Problem with measuring voltage in a differential op amp design

Hello ,

I have a problem with using a Diff OpAMP (using LM324) in unity gain. I need to measure voltage from 0~2.5V from an electrochemical cell.
I used a mV source to prelim check the equipment. I was able to read and display the voltage correctly.
But when I connect the electrolytic cell , the voltage is displayed as maximum.
I tried disconnecting the R76 ( 1M) pullup but the problem still remains. Pad8 and Pad9 are shorted.I cannot pull S11-1 directly to ground. This because the ground path has a current sense resistor connected in another part of the circuit.
All unmarked resistors in the circuit are 4k7.
Any ideas would be appreciated.

op-amp problems

You put only a partial circuit diagram. You have to add an equivalent circuit of the measured element.
Are you aware that the inputs of the LM124 have bias current and need to connect somehow to ground with not too high resistance?

lm324 input resistance

Hi,
Disconnect all connections to IC18B. Ground input of R72. Remove R75. Connect pin 1 of S11 to ground through 10k resistor. It should work for you. You can take the output from IC18A only.

Regards,
Laktronics

guarded shield opamp

Hi Strabush,
Thanks for replying.
The input is an electrolytic cell (Voltage source) with a high o/p impedence. I cannot estimate the impedence , but I can measure the voltage with a multimeter.
Could you tell me which pins of the OpAMP need to be grounded and what would be resistance in the path.
In the present setup , only Pin5 of the Opamp is directly connected to ground.
I am using LM324 quad OpAmp and not LM124.

Added after 9 minutes:

Hi Bharath,

Thanks for the idea. I will try it out. I had wire wrapped this circuit and tried it.
It worked fine,but now does not work after making the final PCB.I do not have the wire wrapped board to cross check now.

opamp not working

There's a lack of clarity with your circuit in various points, to my opinion.

- I don't see if and why a differential amplifier is actually required. It may be the case, if your sensor has another current path to the circuit or any external part connected to it, e. g. by a liquid it's immersed in.

- I wonder what's the sensor impedance. Some eletrochemical cells have a several 100 MOhm or even higher impedance, requiring special FET electrometer OP. Many would like an OP with lower input current than LM324 anyway.

- What's the LM324 supply voltage?

- Why one sensor terminal is connected to +5V by 1 MOhm resistor?

operational amplifier problems solutions

Hi FvM,
Thanks for your intervention.
The cell has 3 electrodes as set up in the diagram.There is a variable voltage impressed across electrodes A and C to GND through a 1E resistor.
The voltage across B and C is to be measured and kept constant by varying the voltage across A and C ( 0 ~ 24V). The measured voltage would be upto a maximim of 2.3V.
The 1E resistor is for measuring the current flow( 0~25mA) and logging it during the test.
The supply voltage for LM324 is +/- 5V referenced to GNDA. I am unable to source any JFET quad opamps specified for +/- 5V,like TL054.
The 1 Mohm resistor is for sensing that no electrode is connected and displaying "OPEN" message if the voltage goes above 2.3V.Anyway I removed it with no improvement.
As I mentioned I had tried this circuit and it worked fine.It still works if I give an external voltage input from a mV calibrator or if I connect a potato cell ( about 0.8V ).I think since I have a new electrode now, the impedance is higher than the older one I have tested prior.

opamp problem and solution

Connect R75 to ground instead of +5V and measure voltage output of IC18A (pin1) to ground with voltmeter. What do you get with cell connected and disconnected?

Connect voltmeter across points B and C with power supply connected and disconnected. What do you get? Connect 1M resistor across B and C and repeat measurements. Having in mind typical voltmeter input resistance of 10M you should be able to deduct cell's resistance.

What are electrode materials? What is solution in the cell? Are electrodes coated with AgCl?

What are cell walls made of? Are there any other objects in the cell beside your electrodes?

FVM already pointed out that with jumper PAD8-9 most of the ciruit is useless. It just works as a buffer of SL11-2 in reference to GNDA.

In order to design any meaningfull circuit you have to know what you measure. If you need help include as much detail as possible. Picture of cell setup, whole circuit including part where you regulate voltage of the cell, measure current, power supply section...

It is very likely that your cell impedance is much higher than 1M and you need to better circuit. In that case you would have very small currents involved and you will need to pay attention to every detail.

op-amp problem

Sinisa , Thanks for your reply.
I will clarify the test setup further :

There are 3 electrodes as shown in my previous post.
Electrode A is platinum of 5mm sq.
Electrode B is a calomel electrode
(https://en.wikipedia.org/wiki/Saturated_calomel_electrode)
Electrode C is a steel plate of 10mm sq.
The electrolyte is a common salt solution 3%V/V in a glass jar,wherein all the three are immersed. There are no other objects in the cell .

I connected R75 ( 10M,1% ) to ground. I used a 10M instead of 1M to reduce the loading on the calomel electrode.
With no power given to the cell ( i.e. no voltage to Electrode 1),
At pin 1(IC18A) I get 0.09V , At pin8 (IC18C) I get 0.094.
The LCD display on my unit varies from 0.103~0.086 .( If I short the inputs my display is steady at 0.004 ~ 0.010.)

With the power connected and voltage set to 1.5V , the circuit maintains 1.484 ~ 1.531 as measured by the ADC.This variation "seems" to be more due to noise than the effect of the control action.If I touch the multimeter probe the voltage drops to 0.9V before the control action again boosts the cell voltage to maintain the 1.5V.

The drive circuit for powering the cell is discussed here :

op amp current setup

You have problem in overall design.

C electrode should be at your 0V potential, while current should be read off 0.1E resistor as negative voltage in your current measurement.
You don't need diff amp as FVM pointed out. But you need high impedance input.
You just need single buffer and here is how you should do this:

Use LMC6062 or simillar.
Use only one amp instead of your IC18B
Place short instead of R59
Remove diode from input
Move resistor R75 before R57, connect it to 0V an change value to 100M or higher
Connect capacitor to 0V instead of +5V, use film capacitor read data sheet and select one that has very low leakage, see WIMA web site.
Output is output of this buffer.

You can use second amp in LMC6062 for current signal.

Make sure you create ring guard arround your input from output of buffer, also connect cell with triaxial cable that center conductor is connected to B of cell and inner shield is connected to output of buffer through resistor of few hundred ohms. outer shield should connect to 0V.

Read up on guarding techniques and electrometers if you are not sure how to do this.

resistor 1e

Hi Sinisa,

I hope this is right ? I could use CA3140 as it is avlbl locally here.

I did not understand your suggestion on the negative current measurement part. Can you point me to any circuit idea for that ?
I am using a 1E resistor on the low side for current measurement.Would this low side current sensing not work in a setup of this type ?

current sensing resistor operational amplifier

I guess CA3140 will be fine. Guard should incircle arround R4 and all Electrode B connections on PCB and should be connected to inner shield of triaxial cable connecting board to cell.

You dont need resistor between electrode and 0V.

For current sense resistor:

You have two power sources, one powering cell and second symetrical ±5V for your electronics. Power source that powers cell should be connected to electrode A and resistor 1E. Only connection between two power sources should be at electrode C where you connect 0V from your ±5V power source. That way voltage read of the sense resistor would be negative and by using inverting amplifier you can get voltage readable by your ADC. This inverting amplifier should be low offset and low noise. If you don't mind 25mV error you can use connection as you did before. Maybe you could substract error caused by sense resistor in micro so you don't have to use above described method. So Ucell=Ub-Ures.

solutions for operational amplifier problems

Hi Sinisa,

Actually the resistor between electrode and ground is 0E ( a dummy resistor) just put to cheat the PCB package as it does not allow dual net names for the GND net.

But of course the 1E sense resistor is there as shown in the previous cell diagram. I am measuring the current and found that for a 1.5v setting for my cell the current is about 1mA~10mA (i.e 1mV ~ 10mV drop ). So I guess that value can be added dynamically to correct the instantaneous B cell voltage.

The grounding is a bit suspect currently.I have 2 ground planes viz. a AGND where all the analog part is connected.This includes the +/-5 grounds for the opamps and the cell power supply ground. There is a DGND ground with all the micro and digital part (keypad,LCD etc etc).Both are connected ay one point under the ADC.
For the cell,I pull out a single wire from upstream of the sense resistor to electrode C.
I am getting a variation of +/- 30mV, I still need to check if the variation is due to the control action or noise in the setup.
The allowed is +/- 50mV so I guess I am doing ok, but one always likes to tidy up as much as possible.

op amp problems

Just to add, I had typo, shield should encircle B elecreode not C...

Diff OpAMP problem

The design of my Masters project operational amplifier is uploaded here. I have a gain of 71 db, PM of 45, and bandwidth of 700Mhz.

https://jennisjose.webs.com/operationalamplifiers.htm