Paul & Sabrina's cheap DIY 144v motor controller

[ev59rag]

Renny,

Too late... Paul beat me to it... lol...


C

[MPaulHolmes]

hahaha!

[bga]

I've seen these things at auto supply shops (Repco in Australia) they're designed to replace or augment a coolant pump, although the one I saw looked too feeble for big motor. I have thought that these would be ideal for EV use and in the correct pumping power range to cool a motor controller nicely. Presumably, they don't mind having hot water in them.

Another possible source is a marine water pump, such as the 'Whale' type.

My only concern with these sorts of pumps is to make sure that the plumbing is arranged so that the pump will prime reliably.

One fellow in WA uses a 4 Litre plastic (oil?) container on a Zilla 1K controlled system for his 'radiator' and has found that to work well.

[edown]

Originally Posted by NiHaoMike

For those making water cooled controllers, instead of using an inverter with a fountain pump, what about a bilge pump with a simple PWM controller?

I chose the fountain pump (actually it's for a parts washer) for my water-cooled version of Paul's controller, because it is maintenance-free - no brushes, silent, last forever, and it was inexpensive (I already had it... ). Also, I was planning on using the inverter for powering some transformers to create some isolated 12 volt DC power supplies for my digital instruments - so using the inverter for other functions was relatively simple.
I'm also thinking of running my oscilloscope in the car during a road test.

EDown

[mcudogs]

I am looking at including a traction battery monitor on the controller board.

Here are 2 options

1. Monitor the whole battery pack voltage using a linear optocoupler cct fed to an a/d input of the atmega. This would be used to shut off the main contactor or reduce motor current when minimum pack voltage is reached. This would not take into account low voltages on individual cells.

2. Use a commercial battery management system with an alarm output that operates a digital input on the controller to reduce motor current when any cell reaches it's minimum voltage.

What does everybody think?

[adamj12b]

Why not do both? The pack voltage sensing could also be used for automatic pre-charge.

Do you have a picture of this linear optocoupler circuit you mentioned? I have done some looking but haven't come up with anything.

-Adam

[NiHaoMike]

Originally Posted by mcudogs

Monitor the whole battery pack voltage using a linear optocoupler cct fed to an a/d input of the atmega. This would be used to shut off the main contactor or reduce motor current when minimum pack voltage is reached. This would not take into account low voltages on individual cells.

There's an implementation that does not require a linear optoisolator. Simply have a small capacitor charge from the battery through a resistor and connect a diac circuit that periodically discharges the capacitor. Use a regular optoisolator or a signal transformer to couple the pulses and then measure the frequency.

You can build a battery balancer using either charge pumps or a multiwinding flyback transformer.

[MPaulHolmes]

Did everybody get all that? hehe

[mcudogs]

Application note for a linear optocoupler circuit is http://www.clare.com/home/pdfs.nsf/www/AN-107.pdf/$file/AN-107.pdf. The problem with using this is that you have to derive a low voltage supply from the pack voltage to run the op amp on the battery side.

NiHaoMike, so the diac discharges the cap through the LED of the opto when the voltage across the cap reaches the diac breakdown voltage ~30vdc? The time taken to charge the cap depends on the battery voltage. Therefore the frequency of the pulses through the opto is relative to the battery voltage.

That is certainly a simple circuit. Do you have a working circuit we can borrow?

[NiHaoMike]

Originally Posted by mcudogs

Application note for a linear optocoupler circuit is http://www.clare.com/home/pdfs.nsf/www/AN-107.pdf/$file/AN-107.pdf. The problem with using this is that you have to derive a low voltage supply from the pack voltage to run the op amp on the battery side.

NiHaoMike, so the diac discharges the cap through the LED of the opto when the voltage across the cap reaches the diac breakdown voltage ~30vdc? The time taken to charge the cap depends on the battery voltage. Therefore the frequency of the pulses through the opto is relative to the battery voltage.

That is certainly a simple circuit. Do you have a working circuit we can borrow?

Start with a simple neon bulb blinker, replace the neon bulb with a diac (lower voltage for better linearity), and add an optoisolator or pulse transformer in series with the diac. The battery with the series resistor can be approximated as a current source, which charges the capacitor until it reaches the diac trigger voltage, causing it to discharge the capacitor through the transformer or optoisolator. The frequency would vary depending on the battery voltage.