Precise High Current Sensing at 12.5 mV/Amp

[wizpic]

hI KlausST

I've made the changes as suggested, Well I think I understand in what you meant. Please see Picture.


Thanks for your input

 

[KlausST]

Hi,

You may omit R13

C1, C11 should be ceramics. No electrolytics

R14 should be 10R, not 10k

R5 now generates the offset, therefore you may omit R5, R7

You currently use 2,5V VRef, isn't a 4.096V better? It generates a 2.048V and thus symmetic for the ADC input.

Klaus

 

[wizpic]

Hi,
I've made the modifications, I will temporary layout a PCB for it and get it made on my CNC machine as I use SMD parts. I've done a quick simulation in ISIS and it looks good. I know you can't really trust these programs and it does not give the true reflection compared to real life testing but I do it as a bench mark.
KlausST Can I use the same calculations for the Op-Amp as I did before with it set up my old way on the newer design ?.
Based on this equation. Looking at my design. mV & mV1 are the inputs to the op-amp.
((1+(R1/R4))/(1+R6/R8*Mv)-((R1/R4)*Mv1)= Vout
This is so I can convert the data to decimal on the LCD.
Regards,
Wizpic

 

[KlausST]

Hi,

Gain of your difference amplifier = 10k / 1k = 10
Output offset = 2.5V / 2 = 1.25V

Formula: V_out = V_in_ diff × 10 + 1.25V

Klaus

 

[wizpic]

Just thought I'd share another circuit, This one was reversed engineered from a commercial product.
This product used a 60A 60mV shunt, it can measure current in either direction with out using a reference voltage with no current flowing the output is 0.00V. It produces 0-3V output no matter which way the current flows. I have built in the past and it works really well, The only draw back you don't know which way the current if flowing I.E in/out. But I added another little circuit that either gives an HIGH or LOW signal depending on which way the circuit flows so I could see it on the display. The only thing I've added extra is the low pass filter on the output to P3
Someone may find it useful, Again any comments are welcome.

 

[KlausST]

Hi,

Your headline says "precise", but I doubt the rectifier is precise.

My opinion:
A lot of effort with reduced precision .... I don't see the benefit, because the rectifier could simply be done with software...this is more precise, more flexible and it's easy to calibrate.

Klaus

 

[wizpic]

Just to give an update really,
I've made the PCB on my CNC machine, Picture attached. Now this one was only a quick method because it will only do isolation milling, could of done a better job with the polygons but this takes to long to remove more copper. Just wanted to prove that the schematic was correct with all the connections basically.
Has for the shunt I've used 2 X 0.01R 5W resistors to give 0.005 (measured with milli-ohm meter), Calibrated it in software with the reference voltage so that with no current the display reads 0.00, Then also got a calibration routine to calibrate the current. I calibrated with 1.00amp flowing through and then connected it to a 10amp battery charger, Has you can see for the photo it's spot on and very happy with the results and readings are very stable and does not fluctuate at all. I've also carried out a test where I changed the resistor values to divide the voltage and can use the same circuit as a voltmeter and measure voltage up to +/-99.99V, Again with stable readings and as accurate my calibrated fluke meter. In the photo you can see my test PCB, looks a bit rough as I made the mistake in trying to cover the copper with solder which makes it look as utter mess but only wanted it for testing, I shall make a better PCB now knowing that it works.

Thanks Klaus for your suggestions and improvements it works loads better then it did before more stable readings and the Vref voltage is the same every time I turn the meter on 0.00v where as before the display would read 0.01-0.05V now it reads 0.00 all the time.
PCB layout


Photo
**broken link removed**

I guess I could go to the next step and get the scope out to compare the old circuit against the new circuit. I did try the ACS712 but found this fluctuated to much so not using them any more

 

[pradhan.rachit]

Hello Experts,
An extension to this post...

There have been instances on the forum where some people have mentioned the use of auto/self calibration routines on the software side.
Looks like these routines are used to cancel out the effects of drift in the reference voltage and sensitivity due to varying temperatures and other operating conditions.

Could someone please put some light on it?

Thanks!
- Rachit