Interfacing HX711 with DC Shunt

I'm trying to read DC current consumed by a DC motor with the help of a Shunt resistor ( 0 to 75 MV output for 0 to 100A) . The signal output is interfaced with HX711. A panel amp meter is connected parallel to calibrate the results obtained. After successful calibration, the Arduino values are exactly matched to panel meter values by running a test motor with constant load. The same test was repeated and consistency was ensured for this test conditions. The strange thing began when I changed the applied load on the DC motor.( Running the test motor at 6V with 0.5KG load on shaft and running the same motor at 15 V with 0.1 kg load on shaft will consume same current ,Ex- 10A, and shunt gives a constant millivolt output, 9.4Mv for both the conditions) The panel meter always showed a constant value for a specific millivolt input irrespective of applied load on test motor. But HX711 sensor gives different values for the same millivolt input. (Note: Applied DC volt on motor is no way connected to Panel meter or HX711). Any idea why this is happening? How HX711 interprets the calibration constant?

75 MV!!! :o

ravindran:
(Note: Applied DC volt on motor is no way connected to Panel meter or HX711).

Needed, because the shunt has to be within the common mode range of the HX711.
I like to see a diagram of how you have “floated” the shunt at about 2volt above HX711 ground.

Are you using a filter to smoothen motor current spikes. R3,4,C4 in the HX711 reference schematic.
You might have to increase the cap value.
Leo…

U might consider moving UP to Hall effect current sensing..much much easier and totally isolated

http://www.ebay.com/itm/New-design-30A-range-Current-Sensor-Module-ACS712-Module-Arduino-module-/181847127383?hash=item2a56eef557:g:~cUAAOSwNSxVLO2D

Wawa: Needed, because the shunt has to be within the common mode range of the HX711. I like to see a diagram of how you have "floated" the shunt at about 2volt above HX711 ground.

Are you using a filter to smoothen motor current spikes. R3,4,C4 in the HX711 reference schematic. You might have to increase the cap value. Leo..

Thanks for your reply Leo. Yes you are right. Even though motor power supply(DC Fully rectified 30V /100A) is a separate one( only used to drive the motor)), Signal output from the shunt and motor power supply have common ground. Signal output from Shunt is connected to A+ (in HX711) , A- (in HX711) to common ground (motor Power supply) . What i was trying to mean was,there is no way for the varying voltage(only Voltage) in the DC motor power supply to influence panel meter or HX711( Because what ever the voltage is, the shunt's output is only proportional to ampere consumed) . There is no way to "float" the signal. Excuse me for my confusing/misleading explanation.

Yes, i'm using 100 nF Cap to arrest the spike.Here Going to higher value further increases the response time.

But my question still remains, why would a HX711 behave differently for the SAME INPUT SIGNAL generated under different Voltage/load combinations , but behaves consistent for the same Voltage/load combination in repeated tests.

//U might consider moving UP to Hall effect current sensing..much much easier and totally isolated//

Thanks for your suggestion backwoodsjack. I'm getting fairly accurate results with Shunt sensor than hall effect sensors.

ravindran: Signal output from Shunt is connected to A+ (in HX711) , A- (in HX711) to common ground (motor Power supply).

Forget the motor (and motor ground) for now.

So the shunt is connected between A+ and A-, but "floating" somewhere between the HX711 analogue supply (E+) and HX711 ground. If you're lucky.

No direct experience here, but I think this can't be good. I would use two 1k resistors. One between E+ and A-, and one between A- and -E (HX711 ground). Then A- is fixed in the middle of the common mode range of the HX711, and A+ also is through the low resistance of the shunt.

ravindran: Yes, i'm using 100 nF Cap to arrest the spike.Here Going to higher value further increases the response time.

Did you calculate the response time of 200ohm/100n. How does that stack against the 10Hz or 80Hz data rate of the chip. Leo..

Wawa: No direct experience here, but I think this can't be good. I would use two 1k resistors. One between E+ and A-, and one between A- and HX711 ground. Then A- is fixed in the middle of the common mode range of the HX711, and A+ also is through the low resistance of the shunt.

That a good idea. HX711 is powered with VCC & Gnd. I'm not using E+, E-. I'll try and update. Thanks Leo