730kHz Active bandpass filter

[frankl]

Nasically i currently have a design working on MultiSim (if anyone knows it) where i have a very nice 12kHz bandwidth around a centre frequency of 730kHz. The filter itself is a 4th order Biquad configuration with two stages, each with 3 op-amps in a feed-back configuration. Even the parts in simulation are the parts i'm also using in real life, Burr Brown/TI OPA658P, a 900MHz current-feedback op-amp (900 MHz is to accomodate future uses). However, when i wire them up on a test board, I get oscillations on the output. I realise that there may be stability issues with the feed-back, however only negative feedback is used. No oscillations are noticed in simulation and when i hook up a simple inverting amp, it works ok. Anyone with any tips as personally i'm at my wits end, cant find anything i can test any further. Any help would be gratefully appreciated.
Thanks,
Frank

 

[melone]

In your simulations, did you take into consideration the impedances (R, L & C) of your connections? Depending on how you hook up your circuit, you might have some signifigant impedances. For example, if you wire up 2 of your op-amps on one protoboard and the 3rd on another and connect the circuit via a ribbon cable (BIG L!!!), then you need to modify your simulation.

Just a thought!

 

[frankl]

Thanks for that but no, i havnt taken them into account. I also dont really know how i would start to do that, any tips would be good. Multisim is pretty advanced and does take leakages into account but i'm not sure to what extent. The circuit in simulation is pretty tolerant of value changes, i also have the whole circuit (all 6 op-amps) on the one breadboard. I realise that breadboards are bad grounding wise, however i've been twisting wire pairs as often as possible and have pretty decent power supply decoupling.

 

[cbarn24050]

Hi, I dont know the IC's you are using but they seem pretty fast for you application.High speed amps are notorious for cscillation and a bread board is the last place you should be using them.You must have them on a ground plane to have any hope of stability. Can you use an op amp that has a speed nearer your needs? if so try that first.

 

[Lewish]

First question, why in heaven's name are you using a 900 MHZ part for a 730 KHz application?
You are attempting to use a current-feedback opamp in a voltage-feedback situation. Your simulator doesn't know the difference.
You probably have the part on a perf-board or some type of dead-bug arrangement. If so, you can bet your sweet bippy that it is going to oscillate. You have all kinds of parasitic positive feedback paths.
Try selecting a voltage-feedback opamp with a gain bandwidth of no more than 10 MHz and using it. I think you will be much more satisfied with the results.
I hope this helps.

Lewis