[wap32]

I'm wasn't sure this would be worth creating a new thread, but I didn't want to hijack other threads.

I was simulating an amplifier circuit in NI MultiSim (basically an opamp gain stage with class A biasing followed by a mosfet buffer) and while testing it with an AD843 I noticed I was getting a high DC offset (~40mV).

First I assumed the problem was with my circuit, but even in a basic configuration such as the following one, I get an offset of 22mV (supply is +/- 12V).

If I just replace the AD843 with say, an OPA132, the value drops down to ~50uV, even less with an OPA627.

Am I doing something wrong (most likely) or is this typical of the AD843?

 

[cobaltmute]

What are your settings for XFG2?

I've seen the same thing in Multisim - I believe that some of the SPICE models in it are not as good as some others. You can see a similar thing from OPA227 to OPA627 even though in real life OPA227 works good.

I should add that simming your circuit in my multisim only got me 10mV offset.

 

[diditmyself]

Look in the data sheet and you'll see that the input bias current is quite high when warmed up. 50 mV DC offset is what to expect from your circuit. If you want to reduce it you have to balance the input currents, either with "the right" resistor values (R15/16/21 in your circuit) and an input cap or with a current source to the input.

 

[wap32]

Well, here's the thing, I'm getting different currents in the inverting and non-inverting inputs and both are way off-spec: 5.5nA and 4.5nA respectively.

Looking at the datasheet, the opamp would have to be running close to 90ºC to reach those values, even though the operating range for AD843JN is 0-70ºC.

 

[diditmyself]

I used tangent's calculator Electronics Calculators and 21.17 mV came up when the "typical" values were choosen. With the max values it reads 54.98 mV. This seems to be the nature of this opamp and not an error in Multisim.

You either have to balance the bias currents, use another opamp or be happy with some DC-offset.

 

[wap32]

Well, the 21.17mV you got is with a cap between the input and the pot, I get about -38.53mV with that configuration.
Trying to equal the voltage on each input using IB*R15=IB*(R16//R21) (resistors from the posted image) and the typical IB (0.8nA) yielded no results, changing the offset ever so slightly.

Either way, I guess the AD843 is just a hard chip to handle and requires careful design, not the task for a beginner like me.
I had read Tangent's notes on it and was hoping to get a great sounding opamp while avoiding OPA627 prices, but I'll guess I'm better off designing for something like an OPA134 which costs 3 times less and has 400 times less offset (in Multisim at least).

 

[12Bass]

FET input op amps have much lower input bias currents, thus much less DC offset. Something like an OPA827 will often have none at all. With bipolar inputs it is necessary to either balance the bias currents with resistors or use a capacitor at the input (or output).