Thanks all for the input on my Denver thread- please keep it coming.
I went for a bike (bicycle) ride today and put some thought into the cell balancing problem.
I work occasionally with a battery pack manufacturer in Denver (Nexergy, Inc.) and have a little background on cell balancing.
The tall stack of cells in the Vectrix is indeed tricky. However, I think a reasonably useful active cell balancing module can be created on a per-cell basis, i.e. 102 little circuit boards, each one across an individual NiMH cell.
Components would include, 1.2V->3V boost converter for powering the electronics, an MSP430 ultra-low power MCU, voltage reference, MOSFET, and power resistor. The quiescent current for such a unit should be well under 20-30 uA, and in the noise compared to the self-discharge rate of these cells. This would be a tiny board and I think it would fit right on top of the cells, based upon the pictures I've seen.
The idea is that when the cell voltage exceeds ~1.40V (or whatever the forums deems most appropriate) the MOSFET shunts about 1.5A-2.0A of charging current around the cell through the resistor. This energy will still be dissipated as heat, but at least it won't be going into the cell and boiling the electrolyte. The MCU can PWM modulate the MOSFET to achieve lower average shunt currents if needed.
This technique won't balance a stack at once- instead during charging the full cells will have their current shunted, preventing heat/electrolyte damage, while the weak cells continue to charge. Over time this should bring the stack into balance, and help maintain balance. A severely unbalance stack probably still couldn't be brought back with this, but would probably still be helpful.
The main benefit to this setup is backwards compatibility with the existing charger and cells, assuming the board is small enough to nest in between the terminals, etc...
Has this been discussed before? I've been pouring over this site for a day now and haven't seen anyone try to address this.