Ebike scooter project QS 4kw V3 72v, Kelly KEB72601

Are you planning to have your CA have control over the KEB? I recently tried to do this with mine, but none of the wiring diagrams I saw seems to complete the wiring to make this happen. I gave up on it for now till I have time to relook at greater detail.
 
So this weekend I got my LJ MJ1 modular battery fixed up

It's the 20s10p so has potential for 100amp continuous

Recently I've been having issues with the CA and programming the kelly controller. The guy who sold me the kit from QS motors was keeps insisting that my controller doesn't require programming. It's a 6kw controller and the wheel can handle 7kw peak. A fully charged battery would be 84v at 100amp would be 8.4kw output. Recently I have been putting it on 40% battery current (max is 70%) because the The 72v lifepo4 bms keeps cutting out.

Also I think the CA I have is telling me wrong numbers

For starters I have a 100amp circuit breaker which trips at 135amps. Thus far it has never tripped. So yesterday I put the controller at max settings 100% motor current, 70% battery current. Used the 72v 35ah lg cells modular battery for some tests

My God this is what I was originally expecting. It's like a completly different bike. The torque is muuuuch better and the top speed all I was game to test in flip flops shorts and t shirt was gps speed 75km/hr I'm sure it would reach the QS specs of 85km/hr

So I kept doing full throttle bursts then slow right down for a while then afterwards tested the motor with the back of my fingers. Not even warm. The battery wasn't warm or hot either

So that solved it for me
 
the breaker will trigger much later. it wont go at exactly that current. you need to sustain more then that current for a few seconds, the higher the current the faster it will pop.

personal feelings? just make a direct battery-controller connection and bypass the bms for the controller current but keep it for the rest. that way the bike will still cut out as you lose the control voltages for the controller and it will instantly die if the bms triggers a low voltage or something but it wont trigger on current anymore. that is regulated by the controller anyways.
 
Scottydog said:
Hey just noticed this build, really nice reading, will have to go over a few more times! I am using a KEB72801 currently myself with a 5kW QS motor. From what I have read you won't be able to program in wicked torque from a standstill on the KEB's, they have a soft ramp up. There was a fellow who had a post on his scooter a few years ago who used the 12kW version and got frustrated by this.

I have just swapped out my 20 cell GBS battery set for a 72V Chevy Volt module and holy smokes it is a totally different beast!!! Before I would only see maybe 135 amps on a warm sunny day "hot off the charger", but now it peaks at 275 amps and showed 18kW on the CA!!!

But I did notice at the lower speeds, under 12 - 15 mph if I give full throttle the bike cuts out abruptly then repowers as if a protective mechanism is kicking in. I'm going to look at my Kelly settings and see if there is anything apparent causing this issue. Reading your post on the settings might help me understand better.

Hey I was reading over some of these old forum posts and found one where you're saying the KEB72601 has a soft ramp up. Since the starting torque on this still annoys me.

You said you had the same controller with a 5kw motor (I'm sure thats rated watts so maybe 8kw peak?)

anyways- you said adding the Chevy volt battery made a huge difference on the above setup with the KEB. I was thinking of getting some different batteries higher powered.

My current pack is 20s10p with LG MJ1 they are 10amp each so 100amp for my pack- I'm wondering if thats my issue. I was thinking of upgrading my cells to

Sony / Murata US18650VTC5A 2600mAh - 35A
20s7p would be 245amp
20s10p would be 350amp

What were the specs for peak amps and cont. amps for your Chevy battery? Did you notice a lot more starting torque?
 
Well its been a while since I made an update to this bike, but I'm doing a couple things.

New motor!
After much contemplating I've decided to order the high-torque version of my QS motor. I was wanting the torque from the start, and I don't want to go 85km/hr on this plastic ebike. 60km/hr top speed with high torque is fine for me. Speaking with Robert from QS he has told me in the past it can handle 4kw nominal and about 7kw peak. He also recently confirmed the motor can take around 100amps which at full charge 84v = 8400watts. Since this is only for a few seconds when taking off (especially with the high torque motor) he said this should be fine

Upgraded battery
I've decided to give lay the modular battery blocs battery to rest. It added quite a lot of weight to the battery 6.84kg! (15.1 LBs)
Since 200 cells is approx 9kg (19.84lbs) this is a huge amount of weight to add (43.18% of total battery weight is just battery blocs)
View attachment 6
IMG_0974.jpg

I've was never sure about the resistance of the magnets and the plates. Also now that I've taken it apart, I've noticed the 3D printed plastic has warped since the cells generate some heat, over time the plastic warps, some of them have split (problem I had from the start). I really expected a higher quality product, but I guess saving money on injection molding is worth it for him. Oh well- i'll probably just give them away to an enthusiast in Toronto, they can be fixed with heatgun I'm guessing, probably about $400 in total hardware there.
View attachment 4
IMG_0972.jpg
IMG_0975.jpg

So the new battery will be using nickel strips and spot welding. My current 20s10p pack of LGMJ1 3500mah cells (200pcs) I'm going to add 100 cell (Panasonic PF 2900mah)

Both these cells are rated at 10a each. Although the capacities are different, I'm going to add the extra cells to each parallel group so that each group will have the same capacity and power output.

The new battery will be 20s15p
10 cells of MJ1 + 5 cells of PF (for each group)

I'm going to make them as two 36v 10s15p packs connected in series because I'm not going to add a BMS. I will balance charge each pack with my iCharger 1010B+ balance charger when they need balancing
I mocked up cell spacers the configuration so I could get the sizing right. Each layer is 36v 10s15p, going to connect them together with some 6AWG and PRC8 connectors. Will put a sheet of acrylic between them + tape & heatshrink etc
IMG_0968.jpg

Kelly controller settings
So it occurred to me I was never getting the full power from my battery because of my Kelly controller. The battery current limit in the settings of this controller max value is 80% which means I'm only every getting 80% of the power from my pack

my 20s10p pack could do 100amps, but I was only ever getting 80amps

Also I noticed my cheap chinese 84v charger had somehow gone down to 80.7v

So if the charger was working properly 84v at 80amps = 6720watts for the "balanced" version motor (balanced torque and max speed is 85km/hr)
But really I was getting 6456watts

aaaannyways. My new 20s15p battery will require me to program kelly max battery current to about 67% because new battery can output 150amps. As per Roberts recommendation I don't want to send more than 100amps to the new motor

Bonus 1: I have a larger pack 49.5AH
Bonus 2: Each cell has to work less - instead of doing their max 10a per cell. To make 100.5a they only need to do 6.7a each.
Bonus 3: This battery can be used with higher power setup in the future

So after extracting all the batteries from the batteryblocs one I'm balancing every single cell before I build. I made a little 10s balancer holder. The left box is staging area for each battery group - you can see the PF 5 cells at the end.


Bike re-wiring
I thought I was being super smart by using a shitload of little connectors when doing all the bike wiring. Turns out I was wrong, when it rains water drips down the small wires into the small connectors and causes corrosion. This would mean sometimes my lights wouldn't work, sometimes horn would't work. Sometimes my 12v key switch to precharger to contactor wouldn't start the bike!!

Solder and heatshrink going forward
 
What winding did you get?

Ps: make the battery 22S and slighty undercharge it to 4.05V or 21S for full capacitt. That will nudge you up to 90v limit of the controller and give a tad more power and makes you be able to hold your max rpm longer.
 
Actually I do have 10 more spare cells

- So I have 310 cells altogether
- I could make a 22s14p = 308 cells
- Charging to 4.05v would be 89.1v

that means 140amps output would mean kelly controller at 72% = 100.8amps

The only problem is splitting this into 2 packs = 11s per pack, my balance charger is only 10s

Another option 21s15p = 315 cells as I have 210 mj1 cells + 110 PF cells

3 packs of 7s
4.2v x 21 = 88.2v

I wonder if 4.2v more is going to be worth the hassle

Another option would be to go
- 24s13p = 312 cells (I would have to get 2 more cells from somewhere)
- then I could balance 4 packs of 6s each

issue with this would be charging these to 3.74v = 89.76v
The lowest I go for 18650s is like 3v which means I need to recharge when pack gets to 72v



I'm not sure 18650s chemistry are good for these voltages? Seems like I would have quite a limited voltage range?

Keep in mind I don't really care about top speed, mostly just torque. Higher the voltage the more speed (max rpms right?)
 
4v more at the same 100A limit means 400W extra watts off the line.
Controllers dont care about volts, they care about amps. So if you care about watts then you need to maximize both in order to get the watts you want.
It also helps to reduce amps during "normal" operation so your cables dont become "melty" and stuff.
It also helps so you keep your max rpm over the entire voltage range so you dont lose speed when the battery drains and you keep a higher output when the voltage goes down.
 
As much as I’d like to send more watts to the motor QS Motor guy said 84v at 100a was ok for a few seconds like for takeoff from stop (8400w) but no sure abou 88.2v at 100a (8820w)

My only viable option is 20s or 21s

If The motor can take 8820w then I’ll do it, but it doesn’t have a thermsistor unfortunately
 
I got the same motor and i have no issues pumping 150A battery amps into it at 90V with the motor amps around 250...
 
wow 150amps at 90v is 13,500watts!

You definitely have this exact motor? 10inch 4000W Electric Moped Wheel Hub Motor V3 Type
http://www.qs-motor.com/product/10inch-4000w-v3-type-e-scooter-hub-motor/

Which controller do you use? Do you have a link to your build?

If I do 21s15p pack and don't limit with kelly controller the most ill get is 80% battery current (my controller battery current max setting)

= 120amps at 88.2v = 10,584watts
 
yes, that motor in a 4T winding.

mind you, it is tricky to keep the front wheel down at those power levels. :mrgreen:

i use a 300A KLS7230.
but honestly i dont drive it with those power settings simply because its undrivable almost. its just too much in the city. but when doing drag races its lots of fun. i just mind the temps and wait a while between runs so it can cool down properly.

right now i am building one for a client that wants a 110v max voltage and with the same motor. i can make some pics when that is presentable as he is buying my motor. my current scooter is disassembeld to tackle rust and upgrading the frame to take a 13" rim and a bigger 144V battery for a 8kW 13"hub. should be able to outrun a zero or tesla with that.
 
So Robert got back to me and said 88.2v at 100amps is okay 8820watts

He also said the motor was 4T, which seems low to me for a "high torque" motor and "balanced" version is 3T. After reading through some posts I’m not sure how much difference this will make because Kelly’s controllers have slow start.

Although he said that my controller limit is about 6000w for the KEB72601
I'm wondering how he calculates that from these settings
•Peak Phase Current, 10 seconds: 280A.
•Continuous Phase Current Limit: 110A.

If its max voltage then 6000/90=66.66 - I wonder if its 66a cont. battery current? I'm guessing here

If this is the case and my controller will never send too much power to the motor then I won't set any battery current limits etc.
I just wish the controller specifications, and motor specifications were more accurate. There is so much guesswork

I might as well go 22s pack (get 10 more LG MJ1 cells) then charge to 4.05v (89.1v) like you say This will give me more cycles from the pack too- although there would be no easy way to balance this (11 packs of 2s??) I don't know any balance charger that does over 10s

Actually do-able pack at the moment for me is 21s15p, just gotta get some 7s balance connector wires

I was reading a bunch of posts about battery current vs phase current and its anything from 2.0x to 2.5x to 3.0x its a varying thing based on so many factors. So if peak is 280a phase

maybe I should build a bigger pack like 21s18p = 180amps @ 80% (kelly max) = 144a

Why not 21s28p = 280amps lol

Having said that if my controller is the bottleneck at 6kw, then there isn't much point building a bigger battery that can produce more than 68.02amps?? (other than the range of course)
6000watts / 88.2v = 68.02amps?

I have a Cycle Analyst thermistor, which I will install into the new motor. At what temperature should I stop using the motor? According to the spec page
Working Temperature: 70 degree, Peak 120 degree

So if its going above 120 deg. C then I should stop

What temperatures do you look out for?
 
check the bluetooth bms topic, there are a few good options there that you can program to balance at whatever voltage you want.

and the reason for "bigger is better" is a question of load per cell. if you add more cells but keep the total load the same you lower the stress on each cell increasing lifespan for the whole thing. it also prevent more sag under high amps.

trust me, that motor is high torque as you will have issues keeping the front wheel on the ground. that kelly's tend to be a bit droopy off the line is probably the only thing that will keep you from wiping out.
 
Battery build
So I decided I'll just go with 21s15p = 150amps limit by 80% kelly controller = 120amps @ 88.2v = 10,584w
Going to set the kelly controller lower and monitor the temperature of the motor under different conditions.
I've read on this forum that QS motors can handle anywhere from 2x 3x or even 4x their rated watts power so this 4kw motor should be able to handle 8kw or even 12kw maybe. Flippy has the same 4T motor and has had no issues running 150a @ 90v = 13,500w through it.

I could have gone 21s16p but I would have had to buy extra cells (more $$) and I couldn't have 1 string of 15p - because of battery box dimensions (I'd have to double up making 2 rows of 8 on top of each other for 16p)

Also I have a 100a solar circuit breaker which trips at about 130amps or so 128amp from 16p would have been cutting it fine

Spacers
This is about all I can do at the moment with my current spacers. Waiting nearly 2 weeks for 4x5 spacer sheets from "USA" - would have been quicker to ship from China ... sigh...

So since they are 7s I'm going to use the 4x5 sheets lengthways for the 15p (first 4 cells) then use the single spacers for the others (last 3 cells)

Wires
Also waiting on some 7s balance wires from Progressive RC. Seems as though I have the exact right amount of PRC 8 connectors - I'll just have to snip 2 of them in order to separate so I can put 3 of the 7s packs in series
 
Ok a lot has happened

Battery is finished-
early discussion details here: https://endless-sphere.com/forums/viewtopic.php?f=14&t=99374
Latest details below

Ended up going with 3x packs of 7s to make 21s15p 49.5AH pack
- charged @ 4.2v = 88.2v, but will charge it lower 80% at 4.05v = 85.05v for max charge cycles.
- It has no BMS will be monitor cell levels with CellLog device

Each 15p group:
- 10x LG MJ1 cells 3500mah @ 10a cont. per cell
- 5x Panasonic NCRB cells 2900mah @ 10 cont. per cell
- 49.5AH and 150a output
- painfully balanced each individual cell for the 15 cell group
- then balanced each of those groups 7s

Previously I though my kelly KEB controller max battery current was 80%, but its actually 70% which means I can only draw 105a out of this pack. At a 100% charge ill be able to get 9261watts (88.2v x 105a), I'll mostly be charging at 80% so 8930w which is fine for this motor.

I embedded a temperate sensor into the middle of the battery with waterproof connector
- used PRC8 connectors to connect each pack
- each 7s has a balance connector
- battery wrapped in foam and plastic sheeting
- spent too much time sewing a handle with webbing harness
- mixture of nickel strip and also 16AWG wire as advised (see above link for discussion). Also 10AWG naked along POS and NEG of each 7s pack (see photos)

IMG_1090.jpg



you can see the wire for the temp sensor
IMG_1089.jpg



temp sensor LED to be wired into the dash
IMG_1088.jpg

balance wire
IMG_1091.jpg

10awg to 6 awg for POS and NEG of each 7s pack
IMG_1084.jpg

needed my 300w soldering gun for this one
IMG_1115.jpg

foam and tap
IMG_1127.jpg
IMG_1128.jpg
IMG_1129.jpg

protecting
IMG_1152.jpg


Thick plexiglass between each pack for safety
IMG_1153.jpg
IMG_1155.jpg
IMG_1154.jpg



checking voltages with cell log
IMG_1156.jpg


checking whole pack voltage
IMG_1157.jpg


more plastic
IMG_1172.jpg


finishing up
IMG_1173.jpg
IMG_1174.jpg
IMG_1179.jpg
IMG_1180.jpg
IMG_1182.jpg

labelling each of the connectors and balance wires
IMG_1461.jpg
IMG_1462.jpg
IMG_1463.jpg
IMG_1464.jpg
IMG_1465.jpg
 
New motor is broken (not completely)

So I wanted to install the CA 3 thermistor into my new 4T higher torque motor. After watching a lot of QS motor videos I saw them banging the axle on the ground to get the hub covers off. Since I have learned that banging not too hard with a rubber mallet, then levering up the plates a bit. Also used a 10" gear puller to get the covers off without too much impact.

Over time this has broken the already delicate hall sensors, even though this motor has 2 hall sensor plugs, the combination of my taking the covers of multiple times, and zip cynching down cables (so they dont rub on the cover) has snapped 2 of them off (1 on each plug of course). I tried to follow suit as I saw white silicone being used to hold and I thought strengthen wires together, but to no avail.

Also a lot of silicon bits and grime has found its way in there, - also added statorade so this acted like a grime magnet

also noticed one of the magnets is chipped, but could not find the piece inside the hub motor anywhere, I wonder if it was always like that.

So I'm going to go back to my old motor in the mean time (so I have a working bike) which I haven't opened up, and eventually repair the new one once I figure out where I can get the right hall sensors from (assuming they are all different with different sizes and specs etc) waiting to hear back from Robert at QS about it

IMG_1413.jpg
IMG_1446.jpg
IMG_1447.jpg
IMG_1448.jpg
IMG_1449.jpg
 

Attachments

  • IMG_1445.jpg
    IMG_1445.jpg
    241.9 KB · Views: 2,161
Re-wiring everything

The rats nest. I thought having lots of connectors was a good idea, turns out I couldn't have been more WRONG

In the rain, since this bike is not very waterproof, rain would drip down wires, and go inside these blue connectors. The would corrode and cause connection issues. I even had relays that were controlling the horn that were completely rusted out

So as painful as it was, I replaced every single blue connector with a soldered, heat shrinked connection.

Its still a rats nest, but I started to use a labeler to make things a bit easier

IMG_1293.jpg
IMG_1294.jpg
IMG_1295.jpg
IMG_1296.jpg
 
CA3 to kelly controller wiring and throttle settings

So my post from here actually proves useful
https://endless-sphere.com/forums/viewtopic.php?t=84739

CA3-KEB.png


The thing is my CA3 was turned into the HC (high current) version because I already had external shunt from Grin.
So following the diagram I used wiring from the plug, connected everything and

nothing... no power

I plugged 12v into the DC connector, and boom I had power.


okay so... what gives....

turns out the HC mod, re-routes all the wires from the main plug (shown in diagram) to another cable containing the ring connectors

*sigh* ok....

so its working now, I just need to configure. Lots of reading the "unofficial manual" in the ES forum, but has old terms as I think they have updated the menu.

Having a faulty hall sensor was not helping here, as I was trying to find the throttle voltage thresholds, and the motor was making weird noises

IMG_1397.jpg
 
This is Shawn McCarty the developer of BatteryBlocs. I agree with your comments, that the PLA plastic can deform under heat. I first became aware of the issue when a customer in Arizona left a battery in the dash of his car and returned to deformed BatteryBlocs. In response, I switched to using ABS plastic which has a glass transition temperature of 105C vs PLA's 60C. That's 221F vs 140F.

Yes, PLA plastic is brittle and becomes more brittle over time. I noticed this on my own ebike with some six-year-old BatteryBlocs. I was replacing cells and cracked a few Blocs in the process. ABS Blocs do not degrade, they are not biodegradable- but they are recyclable.

Rebuildable/repairable systems do weigh more than spot-welded systems. Looking at the pics, there are extra busbars to handle the high current, and lots of wiring that are not part of the Blocs, and the cases are also included in the weight. Please note that the battery did not fail, the reason for its replacement seems to be because of it's overall weight, which is understandable.

By the way, it looks like an excellent build job, both with the BatteryBlocs and the spot welding.

I will replace old PLA Blocs with new ABS Blocs if you are having issues. This is just the plastic part, not the entire assembly. At the time I began making BatteryBlocs I did not have any experience with the long term degradation of PLA plastics, nor the thought that the batteries might get above 140F in common use. Using ABS plastic, though more difficult to manufacture, solves this valid issue.

3D printing allows me to manufacture a number of different parts for different cells, from 18650's through 26800's. Designs are easily modified for custom batteries. I do not need to stockpile parts, just data and plastic. Creating an injection mold for each part is prohibitively expensive. Hence the business model.
 
Back
Top