help please with design problem

[KTP]

Hi guys,

I have a bit of a problem.  I want to build up a light battery to use with a Minn Kota Riptide saltwater 55 pound thrust trolling motor which has a digital maximizer control instead of using speed coils (in other words, it pwms the motor instead of dropping voltage in resistance coils like the cheaper 5 speed models).  The problem is I want to use some of the many Lithium Manganeese sony US26650VT cells I happen to have (4.1V fully charged, 2600mA, capable of 50A discharge) to build up a pack.  The LiMn battery voltage drops from about 4.1V full to about 2.5V empty (don't discharge below this voltage) so if I make a 20 cell ~200Wh using 5p4s pack it will have a fully charged voltage of 16.4V and a fully drained voltage of 10V.

The problem is I cannot get Minn Kota to tell me if the Riptide 55 pwm electronics would be tollerant of 16.4V, and the electronics are all potted (for salt water protection).  I assume since the trolling motor is normally connected to a battery which may be charging via an alternator at something around 14.5V that the components are at least rated that high.  I could build the pack as 6p3s with a 12.3V full and 7.5V empty, but I don't think the motor would work well below about 10V.

Since the battery would only be at 16.4V for a few minutes until it has drained somewhat, I could run the output through a series string of 3 honkin diodes, dropping the output to under 15V, and then maybe have a circuit to bypass these diodes after the unloaded battery voltage has dropped below 15V?  I guess this way I would only have the wasted power (50 amps max times 1.8V = 90 watts) for a few minutes, then have a low Rds on mosfet bypass the three diodes.

Any suggestions for a better way?

[Zad]

Any idea what frequency the PWM runs at? I'm thinking you could feed the output through an inductor and drop the voltage using impedance rather than resistance. The power delivered would be the same, just at a lower peak voltage.

[rossmoffett]

Test at 7.5V and get back to us.  If it's the easiest way, don't just trash it because of unknown variables.