Current Sensing Problem

[dreamyboy_999]

I am using a current sense amplifier in a solar harvesting circuit which uses a SEPIC converter in order to adjust the amount of power drawn from the solar panel. I am using a 28m sense resistor and the voltage drop across the resistor at some operating point is 1.2 mv. I have used a multimeter to measure the voltage drop which I guess gives the average or DC component. when I multiply it by the gain of the amplifier , it is different from the output of the amplifier (again I use a multimeter to measure the output). I have tested with current sense amplifier with a pure DC signal and it is quite accurate...but I have this problem in my solar harvesting circuit.....

 

[stanislavb]

Hi,
When you tested with "pure DC signal" it was also 1.2mV as input? I guess it cause due some non linear characteristics of amplifier. If I am right, you should divide yours current range into two separated sub ranges: low current with higher sense resistor and high(er) current range with lower sense resistor. This is trick may improve current measure capabilities
Regards,

 

[kam1787]

is this before the SEPIC converter?

 

[mtwieg]

Might be due to poor input offset voltage or CMRR on the amplifier. Or maybe your DMM isn't so accurate. Hard to say without a schematic.

 

[dreamyboy_999]

- - - Updated - - -

is this before the SEPIC converter?

Yes it is before the SEPIC converter, it is a high side current sense amplifier

- - - Updated - - -

Hi,
When you tested with "pure DC signal" it was also 1.2mV as input? I guess it cause due some non linear characteristics of amplifier. If I am right, you should divide yours current range into two separated sub ranges: low current with higher sense resistor and high(er) current range with lower sense resistor. This is trick may improve current measure capabilities
Regards,

Yes with pure DC signal I have tested for different ranges...even less than that....It is pretty accurate...

 

[dreamyboy_999]

Might be due to poor input offset voltage or CMRR on the amplifier. Or maybe your DMM isn't so accurate. Hard to say without a schematic.

If the problem was with offest or CMRR , why is it completely accurate in DC mode ? (even if with very very small sense voltages) ? what do u exactly mean by the accuracy of the multimeter ? u mean the resolution or its sampling rate ?

 

[chuckey]

Is the input to your converter DC or pulses? No reason to believe that a meter on DC accurately measures a pulsing waveform, if anything I would expect it to have a LPF so it ignores noise, did it read low? Use a wideband RMS meter.
Frank

 

[iop95]

Move sensing resistor between load and ground to reduce noise.

 

[dreamyboy_999]

Is the input to your converter DC or pulses? No reason to believe that a meter on DC accurately measures a pulsing waveform, if anything I would expect it to have a LPF so it ignores noise, did it read low? Use a wideband RMS meter.
Frank

it is actually a solar panel, I have figured out one thing...the corss talk from the switching signal is so high in different parts of my circuit..I am trying to measure a voltage over sense resistor which is of order mili volts, but the cross talk it self is about 100 mv!!

 

[chuckey]

As the input impedance of a DVM is normally 10 M ohms,I would put a low pass filter in series with the meter leads. i.e. put a 10 K resistor on each end of the current monitoring resistor, a .1 MF capacitor across the free ends of the resistors and measure the DC voltage across the capacitor. You can compensate for the additional resistances, they will reduce the voltage by 20/10000 ~ .002 or .2% .
Frank