Pressure Sensor(Bridge) Signal Conditioner With Adjustable Offset

Thread Starter

witssq

Joined Mar 29, 2009
48
I would like to make precision amplifier circuit for differnetial pressure sensor. I make a circuit roughly. Initially the adjustable offset is made like
this: First, I connect scope at AD8221(U1) output, then adjust VR4(trimmer) to shift U1 output to zero(0 V). then adjust gain to needed range. input range is -10V ~ +10V, differential input range is -500mv ~ 500mv.
But, the attached circuit does not operate acurately. When I change input value. U1/REF pin value is changing and output value is also inaccurate.

Good members would like to be joined for help to complete this project.
Thanks.

SunSung Hwang
 

Attachments

t_n_k

Joined Mar 6, 2009
5,455
I'm guessing here.....

I would imagine the OP1177 output is meant to be setting Vref for the AD8221. Something looks strange about the Vref control (at least in the circuit as drawn). I would imagine the OP1177 is buffering the supply center tap (off R11 wiper) to fix Vref as the mid point of the +10V/-10V rails of the DC supply. I would assume that this would normally be done by running the OP1177 as a unity gain non-inverting buffer with the output running direct to the AD8221 reference input pin 6 [Vref]. I'm surprised this doesn't appear to be the case and there is a curious feedback connection from the AD8221 differential amp output pin 7 [DC out] via 10k resistor R1 to the OP1177 -ve input. Capacitor C32 filters this feedback in conjunction with R1. Shouldn't the drive end of R1 be connected to the Vref pin (or simply the OP1177 output) rather than the DC out line? Also what's the purpose of C22? Is it filtering the center tap pot wiper DC signal? Perhaps I wouldn't have it there at all.

Given the actual configuration is what you have shown in the circuit, then I'm not surprised you are having problems with Vref stability and the diff amp gain stability.

Good luck.
 

Thread Starter

witssq

Joined Mar 29, 2009
48
Many thanks for comments

The revised schematic is tested and attached.

My calibration procedure is as follows.

1. whatever differential input is, First, I adjust R11 to make U1/Vout to 0V
(this means current differential input is offset to output 0V)
2. I make differential input as maximum(500mV) or minimum(-500mV)
3. Adjust VR4(Gain) to make output voltage to max or min (+10/-10)
respectively

On current test, the above 1 procedure is taken, then I changed
input value, then read scope

differential input current output required readings
-0.15V -0.417V 0.480V
-0.5V 0V 0 V
0.5V 0.480V 0.480V
0.25V 1.440V 0.960V
0.35V 1.940V 1.440V

It seems that center of gain(offset voltage) is not applied.
The current problem is that Voltage output is not linear.

Thanks,

SunSung Hwang
 

Attachments

t06afre

Joined May 11, 2009
5,934
Just something worth mention . If your pressure sensor is relative to atmospheric air pressure. Zero point adjustment is quite worthless because the atmospheric pressure is changing constant. I always use software for this type of calibration, both in pressure and load cells. It makes the hardware more robust. Less pots etc.
 

t_n_k

Joined Mar 6, 2009
5,455
Can you supply circuit details for the input side signal source referencing - is there a return path to ground for input bias currents?
 

Thread Starter

witssq

Joined Mar 29, 2009
48
Thanks.

Attn: t06adre

Then air pressure sensor used is applied to differential pressure. As to situation, the different sensor with different range could be used. Or mechanical part to be assembled with sensor make some variations at
best voltage range. Therefore I need somewhat flexible adjustment.

Attn: t n k

Then sensor spec is attached. 1-100 psi at specification sheet would be used.

Currently I use voltage generator for input voltage IN-(GND), IN+(GENERATED VOLTAGE). And C16, C9, C8 is currently not applied.

I also tested with (IN+ - 1MOhm(R23) - GND, IN- - 1MOhm(R22) -GND). but result is same.
I don't know about bias return path. Please let me know how I could make Bias return path.

And I would like to know whether it is correct that AD8221 REF is changing proportionally when changing AD8221 differential input voltage after I offset AD8221 REF using R11. Without AD8221 REF voltage change, input voltage only change can produce REF voltage change. Is this right?
 

Attachments

t_n_k

Joined Mar 6, 2009
5,455
I don't know about bias return path. Please let me know how I could make Bias return path.
Your latest schematic appears to allow for this matter - either in the input differential pressure sensor having a ground point or in the 1M resistors R22 & R23 tied to ground.

And I would like to know whether it is correct that AD8221 REF is changing proportionally when changing AD8221 differential input voltage after I offset AD8221 REF using R11. Without AD8221 REF voltage change, input voltage only change can produce REF voltage change. Is this right?
Hopefully I am understanding what you are saying but VREF should be "rock steady" once you set it - irrespective of signal conditions at the input terminals of the AD8221 [NB inputs should be within the common mode range]. So after initially nulling Vout using R11, Vout should just be Av*(±ΔVdiff), where Av=20. The drive for VREF comes from the OP1177 output so this should be solid unless something strange is causing the positive input voltage at pin3 to shift - which is unlikely unless either of the the power supply rails are changing. I suggest you need to do a complete static test in which you note the DC values at all key points for a range of differential input values. Ensure everything makes sense for each observation. If you really are using an oscilloscope (?) to make observations I suggest you would be better served by using a good DVM of sufficient resolution and accuracy.
 

t_n_k

Joined Mar 6, 2009
5,455
Perhaps another thing you might like to consider as a straight check of the diff amp part is to temporarily remove the VREF drive from the OP1177 and tie VREF direct to GND. I'm assuming GND is the +/- 10V supply common point potential. With VREF tied to a known solid GND, measure the amplifier performance. The gain doesn't have to be accurately set for the purposes of the test - you should be able to observe commensurate changes in the amplifier output for known changes in the differential input. The values of Vout should all have the same constant offset you observe with the differential input set to zero.
 

Thread Starter

witssq

Joined Mar 29, 2009
48
Many thanks for your help.

OP1177 Vout is now stable(solid) with 1-2mV variations. The problem is that IC pins is tied wrong.

Then I tested calibration procedures as follows.
1-1. I set gain against 0v differential input(Gain Reference Point) - OK
1-2. Offsetted to the desired voltage - OK
1-3. Gained any value as I want against 0v differential input(Gain Reference Point) - OK

The above is good, but the another calibration procedure is needed.

2-1. First input value with any level is taken.
2-2. Make corresponding output value(AD8221 Vout) to 0V using offset(R11)
2-3. Maximum input value with any level is taken.
2-4. Scale corresponding output value(AD8221 Vout) to MAXIMUM for
example 10V using AD8221 Gain Resistor(VR4)

Major difference of two procedure is the Reference Point(V) of Gain, the first is diffential input 0V or the second is corresponding input value on the 2-2 procedure.

Because I have not good experiences, It is difficult to decide and find solution.

Please suggest how I can solve this problem. The major consideration is precision and accurary than cost. And I would like pros and cons about my expectation. If you have any other idea. Welcome!

I think that main issue is how I can intercept AD8221 between Differential Voltage and Gained Voltage. AD8221 Ref Pin seems to be at after Gain Applied. Is this right.

My expected solutions are
1. 2 Step Conditioning
1-1. First stage: Fixed Gain Or Unity Gain AD8221 And Offset
(Current Schematic)
I think that Fixed Gain is better than Unity Gain because of very
small input value.
1-2. Second Stage : Desired Gain Setting, More AD8221 or OP1177

2. Single Step
To find IA IC to be intercepted Before Gain stage.

Please kindly advise me.
 

t_n_k

Joined Mar 6, 2009
5,455
Hello witssq,

Glad to hear one of your problems has been solved.

In all of your posts so far it's not clear (to me anyway) what is your goal in the overall design.

I understand what will happen if you null the amplifier output with some initial signal offset (second option).

For example - suppose the initial differential input signal is +50mV and the amplifier gain is 16. To null this condition would require VREF=-0.8V. So everything in the output is offset by this value as far as absolute pressure values are concerned. But as far as relative pressure changes with respect to this initial nulled value are concerned, the output changes should be linear for +ve or -ve deviations away from the initial value. As to6afre points out this initial nulled condition is likely to change and you still end up with a problem unless you redo the null.

Rather than getting involved with the various details of the null / calibration process, are you able to give a clear statement of your design goal(s)? Include some numbers for input signals vs output signals which might clarify what will happen with your expected typical sensor outputs and your envisaged physical setup.

Keep in mind also for the AD8221, that the output cannot swing over the full +ve to -ve rail supply range. Check the data sheet for indicative output voltage swing values for typical output loading.
 

t06afre

Joined May 11, 2009
5,934
Could you please give more details about your application. Like the the type of the pressure sensor, but also what will happened to your signal the it leave your board. Will it be feed to some A/D unit? And is this front end unit programable say a PC with a program made by you?
One thing I do not like with your design is that if you change the gain, you will also move the offset point. Such a system can be a pain to calibrate. You should use a fixed gain on the AD8221, and move the tuning of the final gain to a opamp stage. And perhaps also ad some filter stages.
For improved accuracy you should use a voltage reference as source for the offset regulator. And perhaps also a voltage reference for excitation of the pressure transducer. As it now both the offset adjustment and the pressure transducer are dependent on the quality of the +/- 10 volt source. How far from the power supply is your application meant to be?
 
Top