Author Topic: 10MHz bw, +-15V >CMRR <offset unity gain INA?  (Read 1676 times)

0 Members and 1 Guest are viewing this topic.

Offline max_torqueTopic starter

  • Super Contributor
  • ***
  • Posts: 1272
  • Country: gb
    • bitdynamics
10MHz bw, +-15V >CMRR <offset unity gain INA?
« on: June 27, 2017, 05:03:54 pm »
I need a differential, unity gain instrumentation amplifier that accepts +-15 inputs (or just under when using a +-15v vcc) with high Common Mode rejection, and low(ish, say 10's of uV) output offset.

Do i

A) roll my own from descrete opamps and trimmable descretes (to get high CMRR and low offsets)

B) Use an existing off the shelf INA solution  ( i currently use an INA128, but it's not fast enough for the latest application and has horrible non symmetrical slew rates and a nasty gain hump at high frequency)

The input to the amplifier is a 100:1   (5Meg/50k) capacitively corrected resistive divider, so i need a relatively high input impedance and low input capacitance.

Any suggestions for either solution?
 

Online Marco

  • Super Contributor
  • ***
  • Posts: 6694
  • Country: nl
Re: 10MHz bw, +-15V >CMRR <offset unity gain INA?
« Reply #1 on: June 27, 2017, 09:08:46 pm »
That voltage range makes things hard.

Is the common mode voltage low frequency? If so you could combine an INA188 with an AC coupled AD8129/8130 as a two path inamp. With the INA188 setting the DC offset and the AD8129/8130 providing the high frequency path.
 

Offline jbb

  • Super Contributor
  • ***
  • Posts: 1128
  • Country: nz
Re: 10MHz bw, +-15V >CMRR <offset unity gain INA?
« Reply #2 on: June 27, 2017, 10:24:26 pm »
This sounds similar to an oscilloscope active differential probe.  You could perhaps look at what people have done for them.

The input dividers could cause some CMRR issues - are you buying nice precision-matched parts or doing it yourself?  If doing it yourself, you'll need careful input trimming for CMRR anyway, so deploying your own might not be so bad. (I think you'll need DC trim on at least one leg and AC trim on both legs, but I'm no expert...)

Some vendors (e.g. Linear Tech LT5400) also provide precision resistor blocks specifically for making the differential amplifier stage of an instrumentation amp. They are a bit expensive (laser? trimmed at the factory) but would provide higher and stable CMRRs. You don't want a sensor solution which needs to be recalibrated morning noon and night.
 

Offline danadak

  • Super Contributor
  • ***
  • Posts: 1875
  • Country: us
  • Reactor Operator SSN-583, Retired EE
Re: 10MHz bw, +-15V >CMRR <offset unity gain INA?
« Reply #3 on: June 28, 2017, 12:04:40 am »
Google "diy differential probe", lots of activity, even layouts.

Do same on eevblog, several designs here.


Regards, Dana.
Love Cypress PSOC, ATTiny, Bit Slice, OpAmps, Oscilloscopes, and Analog Gurus like Pease, Miller, Widlar, Dobkin, obsessed with being an engineer
 

Offline max_torqueTopic starter

  • Super Contributor
  • ***
  • Posts: 1272
  • Country: gb
    • bitdynamics
Re: 10MHz bw, +-15V >CMRR <offset unity gain INA?
« Reply #4 on: June 28, 2017, 09:58:55 am »
Yeah, i've looked at other solution, and taken apart the popular 100:1 HV diff probe re-branded and sold by various peeps including dave.

But i wanted to also approach from first principals, and ask what others have done or would do.  I have the 100:1 compensated attenutator from the existing design, which i've tuned up, and now has roughly a 20Mhz bandwidth (-3dB), so was wondering if there was an off-the-self INA that i can just slap on after it, or if it's worth the trouble to role my own.  Component Cost is (relatively) un-important.

The Common mode signal is typically a square wave at around 20 Khz fundamental, but of course, having fast edges (think 500V in <1uS) it contains a significant high frequency energy. It is this energy that is currently causing issues with the INA128 i use at the moment, driving it into saturation / overshoot and then the long (relatively) recovery period during which time it's output cannot reflect the true differential input signal.
 

Online Marco

  • Super Contributor
  • ***
  • Posts: 6694
  • Country: nl
Re: 10MHz bw, +-15V >CMRR <offset unity gain INA?
« Reply #5 on: June 28, 2017, 06:31:45 pm »
It's silly to expect 150 dB high frequency CMRR after two "matched" 100x dividers, regardless of the amplifier ... what's the point in this exercise? It will not work.

Isolation amplifiers exist for good reasons, most of them apply to your measurement.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf