jason9

modified 7 years ago

Rail-to-Rail OP-Amp with Few Stages

229

02:04:39

This op-amp has an offset voltage between the inputs of 1.22mV. It has a gain of 15.3mA/mV (it outputs 15.3mA for every 1mV difference between the inputs). It can output up to 20mA push, but can pull up to 2A, so if this were made in real life, you have to be careful not to overload it.

published 7 years ago

hurz

7 years ago

Offset compensated by design, 1mV in / 11mV out http://everycircuit.com/circuit/6611294236180480

deadall127

7 years ago

that's more an OTA than an op-amp isn't it?

gasboss775

7 years ago

Yes it's like an OTA but without the bias input to vary the transconductance which is here fixed at 15.3 mA / mV

jason9

7 years ago

@hurz. Oh wow, I guess the inputs must push some current out. I’ll see if I can fix it.

jason9

7 years ago

It can be fixed with a bunch of current mirrors and stuff, but that makes it more bulky and prone to self-oscillation, although I haven’t tested that. I’m not sure what to do with the offset problem, but I’ll see what I can do about it.

jason9

7 years ago

You can fix the offset by adding two 10ohm resistors to the emitters of the two NPN transistors at the bottom of the differential pair and then connecting either end of a 100ohm potentiometer to the emitters of the transistors (before the resistors) and connecting the middle (the wiper) to the negative supply and adjusting the potentiometer.

hurz

7 years ago

Right, this way is like a laser trimmed silicon and reduces the offset problem to a minimum. But its not fully gone. Anyway, just want to demonstrate you how importante resistors are as it looks like you dont like them 😁

hurz

7 years ago

One word to the standby current, because you pump 20mA into output stage AND use it also 20mA for the differential amp stage which needs only a few hundreds of microamps. This needs an update for the output stage and less current for the differential stage.

jason9

7 years ago

I fixed the standby current.

jason9

7 years ago

Never mind. I un-fixed it. When I had fixed it it had oscillated horribly.

eekee

7 years ago

Open loop gain is hard to test because without feedback the output wanders, but it appears to be in the region of 10MV/V. That's a little bit spectacular, that is. :)

eekee

7 years ago

Input range is good, -14.1V to +13.7V. (Test with unity-gain non-inverting setup.) My op-amps' inputs go to negative rail, but for a while the upper limit was less than 6V (with +-10V power). I got it up to 9V by tweaking the LTP, but it dropped to 8.93V when I increased the gain.

eekee

7 years ago

Input range also affects inverting Schmitt triggers -- the kind where the source doesn't supply current to the feedback network. With one of my amps, I saw weird behaviour when hysteresis exceeded input range; switching point became dependant on rate of change of the input. May not affect this amp, it does the opposite to mine outside input range. Mine clipped, this goes to the nearest rail.

eekee

7 years ago

Yawks! Current into the inverting input can reach 20mA when the input goes below the negative limit. One for the "don't do that" list. :) No such excursion for exceeding the positive limit.

jason9

7 years ago

Open loop gain is best measured in mA/mV, not V/V, since this appears to be a transconductance differential amplifier that outputs like a constant current source rather than a constant voltage source with the output current dependent on the difference in voltage between the two inputs. It’s gain is 15.3mA/mV as stated in the description.

eekee

7 years ago

Gotcha... I think. Apologies for the misinformation, although I do think this makes a decent op-amp too.

jason9

7 years ago

Actually, in real life, the Early effect would limit the open loop gain in V/V.

add comment in editor