Hi just a quick video on a little circuit that I thought I'd uh build up in Breadboard Test here it is a Precision 1 Aamp Uh current generator and well I thought well you know in theory it should work, but in practice I thought there might be a few issues. So I thought I would uh, just lash it up on the breadboard and um, you know, see how well this thing actually performed. Now what I'm actually trying to do is get a as I said a precision 1 amp output which is pretty high for a Precision constant current output. and by Precision I'm talking you know 0 better than 0.05% um and absolute so it's not trimmed or anything like that.

And the way I'm doing that is using a Precision uh LTC 65- 1.25 volt voltage reference and that's uh, normally 0.025% accurate. so you know, uh, really, quite a decent chip. You know it's like $10 or12 in oneoff, uh quantity. and basically I'm using that as a Uh Precision voltage reference driven across a Precision resistor in the final product I was going to have a 0.

uh, 2% uh sorry, 0.02% uh Precision resist in here of 1.25 ohms and of course Ohms law. 1.25 volts across 1.25 uh ohms, you get precisely 1 amp. Now what you can do with these Um series? uh, voltage references or um, you know, uh, some of them you can actually the ground doesn't have to be ground so you can actually float that above ground. And what I'm doing is using a Precision Op amp here I've chosen Opa Uh 376 for the job.

You know just a really low offset voltage so that the error the offset error of this thing is less than you. You know the Uh 0.025% of the 1.25 volts I want here and I won't go into the math there. But anyway nice Precision uh opamp in there and basically what I'm doing like normally this these Precision references can only source you know 5 10 milliamps at best kind of thing. so you know it's not like you can get one amp out of it.

So we put in a series pass transistor and MPN one and just so happens in the data sheet for the LTC 65 it shows you exactly that configuration a boosted output current configuration with an MPN transistor. It also shows you another one boosted output current with an A Uh, P&P transistor and um, some other manufacturer data sheets used a PNP one as well, but this one shows you know look 2 n22 uh IMAX the maximum output current is set by uh you know the current Cape bill of the MPN transistor. So I thought okay I'll put in a Darlington because you know a dou two2 is not going to be able to do an amp and this chip I think from memory can only draw like Source like 5 milliamps maximum. So really the gain of uh, this transistor here isn't going to give us our 1 amp output here.

so I thought I'd choose a Darlington Npn a KSD 1692 which looked like it would do the job I'm not going to go into datter sheets and explain uh, why I just want to show you the uh the result I'm actually uh getting here and basically the good thing about this voltage reference is that it has a um, a force output and a sense output. Now most series voltage references don't have that, they just have a fixed voltage output. you know, 1.25 volts, 2.55 Whatever the output voltage is, this one has a sense one that so you can connect it directly across like a Precision four terminal resistor and you know that's that's. really quite nice and that's what allows you in this configuration to use this MPN transistor and then feed it back back.

And of course you got to have a minimum output capacitance on here as well. Uh, for stability because the voltage reference has an internal error amplifier and I knew that and of course error amplifiers. Uh, just like uh, linear regulators and low Dropout Regulators they can oscillate if you don't have the correct you know, uh, bypass capacitors on the input and the output. So that's why I wanted to breadboard this thing up I wasn't sure I had a funny feeling that you know look it's in the it's in the data sheet.

You know it should just work but does it? h Yeah, let's find out now. I Forgot to finish. Uh, explaining how this thing actually works? Um, this uh, Opamp here sensors the output voltage across this uh output Uh current shunt resistor here and then you know and it just basically it's a buffer that's all it is and buffers that voltage to the ground pin so it raises the ground up instead of being connected through the normal uh ground. It is the output of the Op amp here.

So you raise that ground signal up. So the difference between the output sense line here which will be tapped you know on a in in my final circuit tapped off like a four terminal current sh resistor like this. then 1.25 volts is going to be present directly across that resistor. and by Ohms law you know that's we're going to get a constant current out of here of 1 amp and of course all the current flows.

or almost all of it flows through the Darlington Npn transistor up here with only a tiny you know, a couple of Milli milliamp or two. this KSD 1692 I chose actually for 1 and A2 amps collector current I Think it's like only 1 milliamp base current or something like that, so you know it's really nice. High Gain Darlington Transistor It should in theory work well, but um yeah, so this one amp output here, you may be wondering what this Diode's uh doing down here. Well, even though this op nice Precision one is a rail to rail.

If you are feeding this into a direct short which you may want to do, you know you're calibrating a multimeter or something like that. It's got a tiny current Shun in there right then. Uh, you're bringing this input here right down to ground. And that and also the output right down technically to the ground.

Rail here doesn't work that well. You even though they're called rail to rail, you are still going to get, you know, tens of molts, uh, offset voltage. And of course that just ruins your day. So what you do is you just put a series diet here.

It's dropping. You know it' be a shock key type even though I haven't drawn that you know 0.3 volts or even standard diet at6 volts and then it just raises effectively. Uh, the input. Uh, to? well, it raises the voltage here.

sorry, um, in on the noninverting input at least. 3 or6 volts above. So it's just above. Uh, so it's well above any, uh, ground sense issues with this thing.

And so you re so you're avoiding the offset voltage issues right down at the low end of the rail. So that's why. Anyway, that circuit should in theory work. but I had a thought that H this thing could actually oscillate and it may not work that well.

And sure enough, I built the damn thing up and well, no, it was horrible. So what? I've peeled it back to. Let's go over here to my breadboard now. Unfortunately, the voltage reference only comes in one of these tiny little pain inth ass.

uh, 65 mm pitch? um mop uh packages. Oh, just awful really. But uh so I put it on an S So8 adapter here. pain in the ass.

So8 adapter had the wrong body width on it. this I think I got these in the mail bag. the wrong body width. so I've had to actually bend the pins down in there.

Absolutely horrible. Anyway, there's my Op hand, which is an So8 and this thing. even though it's um, 65 mm pin pitch, I was able to squeeze it onto uh with a bit of uh, you know, trickery down there. Onto the S So8.

I might get the macro lens and show you that up close. So there you go. Even though it's an S So8 adapter, the two middle pins can solded directly on there. and what I did is just lift up the outer pins and then just you know, put some little jumper wires over there like that.

So that's how I can fit a different a smaller pin pitch one. Ono standard S8 adapter because I didn't have any adapters of the right pitch anyway. so this damn thing didn't work. So what I've done is I've basically take I've disconnected the Op amp here, so just imagine that's not there.

and I've just got my voltage reference with a 2N uh, W22 transistor in there. I've got a load on there of uh, two 22 ohm resistors in parallel, so you know we're talking. this thing should generate about 110 milliamps or thereabouts. you know, Well, under the amp.

uh one amp I want? um, but I just scaled it back to get exactly the same circuit that they've got in the application node here. and you know the bypass. Yeah, I've got no bypass caps directly on there, but the whole idea of this was that, um, look if it worked on a breadboard like with this, with the bypass caps, sort of, you know, like a centimeter or or two away at most kind of thing. with you know, going through a breadboard, if it's stable there, then you can be pretty confident.

when you tighten the loops off and everything on your final PCB layout, it's going to work a treat. So you know I really wanted, uh, good confidence that it worked on the breadboard. but you guessed it, it doesn't and that is my 1.25 Vols output voltage. and as you can see, 200, 400, 600, 800, uh 1 Vol 1.2 You know it's near to there, but look how fuzzy it is, it's just got all crap on it.

and if I have a look at uh, this is the sense output by the way. And if I have a look at the the force output pin, look at this hey oscillation There we go. Look at that. a Bob's your uncle so that right there is 20 MTS per division AC Coupled, that's my sense output.

look. it's just all absolutely horrible a man. so just ignoring all that. I've just got their basic circuit in there I've got a 10 microfarad ceramic uh output cap I've got, um, a ceramic input bypass cap albeit yeah, they are on the breadboard but you know, two in22 they make it look like this thing will just work and you know IMAX You know they don't even give you sort of, you know, a maximum recommended output so this thing just doesn't work in the application.

Note: maybe if we tighten the loops up a bit. oh, but jeez, it's you know it's a bit touchy. So I actually, uh, expect the PNP one possibly to be more, uh, stable in that respect. but this, for example, like, you know, 35 milliamps maximum output.

So I don't really want to, uh, press that one into service to try and get my 1 Aamp output I wanted to, uh, you know, persevere with this circuit and see if I could get it. So that's my sense output there. Watch what happens if I disconnect the input byass resistor and I got this coming from my uh Ryo linear bench Supply 5 volts. By the way, the input uh voltage really doesn't make a uh make a difference at all.

Let's remove that and you can not that that's the input bypass cap and that's my sense output. You notice the noise went up now I'll just whack in a another bypass cap in there to replace that. Oh, look at that. There we go.

Woohoo! Now, one interesting thing with this: RAR scope look I'm not sure I can't remember which firmware version I'm running, but I may have found a firmware bug. Look at this right. You know it's like it's not uh triggering properly. It's jumping all over the place.

Now if we stop that, of course we should get a nice clean waveform but we don't look. it's furry and like it's You know it's jumping all over the shop just like you get on that now. I'm not sure if that's a feature now when you change the time base, but you know it works So that's what I would have expected when you press the stop button, but no, it doesn't I don't know feature or bug, you tell me anyway. so that's how sensitive this thing is just to.

you know, the bypass cap like that in there. So that input uh, bypass cap seems, uh, pretty critical. So what I might do is actually go in there and like solder it directly on the pins of the uh adapter. or something like that.

Maybe not directly onto the chip itself, but maybe onto the adapter. and well, that might improve things all right. So what I've done is solded bypass uh input cap directly across that and let's see what we get. Look at that absolutely awful and that's directly across the pins of the S So8 adapter.

How close does this bloody input bypass cap need to be? unbelievable? So if we put the original one back on there as well, look there we go. That's knock that down. I Mean it's still bad. I Mean it's still oscillating, especially for a Precision reference.

It's absolutely useless. But there you go. that is. Loop stability of the error amplifier inside this bloody voltage reference.

it's all over the shop now. of course, if we put no load on there, it should be okay. and yep, there it is. 1.22 Vols And yeah, I have actually, uh done the I have configured it without the uh um series pass transistor in there at all.

So just the voltage reference V in V out tied force and sense together and yeah, it's You know it's exactly spot on like that, so it's okay. but once you put any sort of load on this thing, it's useless. It just takes off. so there it is normally and let's hook up a 40 uh, 4 ohm resistor which is about 28 milliamps or something like that.

Here we go there, we go. Look, you can see it. Not that you know for a Precision um reference, it's just no good. It's hopeless.

and if we swap the transistor for our High Gain Darlington KSD 1692 and let's have a look what happens. This is with no load at all. There we go. and that's the force line.

There's the force line. Woohoo! Look at that. and of course this is going to be a function of of the output capacitance. Well so I've got a 3.3 mik uh poly cap in there instead of my well in parallel with my 10 mik ceramic and I can actually you know, take those out and I've had to play around with it.

And here we go. We change our config. We change the oscillating frequency. look at that and the type totally.

but the thing is still oscillating and that is the sense output that is not not the Uh Force output. So here's the force output. Look at that. W Man, there's a shocker.

Woo! So sorry for those who like a happy ending on their video. I Said this one will be quick so there is no happy ending I Haven't finished playing around with this thing at all. I Just wanted to show you that uh, you know, don't just trust the application. Uh, notes like that in the application circuits build them up and verify that they actually work because this one oscillates like a so sure yeah.

If you build this up on a PCB you know it's going to be. The loops are going to be uh Tighter And of course you choose your components carefully and it's going to be better. But the fact is, you know you may not get this on the first spin or the second or the third. You know, pain in the ass.

So if you can't get it, just, you know, at least um, semi-stable on a breadboard like this, just using, uh, you know, um, when you're trying different types of input and output caps and things like that and it's just all over the shop, then you know you're going to be in trouble. You know you just don't have good confidence in this thing building up on a breadboard. And yeah, like try doing an amp on this thing and other stuff like forget it. You know that's just the Um transistor you know, on its own with no load and it's oscillating like that.

It's just crazy. but hey, you know that's what I suspected it would do and it certainly did. So I'm glad I didn't go straight to PCB on this circuit and so just be wary of these sorts of things. Don't trust those app Notes H especially with stuff like this with error amps.

And yeah, this is going to. There's going to be a lot of factors involved here for the uh, stability of, uh, the Aor amplifier inside this uh reference. So you know, really it, there's a lot involved in getting this thing right and it's not going to happen today. But anyway, I hope you enjoyed that video.

Catch you next time.