Sallen-Key output gain surprisingly stays at unity.

Go To Last Post
10 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi,
I am filtering out the output of
MPXV5004DP differential pressure sensor.
I played a little with 324 and currently MCP6004.
MCP is powered by a clean 5V, I have designed a few Bessel filters with Sallen-Key topology already and all work well on the same board.

The problem is surprising:
using Ti FilterPro software,
somehow the output of the filter stays as if there would be no resistors R3 and R4 (i.e. short R4, unsolder R3).

2 remarks:
Capacitors are SMD 0805, so max 1uF (no space on board). As a result I got quite large resistors.
No noise observed, however.

Design goal:
Using Ti FilterPro, click:
Low-pass, Sallen-Key, Bessel, C2 set to 1uF, C1 set to 470nF, gain set to 2, 2 poles, 1Hz cutoff frequency, hoping to bring the useful input from 1...2.5V to 0...3V (have 10bit ADC, unable to get different sensors, no space on board for 12 bit MCP3208, VREF only at 3V because there are other sensors at range 0-3V).

Problem:
Works excellent, but I was not able to achieve any gain (not to mention about removing the offset), the input is 1.085V as expected from MPXV5004, but the output stays the same. Dynamic characteristics as expected, indeed looks like 1Hz cutoff. The layout looked cool according to my introductory book on opamps, unless I misread something.

What went wrong, what should I learn apart from 400-page book?

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

the 22K resistors do not effect the DC gain!

At low frequencies (ie, below 1Hz). the caps are effectively open circuit. So, imagine them as absent. That gives you unity gain to the op-amp output node, no matter what. The only gain variation would occur in the effective voltage dividers made up with the second 22K and the 100K and 200K resistors.

So, it looks to me like the behavior you describe is exactly the expected behavior.

Should also point out that, in this application, the 22K feedback resistor has NO effect on gain.

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

If you add a resistor R between the inverting input of the Op-Amp and 0v that will give you some gain, something like (R+22K)/R. When not saturated (right supply voltages etc) the Op-Amp will keep both the inputs at the same voltage. R and the 22K set up a pot divider, so the value in the middle of this divider = ip voltage, therefore the output voltage must be that ratio larger, ie there's gain. Without a pull down res the op just tracks the ip, unity gain.

<º))))><

I am only one lab accident away from becoming a super villain.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Dren, thank you.
And how about removing the offset? Before the opamp or after? I don't want to separate ground, since the 3 remaining opamps in the same MCP6004 need 0V floor level.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You could use a summing point at the inverting input (another R with a DC voltage) to subtract the offset.

Basically when the OpAmp is set up OK, the output (whatever V that is) will do it's best to make the inverting and non-inverting inputs the same voltage.

So 'knowing' that the sum of the 22K (-ve feedback), Rx to 0v and Ry to a DC level = Vin. You can calculate the values to give the right offset and gain. What values were you trying to obtain ?

<º))))><

I am only one lab accident away from becoming a super villain.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Not sure if I can see the problem,
Filter pro response looks OK
The DC gain is +2 (6dB) as set by R3, R4
Low pass filter response is as designed

Your changes:
R4 must be connected to your a ground to have gain other than 1 (as your short would do)
If you want to offset the DC output voltage by some value, the virtual ground to which you connect R4 and the input signal ground must be low impedance, not the high value resistors you have in your circuit which just adds to R4.
If you shorted R6 to ground it will have a DC gain of 2 (ie it will amplify the offset voltage of the source).

Check your opamap has split supplies so that negative going input volatges can be amplified/filtered.

Klave

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The gain is not set by R3, R4.

R3 just goes to the Opamp non-inverting input, (assuming that it is a decent high impedance ip) then precious little current is flowing through it so no voltage drop and R3 acts a bit of wire. Vin=Vout ie unity gain. Changing it's value will do nothing.

R4 is just a load on the OpAmp output, so changing it's value will just make the OpAmp work harder squirting out current (to be wasted heating up the R).

It's only if R4 is moved to the other side of R3, then the ratio sets the gain.

<º))))><

I am only one lab accident away from becoming a super villain.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Dren said:

Quote:
It's only if R4 is moved to the other side of R3, then the ratio sets the gain.

And he is correct.

The circuit shown in the TI file shows the correct location of R3, i.e. from the inverting input to ground.
In this configuration the DC gain from the positive input to the output is 1+R4/R3 ... i.e. 2.

In your circuit you have erroniously connected R3 from the output of the OP_Amp to ground.

This just adds a load to the OP_Amp, sets the non-inverting gain to 1, and stuffs up your filter response.

Get the circuit right, and retest. With R3=R4 the DC gain should be 2.

If you need to apply an offset, tie the grounded end of R3 to the output of an OP_Amp follower and adjust that OP_Amps +ve input voltage. Note that a negative supply may be required if you need to offset below ground.

This way, the offset adjust does not alter the R3/R4 ratio as it is a low impeadance node.
Ron.

Edit: fixed typo

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
In your circuit you have erroniously connected R3 from the output of the OP_Amp to ground.

Ahh, so true. If you read too many books in the same subject the typos are spawning.

Quote:
If you need to apply an offset, tie the grounded end of R3 to the output of an OP_Amp follower and adjust that OP_Amps +ve input voltage.

BTW if there are no 'spare' vfollower on the board,
can I use for example 1K and 4K vdivider (obtaining 1V from 5V Vcc)? or maybe 1K and 2K vdivider using REF193, but this time I am afraid to worsen it's stability (even if it can output 30mA, the load regulation rules apply).

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

No need, just add the +5v into the summing point via a res :)

So the summing point @ the inverting input has three res, one the -ve feedback from the OpAmp output, one to 0v and one to +5v. Now we just need to work out the values to give the right offset and gain. The offset was 1v and gain x2 is that right ?

<º))))><

I am only one lab accident away from becoming a super villain.