What's this op-amp for? (AFG Front End)

[timb]

Hey guys, so I'm reverse engineering the front end of a Tektronix AFG2021 and I've come across two things I don't quite understand. This is unit is a 250MS 10MHz Arbitrary Function Generator.

Here's the schematic:



Sorry for the quality of the schematic. The +/- signs on all the op-amps seem to have disappeared upon exporting the image from LT-Spice; for all the amps the inverting input is on top and non-inverting on the bottom.

All resistor values shown are correct, however capacitor and inductor values are unknown. V6 is a voltage that comes from another DAC to set the offset voltage of the output (+/-5V total offset).

Also, I haven't shown this on the schematic, but I'll go over it here briefly anyway: The signal comes from a 14-bit differential IDAC, goes into an LMH6702 setup as a differential I-to-V converter, goes through some LC filtering and is then passed through 5 relays, setup as selectable Pi-Attenuator pads, then it arrives at R3 on the schematic.

Anyway, I'm curious as to the purpose U2 (the OPA134) serves. It's a relatively slow amplifier (designed primarily for audio applications) which only has a capacitor for feedback and drives the non-inverting input of a much faster (1GHz GBP) CFA.

I get it's a DC servo, I've just never seen a setup exactly like that, so I'm wondering if someone can explain it in a bit more detail. (I.e. the cap between two 1ohm resistors on the inverting input of the OPA134; it seems counter intuitive to hang a cap directly off a transmission line like that.)

Bonus Question: There's, what I assume to be, a bandstop filter on the output (L1, C5) right at the input of the "on/off" relay that's just before the output BNC. Why? What purpose could that serve?

Thanks guys!

[Ink]

you can check out Linear's application note AN-21 that written by Jim Williams.He have explained exactly the same circuit.

[timb]

you can check out Linear's application note AN-21 that written by Jim Williams.He have explained exactly the same circuit.

Right, I understand it's a composite amplifier, I guess I should have been more specific:

The part that confuses me is the input network to the OPA134. In all the DC restoration composite amplifiers I've seen, they generally pick the signal off using a resistor value that make senses when compared to Rfb of the main, high speed amplifier (as shown on the first page of the app note you posted).

In this case they've got two 1Ohm resistors in series, with a cap between them. What purpose does that serve, exactly? Also, at high frequencies most ceramic capacitors have pretty low ESR, right? So, with such a low value input resistor, you're basically hanging that cap right off the input of the high speed amp, which I'd imagine would degrade the signal. (Though, it is at the summing junction, so there's not really a very big voltage swing there I suppose...)

[bktemp]

I doubt both resistors are 1 ohm. Are you sure it is not 1k or 1M?
Using 1ohm it will both attenuate high frequencies going to the main amp and make the OP134 instable.
But using much higher values makes totally sense: The first RC attenuates high frequencies the OP134 can not handle, the second R with the integration C around OPA134 serves as the DC servo integrator.

[timb]

I doubt both resistors are 1 ohm. Are you sure it is not 1k or 1M?
Using 1ohm it will both attenuate high frequencies going to the main amp and make the OP134 instable.
But using much higher values makes totally sense: The first RC attenuates high frequencies the OP134 can not handle, the second R with the integration C around OPA134 serves as the DC servo integrator.

They're labeled 010, which according to my Secret Decoder Ring is 1Ohm. Being a much higher value the circuit would make more sense, yes.



Edit: Son of a bitch, looking closer at that image, I think it's actually 01D, which is 100k. *Palm*

[bktemp]

Yes, I would also read it as 01D, so 100k which makes totally sense.

[timb]

This is why I hate EIA-96 codes. Either put a damn slash through the zero (which is hard to do on small parts) or don't use letters that can be read as numbers, it just causes confusion!

So yeah, the circuit makes about a million times more sense now.

Now my only remaining question is about that series inductor/capacitor on the output. I'm guessing it's to help reduce peaking or overshoot or something?

[danadak]

Whats the resonant f for the trap ?


Regards, Dana.

[timb]

Whats the resonant f for the trap ?


Regards, Dana.

That's a good question. I can't remove the components to test for their values and in-circuit testing won't work very well either... The best I can come up with is to maybe inject a signal at that point and sweep it until I find the sweet spot.