. Um, so yeah. Look, by the time you get like three in series, the improvement is only 1.25 times as opposed to just adding a sec. Oh sorry.

full stage At in two stages you get like 6.4 times improvement. Um, in the bandwidth based just by putting uh, you know, times 10 and times 10 instead of times 100.. Anyway, there's a formula up the top there for those playing along at home. So yeah, that's an interesting video.
So what I wanted to do is actually, um, have a look at what the current state of the art is because I haven't looked for many years now. I mean, back when I was looking, pretty much the max 4239 4238 was pretty much the duck's guts in terms of Op-amps because pretty much there's one major criteria. Well, there's two major criterias. The first one is the offset voltage of the Op-amp because when you're measuring a low currents like this, you don't want, uh, the offset voltage you Op amp to like have a residual offset on your multimeter that you're using to measure the output.

Sure, you could null your multimeter and null it out and things like that, but it's not. if it's nice if it's actually, uh, trimmed and calibrated, or in this particular case, we don't trim and calibrate. We actually, um, I, I actually use like this 10 ohm and this 10 milliamp resistor in here, for example, is like a 4 four us dollar resistor. It's really expensive because it's not point one percent, ten milliohms and these are you know, 0.05 percent so we don't have to trim anything and we use, uh, pretty much, well, what what was at the time one practically the lowest offset voltage Op Amp on the market.

So it had 0.1 micro volts a typical offset as uh, will go into and that changes a little bit when you uh, cascade them in series like this. But yeah, anyway, like so pretty much the micro car. The good thing about it and the reason why uh, it's it's relatively expensive for the type of parts in is because these parts used in it are quite expensive so that you don't have to do any calibration, trimming or anything like that. So yeah, we're going to do a video.

Just looking around the market for Op amps, I'm going to try and find the best. In this particular case, it's a chopper amplifier I won't go into like that's how they get the low offset voltages that chops at a certain frequency. In the case of the Max 4239, it's about, you know, 15 odd kilohertz or something. So there's a little pole response in there of you know, so it has disadvantages.

But chopper amps give you extremely low offset current. Almost negligible. But yeah, let's take a look. Is there anything better these days than the Max 4239? Oh, haven't looked.

So that's why I'm going to do this video. A lot of people say this is one of the most boring aspects of electronics design. I disagree. I think it's one of the most exciting aspects because it's like, uh, looking up something on Wikipedia.

Like you go look up a couple of minutes, look up something on Wikipedia and before you know it, you go down the Wikipedia link rabbit hole and you're learning quantum theory and I don't know, goat farming or something. And the same thing can happen with chips here. as when you start browsing, doing parametric search and stuff like that, you often find very unusual parts. You go, oh wow, I didn't know that part existed.
Whoa. And then your brain starts thinking and you get another idea and then you get sidetracked. Oh, I got a new project idea and then you forget that project idea and go on anyway. Yeah, it's one of the most exciting aspects of electronics design in my humble opinion.

Let's go just a quick design explanation why we need a low offset voltage here. In this particular case, like in the order of 0.1 micro volts nominal which is what the max 42 39 has is because well that that sounds like bugger or like like what's wrong with even you know, 100 micro volts or something like that. Well the thing is the gain. You've got a total gain times 10 times 10 for a times 100 gain here.

So any input offset voltage on the on the effectively. Here's your input if you have 0.1 micro volts typical offset typical and go into that typical offset voltage here. you multiply that by 100 and eventually you're going to get out of here a typical offset voltage of 10 micro volts, and well, once again, that doesn't sound like a lot either. But think in terms of your typical four and a half digital 50 000 count multimeter on your millivolt range, which you might be using to actually measure the output of this thing.

Or you might be using some data logger output. or you know, whatever it is, then that 10 micro volts is 0.01 millivolts and that's the least significant digit of a typical 500 millivolt range On say, a 50 000 count multimeter. So you want the offset voltage to be, you know, basically the least significant digit of your measurement device. So we're targeting that nominal 0.1 micro volts offset voltage.

so we're looking for anything better than that. If there's nothing better than that on the market, or you know, maybe we can get something that's the same, but it might be cheaper for example. So there might be benefits, or it might be higher bandwidth. because I said they didn't finish my thought before.

There were two design requirements: One is the offset voltage, two is the bandwidth, and this one, we increase the bandwidth. What is the bandwidth in microcurrent? I forget. it's like 300 300 kilohertz or something like that. So if we could push that to 500 or even a meg.

Whoa. Now we're talking right? So we've got the max 4239. This is the latest data sheet I believe. Anyway, here we go: Dc performance Ideal for high precision sensor interfaces.

blah blah blah. 0.1 micro volts Typical. They don't say typical, it just says ultra low 0.1 micro volts. Never believe the banner specs always go in.

They're actually very nice here and they tell you 2 micro volts maximum at 25 degrees C. But I've done extensive testing on these chips over the years and if we go further in data sheet we might see some bin in data. A good data sheet will have been in data for an important you know. Basically one of the main benefits of this chip because it's a chopper amplifier is or they call it a they don't call it chopper amplifier up here but that's the technique it uses to get your ultra low offset voltage.
So anyway, um yeah so it you know it could be as much as 2 micro volts at 25 degrees C. I mean you're not going to use this thing at 85 degrees C to minus 40. so you know you take your 2 micro volts. Worst case.

but I've done extensive testing over the years and it is pretty much in the order of 0.1 microvolts. You know, 0.2 something like that. It is real. You know that is quite typical.

So anyway so anyway, the 4239 has 6.5 megahertz uh, gain bandwidth a product which is a pretty schmick. so we need um, you know, if we're looking for a better chip we might get. you know we might even want to trade off. You might even want.

This is where this is where you can come. a gutsa. By the way, when you rigidly when you're designing a product, you rigidly set your specifications. I must have this and that's it.

I draw the line right there and I must have this bandwidth and this offset voltage. If you aren't willing to negotiate in there, then you either might not be able to find a chip or you might use a really expensive you might be pushed into a really expensive solution that prices your product out of the market for example. or it's just you know and all sorts of things. Can uh, do that so you might want to go.

Oh, I found this chip but it's got okay, it's got you know, 0.2 microvolts typical, but its bandwidth is like 20 megahertz or something. you might want to go. I'm bandwidth. Hmm, yeah, I think I'll I'll live with the lower offset, higher offset voltage on my product if it means I get a much higher gain, bandwidth, or total bandwidth of my product.

So you just be willing to negotiate in your specs when you're searching for parts. Pro tip: and this one is, uh, up to 5.5 volt supply voltage. A larger supply voltage might be an advantage here because of dynamic range, dynamic range is very important for a product like this that you're trying to measure. You know, if you're a product switching from an amp down to one micro amp and it's switching like that, That's a massive dynamic range.

and that's why you need multiple ranges. So, but if you've got a higher voltage supply for your Op amp, then you can actually get a higher output voltage by a higher dynamic range for a given range resistor, for example. So you know I look, I won't go. I've probably done it in other videos.

I won't go into the examples, but an advantage would be a higher supply voltage. You know, if you run it, run to 12 volts, find one that goes to 12 volts for example. But then again, if it's going uh in, if the output's going into the input of a data logger or a like an Adc or something like that, then you might not care about the high voltage. There's no advantage.
You just have to live with the multiple ranges and lower dynamic range and stuff like that. And then there's noise specs as well. This is 1.5 microvolts peak to peak from Dc to 10 hertz. So that's like only 10 Hertz.

That's just the that's just the low end stuff. You've got to look at the higher end stuff anyway. So there's lots of specs that go into this uh, and a drift as well. 10.0 volts per degree C drift.

So if you're measuring things generally if you're using a product like this, you're generally using the lab you're not going to be using in the field so you know it's not going to temperature's. not going to change that much. So generally you don't have to care about that sort of thing. But if you were designing, say, an outdoor product that could go from, you know, being used indoors and then being used outdoors.

And I had this problem when I worked on seismic equipment that was used on the back of seismic survey vessels. Uh, during testing on the back deck of the boat. for example. Um, in in the middle of the ocean, it could get like 50 degrees like baking hot on the back of this on the back deck of the ship.

and then when they throw it in the freezing cold water. Um, huge temperature variations like 40 degrees temperature variations between actually testing and then deploying your product. And you know temperature drift was a like a big thing. Anyway, as I said, don't believe the Bano specs.

You know 0.1 micro volts. Always go have a look and here it is typical 0.1 maximum 2 micro volts. but it really gets nowhere. Note 1.: Always read the notes.

Oh, come on. Note one: Guaranteed by design, thermocouple and leakage effects preclude measurement of this parameter during production testing and they just went. Oh bugger, this is too hard to measure production. We just guarantee it by design, but there's nothing wrong with that.

Of course that's fine. Um, yeah, because like the reason a lot of these parts precision parts are expensive is because they're expensive and time consuming to test. It's really difficult, especially really precision stuff, low leakage, and you know, low offset voltage, and stuff like really difficult to measure on a mass scale when you're designing production test jigs. So yeah, that's why you can pay like five ten bucks for a real precision op amp, for example.

It's like you wonder why it's the same amount of silicon. Yeah, it's the labor and the complexity that goes into testing and guaranteeing the performance. Anyway, they are screened blah blah blah Okay, let's go now. I told you they might have some bin in Ta-da Here it is down here.

Input Offset distribution. Not all manufacturers will give you this maximum's been pretty generous here in giving well like they've measured. Do they say how many parts? Uh, they don't say oh I I saying it up the top no, doesn't say how many parts I don't Yeah, I'm sure I've looked at these this graph before and they told you the number of parts. Anyway, they've measured a whole bunch of production parts so this is a 0.3 microvolt.
uh bin. So from there to there, so that's plus minus 1.15 micro volts. Uh, and you can see that you know like 42 of them of the parts are going to be in that uh region and some like the production processes. They can drift like this.

so this is not guaranteed. This is just what they measured. They will guarantee it's not more. You know they'll guarantee it's not outside of here and it's not outside of here.

But you know they're just showing you. And once again, this will be like a typical bell curve response. A parametric, uh, bell curve like that. So you know they're finding that like most of them, it's like 70 percent plus of parts.

there are within like 0.45 micro volts for example. So they're running a pretty tight ship here. so they say 2 micro volts maximum up here. But as you saw in the production actual production test results, it's much tighter than that.

But if you need to guarantee your performance for your product, for your customer, for your application, then you would work with the two micro volts. Unless you want to bend them yourself and you can do that, you can actually buy the parts and build up a pro custom test jig and actually measure that, measure it and then bin them. Good bad. And you can do that actually on the component level or on the board level.

Uh, because I designed a little amplifier board once that was used in a sono boy and we actually designed this huge like big like table size. Um, automated because we'll manufacture a couple hundred thousand of these boards and we had to guarantee the performance of each one. Fully characterize them with uh, dynamic signal analyzers and everything. measuring all sorts of parameters and stuff like that and we you know it'd have like a you'd load all the boards into a big like uh you know clip you you know, insert them all in, you put them in and these automated actuators that shoot out the next board and it'd travel along to a certain spot and then a bit of nails probes had come down and then it'll automate, run the test and then lift up and then shoot the board into a like a bin.

them either a pass fail thing or into a graded uh type thing. Anyway, what are we like? 10 15 minutes into this and we haven't even started looking for alternatives yet. Anyway, let's just say that we're looking for you know, in the order of not more by the you know, by the order of typically means like order of magnitude but you know, like we want to be a bit tighter than that, not as tight as a nun's nasty but we want to be reasonably tight so you know 0.1 microvolts ish. offset voltage is what we're shooting for the 6.5 megahertz Gain Bandwidth product is what we're shooting for here.
Um and everything like drifters. If you've got a fantastic low offset op amp like this, the the The Drift isn't going to matter. Noise? Well, we'll look about that later. Um, you know, let's just see if we can base it on those two parameters: offset, voltage, and bandwidth.

All right, who's your favorite provider? Who's your daddy? Uh, let's just go to Digikey first, shall we? Op Amp: 46 000 Op Amps Thank you very much. Now, of course there are two ways to do this. One is through your component supplies, parametric search, your digi keys, your mouses, your Lcscs, or uh, your Rs components, your funnels, Element 14, whatever they call themselves these days. And you know all these those parametric, uh search engines, And they can.

They're not always infallible. Um, but you can also go directly to the manufacturers as well. So I'm I might have started this off by going to the Maxim website and using the Maximum parametric search and we'll probably do that shortly. but for now I don't I don't care who my manufacturer is, you might have your preferred supplier, but I don't care.

All I care about is that offset voltage. So here it is: Chopper Zero Drift Amplifier Type A Pretty well you could say current sense. so I'm going to include because that's our application. So I'm going to include current sense in there just in case.

Because the art of parametric searching is to ensure you don't miss something is to you know, try it many different ways on many different sites to make sure you don't miss anything because there could be some obscure parts somewhere that you're missing. and as always, there could be some obscure manufacturer I've never heard of. If you do know something better, of course, leave it down below. But anyway, um, instrumentation? no isolation? Jfet Limited.

No, No, No. no programmable gain. Nope. Nope.

Nope. No trans conductance. No variable game video. No zero drift.

There you go. It could have been classed as zero drift see look chopper, Zero drift and zero drift. Exactly the same thing, but two different whoever's it was entering the data into this digikey database. They went.

oh, this one. they call it a chopper And then they went. Oh, you had zero drift, but they didn't realize that they already had a category for zero drift. So there you go.

A trap for young players. We could have missed some valuable devices. I don't care about the number, circuits and everything else. I don't care about the gain bandwidth products.

Let's just apply filter. like. We could have gone for broke right here and gone. Oh, we only want the game band with product above.

blah. But anyway, I don't care. Let's go 4 300. Whoa.
This is heavy, all right. Input offset voltage is the column that we want here. so let's sort from smallest upwards input offset voltage. Oh, we got some blankety blanks.

Well, blankity blanks aren't going to help us, are they? Um, that's really annoying and these all happen to be the current sense amplifier type. by the looks of it. no Knights one, Zero drift. too bad if you use that microchip one, it's obsolete thank you very much and Dip package.

That's probably why he's probably still got the same chip. now. I don't want to eliminate any of these, but for now, I just want to see uh, what input micro Nia look. The lowest one we can get is the 0.1 micro volts.

So let's just go say we want you know, 250 nano volts? Something like that. Okay, so there's 59 devices. They're mostly the same, but they'll just be all like, um, different variants. Stuff like that.

Max 99 22. there's a newbie. Um, I don't recall that one before. Typically input offset voltage: 0.1 micro volts, but minus 3 db Bandwidth: 10 kilohertz Um, well, that's not Game Bandwidth product, Is it? Game Bandwidth product is here and it's got nothing in there, so I've never.

Once again. um, this is where you might find new parts. I've never. I don't think I've heard of the Max 9922 before.

but anyway, let's go to the data sheet. All right. Ultra precision, high side current sense amplifier. It's a high side jobby.

Uh, we don't need a high side jobby. I've gone into in my power supply design thing. I've gone into high side design and stuff for high side versus low side and things like that. Um, anyway.

full scale gain actually. Um, plus minus 10 micro volts max. Oh no, that that doesn't sound very good does it? I'm not impressed by that. 0.1 typical.

plus minus 10. That's not nearly as tight as the 42 39 is it? So I think you might be able to. Might might be able to rule that one out here. Here we go.

They got Uni again. No, no, that's gain history. offset voltage. There you go.

Oh yeah, look, some of the bins are up in the microvolt range here. So yeah, look at that too. You know. So yeah, it's It's just not as tight.

Yeah, so we'll go away from accident. Renesas? Um, Isl. Well, they've been buying. Um, zero drift.

Uh, 1.5 3.5 Meg Game bandwidth product. Not as good, but 0.2 micro volts. Okay, I don't know. With Squiz, Here we go: 5 volt.

ultra low noise Auto Zero rail to rail precision. 2.5 micro volts max. Looks like it's very similar. 3.5 Meg Game bandwidth product is that similar to the 4238.

So the 4239 still has the Um edge. There there's our Vos histogram. Fantastic. Our number of amplifiers.

There you go. They tell you they're sampled 1400 amplifi? Well, a couple of thousand amplifiers. That's each individual bin there. and each individual bin looks like it's uh, 0.25 microvolts.
so a bit tighter than the 42-39 so that that is an equivalent. I'd deem that to be an equivalent part apart from the Uh Gain bandwidth product there. So yeah, apart from that. Without looking into further details, you know, noise and and drift and stuff like drift up here.

we don't. Uh, at one kilohertz? Uh, one Hertz did 250 nano volts. I can't remember what it was over there, but you know we're not too fussed about that. uh, common male rejection.

and the common ram? The common mode rejection ratio? Uh, 135. I think it's uh, is that slightly better? Doesn't matter. We're not too concerned about that. Uh yeah.

2.5 microvolts. That's That's pretty much the same as what we're getting. Is that that's that's tighter over temperature, isn't it? So there you go. You know that's that's worth considering.

And interestingly here here you go. Let's have a look at the gain phase response and you can see the switching frequency very similar. That's that little dip in there like that. So the switching frequency around about, you know, 15 kilohertz? Very similar to the max.

40 to a 39 and you can potentially come a gutsy with that. Um, you've got to be careful if you're actually using it at that specific frequency. the amplifiers can, you know start to do some wobbly business. Um, and you can come a gutser.

Um, but it's got to be at the specific frequency which we're which is not tight by the way. It drifts with temperature and all sorts of things. That's like they probably won't even tell you. Um, the switching frequency apart from, you know, nominally 15 kilohertz or something like that.

So it's rail-to-rail input of course. Uh, we don't need a rail-to-rail input because we're measuring um, just from A unless you had a single supply and you wanted to measure right to ground. But rail-to-rail input is not necessarily ground sensing, so just be. you know you can.

They can be different things so you can come and guts are there. But anyway, the railroad input and output. That's quite nice. So yeah, that's that's not a bad part.

Um, and that's been updated March 12th, 2020. So that's update. Is this a new part? I don't actually recall it before, but uh, yeah, it's um. they've got a spice model as well.

Um, competitor? A? Ah, come on, come on, tell us. put the part numbers in here. Come on man up, tell us. Oh, they got a spice schematic as well.

Whoa. Go Renesas for the win! Wow. Spice aficionados getting all moist right about now. Oh no, it's an old part.

2011. There you go. As old as the hills. Aha Looks like they've updated all the stuff.

is that when uh, have um, Renises just recently bought? Sure enough, Renisosa acquired Intercell Um which is the Isl Uh partner and the Isl designator on the Uh front of that and that was back in 2017. I don't know I might have covered that on the amp hour. So many companies being bought out we gave up. like trying to keep up with who's buying who in the electronics business.
It was just it's just insane. So yeah. okay. that one's all right, but it's not significantly better.

So you know I'm not going to be writing home to my mum about that one. I'm looking for something better. And right off the bat here. just on the input offset voltage like nothing is showing up here.

It's like, you know, the maximum 90. Oh well. where's the 4238 4239? Why is that not right up there? Aha. Now get on to page three here.

and Ti: Ah opa. uh it 2388 of rings a bell. Um, we're talking 0.25 micro volts? Uh, 10 Megahertz Game Bandwidth product. So it's got.

that's definitely worth the squeeze. Beware, sneaky buggers that, uh, try to put instead of nano volts per degree C Drift? they put it. Micro volts. Why is? because .005 registers your in your brain instantly as better than 0.5 uh, nano volts or five? Sorry.

Five nano volts per? uh degrees C. Little marketing trick there. Yeah, good on your Ti wankers. 140 Cmrr.

Uh. True Rail derail. Low noise. Seven nano volts per route.

Hertz at one kilohertz? Uh. one on F. Uh. from point one to ten Hertz? Yeah, I think they're all the same.

Um, anyway, game Bandwidth product: Ten megahertz, a single supply. same as the 4239. Uh, true rail-to-rail input and output. Our Emir if I feel, um, filtered inputs, that's nice.

Current shunt measurement. Geez, that's just what we want. Ultra low noise, fast settling, Zero drift, blah blah blah blah blah. 0.25 microvolts plus minus 5 micro volts so you know you got nothing.

It's not as good. The max 4239 is still better, because once again, if we had that and it was typically around that and we used it with a four and a half digit meter, you'd get, um, a plus minus two and a half least significance digits flapping around in the breeze. Well, they're not flapping around in a breeze. They'd be, uh, you know, you could null it out.

But anyway, we're after a precision Op amp. Um, so not not quite as good. But if bandwidth mattered, you know, if that could give you, uh, some benefit over somebody else's product or something like that, then you might go. Hey, you know, I'll sacrifice some of that offset goodness for, uh, bandwidth.

Oh wow. The input offset voltage distribution here. That's why does the harbour bridge? That's that's a bit how you're doing. Um, yeah, that's not quite as good.

Is it not quite as quite as sharp? So much larger specification there. An offset voltage drift. Ah, we don't really care about that. But yeah, not as good as the Maxim.

So really, um, out of all those, we're only. that's all we had was. uh, three pages. Um, pretty much only.

Uh, three major parts there. And really, um, nothing better than the Max 4239. But why isn't the max 4239 there? That's interesting. Search for Max 4239? It's there.
What's what's uncategorized? One new product. What's that? Oh, Marketplace product. Unable to receive pricing a stock at this time? What's a marketplace product I haven't seen online? And Digikey Marketplace provides access to a broader range of supplies and products. Uh, oh, ding-ding-ding-ding Warning: Will Robertson Warning: Um, yeah, where are they getting it from? Um, Whoever they can find at the Shenzen market this week, Marketplace products shipped directly from their supplier.

Uh, yeah, from their supplier who you know from under the table in the Shenzhen market and it's still at the gym's in market. It may incur a separate shipping fee. Service typically Valve may not be available. Yeah, you would not be touching any sort of product that didn't come from direct from the manufacturer.

So you can really come as a supplier. There you go, it comes through, and then they tell you. Rochester Electronics? Well, I didn't know Digikey could actually source them, but they're out of stock. but presumably they can source them from other suppliers.

That's interesting. Is that new? I? I haven't seen that. Maybe it's maybe it's been there for a while. But hmm yeah.

Buyer beware. Yeah, there it is. The Max 4239, aut and all those letters after the like. The plus means it's Ros compliant and stuff like that.

and uh. and anyway, let's uh. 0.1 microvolts. So yeah, why didn't that show up on a general search for low input, offset voltage Op amps? This might be a good example where you could actually come a gutser doing this.

So let's actually go back. And the good thing about this is you can go back in your parametric searches like this. and there you go. Input Offset: 0.1 micro volts.

Like why doesn't Why doesn't the 4239 show up in that one? 0.01 microvolt input offset. That's interesting. So if we clear that and we just apply filter now, I think we need to go back a bit. Maybe it wasn't under the amplifier type here.

I I think it might have been categorized as a different amplifier type. So let's go back here. Okay, let's go. Oh, we could just clear the clear all selections right? and but but here's where we selected these.

Let's actually clear those selections. Okay, let's get rid of those and let's just now. go apply filter and how many? uh, response. Forty six thousand, Forty six thousand results.

But that's okay. We can. We can go input offset voltage, go from Nafl right and then choose because it might have been under like general purpose. I think it was under general purpose.

Aha see, we've come a gutsy already. So yeah, don't follow dumbass Dave. I guarantee you that's the problem. You search for input offset voltage like that and you search for filters and search for 0.1 to 0.25 microvolts like we did before, but we don't have a category selected then.
Bingo. Oh, 4288. What's the obsolete Ic buffer? Hmm, that's interesting. Anyway, Um, 4238.

There it is. 42.39 We found it because it's under general Purpose Op-amps Yeah, there it is. general purpose. So yeah, we got my guts are doing that anyway.

So this has given us another. But once again, there's nothing better. Like, like, like we showed here. there's nothing better than 0.1 microvolts.

So really, um, yeah, nothing better exists according. Searching every single Op-amp on digi-key nothing's better than 0.1 micro volts. Just not going to get it. But we might get something.

Say, let's just go save for that and go for the Game Bandwidth product. Well, let's apply filter for that and then we'll filter for game Bandwidth product. Where is it? Is it Not there? It is. 10 megahertz.

10 Megahertz is the best. We'll get that Ti. These two here ones will get that Ti part and the Maxim parts if we select those. and Bingo 4239 and the Texas Instruments part down there.

That's it. It's nothing better. Does that end to video right? So what we can do is just go to Ti.com here and go to their parametric search products. Haven't done it in a while at Ti.

Where is it it's going to? Amplifiers featured products. Once again, you can do parametric cross-reference here. Search any supplies Op-amps to find similar devices by part number. Let's actually try that.

Let's put in the max: 42. Can we? other part number: Max: 4239? Can we do that? Select Vos? There you go. Two micro volts, Five micro volts. Okay, oh.

one. One microvolt. Hang on. What's the Tlc 2652 Precision Chop and Stabilized Operation amplifier? That one didn't show up in our digikey search, did it? Hmm, that'll do.

Plus Minus 5 volts too. So there you go. If you wanted to, uh, needed a larger dynamic range, that might be, uh, better than the other alternatives data sheets. The easiest way to see it.

One micro volt max. See. Technically, that is better than the 4239 one micro volts max, whereas it's 2 micro volts max and the 4239 and right up the top. Here you go.

we've got our distribution. Uh, nicey. They're down like a good lot of them. Uh, like some are down at the minus one micro volt level, so that's not.

That's not terrific. Compare that to our 4239 over here. Oh, look, you know good bulk of them. that's tighter.

That's tighter so they might claim a lower max. But uh, in in practice, you might find that the 4239 is still better. Once again. Not going to write home to my mum about this and say oh, this is one microvolt and therefore it's better than the 4239.

But look at the spread. I mean the spread on the high side here is it. But that's just how they happen to test it on that day, you know? So anyway, no, no, no, no, no, no, No. So you know it's probably a reasonable alternative part, especially if you needed that higher voltage, uh, rail and better dynamic.
Oh die shot, Thank you very much. Um, it's you know it's bonding because you can probably buy it as a as a bare die and then bond it yourself. Um, for hybrids and stuff like that. Um, it's very important to be able to pretty much any manufacturer.

If you go to them and say we need a bear diversion, they will send you a copyright 1988. that's ancient. Wow. Anyway, that's been away, that's been around for a while that jobby.

Anyway, they'll um, yeah, they'll sell your Bare Die for a price, even if they don't have it in there. Uh, as their standard ordering thing, I'm sure they will do it. They'll bend over backwards for you. anyway.

Let's go into Precision Op Amps here. 343 different Op-amps and that'll be all parametric. Uh. difference.

Low Offset: Less than 500 microvolts. Um, and this isn't necessarily the proper parametric search, but I'm just going uh, into Oh Zero Note. I saw it. Zero drift.

Zero drift. Go back. Zero Drift. Less than 20.

So low offset. Find products. We are in our parametric search here, but once again, they could miscategorize them. but it's I that's extremely rare for that to happen on a manufacturer's website.

You'll get that on the catalog suppliers websites, but not on the manufacturers. So yeah, that's fine. So we should be able to search sort not game bandwidth product Vos There you go. Hang on.

That's and 0.5 millivolts. No, we're after micro volts, so we're in the wrong category current sense amplifiers because that's exactly what we're doing. We're doing a current sense, so they're going to be optimized for low offset voltage, but some of these are obviously. look, you know, high voltage? You know, 40 volts, 110 volts? You know, high side, um, current sensing, and stuff like that.

so not necessarily the best thing. Can we just go to Operational Amplifier? And where's the parametric search for just all? Search every single Op amp 1520 of them. Search all precision Op amps, offset voltage, millivolts. Did I have that around the wrong way before? I might have been searching sorting that the wrong way.

Chopper stabilized point. Anyway, one micro volt job. There you go. The Tlc 2652 already looked at that.

The Opa 189. Oh, look. 14 megahertz Mux friendly. wonder what makes it mux friendly? Um, that would have to be some dynamic, uh, switching performance thing, surely.

And the 2652. They actually list that differently to the 2652a up here because the A version is just tighter tolerance. Um, yeah, that one's tight as a nun's nasty. This one's a bit loosey-goosey Anyway, the one Eight nine I think I remember that you know that's been around the traps for a while.

Three micro volts maximum? guaranteed. That's okay. there it is. Robust Max: Friendly Inputs: Rfami Um, that would be uh for a transient, uh, switching on the input.
Um, so that it didn't so it's probably got faster. Fast settling probably got fast, uh, recovery from input. Um, that could be handy in your design. They've obviously had like some customers with that problem actually request or you know that sort of thing.

So yeah. Robust mux, Friendly input inputs without anti-parallel diodes. There you go. So any parallel diodes I want to find settles quickly and maintains high input impedance when switched.

That could be handy for something like what we're doing with the microcurrent for example. especially like an auto ranging type version, you want them, you don't want them to be overload. because overload recovery time. If I done a video on overload recovery time, I'm sure I have somewhere rings a bell.

I've done like 1500 videos or something, can't remember Anyway, Um, yeah. recovery time. uh, can be a big thing when you overload. uh.

the Op amp. you go full scale input and you clear and the output clamps hard. It actually takes time to recover and then you can actually lose data in there when you're like trying to sample what's actually happening because it takes a certain amount of time to settle back down. So yeah, that could be a big thing.

And competitor high voltage amp muxpending prevents loading of source. Oh, that prevents. Okay, okay, okay, so what they're doing? No, it's not nothing to do with the settling, it's the loading of the source. Hence, why they don't have the anti-parallel diodes.

Because when you've got a mux, you've got to remember a mux is just a a basically a resistor in series. Um, with it, and like a, you know, a selectable mosfet which has a serious resistance, effectively, series impedance. and then it switches it through so it can load down. If you have these anti-parallel diodes, when you're switching, it can actually load down your loads and stuff like that.

So yeah, okay. see now. This is kind of an example of where even when I was still searching within Op amps, you might find an obscure device like this and go. Oh yeah, I had this other design I was thinking about, but uh, you know I come a gutser when I was switching much.

You know it's this has got mux friendly inputs, so you know you lock that away mentally up here and go. Oh yeah, I've heard of that. Um, when another design application comes along. Oh yeah, these mux friendly input ones.

I remember those point four micro volts of offset voltage. If that's their banner spec, then you know it's not as good as the 4239. plus minus 0.4 plus minus 3 microvolts max. Yeah, it's It's not as good.

So unless you wanted, uh, the game bandwidth. uh, the increased gain bandwidth product. Uh, the common rejection ratio. That's not too shabby.
is it a plus? Minus Plus minus 18 volts? Wow. If you wanted your headroom. Wow. Wow.

Really Yeah. 4.5 volts to 36 volts? Wow. If you wanted your dynamic range, this job has got you covered. Wow.

Okay, that's impressive. Imagine having plus minus 15 volts of dynamic range. So yeah. mental note on that one.

Um, Mux friendly and a huge supply voltage range up to 36 volts? Um, not necessarily what I'm looking for And a 14 megahertz game bandwidth product. Wow. How much is that? Jobby? Opa? Uh, 189. Geez, that's all right.

2 bucks 50 for one off. Uh, buck? 12 in. Uh, 2500. Real quantity? That's pretty good.

Okay, let's go to Analog devices here. Amplifiers: Rs specialty amplifiers, Precision Op Amps Vos less than one millivolt. Yep, that's what we're Zero drift. No, we don't want a zero drifter.

Zero drift? There we go. Zero Drift Op Amps: Uh, combine auto zero and chopping techniques to continue and I was interpreted less than Point Four. My, uh. less than four micro volts? So it looks like analog devices.

Um, and two microamps? Oh, they've got a 36 volt Jobby too to chop Or zero? That is the question. So there you go. You come across interesting articles like this: because a chopper amplifier is better at like lower frequencies and you know stuff like this, an Auto Zero? uh techniques. And you know.

here's some of the Auto Zero techniques. I'm sure I've gone over these uh, before and how they switch. you know the thing in and they they constantly. Basically basic topology is they're constantly correcting for the input offset voltage.

So um, that's that's what they do. So Chopping technique? Oh, there you go. You know it's a bit complicated there, but uh, and combining Auto Zero with chopping? There you go. So you know that's kind of what they're doing.

Um, all these different phases inside. Um, the Op-amp So that's what's you know. Chopper stabilized and so there you go. different.

Auto Zero uses sampling to correct offset, whereas chopping uses modulation and demodulation. Sampling causes noise to fall back into the baseband and you don't want that you can come and get. So Auto Zero have more in-band noise to suppress more blah blah blah and yeah. um.

Pros and cons both ways. application suggestions and things like that. So that's a nice little article, you know? Okay, parametric search Vos sort 2.5 micro Ada 45 28 Um, Oh, they got high voltage parts and things like that. That's quite neat.

High voltage working with a strain gauge? Um, anyway, it looks like the best they got here is. uh, whereas Game Bandwidth product 3.4 Meg 2.5 mic I don't even need to check it, It's not really happening. It's not blowing my socks off. The 423 max 4239 is 6.5 meg.

Game Bandwidth product? Nothing's You know they've They've got nothing there that stands out. Um, so yeah. no Ti? Nope. Uh.
Analog devices? Nope. I mean, we can look at more. I mean, mouse is going to have the same thing. I greatly doubt that mouse is going to have anything else.

Operational Amplifiers, Input offset voltage. That's no good. One one micro volt. Oh, here we go.

Okay, apply filter and on semi. Uh, that's National Semiconductor part number. But here you go. We've got an interesting one down here.

This is where you know you find this interesting stuff. I don't know why they've only got three here. I've goofed Amplifier: Ics, Operation amplifiers, Vos input offset voltage, multiple values. Okay, so maybe they didn't have all the data.

Maybe it's not as good as digikey. Anyway, we found a manufacturer who is Ablik. I've never heard of Ablic, but they make zero drift Op amps so I'm going to give them a go. Let's have a squiz who is Ablik.

It's Ablik.com This is one of the fascinating parts about looking for stuff world's thinnest hall effect. I see: create comfortable. what do they make? They make real time wake up interval counter Ics Power sequences, Wireless Power Ics Squared Prom Power Management Ics Okay, I probably don't. Uh, mind mine, be a Mitsumi Minbia Mitsumi What? Who are they Ablik? Who's who or what is that? I'm sorry.

It isn't an ad. like is it a parent? Is it a Um company? Begin selling masks on Amazon? What? what? Selling masks? What it's It's oh god's an ad. I've been mask Rick rolled. I think the world's lowest self-current consumption.

industry's fastest detection world said. well. I missed it often. These banners up the top.

You know they they can highlight new products and stuff like that. Worth? uh, we're squeezing through them. Industry's fastest detection of voltage malfunctions. Really? Wow.

There you go. I guess in self-driving cars, that's important. You know these days? True autonomous cars. Decades away? it's like nuclear fusion 20 years away.

Yeah. Antenna diagnosis. Ics: Wow. Composite Ics, Lithium ion battery protection.

Oh, they do well. They do a whole bunch of stuff, but look. Operational amplifiers, Ultrasound transmit pulse and digital transmit pulses. Wow.

Look at that. That's uh, that's relevant to what we. um, we've done recently. Tablet quad Octal 16 Uh.

channel pulses integrate logic and mosfets and stuff like that. Oh wow. there you go. They make specific ultrasound If you want to know who make these ultrasound switching front ends that could be used in.

you know, the new Philips Ultrasound probe that I've looked at in a recent video. Um, they'll look to know one, but they've got new ones. Ah, then yeah, there you go. Ablik Public and they make they make operational amplifiers.

Let's go and have a squares just for Kr. Yeah, is that it? Yep, they got three parts. Automotive Zero Drift operational amplifiers supporting Aec Q100. Don't know what that is.
Obviously some standard is a failure mechanism based stress test qualification for packaged integrated circuits used in automotive applications. Okay, so it's an automotive application qualification for Rf immunity. Basically, I think that that seems to be what it is. Devices pass a specified stress test guarantee a certain level quality, Reliability Grade: Zero, blah blah blah.

Anyway, so you can find interesting stuff like this that you didn't know about before. Yeah, we're we're only talking input offset there. Only talking 50 microvolts here. So yeah, it's not.

It's not that great. Rev: 1.100 Yeah. 50 micro volts max. You know.

so we're not even in the ballpark. Um, but there you go. Um, interesting because these will be automotive. What's the what's the collective term for diodes? They've really gone to town.

There you go. Um, so yeah, they've got the anti-reverse uh, parallel ones. They're They're like, um, they're anti-marks You remember that? Um, so yeah, it's so we've got. I'm just protection everywhere.

So we've got one diet right across Vdd and Vss right across. and the inputs and outputs and then geez there. Yep, they're really going to end. They're probably beefy too, being you know, automotive type thing.

they really want to, uh, clamp that protection. So anyway, yeah, that's not going to work. Set the world on fire. but hey, you know, automotive joby.

So anyway, um, I don't know what's up with Mouser here. I've I've obviously done something dumb, haven't I? Oh no. Why? Why didn't I see this category? 0.25 micro volts? 4. 0.1 micro volts is down.

No, they might have. Be careful. You can come a gutsy here. Kamagutsa.

They might categorize these things wrong, or they might have no category entry whatsoever and you could miss it. I don't think we're going to see anything here that we haven't seen before. Linear Tech 2064 Input Offset voltage: One micro volt. No, I don't.

Oh well. no. Okay, let's let's go have a look at that. Okay, five micro volts maximum.

Okay, so it's not going to be as good. Let's go down to our parametric spread duty cycle: low, lower system Power Input Offset Voltage: typically one plus minus Five: Once again, yeah, it's not as good. Uh, what's our game? Bandwidth: Uh, products. It's not here.

It's gamebound with product 20 kilohertz. Okay, yeah. nice. Not even in the ballpark is it? And there's our distribution.

233 Typical units. Well don't overdo it there. And then I look. plus minus snare two.

Nap that? no. So it looks like, um, that's that's basically it. I don't think that we can beat the Uh 42 38, 42 39. You know I could spend another hour or two doing this, but for the purposes of this video, I think we're going far enough.

Please, if you know of a better part than the 4239 for typical offset voltage, a typical and max offset voltage, and game and or game bandwidth product. Uh, please leave it in the comments down below. I do read all comments or try. I try to, especially in the hours and the days after the video is, uh, released.
But yeah, I do try to read all the comments and I try to I, you know, when I part them and i, um, thumbs them up and I do all sorts of, you know, and I pin them and all that sort of jazz. Anyway, um yeah, if you know of a better op amp, uh, please lift it. But please list it down below. but it looks like, um, you know, just off the bat here.

Um, the max 42 38 42 39 is still pretty much the duck's guts for this sort of application. So yeah, there you go. Nothing's changed in like, you know, 12 years at least. so that's interesting.

Anyway, videos waffled on long enough. I hope you found that interesting. If you did, please give it a big thumbs up. Yeah, I've said comment down below and over on the Eev blog forum.

That's where everyone chats about this sort of stuff. And yeah, I'm on all the platforms, the libraries and the bit shoots, and the vimeos, even and all sorts of stuff. So yeah, hope you found it interesting and useful. Catch you next time you.


Avatar photo

By YTB

19 thoughts on “Eevblog #1318 – what’s state-of-the-art in current opamps?”
  1. Avataaar/Circle Created with python_avatars SnrC says:

    Why does Dave hate nuns ? Catholic school upbringing perhaps ?! 🤣

  2. Avataaar/Circle Created with python_avatars Steven Bliss says:

    MAX9922??? Was that 10uV at 100x gain???? I think you may have missed that rather important detail because if so that would make it 0.1uV

  3. Avataaar/Circle Created with python_avatars The Noisy Electron says:

    When your employer hands you a requirements spec as follows:
    Bandwidth: All of it
    Temperature range: Dark side of the moon to surface of sun
    Gain: Yes

  4. Avataaar/Circle Created with python_avatars movax20h says:

    How could they claim that this TLC2652 is 1uV input offset max? From the graph it is clearly 2uV input offset max.

  5. Avataaar/Circle Created with python_avatars Gacheru Mburu says:

    👍

  6. Avataaar/Circle Created with python_avatars Željan Alduk says:

    why not use "hand-picked" op-amps with "0 µV" offset in your product?

  7. Avataaar/Circle Created with python_avatars Asger Vestbjerg says:

    Nice search 👍
    Thanks for sharing 👍😀

  8. Avataaar/Circle Created with python_avatars Yaghiyah Brenner says:

    Bring back fundamental Fridays, we watch more of your old content than the new stuff.

  9. Avataaar/Circle Created with python_avatars ElektronikLabor says:

    In the industry an engineer can simply contact his distributor and ask him to look for a newer and better part; you just tell him which specs that interests you.
    In that way I often got parts that you cannot find with catalog distributors like DigiKey or parts that are so new that they are not officially available yet.

  10. Avataaar/Circle Created with python_avatars Kon Berner says:

    Odd.. I was watching this video and, before I knew it, I was reading this: "Goat meat can be prepared in a variety of ways, including stewing, baking, grilling, barbecuing, canning, and frying; it can be minced, curried, or made into sausage. Because of its low fat content, the meat can toughen at high temperatures if cooked without additional moisture. One of the most popular goats farmed for meat is the South African Boer, introduced into the United States in the early 1990s."

  11. Avataaar/Circle Created with python_avatars muctop says:

    SW1 ! The current is running through that switch! What if contact resistance isn´t reliable and stable?
    Uh oh…

  12. Avataaar/Circle Created with python_avatars Kerbal Launcher says:

    Anyone have recommendations for high speed (slew rate) current sense or instrumentation amplifiers with very low input offset? I'm looking to measure lowside phase currents with a minimum sampling window of 1.5uS. I was originally using the INA240A1, but its 2V/uS slew rate was too slow for the high duty cycles required by my application. According to the graph and my measurements, it wouldn't reach 90% of the target value until just after 1uS. I've been looking at the MCP6N11-010 which is less accurate, has much higher input offset and offset drift, and a greatly reduced negative common mode voltage, but has a very fast settling time in comparison to the INA240, additionally it requires me to externally drive the reference voltage with a dedicated op amp due to the low feedback resistor values that the datasheet recommends. I need something that's rated down to -1V common mode operational because I need to be able to measure at least 150A, and parasitic resistances can very easily pull the common mode below -300mV.
    I can't do inline phase sensing because I need to support up to 100V operation. Isolated sensing topologies will price my product out of competition, and even then, I implement a signal injection scheme to resolve motor position at zero speed with no sensors. This requires extremely accurate measurements and a high fidelity data acquisition stage.
    Don't ask, my customers basically want a motor controller with the power density of a nuclear reactor that comes with a set of black magic features at chinese level pricing. .__.

  13. Avataaar/Circle Created with python_avatars Amir says:

    Wankers at TI

  14. Avataaar/Circle Created with python_avatars TheDrunkenMug says:

    Wow, that description of Dave as he goes on Wikipedia… That so much resembles my own experiences 😂😀

  15. Avataaar/Circle Created with python_avatars wow man 9999 says:

    Tight as a nun's nasty. That's a new phrase.

  16. Avataaar/Circle Created with python_avatars M Stjerning says:

    With regards to the distributions presented in the datasheets I have two questions: 1) is it plausible that some manufacturers would sacrifice yield to allow them to cut off the lower and upper tolerance intervals, allowing them to present tighter tolerances than their competitors? Of course, this would imply testing and scrapping during manufacturering, but I imagine that this added cost would then simply be added to the sales cost of the device.
    2) Does the distributions take batch-to-batch variations into account, or is it possible that the tested devices are all from the same batch? This would certainly yield different results and would look bad in the datasheet.
    Has anyone experienced these kinds of creative statistics? Any particular brands?

  17. Avataaar/Circle Created with python_avatars Alyx says:

    wow 2014, i have used a microcurrent countless times since then, all the gold is basically rubbed off
    i recently needed current amplification in a project and i used an AD8554 with 2 sets of 2 series amplifiers, it worked wonderfully but was a bit overkill, could cal out in software

  18. Avataaar/Circle Created with python_avatars PerfumedManatee says:

    I love the way Dave's eyes twinkle looking at all the goodies…

  19. Avataaar/Circle Created with python_avatars EEVblog says:

    UPDATE: The ADA4528 with a nominal 2.5uV offsett actually looks like it has a tighter spread than the MAX4239, and unity gain stable. But lower bandwidth.

Leave a Reply

Your email address will not be published. Required fields are marked *