EEVblog Electronics Community Forum

Electronics => Projects, Designs, and Technical Stuff => Topic started by: SArepairman on September 03, 2014, 02:03:54 am

Title: op amp current source feedback filter?
Post by: SArepairman on September 03, 2014, 02:03:54 am
A dual op amp current source (buffered high impedance howland current pump) is driving a wire with a current going to a sensor. This wire runs in parallel with audio frequency (of 100mAish rms power) lines for about a meter.

Does it make sense to put a sallen key filter in the feedback path after the howland current pump feedback buffer?

(http://s17.postimg.org/pyt31w4cf/32432342.png)

I saw this in an app note before. Will this sort of current source turn coupled noise into current?
Title: Re: op amp current source feedback filter?
Post by: T3sl4co1l on September 03, 2014, 02:17:40 am
"Audio frequency" = < 20kHz BW, or not many times more than that.

1m of most any kind of cable isn't going to be more than ~300pF, or more than 26kohms reactance (or not many times below that).

So, if you current source is much better than 26kohms, you're wasting your time -- or alternately, if the sensor can't tolerate under 26kohms of compliance or crosstalk.  (Which means, for example, a Howland CCS built with 1M 5% resistors will still have a worst case good enough for present purposes, if I'm estimating that right.)

What is this, current bias for a voltage on the sensor?

Tim
Title: Re: op amp current source feedback filter?
Post by: f5r5e5d on September 03, 2014, 03:25:14 pm
Howland can be a poor choice for a regulated unipolar current source - a fet or a bjt (with base current compensation if needed) can have higher output Z to higher frequency, only one op amp needed for adding accuracy to the transistor ccs

and you shouldn't in general "add a filter" to a feedback connection - feedback loop performance has to be designed for