Measure USB Current up to 3A Without Adding Series Resistance

[ngmedaboard]

I'm looking to build a USB current measurement circuit that doesn't add any series resistance. Even a 10mOhm shunt is going to create voltage drops that affect my testing more than I'd like.

I'm trying to hit accuracy and precision to within about 5-10mA.

I'm guessing I should be looking at loop based current sensors, but was looking for some specific recommendations.

0 - 3A Current range
5 - 10mA on Accuracy and precision
No series resistance added (okay 1 - 2 mOhms is fine but I don't want much more than that.)

Your thoughts and ideas are greatly appreciated.

Thanks,

-NGM

 

[crutschow]

How about a Hall Effect current sensor.

 

[BradtheRad]

Those are tough specs to meet.

I wanted to make an instrument to measure current from my solar panels. Over 15 A.

I used a few inches of 12 or 14 gauge wire. That's only a tiny resistance.

The voltage across it was a few mV. Maximum 25 mV. I wanted to amplify it up to 5V maximum. I used an op amp at gain 200.

 

[ngmedaboard]

How about a Hall Effect current sensor.

That looks like the path I'm going down now, found one with a 1.2mOhm path resistance, way more tolerable than a 10mOhm shunt. And 185mV/A. **broken link removed** My scope's best resolution is only 2mV / div, since it's DC, I could use a high precision DMM, but would like to be able to read to within 1-5mA of accuracy and precision the current through my scope.

I'm not sure this will get me there without another amplifier stage, no? And I'm not sure whether having a 10X probe is of benefit since I think the scope adjusts such that you still only have 2mV / div not 20mV / div.

Any thoughts?

 

[RCinFLA]

Since normal USB is limited to 0.5 amps you must be looking at custom (manufacturers) USB charger application. You will likely find some of these charger to be current pulsing, not necesarily pure D.C. Look for good frequency bandwidth, like 100 kHz to 300 KHz so you can observe the pulse shape and duration.

 

[pplus]

Circuit design in analog form with you specified measuring accuracy (0.01A/3A)*100=0.3% can be attributed to precision.

How about a Hall Effect current sensor.

The use of hall sensors to achieve the required accuracy will require a microcontroller

 

[Nick_F]

Those are tough specs to meet.

I wanted to make an instrument to measure current from my solar panels. Over 15 A.

I used a few inches of 12 or 14 gauge wire. That's only a tiny resistance.

The voltage across it was a few mV. Maximum 25 mV. I wanted to amplify it up to 5V maximum. I used an op amp at gain 200.

Hi Brad,

As far as I know, if you didn't use some special resistive wire, then your measurements could be significantly affected by variations in temperature. The current going through wire can change the temperature and also the ambient temperature can affect it. For example, for copper, an increase in temperature from 20 celsius to 30 celsius causes an increase of the resistivity of almost 4%. If more precision is required, probably correction can be applied by adding a temperature sensor to close proximity of the wire and doing some calculations. Special materials such as manganin and constantan have very low temperature coefficients (see http://www.endmemo.com/physics/resistt.php)

Regards,
Nicolae

 

[BradtheRad]

Hi Brad,

As far as I know, if you didn't use some special resistive wire, then your measurements could be significantly affected by variations in temperature. The current going through wire can change the temperature and also the ambient temperature can affect it. For example, for copper, an increase in temperature from 20 celsius to 30 celsius causes an increase of the resistivity of almost 4%. If more precision is required, probably correction can be applied by adding a temperature sensor to close proximity of the wire and doing some calculations. Special materials such as manganin and constantan have very low temperature coefficients (see http://www.endmemo.com/physics/resistt.php)

Regards,
Nicolae

Yes, that was a project I did years ago. It didn't occur to me that copper would change its conductivity much, within a temperature range of my basement.

However since joining this message board I saw a thread about shunts, where someone pointed out that copper is unsuitable as a precision resistance.