Current Limiting Circuit

[jamesv]

Hi Guys, I've finally decided to go for my electric bike build for a school project as I can use the workshop tools. My setup will be an alien C80100 rewound to 40kv, a turnigy dlux 160 amp HV ESC driving the freewheel cranks. I know my main problem is going to be limiting the current as the RC controller wont. Have any of you seen this circuit

from here http://www.recumbents.com/wisil/tetz/CurrentLimiter/default.htm
The circuit uses a shunt that drops 85 mv with a 10 amp load, as I need it to limit the current to around a 100 amps to be safe could I just replace the shunt with a 100 amp 75 mv and the just use the pot to fine tune it or do I need to change more components?
Thanks
James

 

[johnjcbs]

Looks like Fechter's circuit.He posted it about four or five years ago.

 

[johnjcbs]

I came across that circuit last year,looking for the same thing as you.
http://endless-sphere.com/forums/viewtopic.php?f=28&t=7675

 

[jamesv]

Thanks for finding that John hopefully he will come across this and be able to give me a hand. I get the general idea of how it works but don't completely understand it which is why I am having trouble working out how to make it work for my set up. Did you end up using it? I haven't found another option.
Thanks
James

 

[teklektik]

The easy (but more expensive) solution would be to use a CA-V3 and ditch the servo tester and special electronics. The V3 will provide the throttle-to-PWM conversion. That would give you a standard throttle, current limiting, monitoring, and other stuff - V3/ESC hookup is described in the V3 Guide.

Otherwise - your proposal to change the shunt sounds okay, but you may need to diddle the feedback cap if the current limiting oscillates. You might PM John to look in on this thread if he gets a chance to help out...

 

[alsmith]

why not pm him and ask him directly? Hopefully he'll have the time to help.
Up on the top bar it has a little Members button- click it. There's lots of members and the list isn't in any useful order except he's been here so long that he's on the first page.

 

[fechter]

Ah that's the circuit I did for John Tetz years ago.

Yes, you can get any current you want by changing the shunt and adjusting the pot.
One problem when using a RC controller is the latency, or delay time, between a control input change and the motor response. This delay raises hell with the loop stability and makes it prone to oscillation. Tetz had to slow down the response of the limiter to avoid oscillation (done by adding the 5uf cap between the input and output of the amplifier). Slower response will reduce the tendency to oscillate, but also may allow significant overshoot. Reducing loop gain (the 100k resistor across the capacitor) will also improve stability. Oscillation will cause a surging in motor speed. Once the current drops below the limit, the circuit is no longer part of the loop and can't cause oscillations.

I tried all kinds of approaches to stop the oscillation and never came up with a good solution. The same circuit used on a more standard sensored motor controller works great since there is much less latency.

 

[jamesv]

Thanks for the reply fechter it has helped allot, and thanks for designing that circuit! I am aiming to just use the limiter as a safety device to stop the controller frying itself I think I will keep it like the first design as to reduce the 'overshoot' you mentioned. From your vast amount of experience how much power do you think is necessary to cruise at 35 mph (I am pretty light 60kg) the batteries should only weigh 12kg even if they are lead acid.
Thanks
James

 

[fechter]

You can play around with the value of the capacitor and 100k resistor to tune things in. Every controller/bike/servo tester combination is going to behave a bit differently.

The idea is to make it respond as fast as possible without causing oscillation.