AnsweredAssumed Answered

AD8237 Gain/Bandwidth problem

Question asked by m3atwad on Sep 27, 2014
Latest reply on Sep 30, 2014 by Emman.A



I'm trying to use the AD8237 amp in a design focused on taking small differential sensor output and amplifying it.  My first stage is to take the sensor output into the 8237, apply some gain (no more than 10) and pass it on to a low pass filter and then a non-inverting gain amp.  These are all ADI parts and everything seems to be working well except for the 8237. 


I'm only working with "up to" 20kHz input signals/pulses.  The inputs are cap coupled as shown in the picture. I have a development board (so everything is on copper and cleaned) that allows me to pull a jumper (J11) and disconnect the 8237 circuit from everything so I can just analyze that (which is what I'm other amps circuits involved).  I am not using the Vref pin and I only go up to a gain of 10.   In the feedback path is a AD5272-20k and the FB pin node is also connected to a 2k to GND yielding a max of 10 for my gain here.  The input caps are 0.01uF with a 100k pull down resistor on the +/- inputs of the 8237.  It is powered with 3.3VDC. The input signal is only 50mV.


What I"m seeing.

I power this up and on power up the AD5272-20k pot sets itself to 10k so without doing anything really this powers up with a gain of 5.  Now at low frequencies 1k-10k I seem to get the correct gain.  After 10k it is attenuated very badly.  I need to pass up to 20kHz (the LPF in the next stage is designed for that) but when I push the 8237 above 10kHz my magnitude (of my output) drops off sharply.  Now, if I jumper the BW pin from low BW to high BW its fine.  The issue here is, I thought per the data sheet it said that you could get a gain of 10 at 20kHz.  I was banking on this with this design, but it doesn't seem to perform that way in the lab. 


I assume the misunderstanding is on my end so does anyone see what I'm missing?  My problem with the BW pin is that I don't have the means to toggle it or control it.  I was hoping to be able to tie it low and just not exceed the A=10 @ 20kHz limit which is fine on my end.  I just don't seem to be getting that performance



Pic of the schematics - specifically just the circuit I'm working with.

Circuit visual aid AD8237 GBP problem.jpg

Pic of the input (yellow - I know my signal generator probably isn't the best) and output of the 8237 circuit (blue)with a 10kHz frequency.



Pic of the same inputs and outputs at 20kHz..nothing else changed


Here is a screen of the part of the data sheet I'm looking at