Calibrating Shunt

[J.W van de Poll]

Hi All

I have a 100A PV shunt and a 250A Load shunt.

The 250A shunt is suspect in the tolerance value of the device.

A load current of 350mA when it should not be more then 60mA.

The accuracy of the SOC (after first 100%) is totally dependent on the accuracy of shunt.

Have had to adjust the mOhms on the 250A to make it agree with the solar current coming in.(with no loads)

So some questions.

1. Is it the general consensus that regardless of the actual shunt resistance, the device WILL be linear ?

2. What methods have others used to calibrate the shunt.

     I was thinking of a Fluke in series with loads up to 10 Amps, which if the device is linear will be ok.

     If I had an accurate current source then I would use that.

3. There is the unknown of current drawn by the SBMS0, stated as between 300mW & 800mW with backlight and Wifi On.

     I don't have Wifi and assuming no backlight what is the current drawn by the SBMS0 ?

     But as that is drawn before the shunt it doesn't affect the shunt reading.

     Still I like to know what it is.

Regards

[Dacian Todea]

Sorry you did not mentioned the complete spec of the shunts.

Say both shunts are 50mV then the 250A shunt will have 50mV/250A = 0.2000mOhm and that is what you will set in the ADC menu in the SBMS0 and if the 100A is also a 50mV shunt then 50mV/100A = 0.5000mOhm and you will set that for the PV shunt.

If shunts are OK quality below 1% tolerance nothing else should be done and displayed current should be withing 1% of the real value.

There is a zero offset calibration during power UP and since a properly installed SBMS0 should be able to turn OFF all loads and all charge sources there will be zero current trough the shunts during power ON and the automated zero current calibration will be done correctly.

If your shunts have lower quality so maybe even 3 to 5% tolerance and you want to correct that or you want better than 1% accuracy then you will need a multimeter that has a current measurement better than the current shunt tolerance else you can make things worse.  You will first calibrate the battery shunt and then based on that you can calibrate the PV shunt.

[J.W van de Poll]

Hi All

Dacian the shunts are 75mV.

Do you think regardless of the tolerance that they will be linear ?

Does this offset calibration include the 800mWatt of the SBMS0 ?

Regards

[Dacian Todea]

Are you sure that is a 250A shunt as the 75mV low cost China shunts are multiple of 100A so 200 or 300A as they use 2 or 3 100W blades in parallel.

Those shunts have higher tolerance maybe in the 5% range so if you are happy with an up to 5% error then all you need to do is caclualte the correct shunt resistance from spec so 75mV/100A = 0.7500mOhm now for the supposed 250A one I will make sure is not a 200 or 300A one so maybe post a photo of that one.

The current shunt amplifier and ADC are very linear and shunts usually are also acceptably linear as they do not get that hot to make much of a difference in resistance is more about the spec that can be for the cheap ones up to plus minus 5%

The shunts do not measure the SBMS0 self consumption that is just added as calculated by the SBMS0 so one value for WiFi ON and one for WiFi OFF and then just added there.

Depending on battery voltage you will see a calculated negative current displayed for battery current when you have no load or charge source present. It will be maybe at most -50 to -60mA with a 12V battery so you will see something like -0.060A with zero current trough the shunts. So in this case if you add a 300mA load you will see -0.360mA as it adds the calculated self consumption of the SBMS0 to the display but the SBMS0 will actually measure just -0.300A

Same if you have say 1A of charge current and nothing else then PV shunt will show 1A but battery shunt will show +0.940A as it subtracts that calculated SBMS0 self consumption.

[J.W van de Poll]

It is a two blade shunt.

As I am confident of the PV shunt , I simply adjusted the value for the Battery shunt until it read the same (except for about 60mA).

The value was 0.3175mOhms.

As I can't measure anything much above 10Amps, I will wait for my trickle charge PV panel (50Watt)  to arrive and then tests again.

Regards

[Dacian Todea]

That seems like a fairly useless shunt as it should be a 200A 75mV shunt and trimmed to correct value. They seems they did not trimmed the shunt at all or just made a small attempt to look like it was calibrated.

By cutting out material you can increase the resistance but you can not do anything to decrease resistance.

A proper 250A 75mV shunt will have 0.3000mOhm but even based on your crude calibration it is at least 6% out of spec so it was imposible for this to be a 250A shunt as it was designed to be a 200A shunt that needed calibration by cutting in material to increase the resistance to 0.3750mOhm

It will have looked something like the ones below if properly trimmed to 200A

So you now have maybe something like a 236A 75mV shunt. I was not aware that they sell the same 200A designed shunts as 250A 

 

[Barry Timm]

I replaced both the shunts I bought on Amazon with Riedon shunts, after I measured huge calibration errors on those cheap shunts. The Riedon shunts were much better and pretty close to spec. I had to make very slight calibration change in SBMS0.

[J.W van de Poll]

Hi All

Today I got my trickle charging panel and when wired produced different results again.

When the panel was unplugged and there was no charging current from anywhere, the PV current showed 0.00

BUT the Battery current showed + 0.650 A ??

When I saw that I decided that I needed to get better quality shunts or measure the existing.

I remembered I had a milli ohm meter amongst my gear (see pic).

30 years ago I had someone translate the instructions on the back from French.

So, now all I need is a 2vdc supply to operate it and it should read the resistance within 1/10th of a milli-ohm.

Regards

[Dacian Todea]

The fact that you see a +0.650A means when powering the SBMS0 you had a 650mA load that was not controlled by SBMS0 as SBMS0 needs to be able to turn ON/OFF all charge sources and all Loads.

If all is correctly installed the SBMS0 will turn all things OFF and do a zero offset calibration for shunts expecting that there is 0A trough both shunts. But since you had a -650mA load it considered that as zero point and now that you disconnected that it shows a +650mA charge source.

So just power cycle the SBMS0 making sure all loads and charge sources are connected to an EXT IOx and zero offset will automatically be done correctly.

   

[J.W van de Poll]

Hi All

The only loads NOT switched by the SBMS0 is the supply to the SSR's.

ie: EXTIO3 ,4,5 which would be less then 30 mA. (otherwise I can't switch them)

However I may have had things connected during testing, but will verify.

At the moment I have the system shut down and the shunts removed to accurately determine the values.

Regards

[J.W van de Poll]

Hi All

I am now officially obsessed with getting the correct shunt reading.

Using the above instrument and a 2vdc supply have measured the shunts.

The 250A shunt measures at 0.00028 ohm (Should be 0.0003)

The 100A shunt measures at 0.00074 ohm.(Should be 0.00075)

The resolution of the instrument is 10 micro ohms.

But I have another card up my sleeve to calculate micro ohms by

using a 20 bit Load cell amplifier to measure the voltage drop across the shunt.

However the accuracy and resolution of the supply Voltage plays a large role there.

Regards

[Dacian Todea]

If you have a constant current lab power supply all you need to do is set it to 2 or 3V and 5 or 10A and if not sure about how accurate is the supply then measure with the multimeter the current set to show an exact value say 5A and then use the multimeter on mV scale and measure voltage drop on the shunt.

So say the shunt is around 0.28mOhm then at 5A you will see 1.4mV so if your multimeter is even just a 2000count one and you can set on 2mV then you can read 1.400V say you read 1.401V

Then 1.401V/5A = 0.2802mOhm so accurate enough.

But if you are super obsessed with this then you should use the 5A lab power supply and match the reading on the SBMS as there is also some accuracy in the 20x current shunt amplifier and the microcontroller ADC so then if you know the 5A source is accurate measured with a multi meter then if you set battery shunt measurement to read exactly -5.050A in one direction and +4.950A in the other you know that calibration is excellent (the 0.050A is considerd the calculated self consumption but you will need to note that when you power the SBMS as is based on battery voltage the 50mA was just an example).

Now once you have this perfectly calibrated you can calibrate the PV shunt based on this battery shunt by first using a small current like 500mA that flows trough both shunt is series so exactly the same current.

And say you read considering the same 50mA self consumption you should see +450mA on the battery shunt that will be correct and should see 500mA on the PV shunt but if the shunt has a current shunt amplifier with positive offset (50% will have a small positive and 50% a small negative offset) then zero offset automatic calibration will have worked perfectly and you will see 500mA if shunt resistance value is set correctly but if it has a negative offset that can not be auto calibrated and say that offset is 100mA so you only read 400mA instead of 500mA then you can add that 100mA in the ADC menu to compensate (basically a manual zero offset calibration).

But say you do not know the exact shunt resistance so just used the 0.7500mOhm  calculated then you need to do two measurements at two values the 0.5A and 5A will be a good value that way if you observe the same 100mA missing you know there is a offset problem that you can correct but if you gate different delta  values at 0.5 and at 5A then you know resistance of the shunt needs to be adjusted in ADC menu.

So say you read 400mA and 5100mA

Then 5100 - 400 = 4700 so you are reading a higher value than the 4500mA real current delta so you modify the shunt so that the delta gets to 4500mA then after you did that you note the offset that needs to be corrected if any and add that in to ADC setting to get the exact readings and matching with the battery shunt value as you calibrate against that.

[J.W van de Poll]

Hi All

Dacian just to clarify  "self consumption"

The SBMS0 gets it's power from the balancing connections, so that current does not appear on the shunts.

Then the only other consumption (when there is NO load) in my circuit is about 10 mA  for EXTIO 3 & 5 and whatever current the DEXT16 (EXTIO 4) takes. (Haven't measured that, but will)

These are measured by the Battery shunt.

My constant current power supply is 3A and the tests I have done rely heavily on accuracy of multi meters.

I kept one multi meter in circuit reading Amps, typically 3.018. (That's when the ps read 3.000)

My downfall is that I can't read single micro volts with any of my meters.

The reading is typically 0.84 mV across the shunt.

Therefore a calculated resistance of 0.00027833.

Considering that the only meter I have faith in , is the milli-ohm meter shown above, it is reasonable to assume

that it's reading of 0.00028 ohm is a rounding of micro volts.

I have a PicoScope 2000 and tried to use it on the +/- 50mV range and then reading Average DC Volts.

The DC was incredibly noisy with peaks of a hundred mV.

Adding a few .pF across the shunt dropped the peaks to about 10mV but the amount of spikes was corrupting the average voltage readout.

I could not get a stable micro volt reading.

So, that was a dead end.

I will re-install the shunts into the system and use 0.00027833 as the resistance.

Then as per your suggestion calibrate the PV shunt as per your comments above.

Regards

[Dacian Todea]

Yes the SBMS0 takes his current trough the cell monitoring ribbon cable so not measured by the shunt but the current is manually added in the software as 0.3W if WiFi is off or not installed.

It is important to add that 0.3W as while it seem insignificant it can discharge a small battery after many weeks or months so it is needed there to calculate the battery SOC more accurately.

DEXT16 consumption will depend on how many DSSR20 with diversion you have connected if any.

If you read 0.84mV and 3.018A then you should use 0.2783mOhm in the SBMS ADC menu but yes if that measurement is out by one digit say 0.85mV real then value will be quite different.

But you also know what the SBMS0 will read at 3.018A so maybe that is more important.

Oscilloscopes will not be as accurate on voltage measurements as a multimeter so it will make no sense to use that