Current Shunts (Voltage Clampers)

        While they offer quite a few benefits, the potential for damage to VRLA/AGM style batteries is certainly well within the realm of the average EV owner.   Batteries can be damaged either by driving the voltage too low, and reversing cells, or by overcharging and venting a cell.  The former can be avoided by monitoring the pack and driving judiciously; The latter is harder to do when charging in a series string!  Of course each battery is different, has a slightly different charge requirement and capacity.  And I just got a new (well, used) set of Panasonic 60Ahr batteries… So to avoid overcharging a particular battery, and to keep the pack more in balance, I decided to build a device to limit the voltage at the battery maximum point by shuttling current around to the rest of the string.

        There are a few ways to keep the voltage below the maximum for your battery.  The two main ones considered were based on a power resistor, and a power transistor.  The former is widely in use, either with light bulbs, power resistors or some combination. Advantages are inherent current limiting of the clamper circuit and a nice curve; Disadvantages are that the power is dissipated as heat. The second solution is to simply shunt power across the battery using a power transistor.  Advantages are efficiency, both in terms of most of the power going across goes into another battery and little heat, and compactness.   The disadvantage is that your controller has to be intelligent- both during charging and regen.  The Hughes system is well suited for this approach, and the thought of moving all that heat from the power resistors or light bulbs out of the battery box does not appeal to me- nor does a tangle of wires from each battery out of the box.

        The build in charger puts about 6.5-7A into the pack near the end of the charge cycle.  Initial design has a zener diode controlling a 10A Darlington.  The set point is through a resistor network, including a potentiometer to make small adjustments.  I have mine all set to 14.7v- which is the 70F value for this particular battery type. These are always connected to the battery, which means that there is a small draw at all times.  Fortunately, the idle current draw is on the order of milliamps, so I don't worry about it.  However, if you were to let the truck sit for months at a time, a refresh charge would be required every so often.

        An optocoupler is included in the circuit to allow me to talk to a charger or BMS.  Right now I am just using the clamper, as I do not know how to interface to the charger- but with the currents being so low, I don't really need to.  If I were to use my PFC charger, these lines would be hooked up to the regbus to tell it to cut back when any of the regs went active- else at 4.6kW/12.6A I would be blowing Darlington, fuses and burning the sinks up something fierce! No heat sink, but in place on the battery is shown below:

        Design was done through Express PCB's free software.  It of course is only useful for their ordering process, but is very easy, relatively intuitive, and within an hour I had my first draft of the circuit.  The second draft took into account the size of the cutout on the battery, the clamper fit within the confines of the little rectangular space.  The power side hangs over the side of the battery, and is attached to a ~6" piece of aluminum angle as a heat sink.  Ordering the boards was simple, and within three days I had 28 boards in my hands for $163.  All of the components are from Digikey; I ordered enough parts for 52 boards (was trying to fit two strings, due to the size of the Panasonics that didn't work out!) for a total cost of $400.  So, for about $600 or so plus a lot of spares I have around the house (fuses, 10 gage wire, crimp connectors, elbow grease) I've got a set of voltage clampers.

        The night the boards arrived was also the night that the power went out on my block.  But I needed to get them done in time to be installed over the weekend with the new pack- so there I was, 10PM at night with a candle, bending leads into place.  When the power came back on at midnight, the soldering iron made it's appearance and the first batch of clampers was made.  This was an exercise in willpower and patience- doing one or two boards is easy, when you are doing 28 it gets old in a hurry.  I can't imagine doing 100 like Victor did, no thanks.  One of the local EV guys mentioned an electronic assembly machine- if I were to do it again I would definitely get a quote on having them assembled.   All in all, assembly took 8-10 hours. 

        Initial set points were made before installing the zener diode and Darlington by borrowing a variable power supply set to 14.7v and tweaking the 5k pot to give 2.5 volts at the resistor node attached to the signal lead of the zener.  Then the zener and Darlington were installed in the board, the heat sink and wiring attached, and the whole mess put on a string sitting in the garage.   The PFC20BE charger I had came in very handy here- set the current reasonably low (didn't have an ammeter hooked up, probably should have), set the voltage set point so that all of the lights barely come on.  Then with a voltmeter and screwdriver adjust each module for 14.70v  

        Installation onto the batteries was simple, I used foam with double stick tape to attach them to the sides of the battery.  They are physically held in place by the wiring to the posts, as well as the top and center plastic insulators, so I don't worry about them going anywhere.

        Testing (time to verify everything!).  With the Dolphin parameters unchanged, here's a picture of the green LED's going active as the batteries reach their set point.  Kind of reminds me of lightning bugs back in Chicago.

        Regen.  There has been some concern about the regen parameters, basically that under hard regen you could push a clamper beyond it's limits and blow it up.  While this is possible (if you had all discharged cells except one fully charged one) it is probably indicative of a larger problem.  So far I have had no issues, besides, that's what the fuse is for, right?  In the end, I feel that the system keeps the pack in better balance that it would have been otherwise, and so since it's better matched you don't have so many problems with an uneven pack.

Back to the S10 page
Back to the Home page