AD8015/MAX3645 Oscillating

Have you tried taking out C14 and C15 and checked if it still oscillates? If it does, is the capacitance of the photodiode higher than usual? That might require a wee feedback cap for compensation. I didn't read the whole data sheet but was a bit surprised that this isn't mentioned in the schematic.

Don't. The first TIA I ever built was also singing like a soprano until I figured out that feedback cap issue. Which, of course, hadn't been mentioned in its datasheet either.

Thanks for the reply - and words of encouragement!

I've tried lifting C14 and C15 and i've also tried shorting the input to ground and adding a 2.2pF feedback cap. But i'm still getting what looks like noise but could well be oscillation on the output of the TIA. I strapped the feedback cap from the output+ to the input. However, i've also now noticed that the MAX3645 part is merrily oscillating away! (with C14 and C15 lifted there is no input). The output is swinging the full PECL output in a lovely sinewave! Back to the TIA however, the noise/oscillation is around 40mV and if i had to describe it i'd say it was oscillation with lots of noise on top! Does this mean my TIA is doomed even without an input or am i measuring badly? I'm trying to keep the earth clip local to the signal but it is a lead and therefore subject to rf noise. Someone mentioned a technique for measuring small signals with a 50R coax, but i'm not too sure i follow that.

Thanks again

Andrew

If both chips are oscillating you might not have a good RF-tight prototype. Hard to say w/o photos. Ground clip probes aren't any good here. Switch the scope to 50ohms input, hook up a small coax, cut the end and hook a 1k SMT resistor in series to the center conductor. All really short. Now you can press the shield directly to your ground plane and check the supplies, outputs etc., just not any hi-Z input.

Try a 10-20pF FB cap as an experiment. Should be clean like a whistle. If not you might have to work on your bypass caps (must be SMT).

Don't give up, it can be done. I just used a THS4021 in a similar design. That thing is a real hotrod.

Andrew,

Faith is what makes electronics work and airplanes fly! Als when plan "A" and plan "B" do not pan out there is always "CD&E" the key to your solution. Contact me at snipped-for-privacy@att.net to continue the discussion

Marc Popek