Some Ideas on building a 5000w E-bike

[Wheazel]

So, what's people's opinions on Adaptto products? Apparently, I should be looking into these, for the custom battery I want to build. It'll be a 22s8p of GA cells at first, then maybe a second one for a full 22s16p. The BMS, display, tunability, and size all seem to have good points. I've been looking at the midi-e.

The adaptto controllers' strong point is that they offer a complete and very functional system.
Personally I have been using an old max-e controller at low power for 2 years. The integrated bms solution works very well. And if you brake the mains from the battery the bms will not drain the battery during storage.
(great for my needs)

It is also a big plus that the display offers most info easily accessible. Good for troubleshooting and tuning in general.
Adapttos are best used with hub motors, as the auto tune functionality is questionable at best, and hubs just seem to work better.
My experiences with adaptto + revolt 120pro was an inefficient story. (might very well be due to the motor, but setup was more complicated than a hub)

These functions together with easy charging solutions makes it a great system for bikes with a permanently mounted battery.
For a bike with a detachable battery I think there are better alternatives.

 

[Samd]

Adappto is expensive imho. So is qs.

You can put a cheap infineon sine on a mxus or other 50mm motor and dial up 5kw without too much hassle. Spend the money on good batteries instead.


Sent from my iPhone using Tapatalk

 

[Ianhill]

Spend the money on good batteries instead.

IMHO if budget is tight the highest quality part of the ride should be the cells.
I always look at how much I can afford battery wise then design the rest of my system around the power they can discharge without breaking a sweat, I've never had the money to go mad and get anything decent really but as long as I'm building and learning then I'm happy.

 

[hypertoric_amplituhedron]

We're only talking a few hundred bucks cheaper, though. It's not enough to warrant the cheaper route on motors, controllers, bms, etc. In my case, getting a custom, locally built battery was the major decider on what I can afford to get. It's crazy how much money I've saved by holding off, and doing a few more weeks of research, lol.

 

[hypertoric_amplituhedron]

Here's what I'd like to do with the batteries:

Build 3 separate packs : 11s8p , 11s8p , 22s8p out of 352 Sanyo GA cells
- the two 11s8p in a kilowatt bag
- the 22s8p mounted on the downtube

Link the two 11s8p in serial for a 22s8p pack

Link the two 22s8p in parallel for a 22s16p pack

The goal is to get a 16p pack with 4.3 kwh, which can pull more amps than an 11p and 5p separately (11s11p , 11s11p , 22s5p).

From what I understand, the separate packs all have to have equal Ah (equal p) to be able to be connected in parallel.

* Can these three become a 16p pack with no issues?

* Can I get full use of the 16p, which should be around 80 amps (~2C) on the safe side?

 

[Wheazel]

In theory it is no problem to have two 8p packs and parallel them.
However you need to install them in such a way that the differences in installation will be of no notable impact to discharge equality.
For example, use thick enough wire or keep wires of similar length, and have the packs equally built internally.

Personally I do not like splitting packs unless absolutely unavoidable. I think it complicates battery management and the installation overall.

What do you mean by get full use of the 16p?
The 16p will always contribute to lower voltage sag, shallower cycles for a given energy output compared to a smaller pack.

I do not think you will see 4,3kWh usable out of a 352cell pack.
Of course this depends on your charge and discharge range amongst other things, but I would be more conservative and calculate for ~10-11Wh per cell for the GA.
At 5A discharge from 4,2 to 2,5V the GA gives 11Wh according to the lygte site. And this is under optimal temp conditions and a wider cycle than most real applications will use.
So even if you won't be pulling 5A continuously, a shallower cycle and lower temps can affect the energy output.

 

[hypertoric_amplituhedron]

Personally I do not like splitting packs unless absolutely unavoidable. I think it complicates battery management and the installation overall.

It's just an idea I got, based on how I want to mount the batteries. The top tube pack made by teslanv can hold quite a bit, and it places battery weight in a good place for stability. Since I'll have more batteries than what can fit in the bag, why not make an additional downtube pack, like an external fuel tank.

It would be more complex to wire up, so long as it will work. Good point on equal gauge and length wires. An unequal discharge over the 3 packs is presenting the biggest issue.

What do you mean by get full use of the 16p?

It's a matter of how many amps can I get from the batteries, however I get them hooked up.

If the sanyo GA is rated for 10 amp continuous (ab 2.9C), then for a 16p pack, you'd have 160 amp cont. But, to extend the battery life and keep heat down, you only want to use half of that, like 80 amp continuous.

This provides the greatest amount of power I can get from the 352 cells. If I have to split up the 3 packs to 11s11p, 11s11p, 22s5p, then I can't link them in parallel, and can only get 55 amp cont out of the 11p packs.

I do not think you will see 4,3kWh usable out of a 352cell pack.

Not a usable 4.3 kwh, that's 100% drain. At 80% drain, it's around 3.5kwh of usable, plus or minus some. I won't be testing that range ability at first.


-------------------------


So, it's not so crazy of an idea after all? Just take consideration of wire length, gauge, and unequal discharge?

 

[hypertoric_amplituhedron]

A better question would be: are there any devices out there that can monitor and balance the discharge levels of the 3 batteries? Would that be a feature of a BMS, or something else standalone?

 

[Wheazel]

You can still use one bms for the whole installation. But the sensewires make the cabling more messy. Your split pack basically has longer series connections between some parallel groups.

 

[hypertoric_amplituhedron]

Just one Adaptto BMS will work? That's good to know! I thought I had to buy 3 of something. Well, maybe 3 sets of the sensewires are what I would need. Any suggestions of what else I would need to know regarding this setup? Do you know anyone who had done this build before?

 

[Newagebike]

Last update on your build was back in Jan, how is it coming along?

 

[hypertoric_amplituhedron]

I figured out a crapload of stuff. I took the opportunity to ride an ebike my friend built. Bought it from him in mid March. I put 850 miles on it since then, with a 2 week downtime. This gave me the opportunity to see what 52v 30a rode like, what a 40~50 mile ebike ride was like from a 1 kwh battery.

And, I realized that a qs205 pushing 5kw was going to be too much bike too early on, in my experience. I needed a medium power ebike.

So, I got the EDGE 1500w motor kit from ERT, complete with laced wheel, controller, throttle, 3sp switch, and cycle analyst 2.4.

I'm going to order some panasonic GA cells and get a 24s10p (88.8v 34.5ah) pack built with an HCX-D131. Should be good for 2.4 kwh usable, which allows for 60 miles at 30 mph, 110 mi at 25 mph. I like the idea of high volts, low amps. Should be interesting.

But, I don't know what frame to use. I'm not using my 21.5" Trek 4300. It's too big for me. I've been thinking of the Raleigh Redux 1 frame, mostly because I like how the bike looks, is full disc compatible, and it can fit 240 cells in the medium-sized triangle.

But, I'd prefer a steel frame, which is why I've been eyeing every bike that comes through the shop. No such luck. But, today a friend suggests I look at some of the Surly frames, which I can get through QBP.

So, this is where I'm stuck. I have all the hardware, and electrification parts figured out. But not the frame!

 

[amberwolf]

How about used stuff from thrift stores, yard sales, freecycle.org, craigslist, etc?

Back when I needed more frames and stuff, that's where I got them easily, and cheaply or even free in several cases.

 

[Chalo]

So, this is where I'm stuck. I have all the hardware, and electrification parts figured out. But not the frame!

Nothing wrong with Surly frames, for sure, but you'll have to spend upwards of 400 bucks at a minimum to get a new one, and used Surly bikes tend to be of the thrashed-and-annihilated sort.

I'd look for a generic non-collectible late '80s MTB as a starting point. They're rugged and stable, with all the same general virtues as a Surly, and they've got level top tubes with scads of room inside the front triangle. Specialized Rockhopper, Trek 900 series, Univega Alpina, Jamis Diablo, Schwinn Sierra, Miyata Prairie Breaker, Diamondback Apex, Maruishi or Panasonic or Yokota anything, etc., etc. There are lots.

 

[hypertoric_amplituhedron]

All right, the batteries are on the way. Now, before I order the bms, I have a few more questions on volts versus amps.

Specifics:

- Speed goal is 50 mph
- 26" tire
- Using EDGE 1500 , 11.46 kv
- 40 amp controller

I originally wanted to do 24s10p, for 100v and 30a max power. But, the ebike sim says I won't do 50 because the amps are too low. The wheel can spin very fast due to the volts, but won't have enough torque to reach those speeds.

So, should I be thinking of 20s12p that can run 74v 40a?

 

[hypertoric_amplituhedron]

Anyone have experience with this? I might be asking it wrong, or maybe too vaguely.

Am I throwing away power potential by going with a 24s10p and running 89v 30a ?

Should I instead build a 20s12p running 74v 40a ?

It would seem like a waste to volt the heck out of the motor, giving it the ability to spin very fast, but the amps can't push hard enough to use that high speed range. So, I get a motor that's not as fast, or accelerate the same. Since I won't be ebiking in a vacuum on the moon, I need more amps to overcome the wind resistance.

I think I should step down the volts, and kick up the amps to a 20s12p. What do you think?