Two and a half years ago, I replaced two of the aged out NiMH battery packs in our Dive-X Cuda 400s with 12s1p LFP packs; ‘Alpha’ and ‘Bravo’.
A year ago, after our Cave DPV class, the third scooter was rebuilt as well. As it was our dedicated ‘tow’, it was dubbed ‘Tango’. When cells for the second build were sourced, Headway had refreshed their 40152s series from 15Ah at 50/150A to 17Ah at 30/100A. While the newer cells were slightly heavier by ~5g, their performance difference when short cycled to 3.5v charge was nominal.
Last trip down to cave country, Alpha and Tango started acting up near end of dives. Cell testing showed two failed cells. One was intermittent voltage droop, the other was a complete failure with 0v on terminal. Cells were identified and replaced with shelf spares. This did cause a pack mismatch in cells, but again, these packs are not being cycled from 100% SOC to 100% Discharge.
Reference Voltages
Because I’m sick of doing the math. Cells are rated from 3.65v to 2.5v, but are short cycled 3.5v to 2.91v on purpose to extend life. Actual cycling in practice is even shorter at 3.5v to 3.2v.
SOC Cell Pack
100% 3.65 43.80
90% 3.60 43.20
85% 3.55 42.60
80% 3.50 42.00
float 3.40 40.80
float 3.30 39.60
- 30% 3.20 38.51
- 60% 3.17 38.14
-100% 2.91 35.00
Bench Testing
- Charge pack to 42v, 3.50v/cell at 5A, halt at 0.1A CV.
- Disconnect BMS and record cell voltages.
- Battery is placed on ~8 ohm load to ~35v, .1C discharge rate.
Baseline
Alpha seems as expected.
Bravo pack got used for testing and calibration, thus the lower start voltage. Was not looking for a full burn test, rather the late discharge curve, so did not recharge and reset.
Tango is… not where it should be. BMS triggered a cutoff well before 35v. Charge cycled it, and it’s BMS shut off well before CV charging completed as well. Tango is the pack in our tow scooter, so it really had not seen much active use over the past few months.
Seems Tango drifted out of balance, and the BMS can’t move enough current (200mA) to bring them back into balance without letting it cook for days. End of discharge voltages confirm:
cell: 1 2 3 4 5 6 7 8 9 10 11 12
alpha: 2.89 3.18 3.10 3.06 3.13 3.12 3.13 3.13 3.15 3.10 3.08 3.10
bravo: 3.13 3.12 3.18 3.13 3.18 3.16 3.17 3.04 3.34 2.87 3.17 3.15
tango: 3.26 2.88 2.91 3.27 2.86 3.27 3.27 2.88 3.26 3.29 3.27 3.29
Alpha is extremely low on 1, Bravo is low on cell 10, and Tango is all over the place.
Seems I must pay the Stupid Tax. Again.
Active Balancing
Neey clone 5A Active balance boards were ordered and installed. Operation confirmed via stepping through charge voltages and checking per cell settling voltage. Travel logistics prevented a full burn down at the time, so all three packs were put back into rotation to shake down the new chassis.
After roughly two dozen cycles later, I had time to bench test again. All packs were charged at 5A to 42v / 3.5v per cell, with a CV charge cutoff at 0.1A, and placed on load.
Seems that did the trick. Still 2 Ah short of the rated 17Ah on Tango, but to 10 Ah is within tolerance. It’s not optimal by a long shot, but it gets the job done.
The Cudas needed roughly 20g of ballast weight removed from the nose to address the change from BMS to Active Balancer, and the associated wiring changes.
Dive Planning
Even the ‘lower power’ 17Ah cells with a 35A continuous discharge have plenty of grunt to drive a DPV that will put you 3,000’ back into Ginnie in 20 minutes. That’s roughly 150 fpm vs the cave, and closer to 200 fpm vs flow. Scooter datalogs show they are consuming about 3-4Ah per dive with 40 minutes trigger time. Seems a good ‘rock bottom’ trigger time is slightly short of two hours.
If/when dive plans require more capacity, a higher state of charge to 43.8v is available. Bench tests to confirm ‘full’ capacity will need to be done before executing a plan depending on that higher capacity.
Lessons Learned
- Advertised != Reality. Double so for cheap kit. Stay off the voltage knee unless its a bail out situation.
- Kit breaks. Keep up on pre-dive checks and maintenance.
- Use it or lose it. Bring the tow scooter (tango) and cycle it. Having a pack sit idle for months at a time is not a good thing.
- Cable Lacing is really hard to beat for complex routing of wires in tight spaces.
