Debugging the EV

While I was away in Europe our EV stopped working. My wife was sitting at some traffic lights, the lights went green but nothing happened. So my wife and my daughter had to push the car home, fortunately just a few 100m from where we live. Much to the amusement of the 4 year old sitting in the back! Worse, they were forced to use our other petrol card for a week until I returned home. Once you’ve driven electric, you find yourself going to some lengths to avoid using petrol cars.

The fault was demonstrated to me after I hopped off the plane yesterday. The car would move OK, then after applying the brakes it wouldn’t move anymore. No power to the electric motor. Also the vacuum pump sounded “slow”, or under powered. So I suspected a 12V system fault. After switching car off and jiggling a few wires around near the 12V battery it would run again.

However after a 34 hour trip from Stockholm I wasn’t at my mental peak, so I decided to work on the car after some rest. Australia is a long way away from anywhere, but I find the trips are getting easier with practice. Movies on the plane help pass the time, and I always find nice people to talk to. After two wonderful weeks in Europe attending Cebit in Hannover, visiting Berlin and Stockholm it seems strange not having to put on a big jacket to go outside and I keep wondering “where is the snow”? More on the European trip later.

After a good nights sleep I looked at the EV again. The 12V battery voltage looked OK at 12.5V. I switched the car to “on” (which makes all the systems live) and the battery voltage went to 13.8V as the DC-DC converter started charging it from the traction battery pack (in an EV the DC-DC converter takes the place of an alternator, supply 12V power while driving and charging the 12V battery).

So I poked around with my new clamp ammeter, fast becoming my favorite toy. Makes current measurements so much easier. This means measuring current can be part of basic, every day diagnostics rather than being a special event that requires breaking a wire. Makes current measurements as fast and easy as voltage which leads to another “dimension” in our visualization of circuits. I bought the clamp ammeter for tracking down to phantom loads on my household power, but have found many other uses, like checking the EV charge current.

The wire from the battery to the 12V system of the car showed 0.7A, which was about right – this powers instruments, the dash etc. However the current coming from the DC-DC converter read 0A. When I checked the battery voltage again it was 12.0V and dropping.

So I measured either side of the lug connecting the DC-DC converter to the battery and it had several volts across it, and when I wiggled it could feel it was loose. Bingo – bad connection to the battery, meaning 12V was failing, eventually to the point the main contactor wouldn’t close so the car wouldn’t move. With a gentle tug the lug came off:

Bad solder joint or maybe fumes from the battery affected the joint. The diodes you can see prevent the battery from discharging into the DC-DC converter when the EV is off. I had a bunch of schottky (for low voltage drop) diodes handy so used several in parallel.

Five minutes to diagnose and one minute to re-solder the bad connection. Zero cost. Not bad for 18 months of EV driving. My electric wiring skills aren’t great, and my EV is a one off, hand made prototype, so this sort of problem wouldn’t occur in a factory EV.

4 thoughts on “Debugging the EV”

  1. Actually, if I can see correctly (a big if – stupid breaky glasses), you’re using crimp lug connectors but still soldering? Solder is bad in a car – vibration. Crimping is good, because copper is flexible, but solder just crystalises and breaks when you try to wiggle it.

    Find a good crimping tool and use it.

  2. “I had a bunch of schottky (for low voltage drop) diodes handy so used several in parallel.”

    This rings alarm bells for me. Parallel PN junctions, particularly in schottky devices, are not such a good idea. Slight differences in the forward voltage of parallel diodes will cause one to carry more than its share of the current, which will make it hotter so its Vf drops more so it gets hotter etc until it dies. Then the next one dies, and the next.

    In bipolar transistor amplifier design we’d usually put a very low value (sub 1 ohm) fixed resistor in series with each parallel junction to artificially reduce the Vf as current increases and I reckon you could do something similar here.

    Oh, and +1 on Tim’s crimping suggestion. A good crimp (like a good wire-wrap) results in a series of micro-welds between the lug and the wire so you get mechanical strength along with low resistance. Some suggest also soldering for a ‘belt-and-braces’ approach but I’m not aware of anyone doing that in a production environment.

  3. Thank you Michael and Tim. Yes I agree crimping would be a better approach.

    Michael yeah I was worried about the parallel diodes but so far so good after 18 months. They are all touching each other (thermally coupled) which might help. As with the soldered lugs it’s not a production approach, a DC-DC converter that doesn’t accept backwards current would be a better.

    – David

  4. It would be better to put the diodes in the DCDC converter so that they can be high current and heat-sinked.

    In my experience, diodes and restors bult into leads is never a very successful long term solution. These are usually the thing that fails.

    The converter will have a set of diodes in it already. These things usually use two diodes and a centre tapped transformer. The output of the diodes is fed into a capacitor filter, often with inductors to reduce electrical noise and then to the output terminals. I would expect that leaving the DC/DC on the battery would power up only the voltage comparitor and not draw very much current, probably a few milliamps.
    A relay could be used to disconnect it from the battery when the ignition is off. The DC/DC can probably supply 50 amps or more, so the wiring sizes (gauges) and relay or diode capacity is important.

Comments are closed.