Performance Testing New Cells

LiMn 15ah pack under 3C (45 amp) discharge, shown in thermal

LiMn 15ah pack under 3C (45 amp) discharge, shown in thermal

Lately we've been working on the performance of our battery packs. Though a large component of this performance comes from the cells themselves, the bus wire material and thickness, quality of welds and solders, number of bus bars and resistance of the BMS all contribute to the overall peak discharge performance a battery can supply.

In nearly all cases, though, the way this energy is lost (the sign of decreased performance) is heat. An "ideal" battery pack would not get warm to the touch under any discharge conditions, and the voltage would stay "flat" through the entire discharge and only drop right as the pack has run out of charge.

Such a pack doesn't exist, and so we can only improve our existing designs to "move the bar" in the direction of better performance. Lipo battery packs are pretty good in terms of not producing heat, and LiFePO4 packs have very flat discharge curves, but Lipos lose voltage as they discharge and LiFePO4 only packs about half the capacity of Lipo.

This is why we choose to use LiMn cells - while they have a voltage curve like Lipos, it's much more manageable and they have very low internal resistance so less energy is wasted as heat during heavy discharge.

Of course, the best way to evaluate performance is with testing. :)


Above, the pack was discharged continuously at a 3C rate, meaning the pack when from fully charged to dead in 1/3 of an hour, or 20 minutes. The pack hit about 50C, which is indicative of a heavy but manageable discharge rate.

Cells individually can be heavily discharged when they have the ability to dissipate their heat in all directions, but put a bunch of cells in a pack, where the heat can't be dissipated in all directions (because for some of these cells, there's other hot cells on all sides except top and bottom) and you have to derate your battery's discharge rate. In this case, our new high performance cells are actually rated at 9C continuous discharge (a full discharge in 7 minutes) but due to this heat dissipation issue, we have to derate our cells to 3C. 9C is still very much a possibility for these cells to handle, but only on a time frame that doesn't result in a lot of heat getting built up in the pack.

If my design tests well, I should see the components of my pack maintaining about the same temperature as the cells themselves - meaning the resistance of my bus bars, wiring and BMS don't produce as much heat as the main "action" of the cells, which is the chemical reactions going on inside those cells.

And the test went well! Under thermal, the cells are indeed the main contributors of heat. The BMS has transistors that get hot too, but in this image my 45 amp discharge actually pushing the BMS a little past its rated limit (of 40 amps), so I expected some heat at the BMS. These cells far outperform our last cell manufacturer's best product.

Engineering a more efficient ebike

I'm no stranger to building large battery packs - considering the largest packs we build total around 8.3 kWh. I'm taking a new approach to our next prototype ebike build, though - building an ebike as efficient as possible. We built a 36v 15ah pack special for one such ebike. It will probably actually have three wheels, and be designed to lay down on to facilitate getting the rider out of the airstream as much as possible.

36v 15ah Blue Line pack - we would have named them Odyssey batteries except there's already a company with that name. We like Blue Line, though.

36v 15ah Blue Line pack - we would have named them Odyssey batteries except there's already a company with that name. We like Blue Line, though.

The battery pack is a nominal 577 watt hours (it's actually 37v nominal, 15.6 amp hours but we typically underrate our packs) and weighs in around 7 lbs. This battery, on a trike like Odyssey's Mk. 5, would be good for a range of approximately 25 miles with an energy economy of 20-25 watt hours per mile, but I am aiming for an economy of 10 watt hours per mile initially (a 57 mile range) and upwards of 5 wh/mile (114 miles) after all the cowlings are in place and the vehicle has been tuned to be as efficient as possible. It's going to look like a streamlined street luge, with many of the components onboard 3d printed.

One of the biggest considerations right now is the drivetrain. I've already settled on using a three phase motor (called "brushless DC" because the motor operates as if it's a brushed motor but it's commutated in the controller externally) instead of a brushed motor, but I haven't decided on a method of power transmission yet. I'm considering friction drive - while the transmission of the actual power may be less efficient than a belt drive, the motor can be retracted the second the torque transfer is gone. If a "pulse and glide" concept is utilized, the duty cycle might only be in the 25% range - leaving absolutely no transmission drag 75% of the time, and a very light weight required for the "transmission" on top of that.

I'll write more blog posts as this build continues.

sustainable energy for manufacturing

We recently moved our shop location to a much larger, more well-suited building for our manufacturing. We've been looking at ways to increase our sustainability. I found these panels online recently:

Four 300w 36v solar panels, one Mk. 5 Experimental for scale

The really cool thing about these panels is that they put out 36v @ 8 amps each, which makes them directly compatible with the 36v Mk. 5's battery. That totals 300 watts, and assuming 5 hours of sunlight a day, one of these panels can supply the energy for 30 to 60 total miles of range to a trike, on a daily basis. I foresee these panels actually supplying the shop with power, but it's still a feasible idea that one of these panels could be set aside specifically for charging a 36v trike battery. If this is something you, as a customer, would be interested in, let me know because it will influence how I go about using these panels.

Beyond this, if we end up with two more sets of panels like these, our entire production will be 100% run from solar power.

Madison Trip 1

I took a trip out to Madison to visit a friend from high school, on one of our Mk. 5. I took the long route, though - Starting in West Bend, south on the Eisenbahn, over to the town of Cedarburg, about 30 miles south on the Ozaukee Interurban Trail, then west about 100 miles on the Glacial Drumlin Trail straight to Madison. I ran into a young DNR officer on the way there (he's about my age and I'm 25), who was patrolling the trail - he'd never seen an electric trike like mine before, and we talked for a solid half hour on the possibilities of patrolling on a trike. I charged overnight in Madison and the next day, took the same route in reverse, this time with my cameras rolling. This was on my reverse route, rolling past Brady Beach in Milwaukee, on Lake Michigan. I was about 6 kWh into the total trip's discharge at this point, and maybe 120 miles into the day's ride:

The GDT is a gorgeous trail, though portions of it are unpaved - but even these portions were nicely graded, allowing me to ride at my speed of preference. Downtown Madison was quite the adventure (though Milwaukee is bigger, I have experience riding the Mk. 5 in Milwaukee) and I got lost a couple times, so this portion of the ride turned into more exploration than transit.

In total on this ride, I used about 8.2 kWh (a little over one charge on the 7.2 kWh Super I was riding, counting both directions) and drove just over 300 miles, averaging between 15 and 30 mph on the trails and keeping up with traffic in both Madison and Milwaukee.