Where is the RFI detected?

It's probably being rectified by the base-emitter junction of your PN2222A. ;-)

Kidding aside, as someone who has designed input signal conditioning paths for instruments using similar op amps, I'd say you should expect the op amp input stage to rectify pretty well. You have a setup able to hose RF directly into that (-) input, no? I'd try some RF suppression right at the op amp. Beware: capacitance to ground from that (-) node can lead to instability in the feedback loop. Can you tolerate some added resistance in there? If you want to kill RF above the gain bandwidth product of the op amp, you can add some fairly carefully designed LC filtering that may not mess up the feedback loop too much.

As long as you're biased in a positive way, you should be OK...

Cheers, Tom

its the doides or the opamp...

place 100 pF caps DIRECTLY across the diodes....

if that doesn't work, then

place (2) 1kOhm in series with the opamp input terminals, one for the

• and one for the - right at the op amp itself...

Mark

It's worth noting that a bipolar long-tail pair input stage start rectifying when it sees more than about 20mV of RF, while FET long-tailed inputs remain linear with up to a volt of so of RF hash. There are FET-input op amps around with similar sorts of voltage noise figures to the OP-27, albeit at rather higher prices - the ex-Burr-Brown parts now sold by Texas Instruments include a few.

Almost surely nonlinearity in the op-amp inputs. Keep it out of there. You have a couple of orders of magnitude in frequency and a fairly healthy tranducer signal to play with.

Best regards, Spehro Pefhany

Two clamping diodes anti-parallel to ground? Aha! These will rectify, modulate, delinearize, mutilate just about anything. And if they aren't exactly the same they'll also create a nice AM demodulator for RFI. Even if it was a BAV99 dual diode these often aren't 100.000% identical. Consider running them against the rails or at least some lower (and well filtered) voltage in case the rails aren't feasible.

Such a diode combination can also cause the measured signal value to be off because diodes don't just start conducting at 600mV. They start much earlier and depending on your source impedance that could matter.

Maintain a low impedance shield-to-panel connection at the connector that penetrates the enclosure. Hoping it's a metal enclosure.

A few ferrite beads right before the amp input can sometimes provide a miracle cure. Use #43 ferrite, not #77 or any of that higher AL-value material.

Most likely it's the diodes. But it can be more than that and you'll have to "peel the onion", finding the worst offender, then #2 and so on.

I feel your pain. A few decades ago I bought an audio preamp kit from SWTPC. And as a bonus, it picked up about a dozen AM and FM stattions, all at the same time.

Putting RF RC filters at the phono input sockets didnt help much-- apparently the RF still snuck in between there and the active devices.

. The only thing that workerreally well was putting a resistor in series with the base (gate) leads. A few K ohms, working with the intrinsic and Miller capacitance did a nice rolling off of the unwanted high frequencies. And doing it right at the input lead ensured there wouldnt be any stray pickup.

If you need more and steeper rolloff, the ferrite beads are a good idea. Except for AM band signals, there you'll need some ~1mH RF chokes.

i can't imagine 100Mhz--500Mhz ever being detected and past via a resistive transducer how ever, i can see how the detector it self would react to the signal and use it's own structural capacitance to filter the high frequency and thus produce a low frequency that is a variation of amplitude of the original frequency.

also, taking in light of things, you have miller effects in components (op-amps, transistor amps etc.) that in it self, can produce the same effect and only show you the end results of the variant levels of amplitude. fairly a common practice for cheap detection and also an hindrance in other cases.

just my 2 cents worth.

Bill, Thanks a meg! I replaced the OP-27 with an old LF-156 I had in the lab. Detected RF amplitude is better than 20X reduced. Added to improvement in shielding effectiveness, that will probably resolve the problem, and without a PWB change.

LF156 isn't the final solution, but it demonstrates that the rectification was in the OA front end, and that an FET front end will be much better.

Thanks to everybody who weighed in.

Regards PN2222A NPN (Is = 14.34f Xti = 3 Eg = 1.11 Vaf = 74.03 Bf = 255.9 Ne = 1.307 Ise = 14.34 Ikf = .2847 Xtb = 1.5 Br = 6.092 Isc = 0 Ikr = 0 Rc = 1 Cjc = 7.306p Mjc = .3416 Vjc = .75 Fc = .5 Cje = 22.01p Mje = .377 Vje = .75 Tr = 46.91n Tf = 411.1p Itf = .6 Vtf = 1.7 Xtf = 3 Rb = 10)