[SOLVED] Differentiator zero-point offset

[styojm]

Hi All
I'm suffering from offset problem in my differentiator. The design is pretty standard

Where Rs=100 Ohm, C=1uF, R is a potentiometer with maximum of 20 kOhm, Cf=10nF, and the Op Amp I'm using OP27G from Analog. So when I change the potentiometer R, even if I ground the input, I still see a constant output voltage Vo that increase as R increase. If R=0 and Vo=0, then when R=20k Vo increases to about 10mV. Could anyone help me out with this? Cause I'm pretty confused about this drift and there seems nothing valuable I could find on google. Thanks.

 

[godfreyl]

The OP27 has a BJT input stage, so the input bias current is significant.
Your problem is that the bias current for the inverting input flows through R, and causes a voltage drop across it.

The easiest solution is to replace the OP27 with a FET-input opamp.

If you have to use a BJT-input opamp, you can reduce the offset a lot by adding a resistor between the non-inverting input and ground. This resistor must be the same value as the feedback resistor. For this circuit, that would mean using a dual potentiometer.

 

[keith1200rs]

The input bias current of the OP27 would not explain a 10mV offset with a 20k resistor. Have you looked at the output waveform? Depending on how this is constructed, my guess is general noise pickup. With a low input resistor and 20k feedback the circuit will amplify any noise picked up due to a less than ideal construction.

Keith

 

[LvW]

Styojm, put another resistor between the pos. opamp terminal and ground.
The voltage drop across this resistor works again the offset creating drop across the feedback resistor.
The value of this resistor should be approximately in the range of the feedback resistor.
However, an optimum value can be found by trial and error only because input offset voltages and currents are nvolved which are unknown (tolerances).

 

[styojm]

Hi godfreyl, keith1200rs, and LvW
I can actually tune the offset through a potentiometer across pin1 & pin 8 which are used to set the offset, and it seems to act the same job if I put the potentiometer at non-inverting input of the op-amp and change it when I change the gain (R). Still this is troublesome--- every time I change my gain I have to go back and zero it. And this is actually a part of my PID control which requires tuning back and forth (BTW, I seem also to have the same problem with my integrator). I'm wondering if there's an easy way to solve this once and for all.
To godfreyl: I guess I can try some FET op amps, is there a recommendation on this? Cause I'm not quite familiar with their types and performance.

- - - Updated - - -

Please see above

- - - Updated - - -

Please see above. BTW how can I reply multiple people in one Reply?

 

[FvM]

Styojm, put another resistor between the pos. opamp terminal and ground.
The voltage drop across this resistor works again the offset creating drop across the feedback resistor.
The value of this resistor should be approximately in the range of the feedback resistor.
However, an optimum value can be found by trial and error only because input offset voltages and currents are nvolved which are unknown (tolerances).

OP27 has input current compensation, the bias current has no defined polarity and is equal to offset current. In this case balancing resistors must be expected to increase total offset rather than reducing it.

The input bias current of the OP27 would not explain a 10mV offset with a 20k resistor.

Yes, the offset voltage should be about factor 10 lower. Either it's a bad OP exemplar, or you are facing circuit leakage currents, e.g. if C is an electrolytic capacitor. As long as other leakage currents are larger than OP bias current, it's pointless to think about FET OPs.

Small offsets won't matter in a usual PID design as long as I action is in effect.

 

[LvW]

OP27 has input current compensation, the bias current has no defined polarity and is equal to offset current. In this case balancing resistors must be expected to increase total offset rather than reducing it.

Yes - thanks for correction. I have overlooked the mentioning of OP27.

 

[styojm]

Hi FvM
Thanks for the instructions. Well I'm using a small ceramic capacitor (1u), I'm not sure if there's a leakage or not. Is there a way to test it?
And just asking, do you know of any available temperature controller circuit design whose heater uses heating tape?

 

[FvM]

I wouldn't expect high (100 nA range) leakage with a ceramic capcitor. But the origin of offset voltage can be only identified by measuring the real hardware, I think.

Heating tape is the range of standard heaters, I think. Industry standard PI or PID controllers should be fine for it. It should be noted, that all recent tmeperature controllers are microcontroller based, allow an easy adjustment of controller parameters and mostly auto tune. Analog controllers are feasible nevertheless.