Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Offset / Gain interaction

5u4

Newbie level 3
Joined
May 25, 2020
Messages
4
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
52
Hello All!

In the attached EMG amp design the output stage allows for an offset and gain adjustment( for calibration) however there is an interaction between the two adjustments which ultimately prevents the desired output gain. The circuit design includes a preamp stage, 60hz notch filter, RMS to dc converter, and the final amplifier with gain and offset adjustments.

The theory of operation is the inject a 200hz signal (at C61) at two reference levels then adjust the output to to obtain the required levels.

The calibration fails on some hardware, the failure mode is can not obtain sufficient gain.

The question, how can the offset/gain circuit be modified to reduce the interaction between the adjustments such that when the gain is adjusted the offset adjustment does not effect the gain?
 

Attachments


d123

Advanced Member level 5
Joined
Jun 7, 2015
Messages
1,958
Helped
443
Reputation
887
Reaction score
430
Trophy points
83
Activity points
20,066
Hi,

Could you make use of the +-12V dual supply and another op amp to cancel the offset and so not disturb the gain adjustment part?
 

KlausST

Super Moderator
Staff member
Joined
Apr 17, 2014
Messages
17,313
Helped
3,919
Reputation
7,836
Reaction score
3,781
Trophy points
113
Activity points
114,903
Hi,

I didn't go through the complete circuit...

For calibration:
* Usually first the input is open or short circuit (zero input), then the offset is calibrated.
* After that the gain gets calibrated.

Klaus
 

danadakk

Full Member level 2
Joined
Mar 26, 2018
Messages
146
Helped
23
Reputation
44
Reaction score
16
Trophy points
18
Activity points
790
Not sure I understand whats going on.

The offset circuit is not in the G signal path, and seems confirmed if I do a sim of a roughly
equivalent circuit.

Offset Circuit.JPG

If you examine sim R1 is being swept from 100 ohms to 10 K ohms, and you can
see slope of Vinopamp to Voutopamp is same. That is with G set to +1 by virtue
of OpAmp configed as a follower.



Regards, Dana.
 
Last edited:

5u4

Newbie level 3
Joined
May 25, 2020
Messages
4
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
52
Additional information:
This design is portable device powered with 9v battery.
I posted the proto schematic for ease of understanding the final amp circuit, in the final design the pots are replaced with digital pots.
The 60hz notch filter was proto-type, in the current circuit the filter is removed.
Vref = 4vdc
Vcc = 9vdc

The idea of the transistors; both transistors are in single package and used for temperature compensation.
 

5u4

Newbie level 3
Joined
May 25, 2020
Messages
4
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
52
Better view of amplifier circuit -

The output of U13 (rms converter) for 8uv input is 3.13vdc and for 80uv input 3.14vdc yielding output voltages at U11 of 2vdc and 3vdc

Dana, the results in your graph look correct for the circuit shown, what is needed to better understand the design flaw is a sim of the actual circuit, unfortunately I do not have the tools required.

During trouble shooting my conclusion is the offset adjustment and the gain adjustment cancel each other out to some degree the end result being the gain is almost constant.

Any thoughts ?

Calibration process:

Input 80uv 200hz signal
Adjust gain pot to output target voltage ~ 3vdc
Input 8uv 200hz signal ( at this point the output value is to high)
Adjust offset pot to output target voltage ~ 2vdc
Input 80uv (the offset adjustment lowered the output of the previous adjustment)
Adjust gain to output target voltage ~3v
Input 8uv ( the gain adjustment raised the output for the previous adjustment)

This process continues until the output values track with the input values 8uv/80uv = 2v/3v

However the process fails, the required gain can not be achieved, very close but no go.
 

Attachments


danadakk

Full Member level 2
Joined
Mar 26, 2018
Messages
146
Helped
23
Reputation
44
Reaction score
16
Trophy points
18
Activity points
790
The output of U13 (rms converter) for 8uv input is 3.13vdc and for 80uv input 3.14vdc yielding output voltages at U11 of 2vdc and 3vdc
Yes, the output of the RMS converter is essentially flat, you have a 10:1 change in input, but
just a 10 mV change in its output.

Normally for a straight line f() you set the offset first, "normally" at x = 0, then set the G for the
highest y axis value at the highest x value.

Negative Slope.png

But you have a conumdrum here, you say the output increase with input, yet your amp is inverting ? As input to amp goes from
3.13 to 3.14 you want the amp out to go from 2 to 3.......


Regards, Dana.
 
Last edited:

5u4

Newbie level 3
Joined
May 25, 2020
Messages
4
Helped
0
Reputation
0
Reaction score
0
Trophy points
1
Activity points
52
Dana,

Yes, there is an inversion in the signal path, I was referencing the input to the first stage.

You studied the final amp stage, can you see why the gain can not be increased independent of the offset? The two adjustments seem to cancel each other out, the end result being the required gain can not be achieved, it is always about 1 percent below required. There is some type of negative feed back in this circuit that is not apparent to me. The design looks straight forward however is flawed, and the flaw is beyond my comprehension.


Regards,
John
 

danadakk

Full Member level 2
Joined
Mar 26, 2018
Messages
146
Helped
23
Reputation
44
Reaction score
16
Trophy points
18
Activity points
790
What is the value of Vref and +V ?

Your circuit is a classic inverting amp, trillions used to date.


Regards, Dana.
 

danadakk

Full Member level 2
Joined
Mar 26, 2018
Messages
146
Helped
23
Reputation
44
Reaction score
16
Trophy points
18
Activity points
790
The way to do the adjustment is :

Vout = G x Vin + Voff

set Voff to 0, Set Vin to its max value, and adjust G (using 5 or 10 turn pot)
for corresponding Vout max. Vout max does NOT include the value of Voff.

Now set Vin = 0, adjust for Vout = Voffset

This is a non interacting adjustment.

The first adjustment gets the slope of the response, G, correct.

The second handles the offset nature of the response curve.


Regards, Dana.
 

albbg

Advanced Member level 4
Joined
Nov 7, 2009
Messages
1,200
Helped
406
Reputation
812
Reaction score
377
Trophy points
1,363
Location
Italy
Activity points
9,150
As far as I've correctly understood, before the U13 you have a circuitry that inverts the slope of the signal, that means increasing the input voltage, the input to U13 decreases. Furthermore this circuitry, toghether with U13 add an offset voltage. If this is correct then the signal at the iput of the opamp is:

Voa = -k*Vin+b

Where "-k" is the of the stages before U13 and "b" the added offset. If the op-amp has gain "-G" and offset "Voff" we will have

Vout = -G*(-k*Vin + b) + Voff = G*k*Vin -G*b + Voff

Since you want 2V OUT with 8u IN and 3V OUT with 80u IN linearly, then it has to be:

Vout = 1/72u*Vin + 1.9 that is:

G*k = 1/72u = 13889 and -G*b + Voff = 1.2 as you can see gain and offset are connected one each other.

Your algorithm has to be able to find these conditions. Possibly the routine you posted can do this (I've to think a little bit), but you have to be sure you can reach the gain and the offset calculated above. I mean you should evaluate the min and max for "k" and "b" in order to be sure the two adjustements of your circuits can span the whole range of possible values
 

Toggle Sidebar

Part and Inventory Search


Welcome to EDABoard.com

Sponsor

Sponsor

Design Fast


×
Top