Noise Reduction In Current To Voltage Converter

Certainly I ll reply this after getting the data of signal records tomorrow, but don't you you think it would be useful if you can reveal what the actual value of current and what is the equivalent converted value of voltage.

Regards

This circuit has 2V/nA and can take maximum current of +/-5nA to reach a saturation level.

Actual requirement is +/-1pA to +/-10nA range and hence will reduce the gain by a factor of 2 in my next design and will set to 1V/nA level..

Current schematic under test is shown here.

Normal noise level is shown above and on tapping or injecting vibrations the following signal is due to Micro-phonic vibrations.

Enclosing the instrument case in foam, the effect of vibration is reduced considerably.

Oscilloscope recording Horizontal Time scale is 10ms per division and Vertical voltage scale is 50mV / division. Gain 2V/nA.

Amplified signal for 4nA 1kHz pulses gives 8V output. Effect of cable capacitance 10pF and amplifier response to large current is visible in slow response of the current to voltage converter amplifier.

Your original posting said that there was some trim capacitor somewhere. I don't see one in your schematic.

I don't believe your 10pF cable capacitance value explains your apparent 800 Hz 3dB roll off for two reasons. First, a RG-59 foam dielectric cable has a 16 pF per foot value or 53 pF per meter value. That's only the cable capacitance and its lumped element value changes with length so this thick dielectric, low capacitance cable must be less than 180mm long. That's not including the connectors, any circuit board trace, instrument driving the cable or anything else that will be added to the lumped element total value. The obsolete Op-amp itself has 1 pF of differential input capacitance alone. Second, 10 pF produces a pole of 800 Hz with a 10 MEG resistance. 10 MEG does not exist alone or in combination with any of the elements you've shown.

I suspect the displayed roll off effect you actually have is due to the large signal response of your op-amp not your cabling capacitance. While I don't think that your lumped element capacitances are rolling off your circuit, I'm certain that you have a much larger lumped element capacitance at all macroscopic points of this circuit than 10 pF.

Redfred,

It is a mistake in text writing. It is a trim resistor and not a trim capacitor. Thanks for reminding.

Unfortunately the amplifier was damaged today as my assistant tried to forcefully inserted the power connector in wrong direction and it made contact the pins with wrong polarity of the power supply. Even though connector is polarized, it does make contacts with pins if forced and power was through battery with no safety switch.

OPA128SM died this morning and it is all due to not having protection against wrong power source.

This time I got the amplifier rewired with AD549LH and also added diodes in the power supply line to avoid wrong polarity power.

I found this amplifier a bit slower, and more noisy but did not show much offset.

I have brought down the gain to 200M feedback resistor || 2pF giving gain of 200 million voltage per Ampere input current or 0.2V/nA which is 1/10 of the earlier design.

Waveform recorded with 20Hz current signal is shown here.

Rise time and fall time of 5 ms looks too large. I will reduce feedback capacitor to see how much improvement can be achieved.

Current source is a pulsed voltage source and through shielded 1G Ohms resistor, current was injected. 2.2pF damping capacitor was used to remove the effect of stray capacitance across 1G Ohms resistor. I suspect 0.2pF is the capacitance parallel to 1G Ohms resistor and hence by adding 2.2pF, the effect is damped by a factor of 10. Feedback capacitor also does damp amplified signal but this one to be slightly adjusted.

I now want to try faster I-V converter with 1DC to MHz, 10MHz and 100MHz responses.

I am not yet adding boot strapping signal speeding mechanisms and will try in all these high frequency I-V converters. I am also going to try some discrete designs.

rakesh_semwal:Have you tried any of these designs?
I am going to make some changes in the design for the I-V converter amplifier.Two main changes are to be done.
Increasing the signal response BW up to 10kHz from current 1kHz and this may increase noise also. Converter gain of 1V/nA or even less to 0.1V/nA to be tried in the first stage of the amplifier. Multiple stage gain may be ideal to increase frequency response.
Clamping the over current if input current exceeds 10nA / 100nA / 1uA for different gain settings. Right now current limiting resistor has been used but this design becomes non-linear over large current range due to voltage drop in the resistor.
Other important parameter is to select design parameters to reduce cosmic ray effect, Microphonic noise and EM interference for which mechanical design as well as PCB is to be modified.

Without seeing more clearly the whole circuit, I cannot offer much assistance. Chasing down and resolving analog noise in a noisy environments is a common occurrence for me. It would be helpful to know the required bandwidth of your signal. Since power line frequencies (50Hz) are apparently an in band frequency I do not recommend coaxial input cabling. Coaxial input cabling will promote and not shield magnetic coupling of interference. I recommend shielded twisted pair input wiring and utilize as much common mode rejection as you can design. Common mode rejection will also help in cancelling the tribo-electric effect from the input cabling will acting like a capacitive microphone. Remember, you must methodically identify the interference path for the extraneous signals in order to circumvent them.

Micro phonic noise generated by directly tapping the instrument casing by ball pen. Hit generates all this very intense Micro phonic signalat the amplifier output. See above.

Tapping on table generates somewhat cleaner signal. See below.

200pA 1kHz current amplified signal looks good having output of 2V/nA. See below.

Offset voltage trim capacitor is a mistake. It is a offset trip resistor. It is still there in the circuit. Perhaps I will remove this trim resistor to see the effect on Micro-phonic noise.

Today we have a new version of I-V converter amplifier using OPA128SM and results are very impressive.

Gain was set at 10mV/pA

uncompensated amplifier for 75pA 20Hz signal did show large AC response is shown below. Stray capacitance for 10G Ohms is expected to be 2pF in parallel to 10G Ohms.

Blue trace is input current pulse and yellow trace above is amplifier response.

After adding compensation capacitor of 10pF in parallel to 10G Ohms Traces were repeated number of times to see the reproducibility and it looks pretty good shape.

Vibration induced noise was less than +/-10pA and it did vanish when cable was in stretched condition and in air. Peak to peak random noise was <+/-1pA.

I will like to see now the response of the amplifier with low pass filter and will also like to compare its performance with ICs manufactured by other manufacturers. AD549LH results are not impressive as amplifier is relatively very slow.