jober1979 said:
Thanks for the context. I'm new to all of this, so I'm still getting my head around what all of the relevant considerations are. Is this why more electric motorcycles DON'T use LiFePO4 cells - not energy-dense enough?
Not necessarily, just not those specific ones.
But generally, the LiFePO4 for the same volume / cost is probably not going to supply as much current as other cell types...but it will probably last longer overall, and in certain ways perform "better" (most of it's discharge curve is "flat" so performance stays about the same thru the entire capacity).
There is always a tradeoff between c-rate (current-delivery ability), capacity, size/weight, longevity, charge rate, cost, etc. There are also different form-factors (cylindrical, prismatic (like those), pouch) that can affect performance or how you have to build the pack (which affects it's final size/weight/ratio-of-usefulness).
WHat speeds do you need to reach and maintain, for what distances? (is it the same speed for the entire trip, or different speeds for different segments, etc?)
What are the specifics on the hills?
Do you ever have headwinds?
Do you ever have detours? Of how far?
I'm planning to ride on hilly back roads, and occasionally/briefly on faster 4-lane roads that get close to highway speeds. I'd like the top speed to be in the 75-85 mph neighborhood, if possible. Fast enough to be able to pass safely, but it doesn't need to be faster than that.
Unfortunately, "hilly back roads" won't help you find out how much power you need. You'll need to verify the actual slopes to determine necessary power, or else just build something that can overcome *any* slope you would find on a road (which can be very steep) and still maintain the speed you want, without overloading battery, controller, or motor. That could get expensive, and be unnecessary.
If you have a smartphone, there are apps that can help you map your route and include slope information. I think Strava might do it, and there are others I don't know about. You'd just mount hte phone to your existing ride so it is flat and parallel to the ground, then go along your route(s) the new ride needs to handle, and then you can look at the data later to see what the worst case slopes are and how long you'll be on them. That will then let you use the simulator and other tools to find out how much power you'll need, and how much battery (wh) you'll have to have to complete the trip with capacity to spare.
The speed is important because the faster you go, the more power you need because of wind resistance (and it doesn't go up linearly, it goes up by some power-of-something, perhaps exponentially but I don't recall...the calculators and simulators take care of this for you generally). There are often published Cda / cds specs for existing vehicles you can find that you can then plug into the simulators to help properly calculate air resistance at various speeds. Otherwise, you can assume yours will be worse than whatever it says, and add some margin of extra energy / power to ensure you have enough, just in case.
If you have headwinds, then those add to the speed you're going. Let's say you're going 85mph, and there is a 25mph headwind. That means that as far as power usage is concerned, you're going 110mph. That is a HUGE difference in power from the 85mph requirement, so if your systme can't supply that, your bike will slow down (or overheat, or both) in those headwinds, to whatever ground speed + headiwnd speed is possible with the available power. And because it is taking more power than usual to go the groundspeed you're after, it's eating up batteyr capacity faster, too, so if you don't have an extra margin of pack capacity to account for this, you won't get where you are going.
The motor/trip simulators will help you figure out what you need, but you have to ahve the specific conditions available to plug into it, or it won't give you useful data (or what it gives you won't be sufficient for reality, and you'll find out after you built it and get stuck somewhere).
Again, the LiFePO4 cells seem to work against me here - as you pointed out, it's bulky/heavy enough just getting enough for 1 series, let alone multiples in parallel. I had sized the pack for 96V, on the assumption that efficiency scales with voltage. But if that's not as crucial as I had been thinking, I can go back and re-think my options on battery cells.
Efficiency vs voltage or current depends on the specific parts you use, and their specifications (like what winding (kv) the motor is vs the speed you need it to go vs the gearing to the wheel vs the wheelsize, and what the efficient RPM is for that motor, etc).
Technically it's more efficient ot use higher ovltage and lower current because there is less loss in the wiring, but if you keep wiring short and fat then for vehicle-sized stuff it shouldn't make all that much difference.
Plus it is easier and cheaper to find lower-voltage parts to do the same jobs.
It is also lower demand on the cells (lower current at higher voltages to get the same watts out of a system), so you need less parallel cels of a specific model to do the same job...just more of them in series. It's a balancing act, between cost of all the parts and space to put them and weight/volume on a portable system, and trying to not push anything near it's limits (because the harder you push things the more likely it is they will age faster and fail quicker or more unexpectedly).
If you pick parts (including cells) so you are using them at only half of their capabilities, they'll last a lot longer and perform a lot better than if you use them at the edge of their capabilities. It's not always possible to do this...but when it is, you'll have a more reliable system, assuming all the parts and build quality are good, and the system will have more room for pushing it harder under unusual circumstances, or increasing it's power output later if it turns out to be necessary, without replacing a bunch of stuff.
That's exactly the sort of thing I haven't been considering, thanks. My thought was that regen would be useful if only because there's so much elevation change around my area - I know it wouldn't recapture all of the energy expended going uphill, but it could at least recoup some losses. But from what I gather, on a 2-wheel vehicle it delivers pretty meager returns.
Typically you could get maybe a few percent back from braking and downhills from regen, with the common implementations of it in the average controller. If you had a system optimized for regen in all components for minimal losses and used the best possible methods to generate the regen power from the motor and recapture it, you might get a little more, but it isn't usually enough to build a system around that capability and depend on it to do what you want the system to do, range-wise. COnsider the regen a "free snack" to go with the lunch you paid for already, in your capacity / range calculations.
Depending on how you ride and where, you could get more "back" just by riding differently (not accelerating as hard, riding just a bit slower, etc).
It doesn't realy matter hwo many wheels the system has, either; the more wheels (and perhaps the more motors) connected to the drive/regen system, the more the regen is distributed amongst them, and "shock loads" the system a bit less, and perhaps has more braking power for the energy recaptured vs heat generated, but there is also more mass to stop and thus more to re-accelerate, and likely makes the math a wash.
Just to make sure I've got this: the limiting factor on how much current the battery pack could supply is the amp-hours of the cells connected in parallel (So 160Ah 3.2V LiFePO4 cells connected 30s1P would be able to supply 160 Ah at 96V; whereas the same cells in 15s2P would be 320 Ah at 48V).
Not just the Ah in parallel, but the C-rate of the cells, which is the C-rate number (1C, 0.5C, 2C, 10C, etc) multiplied by the Ah (Capacity). A 320Ah pack of 0.5C cells supplies 160Ah, regardless of what size the cells are or how many in parallel. A 160Ah pack of 1C cells does exactly the same.
Then there is also the total watts you need for a job; that's amps times volts. So a system with twice the Ah but half the V using the same C-rate cells will not supply any more W, so it is no more capable (power), and has no greater capacity (range). 160Ah x 96v = 15360Wh; 160Ah*0.5C x 96v = 7680W. 320Ah x 48v = 15360Wh; 320Ah*0.5C x 48v = 7680W.
So if the BMS detects too much current, it trips the contactor and interrupts the circuit? That makes sense - for some reason though I'm still having a tough time understanding how the BMS is isolated from the amount of current the controller could theoretically demand under full load. Hypothetically, pulling away from a stop on a slight incline could demand more current than a lower-current BMS would be able to handle, right? Using the example above, if I had that 15s2P LiFePO4 pack, would it require a BMS that could handle >320A? A contactor protecting the BMS in that circuit would just cut the power entirely, wouldn't it...? Apologies if there's something really obvious here that I'm missing, lol.
The BMS is there specifically to do just that--cut all power in case of pack danger of any kind (malfunction, empty, overcurrent, etc).
Your controller needs to be setup or chosen so it never overloads the pack in the first place.
That way the controller can roll back power when nearing or at the limits it's set to, rather than just shutting down. It just holds you at the limit that way.
The BMS, if it drives the contactor, would no longer be in series with teh main pack outputs to the controller. Usually it' current-monitoring shunts (if it has any) would be in series with those, so if they aren't it can't detect overcurrents. If you need it to do so, and it is capable of it, then you can use an external shunt placed in series instead, like the controller itself uses for such measurements. As long as that shunt outputs the same mV/A (millivolts per amp) that the onboard BMS shunts do, you can disconnect the BMS sense traces from those onboard shunts and then wire the two sense wires from the external shunt to where those traces connect to the BMS MCU or current-sense hardware inputs. The BMS only detects the voltage across them, not the actual current thru them, so as long as the scale is the same it doesn't know the difference.
