Opamp capacitive coupled input and input bias current showing offset voltage out

[darklife]

I am building a microphone preamplifier where I need the input to its non-inverting op amp capacitively coupled. It's designed around the low noise JRC2068 (type DD) and it works fine but I do notice on both the positive input and the output there is around 600mv with a 4.7 mega ohm resistor for biasing (R1).

Using this example picture to help better describe...


I really don't want smaller values for this resistor (R1) which do lower the output offset voltage because it would load down a hi-Z switchable microphone transformers output.

My question is since the ~600mv hanging on the output is also capacitively coupled to the next stage and this all works fine for my build will having such a large offset voltage on the output increase distortion or degrade overall performance somehow? I am nowhere near the +-15v rails on the audio output swing so the <1volt error is not a problem further up my circuit.

Also I want to comment that I do realize selecting a different op amp here with less offset voltage may be wiser or even a jfet input type, but I am working with this op amp for now and it's performance is fine for me if this error voltage is okay to have. I'm just a hobbyist.

Thanks for any help.

[TimFox]

The 600 mV dc offset at the output will vanish after the coupling capacitor (not shown) into the next stage.
With the small ac signal level here, that offset (especially if positive) might actually reduce the distortion by loading the op amp output (through the feedback resistor network) to bias the output stage away from crossover.
However, 4.7 megohm seems high for loading a transformer — what does the microphone or transformer manufacturer recommend?

[tggzzz]

Make sure you understand the consequences of input bias current, and how the output offset depends on the gain and input offset voltage.

[Kleinstein]

The transformer should not be that sensitive to loading. Otherwise the OP-amp with relatively high current noise would not be a good choice. The amplifer is made for source impedances of less than some 10 K.
Already now and with a low impedance source there can be quite some low frequency noise as the input coupling capacitor may be more than 5 K ohm of impedance for the lower frequencies.

+1 to not expeciting a problem from the offset. The offset is normal from the bias current and rather high resistor value.

[MarkT]

With that high value of resistance you must use a FET input opamp for low output offset.  FET input bias currents are measured in picoamps, rather than hundreds of nanoamps....

[wasedadoc]

If you absolutely really want to keep that high value R1 and are worried about the offset:

1.  Insert a capacitor between the junction of R2/R3 and the -ve input of the opamp.

and

2. Add a resistor of equal value to R1 between opamp's output and -ve input.

[darklife]

The 600 mV dc offset at the output will vanish after the coupling capacitor (not shown) into the next stage.
With the small ac signal level here, that offset (especially if positive) might actually reduce the distortion by loading the op amp output (through the feedback resistor network) to bias the output stage away from crossover.
However, 4.7 megohm seems high for loading a transformer — what does the microphone or transformer manufacturer recommend?

It was originally an 1:20 transformer used in a vintage Gates Gatesway audio console from a radio station so there's not a lot of data about it but given that it ran the high impedance side directly into the grid of a triode with no damping resistor it really does like a hi-z to output like fet or bare minimal what I am using with a bipolar op amp input with minimal resistance so not to affect frequency characteristics. Even with a 1mohm load it's enough to notice a slight decrease in output and some loading characteristics oddly but it was designed with tube in mind so.

I figured the offset would vanish just fine like I stated with the output coupling capacitor as it does in circuit. Interesting that it may actually reduce distortion wow. That's good news lol and the loading of output makes sense. I was thinking the opposite with the residual voltage on the output somehow causing the op amp to increase distortion but that is an area I don't know much about so glad to read your reply suggesting accidental loading can help.

[darklife]

With that high value of resistance you must use a FET input opamp for low output offset.  FET input bias currents are measured in picoamps, rather than hundreds of nanoamps....

Yeah I really want to use a fet op amp for this but I can't seem to find a suitable one. I believe current noise is the biggest issue since voltage noise doesn't seem to dominate. I was originally using OPA827 until I accidentally popped the bugger from a wire clipping on my bench I forgot about brushing up under the board in test (ouch) and that had a nice low noise floor, then tried the OPA2134 but it had a lot of white noise with the transformer input so it seems voltage noise being low isn't the cure there. It's a stereo mic pre and on a low budget so not like I can run out and buy a bunch more 827's as they aren't cheap when buying only a few.
The old tried and true JRC2068 (NJM2068) for some reason seems to be the perfect fit if it weren't for the offset voltage showing up so big on the output but other replies imply it's not something to worry over so far thankfully.

The transformer I am using really wants to see something over a megaohm on its output otherwise it starts having odd ringing characteristics and frequency response. They didn't make them in the 60s like they do now evidently when tubes were king.

[darklife]

If you absolutely really want to keep that high value R1 and are worried about the offset:

1.  Insert a capacitor between the junction of R2/R3 and the -ve input of the opamp.

and

2. Add a resistor of equal value to R1 between opamp's output and -ve input.

Well wasn't sure *if* I should be worried about the offset
It's not an issue as far as the circuit is concerned since I have it capacitively coupled at the output, my concern was if it would possibly
induce some distortion or other odd characteristics.
Thanks for this idea though, I will breadboard a copy of the circuit later and see how your suggestions work out in circuit.
At the end of the day my preamp op amp is in a socket so a permanent solution like you suggest would be a final modification if I end
up sticking with this chip for good and soldering it direct.

[Kleinstein]

A not so expensive but reasonable low noise JFET based OP amp is the OPA1642 (dual) or OPA1641 (single). If not too worried about the low frequencies the OPA1677/1678 could also do.

A 1:20 transformer changes the impedance quite a lot (by a factor of 20² = 400).

[darklife]

A not so expensive but reasonable low noise JFET based OP amp is the OPA1642 (dual) or OPA1641 (single). If not too worried about the low frequencies the OPA1677/1678 could also do.

A 1:20 transformer changes the impedance quite a lot (by a factor of 20² = 400).

Thanks, I'll look into it and see if I can find some that aren't fakes. Being a hobbyist I've been burned from the usual suspects. I really need to investigate if it's current or voltage noise that dominates and it's kind of a tossup because I have the input toggled for either Hi-Z instrument input or mic in through transformer (also Hi-Z but more finicky to loading), so it benefits to have both lowest noise. Also why I am kind of stuck with the capacitor input, otherwise if it was just transformer I would tie it direct to (+) of the op amp because toggling between gives a large voltage bounce at the output otherwise.

[wasedadoc]

If you absolutely really want to keep that high value R1 and are worried about the offset:

1.  Insert a capacitor between the junction of R2/R3 and the -ve input of the opamp.

and

2. Add a resistor of equal value to R1 between opamp's output and -ve input.

Well wasn't sure *if* I should be worried about the offset
It's not an issue as far as the circuit is concerned since I have it capacitively coupled at the output, my concern was if it would possibly
induce some distortion or other odd characteristics.
Thanks for this idea though, I will breadboard a copy of the circuit later and see how your suggestions work out in circuit.
At the end of the day my preamp op amp is in a socket so a permanent solution like you suggest would be a final modification if I end
up sticking with this chip for good and soldering it direct.

Even simpler is a capacitor between R2 and ground.  Offset gets multiplied by 1, not by the AC signal gain.  But I concur with others that the 600mV is probably not something to worry about.

[darklife]

If you absolutely really want to keep that high value R1 and are worried about the offset:

1.  Insert a capacitor between the junction of R2/R3 and the -ve input of the opamp.

and

2. Add a resistor of equal value to R1 between opamp's output and -ve input.

Well wasn't sure *if* I should be worried about the offset
It's not an issue as far as the circuit is concerned since I have it capacitively coupled at the output, my concern was if it would possibly
induce some distortion or other odd characteristics.
Thanks for this idea though, I will breadboard a copy of the circuit later and see how your suggestions work out in circuit.
At the end of the day my preamp op amp is in a socket so a permanent solution like you suggest would be a final modification if I end
up sticking with this chip for good and soldering it direct.

Even simpler is a capacitor between R2 and ground.  Offset gets multiplied by 1, not by the AC signal gain.  But I concur with others that the 600mV is probably not something to worry about.

Oh neat. I did forget to mention I did put a 220uf cap there and kind of just left it in. I had a hunch that has some play with offset but wasn't even the original reason I added it, it was more for high pass filtering. I didn't even mention it because I figured it had no effect on offset.