Where is the shunt, the shunt, the shunt???

[isd88]

Good morning! I'm trying to adapt my kelly kbs controller to european pedelec standard using a gas legalizer with pas input. I want also to read the current to compute power consumption but when i disassembled the controller i couldn t find any shunt or visible hall current sensor! Can anyone please help me finding out a way to measure che current flowing trought the controller in an easy way? I can t add a sensor since the space is contraint. Here is a picture of the board

How the cpu measure the current without visible shunt?

 

[circuit]

It is *possible* to measure current on open mosfet. But hard to get stable measurement.
Also it is *possible* to measure current on a piece of copper trace. Might be OK, if PCB manufacturing process is well controlled.

 

[isd88]

i also think that probably they use a copper trace. Not a good news for me

 

[circuit]

They also use awfully many "discrete" ICs, such as opamps. Quantities like this might be used for current sensing on open mosfet. But that is not good news either.

BTW, acetone washes that black goo off pretty well.

 

[isd88]

i will clean the board as soon as i found a cheap way to do potting, i had some water damage during use in default configuration. Probably i have to use one of my allegro ac752 very very big ic. Thank you for help

 

[isd88]

problem solved!!

 

[circuit]

Ah, so you only needed this to hook up CA? Did you managed to find an actual current sensing method used?

 

[isd88]

yes, i building my own CA. and now i can also measure the regen current!

 

[teklektik]

Just buy a standard CA external shunt - already potted and ready to go with power connection and connector. Then your gadget will work with any controller.

If you are actually building a CA clone and not just a speedo/wattmeter, get the CA V3 shunt instead, it has the throttle and speed sensor wiring broken out and run to the CA-DP connector for you.

In the end, you can break down and buy a CA and just plug it in if things don't go well for your project...

 

[isd88]

thank you for help but:

1)i m not going to build a clone, i m mechanichal engineer and i wont never be able to build something similar

2)I have a folding bike that is carried on the train 4 timet to day with any kind of weather and maybe left 8 ore more hours without any cover exposed to the rain

3) i have the clarence to put just a single box 150*70*40

So i'm modding a standard kelly controller KBS 72051. I don't like trapezoidal wave controllers but i bought this one few years ago. Since the space constraints and the problem due to the water i will put everithink in the kelly's box and seal it with potting compound. No wire except power, front light, handle bar, motor, rear light will be exposed. I want a bike capable of a river ford (up to the batteries, 50 cm) without any damage.

 

[teklektik]

Well - all very interesting although since the external shunt is about as big as 2 packs of gum and is completely potted, I'm not sure I see the objections regarding size or waterproofing. I do see the point if you don't want an external shunt - which is an entirely different objection...

As far as not making a CA clone, you did say you were making your own CA, so ????

Even the old V2.2x CA is not just a wattmeter and needs a speed input and a throttle signal/override output to achieve even the simplest CA-like functionality, so using a similar connector to transfer the same controller signals seemed advantageous, but that's just me....

Anyhow - not important - as long as you get the signals you are looking for, regardless of splitting hairs over what you call the gadget. Best of luck with your endeavor.

 

[fechter]

There's probably a surface mount shunt resistor on the board under the black goo. If you trace the negative battery wire to the source legs on the low side FETs, the shunt will be somewhere between them.

 

[isd88]

i will check. this controller is very different compared to standard china brushless controllers, it is full of discrete ics!

 

[bearing]

I have read earlier on the forum that Kelly does use the low side MOSFET voltage drop to measure current.

 

[fechter]

I have read earlier on the forum that Kelly does use the low side MOSFET voltage drop to measure current.

That could explain why there is no shunt.

 

[circuit]

I have read earlier on the forum that Kelly does use the low side MOSFET voltage drop to measure current.

That could explain why there is no shunt.

And so many opamps. So my guess was correct after all.
I was experimenting on such technique some time ago, but ditched it.

 

[bearing]

I have been researching it as well. The main reason I didn't like it is because Rds_on isn't static, and it's hard to compensate for the variables.