500A d.c. current measurement with an 08M2?

I note there are plenty of useful and interesting suggestions in other threads such as "Current-measurement-How-would-you-do-this" but not much of what's referred to there is particularly appropriate to what I'd like to do.

Basically, I've got two banks of 6 x 12V 110Ah batteries i.e. two 72V 110Ah batteries, each with a suitable current shunt connected to a DMM measuring volts, current, time and other useful things, but the two DMMs are fixed and not easily visible (in a cupboard).

What I'd like to do is to monitor the current from these batteries remotely using an 08M2 sending the data via an NKM2401 to a roving display. Each battery (fused at 355A) is connected to the motor controller via a hefty contactor and either or both can be connected as required. It's unlikely that the current, even when the batteries are combined, will exceed 500A. I don't really need to measure it more accurately than about +/-25A i.e. about 5%, though ideally it would be more accurate at lower currents.

I note from the aforementioned thread, interesting devices such as the LEM part HTFS-800-P available from Farnell but it's a bit OTT for what I need, with 1% accuracy, and is relatively expensive. Because I'm working with 72V d.c. (can be 90V when on charge) I really don't want any direct connection so a Hall Effect type device would be ideal but the ACS758 (now obsolete anyway) despite being Hall Effect, still requires a break in the current path. I dare say I could get a cheap clamp-on current meter and maybe modify it, but would be interested to know if anyone's done anything like this or has any suggestions as to cheap devices which might help please?
 

rq3

Senior Member
I note there are plenty of useful and interesting suggestions in other threads such as "Current-measurement-How-would-you-do-this" but not much of what's referred to there is particularly appropriate to what I'd like to do.

Basically, I've got two banks of 6 x 12V 110Ah batteries i.e. two 72V 110Ah batteries, each with a suitable current shunt connected to a DMM measuring volts, current, time and other useful things, but the two DMMs are fixed and not easily visible (in a cupboard).

What I'd like to do is to monitor the current from these batteries remotely using an 08M2 sending the data via an NKM2401 to a roving display. Each battery (fused at 355A) is connected to the motor controller via a hefty contactor and either or both can be connected as required. It's unlikely that the current, even when the batteries are combined, will exceed 500A. I don't really need to measure it more accurately than about +/-25A i.e. about 5%, though ideally it would be more accurate at lower currents.

I note from the aforementioned thread, interesting devices such as the LEM part HTFS-800-P available from Farnell but it's a bit OTT for what I need, with 1% accuracy, and is relatively expensive. Because I'm working with 72V d.c. (can be 90V when on charge) I really don't want any direct connection so a Hall Effect type device would be ideal but the ACS758 (now obsolete anyway) despite being Hall Effect, still requires a break in the current path. I dare say I could get a cheap clamp-on current meter and maybe modify it, but would be interested to know if anyone's done anything like this or has any suggestions as to cheap devices which might help please?
Well, you could easily build an interface from a 500 amp/50 mV current shunt and a 100 gain op-amp, giving you a uni-directional resolution of about 2.5 amps from the Picaxe readadc10 function. Although I understand your concern about floating 90 volts to your sensor!

Probably even cheaper would be something like the Tamura L03S500D15, which puts out 0.5 to 4.5 volts with 0-500 amps input, with no assembly required, and it's completely isolated.

EDIT: I just looked at the LEM device, which appears to be comparable in price to the Tamura, although the LEM has over twice the current rating (so you'd get less resolution). They both cost about the same as a good quality 500 amp shunt, plus you avoid the hassle and cost of designing and building the gain stage.

EDIT: I just have to ask. What are doing with 72 volts at 500 amps? Those must be some ferocious motors and contactors! Does Elon Musk know what you are up to?
 
Last edited:

premelec

Senior Member
@HairyA... you could strap a hall effect unit to a current carrying wire sensing it's external field - however you would have to custom calibrate this - I'm recalling the clip on gauges used to check car battery currents while starting which simply used a permanent magnet and spring like a d'arsonval mechanics... don't know how accurate you might get but perhaps worth a try a 50 amp unit that you don't actually pass current through but simply put it near a lead...
 
EDIT: I just have to ask. What are doing with 72 volts at 500 amps? Those must be some ferocious motors and contactors! Does Elon Musk know what you are up to?
View attachment ac12 72 volt 300 amp metric peak graph.pdf

It's an AC12 motor, which, as the model no. implies, is an AC motor. The control electronics take d.c. from a suitable battery source, typically anything from 12 - 72 V and convert it to the necessary AC waveform to turn the motor with whatever speed and torque is required. The attached graph gives some idea of the parameters but these are more typical than worst case. The power output from the the motor is of the order of 15 - 20 kW and it should propel our boat once it gets back into the water.

Yes, the contactors also are fairly hefty things - the thunk they make when engaging is quite noticeable.
 
@HairyA... you could strap a hall effect unit to a current carrying wire sensing it's external field - however you would have to custom calibrate this - I'm recalling the clip on gauges used to check car battery currents while starting which simply used a permanent magnet and spring like a d'arsonval mechanics... don't know how accurate you might get but perhaps worth a try a 50 amp unit that you don't actually pass current through but simply put it near a lead...
Yes, that's exactly the sort of thing I was thinking, hoping that someone might have done something similar.

Looking now at the prices of 400 A clamp-on current meters on eBay, I'm currently thinking that buying one and hacking it to get an RF o/p might well be the easiest way to go.
 

techElder

Well-known member
Hairy, in my work life, I found it quite common to be working on equipment that produced 10s of thousands of contact amps with less than 50 volts DC.

It was also quite common for that equipment to simply screw two wires (going to a metering circuit) a certain distance apart onto a current carrying buss to act as the shunt. The metering circuit was calibrated to the shunt by comparison with a known standard.

You might look at what kind of voltage is produced across one of your low-side current carrying leads. The PICAXE would be much easier to calibrate than the methods we used on that equipment.
 

AllyCat

Senior Member
Hi,

+1 The resistance (per unit length) of all "standard" copper wires is well documented, and it shouldn't be too difficult to measure/calculate any non-standard or "bar" connections. I would expect the hundreds of Amps to produce a voltage drop of at least some tens of mV, which is all that's needed for modest resolution with an 08M2 ADC. The internal "FVR1024" (1024 mV Fixed Voltage Reference) gives a nominal resolution of 1 mV (e.g. 1 Amp through 1 milliOhm). Or a dc amplifier (Op-Amp) could be added if the voltages are really small.

However, the temperature coefficient of Copper is quite high, so for better accuracy you could attach a temperature sensor (DS 18B20, a thermistor or even the PICaxe itself) and correct the calculated value accordingly.

But if you already have a shunt, why not use that? If you're looking for isolation, then a small off-the-shelf isolated dc-dc converter from the 80 volts to power the 08M2 and transmitter shouldn't be too difficult to find. Or a PICaxe needs so little power that maybe just a simple "lighting" LED and a small PV (solar) panel would do the job. The RF link of course gives isolation to the display.

However, a couple of potential "gotchas" with your "dc" supply: If it's driving a high power inverter then the ripple current might be quite high, which could affect any (assumed) dc measurements. But conversely, beware that many of the "clip on" current meters measure only the ac component, because they rely on "transformer action" (i.e. a continuously-changing current / magnetic field).

Cheers, Alan.
 

stan74

Senior Member
My 36V brushless electric bike motor controller has a shunt wire in it to measure current limiting. Try measuring the voltage across that.
 
Hi,

+1 The resistance (per unit length) of all "standard" copper wires is well documented, and it shouldn't be too difficult to measure/calculate any non-standard or "bar" connections. I would expect the hundreds of Amps to produce a voltage drop of at least some tens of mV, which is all that's needed for modest resolution with an 08M2 ADC. The internal "FVR1024" (1024 mV Fixed Voltage Reference) gives a nominal resolution of 1 mV (e.g. 1 Amp through 1 milliOhm). Or a dc amplifier (Op-Amp) could be added if the voltages are really small.

However, the temperature coefficient of Copper is quite high, so for better accuracy you could attach a temperature sensor (DS 18B20, a thermistor or even the PICaxe itself) and correct the calculated value accordingly.

But if you already have a shunt, why not use that? If you're looking for isolation, then a small off-the-shelf isolated dc-dc converter from the 80 volts to power the 08M2 and transmitter shouldn't be too difficult to find. Or a PICaxe needs so little power that maybe just a simple "lighting" LED and a small PV (solar) panel would do the job. The RF link of course gives isolation to the display.

However, a couple of potential "gotchas" with your "dc" supply: If it's driving a high power inverter then the ripple current might be quite high, which could affect any (assumed) dc measurements. But conversely, beware that many of the "clip on" current meters measure only the ac component, because they rely on "transformer action" (i.e. a continuously-changing current / magnetic field).

Cheers, Alan.
I hadn't noticed that the cheap current meters are AC only - thanks for that.

The shunts I already have are in the individual battery packs (several feet apart) so I'd need a monitor/tx on each as either or both might be connected. That wouldn't be so bad I suppose, though when both are in use I'd have to mentally add the currents from both (I think my aged brain would cope). Cost wise an extra 08M2 and a few other components wouldn't be a problem, so this might well be the most cost effective option.

It's also relatively safe as the shunts are on the 0V end of the battery packs so I don't need to worry so much about isolation.

The ripple current is a very good point that I hadn't thought of. Some analogue filtering (depending on the frequencies involved) on the shunt voltage might be one way around it?
 
My 36V brushless electric bike motor controller has a shunt wire in it to measure current limiting. Try measuring the voltage across that.
Being AC, I don't think any measurements around the motor itself will be straightforward.

The controller has a CAN bus connection which can produce all sorts of useful information including an accurate measurement of the motor current, though this is probably on the AC side:
Power-section-topology.jpg

I expect that if I studied the CAN bus spec I could get an 08M2 to communicate with the controller, or another option might be to use the (PWM) analog output which produces a 0 - 10V dc o/p which with suitable programming via the VCL (Vehicle Control Language) might be able to output an analog representation of the motor current, which I could then measure with the 08M2.

These options though add a layer of complexity I was hoping to avoid.

Thanks though for all the helpful suggestions so far.
 

Buzby

Senior Member
... The controller has a CAN bus connection which can produce all sorts of useful information including an accurate measurement of the motor current ...
If the controller has a CAN bus, then maybe you could connect one of those OBD devices that support Bluetooth. https://www.ebay.co.uk/itm/Mini-Bluetooth-Car-Scanner-OBD2-ELM327-V2-1-Android-Torque-Auto-Scan-Tool-OBD-II/162617059752?epid=1442054158&hash=item25dcbb69a8:g:SfEAAOSwLWBZhFiy

My understanding is that these devices are fairly dumb, and are basically just wireless interfaces to the CAN bus, all the clever stuff is in the apps.

It will need a bit of Googling to find the details of the protocol, but the results would be worth the effort.

Cheers,

Buzby
 

cpedw

Senior Member
I used an Amploc amploc.com Hall effect transducer at much lower current very successfully. They are powered by 5V, output 0-5V, measure DC and are completely isolated from the measured voltage. Two snags are 300A is the max rating (but my experience of the 25A model shows i works at higher current than the nominal) and not dirt cheap.

Derek
 
I used an Amploc amploc.com Hall effect transducer at much lower current very successfully. They are powered by 5V, output 0-5V, measure DC and are completely isolated from the measured voltage. Two snags are 300A is the max rating (but my experience of the 25A model shows i works at higher current than the nominal) and not dirt cheap.

Derek
Thanks Derek, my present thinking is indeed to go for a Hall Effect device and having found the Honeywell SS49E TO-92 devices at 10p each on eBay, they seem like a good place to start, especially if I use two: one on the +ve lead and another on the -ve lead which should then give me a differential voltage, as each one gives a voltage above or below 2.5V depending on the field polarity. Using two should also eliminate any common-mode magnetic fields, e.g. from our planet.

I'm assuming that if they get overloaded in a high current situation, I can move the devices away from the leads to reduce the field strength.
 

cpedw

Senior Member
especially if I use two: one on the +ve lead and another on the -ve lead which should then give me a differential voltage, as each one gives a voltage above or below 2.5V depending on the field polarity.
That's a cunning plan!. The Amploc device also gives 2.5V at 0A. At small currents any offset can cause significant error.

I'm assuming that if they get overloaded in a high current situation, I can move the devices away from the leads to reduce the field strength.
That wouldn't work with the Amploc devices; they are a closed loop that completely encircles the lead being measured. But I understand that excessive current doesn't damage the detector. And if wanting to measure small current, several turns of the lead through the detector multiplies the reading by the number of turns.
 
Top