UPDATE...
Well, I think I'm finally happy enough with the performance of the latest test unit that I am planning on doing the first run this coming week.
There was an issue that was complicating the testing, which made it very difficult to make sure everything is operating properly, before hooking up the cells/packs. As it turns out, it doesn't really affect anything, with the cells connected, but without the cells, the shunts wouldn't come on. I could test each shunt individually, and they all worked fine, but even if I just tried to double the voltage and test two cells, the shunts stayed off. This is do to some sort of oscillation that was occurring, without the cells connected to slow things down. This made it pretty hard to test the control circuit operation,especially the current measuring section. The fix is to put a small cap (.1uF...) across pins 2 and 3 of the LM431s.
The only other issue we still need to address is power for the fans, for the "boxed" variants. The 30mm fans that fit in the box are 5V, so to run them off the full pack voltage requires a pretty hefty power resistor, or resistors, which can get hotter than the shunts. We can't run them off the existing linear 12V regulator on the control board, because it can't supply enough current. This only applies to the boxed variants, not to the "stacked" versions, which can use the larger 12V 40mm fans. What we are looking at for a solution is to replace the low-power linear 12V regulator, with a switched version, capable of driving the fans as well as the control circuits. This will come in the next "minor" revision. In the meantime, I will go ahead and release the stacked version this coming week.
Another variant I'm working on, which will be more applicable for LiPo-based variants, is separating the BMS elements into separate units. There will be updated LVC/HVC boards that can be embedded in the packs, a separate "simplified" control circuit board, that can fit inside a 1" x 2" x 3" extruded aluminum box. Finally, the balancing shunt circuits are in a larger separate box, with four fans. Depending on the number of channels, different sized boxes are supported. The largest, which is 1.7" x 3" x 6", will hold up to 36 channels. Anyway, the reason for the "separates" is that with LiPos especially, you don't always need to balance the packs/cells, so it may not be desirable to have a full BMS, with the shunt circuits, mounted permanently on the bike. Many times, I just want to do a quick charge, without the balancing, but I do still want the protection of the HVC and "throttling", so that I can make sure no cell goes over 4.15V. Up till now what I've done is just use the whole BMS as an external charge control, and occasionally use the balancing as well. I was never that satisfied with this solution, especially when balancing is not required, which is most of the time, so i took a single control board, and stuck it in a 1" x 3" x 4.7" box, and replaced the LVC-only boards in my packs, with combined LVC/HVC boards. The two-wire opto output simply plugs into the control board, which is basically in series, between the pack and the supply/charger. This has been a better solution, but I knew the control board has a lot of logic that is simply not needed for a standalone charge controller, as opposed to a full-function BMS controller. What I've done now is do a "simplified" version of the control logic, on a much smaller board which now fits in the small box.
Finally, one additional variant of the above "separates" approach, is to use the nifty CellLog units in place of the LVC/HVC boards. These can't drive the charge control logic directly, as the alarm outputs are not isolated, it is fairly easy to use the alarm output to drive an opto output, just like the cell circuits. I got three CellLog units, with my last LiPo order form HC, to play with, and have a working opto proto board.
Anyway, progress... :wink:
-- Gary
