OK, enough waiting - it’s driving me nuts. The zero I did bid on had the bids zeroed out (thanks Copart), and then someone decided they REALLY wanted a salvage title motorcycle (at nearly a new price after all the fees). So that’s out. And that’s sort of a good thing, as in the last couple of weeks I realized why some people say they slow down after going down to about 60% charge. It’s a voltage problem. Zero choose to use 28 cells in serial, which at a charge of 4.15 volts is fine for the 102 volt motors. But at around 50% that pack is down to 100.8 volts. By the end (being 3 volts per cell) it is down to 92.4 volts. Not horrible for around town, but not at all ideal for a racing vehicle.
So, new tact. Things I can actually just buy (including hopefully the pouch cells from Hunan China (yes I’m serious). I’ve decided to take a lowish-cost learning approach and first aim at 206’s. This means I should only need one pack for an entire race day - without charging. And best yet, that pack can also be used with a software adjustment on the controller to move up to 100cc TAG cart speed, just at about 19 minutes per charge VS 42 minutes as a “206”. So I will need a second battery or I’ll need to charge it between sessions. But since there is a 1000 cell minimum order - probably I’ll just make 2 or 3 batteries to take with me when I try that. I will borrow Zero’s cell packaging design - but then add a heat sink to keep them nice and “cool” since I ironically have more room.
Battery - 34S1P pack of 32Ah pouch NMC li-ion cells. Lots of punch without a lot of connections, and “lightweight” - at least as far as batteries go. These won’t catch fire like say LiPo cells, they just go bad when dead shorted or over/under charged.
BMS - Zeva. Well tested parts at a fair price from down-under.
Motor - Montenergy ME1115. A fairly simple air-cooled 96 volt motor. Good torque with a reasonable amount of HP as well. Way over-rated for the 206 application, so it should stay cool without much issue when running it at full voltage.
Controller - Kelly 200 amp 144 volt controller. This way my battery pack will always be above the motor voltage and not have to fall off towards the end like the Zeros. Plus unlike the Sevcon series, it doesn’t require a $900 cable to program with an iffy license renewal process. May or may not be heat-sinked as shown depending on how hot it gets without one.
Plus lots of odds and ends like switches, contactors, fuses, throttle pot, cables, something to record the canbus traffic, and a SOC display (not required to function - but nice for seeing how to tweak the controller settings).
Functionally arranged something close to this. Parts should start rolling in here in a week or so… can’t wait.
In “206” mode the max battery draw should be about 53A. Motor will be about 87% of that. Max current on the motor is 200A, so it will run at a little less than 1/4 max power.
I analyzed Kolme’s latest video, and found that with his 1st gen 75-7 motor with carbon fiber wrap over the magnets and a Sevcon gen 4 size 6 72-80v controller, it appears the effective efficiency under heavy load (> 70% demand) as measured is about a 77% efficiency between the power draw of the battery and the measured motor power. At lower demand the efficiency is much higher, up to about 92%.
Note that this conversion of DC to AC with efficiency losses means the motor power (voltage times amperage) is always less than the power drawn from the battery. The amperage from the battery is normally lower than the motor amperage because the battery voltage should always be more than the motor voltage.
Choosing the cells to make up a battery is of course target load dependent. If you need more power for a short time low capacity high C cells are required. If you need less power for a longer time high capacity low C cells are more appropriate. Of course adding more parallelism will allow lower C cells to deliver high amperage, as will adding more parallelism will allow lower capacity cells to last longer - at the cost of added weight.
Science! This is all very interesting and I wish I had the engineering chops to understand.
In the meanwhile, here’s an apropos song. It is motorsports correct as it begins with Mr. Dolby piloting a motorbike and sidecar. Sorry for the ADHD diversion.
How does battery efficiency manifest itself in performance?
“ How does battery efficiency manifest itself in performance?”
2 ways … The less load placed on each battery cell means 1) Less heat and 2) more availability capacity. Generally speaking, the more cells in parallel, the lower the “heat per cell” and the longer the driving time (all other things being equal)
Thank you. Do electric karts basically run full power and then fall off drastically or is it a progressive decline? I assume the first otherwise it would get progressively slower?
That mainly depends on the programming of the controller and design of the battery pack. For racing I believe the optimal mix is to have it run at 100% race power (although not necessary the max electrically possible) until it hits the minimum battery voltage, then shut down or at least go in a very low power mode. This way you know how much to use so you can judge how much to push and when (at least in Ekart races). Essentially think Formula-E on a smaller scale.
That depends on how fast the battery cells can charge (the charge C rate), how depleted the battery is, and how much power the charger can apply to the battery.
If you had a 3 kWh battery pack that can charge at 1C, and it is at 0% (the minimum set voltage - not 0!), and a 3 kW charger - then you could charge it to 100% in about one hour (as long as you can keep the cells cool enough).
A 3 kWh battery pack that can charge at 2C, at 0%, and a 6 kW charger - then you charge it to 100% in about 30 minutes (again as long as the cells are kept cool enough).
How many packs you need depends on how much energy you plan to extract, and how fast, and how much weight you are willing to put up with. It’s sort of a triangle of the following (pick two):
Speed
Distance
Weight
You can have a light pack that can go quick for a short distance.
You can have a light pack that can go slower for a longer distance.
You can double the weight of the pack and go either twice as far or twice as fast (of course not quite twice as you have to accelerate, turn, and stop that extra weight).
It’s all a trade-off. What do you want most? An larger all-day battery? Or multiple lighter batteries (at more cost). But of course if you make the pack too small, it may not be able to deliver the current you want.
Zero’s work this way. Some models have a single 28S1P pack. This is the lightest, but limits overall current. A “long” pack is 28S2P and can deliver more current, but at about twice the weight. A “monolith” pack is 28S4P and will deliver the same current as a long pack, (controller and motor limited) but for twice as long.
One other point to make about pack size… More weight is a disadvantage of larger packs but the more cells you have “in parallel”, the less strain you put on each individual cell. You also have less heat (all other things being equal). However, the points made about trade offs are spot-on. Currently, is no way to avoid the compromises stated earlier.
Cells for the battery pack (just one for now) have been ordered. I’ll be using 420 18650 Molicel P26A cells in a 42S10P configuration. This will provide a 4.6 kWh pack (using the industry std numbers) - or more realistically, 3.4 kWh in the high-drain racing situation. This will enable me to keep the end of run voltage above the peak motor voltage so that there should never be any reduction in performance during the cell use between 4.15v and 3v with a 120v motor.
I tried to avoid cylindrical cells due to the pain of connecting them up where there are both physically solid connections as well as connections that can carry the current necessary without overheating. But the large pouch cells just didn’t quite do what I wanted, and I decided against using iffy RC pack quality bricks.
What changed my mind on 18650s was a portion of a post on another forum by Spiningmagnets which showed this idea that is both simple and electrically sound. It involves spot-welding nickel strips to the heavy copper (using as much heat as necessary) and then attaching that to each cell - minimizing heat to the cell ends.
After doing amperage calculations I plan to use 0.032" (0.813mm) buss plates for most of the connections, with 0.08" (2mm) buss plates for the skinny section across the top of the top-down view.
The left side is a top-down view of the batteries and plates, the right is the bottom-up view. Bottom will be be an aluminum/UHML plastic sandwitch, and the sides and top will be UHML with aluminum “roadie” case corners for durability. Total weight with cells, bars, cables, etc should be around 56 lbs. There will be no electronics in the battery cases themselves to minimize costs for multiple packs. Power leads will be 350A Anderson connectors and cell monitoring leads will be on a multi-pin connector.
As is, this pack should power the kart at 100cc TAG race speeds for 20 minutes or at 206 race speeds for about 47 minutes. Cells should be here early next week!
That is awesome. We tried several different battery options but ultimately decided on the 18650 cells due to energy density. I also think it is smart to limit the per-cell voltage to 4.15V per cell like you are doing since it will prolong the cycle-life of the pack. Spiningmagnets is a really smart guy. I have bounced ideas off him for several years now. Your project is awesome. Keep the updates coming. Thanks. Gary - VenomKarts
One thing that has been a real pain in the design for this build is working around all the ICE-centric frame issues with the EV parts. So I got to thinking - what if I could start with a clean-sheet and design the kart from the ground up more like a real clean-sheet EV? For EV karts - why strictly stick to the old rules?
This is what I quickly came up with this afternoon.
First, widen the frame slightly to get rid of the crazy battery box(s) and put the cells within the frame on the bottom to lower the CG. Here, one layer deep sideways 18650 fit rather nicely inside a 32mm frame and perfectly center the weight. Of course I’d need a titanium skidplate or something to ensure no punctures from below, but that’s certainly doable.
Next , move the motor behind the seat. This does require moving the rear axle slightly rearward a couple inches, but then allows the motor weight to be perfectly centered.
With the motor off the side, now the seat can be properly centered - no more awkward driving position! Balance out the rest of the components (radiator, BMS, DC/DC converter, H20 pump, and motor controller - for a 50/50 left/right balance. Yes the seat will have to be slightly higher, but this is not that big of a deal and would also allow wider seats for larger drivers.
Front to back there are more cells behind the mid bar than in front, so the desired F/B weight ratio should not be too difficult to obtain. Here, the position of the driver seat F/B will make more of that difference.
Just a thought for now… but this would be nice! Let me know what you think.
Very cool ideas. There’s so many different things you can do with electric. For example, you don’t need that gas pedal either. Think about it… Just a throttle lever on the steering wheel much like a Ferrari or Lamborghini. Very easy to do with electric.
