QS 90 1kW (edit: or maybe 120 2kW motor) with Kelly KLS controller in a Cityel

Hi guys,

sooou I'm looking at replacing the motor in my Cityel (it's got an old, heavy and super inefficient forklift motor from the 1980s) with a brushless motor. I've put the QS 90 1kw in closer consideration, since it weighs 5kg (which should be enough, and also not to much, iron and copper for this low drag, 45km/h, 180kg vehicle, driven in a mostly flat area) and supposedly has an efficiency of >90 %. Also, it's less than 300 € incl. shipping to Germany and VAT, + seems to be of quite nice quality.

Since it's a closed up vehicle, the controller really needs to be sine-wave, so I've already purchased a Kelly KLS 6018S (240 phase amps peak). The battery is 40V li-ion (11s24p, 48Ah).

I've got a few questions regarding this motor: Can someone confirm that it's running well with a Kelly KLS (hall sensor placement wise)? And is it possible to get this motor with a 48V winding (not just reterminated to Delta, as then it won't work with the sine-wave controller anymore due to inaccurate hall sensor placement).

I've also looked at the 3kw golden motor but many people seem to be not happy with it due to thermal conductivity issues, it only has 0.5mm lamminations, and on top of that with 400 € it's even more expensive than the QS 1kw motor (I guess I will have to eat that bullet though if the QS 1kW doesn't work out and I can't find anything better, since so far it's the only 48V motor with about the performance figures I like, that's confirmed to be running well with Kelly KLS controllers).

I'm deliberately planning to let the motor run at a slightly lower voltage than its specified for and gear it up, since that will result in lower stator losses (the Cityel only needs to overcome a drag of ~900W @50 km/h, so the lower the stator losses the higher the range in a flat area).

Since I'm custom building the motor mount anyway and there's practically unlimited space, I can basically put in any motor I want (except for hub motors for a few reasons, mainly high rotational mass).

Any input is appreciated!

Thanks,
Matt

twikis said:

sooou I'm looking at replacing the motor in my Cityel (it's got an old, heavy and super inefficient forklift motor from the 1980s) with a brushless motor. I've put the QS 90 1kw in closer consideration, since it weighs 5kg (which should be enough, and also not to much, iron and copper for this low drag, 45km/h, 180kg vehicle, driven in a mostly flat area) and supposedly has an efficiency of >90 %. Also, it's less than 300 € incl. shipping to Germany and VAT, + seems to be of quite nice quality.

Click to expand...

according to
https://en.wikipedia.org/wiki/CityEl
the original system is a 4.5kw motor. even given the difference between brushed and brushless for efficiency, are you sure that a 1kw motor will be sufficient? (whcih cna for short periods be run at higher power, but heat buildup will be your enemy doing that for longer periods)

it might be, as long as it can take the peak power you will need to get sufficient acceleration from a stop. i'm sure it will be capable of mroe than enough power for cruising at that 45kmh.

as a slightly-comparable example for startups:

my sb cruiser trike is not very aero, and top speed here is slower, but quick accleration times in traffic are needed so cars around me / behind me at a traffic light don't get pissed off. including myself, it's about 230kg, and takes 5+ seconds to reach 20mph from a stop on the flats, at something under 4kw of power, with dual hubmotors (which waste more power doing this than yours with gearing between motor and wheel will). once at cruising its only around 1kw or less, but it takes a lot to get there quickly, and doing this frequently as is common in my commutes heats the motors up a fair bit. can start at 35c+ in summer and end up 60c in a couple of miles, and they're pretty hefty hubmotors mxus 450x. (winter may start at 8c or less, and end up 30c+ in the same time).

Since it's a closed up vehicle, the controller really needs to be sine-wave, so I've already purchased a Kelly KLS 6018S (240 phase amps peak). The battery is 40V li-ion (11s24p, 48Ah).

Click to expand...

is the lvc on the controller programmable? i couldn't tell from the specs on the kelly page for that controller
https://kellycontroller.com/shop/kls-s/
which mentions a preset value for lvc, whch may mean it is not changeable. if it's not, your battery is too low a voltage to run the controller. you'd need a 16s pack for a typical 60v lvc, which would be around 48v or so, about 45v at the lowest.

Hi amberwolf, thanks for your post and research done.

The original "4.5 kW" motor is a Perm PMG 132 on 36V. But that's valid for 48V, at 36V it's only a 3kW constant motor (which is the standard setup it was delivered with; "sporting" three 12V lead-acid batteries, totalling 100kg. I'm using a 14kg li-ion backpack battery now). Also I believe that the motor is conservatively rated at 1kW, and I'm quite intrigued by its 90.5% efficiency rating. There's a power curve in the 7th post here, which makes it seem to run quite comfortably at 2000W mechanical... Battery current limit in this test seems to be 35A (way to low, it should probably be twice that, since according to the results it doesn't push the motor one bit), when fiddling around with the numbers it looks like the phase current limit was ~130A.
Now after a closer look, this test isn't really worth much though: the no load losses (270W) are higher than the losses at the max efficiency point (220W).
Which, quite frankly, is bs.

Btw. thanks for pointing the acceleration thing out (although imo 30C extra really is nothing to worry about!), which made me put the vehicle on a scale. Incl. me, the battery and diesel heater (yea, I know... ) it's about 300kg, when actually I thought it would be something like 230-240kg. That of course impacts acceleration speed, and the power needed for acceleration.

I've played around with the ebikes.ca motor simulator... at 240 phase amps it (theoretically) takes about 8 seconds to reach 50 km/h, which is... fine, for the average German city (most people here drive sticks, and fuel is expensive, so we accelerate slower at traffic lights than it's normal in the US. )

All in all I guess the 120 2000W motor most likely is the better choice, also because it's lower rpm, which makes the drivetrain (either chain or belt, depends on the motor it'll be in the end) less noisy, more efficient and durable.
These last few days I had a little mail exchange with Nick from QS motor, and he said that Kelly KLS controllers are compatible (yay ). Also he said that they would do basically any custom winding... But; he doesn't know what phase resistance is, and he pointed out that his technician (possibly the same guy that ran that test on the 1000W motor lol) said it wouldn't be important (*cough cough*). Which, he said, is the reason why they aren't disclosing that information...
In the last mail I've asked him for quotes on a 1000W 48V and a 2000W 60V motor, let's see what happens... If it turns out they won't custom wind the motors, or are asking to much money for it, the deal is going to depend on if he will give me the phase to phase resistance or not (of the 72V 2000W motor), since that's the most important factor for calculating losses during acceleration.


Regarding the controller, yes the LVC can be programmed that low, to 18V (iirc).

twikis said:

Which, quite frankly, is bs.

Click to expand...

i haven't looked at it, but there's probably an issue with either data noted down during the tests, or math on that data done wrong at some step somewhere. it's unlikely they'd've just made up the data, but that happens too.

Btw. thanks for pointing the acceleration thing out (although imo 30C extra really is nothing to worry about!), which made me put the vehicle on a scale. Incl. me, the battery and diesel heater (yea, I know... ) it's about 300kg, when actually I thought it would be something like 230-240kg. That of course impacts acceleration speed, and the power needed for acceleration.

I've played around with the ebikes.ca motor simulator... at 240 phase amps it (theoretically) takes about 8 seconds to reach 50 km/h, which is... fine, for the average German city (most people here drive sticks, and fuel is expensive, so we accelerate slower at traffic lights than it's normal in the US. )

Click to expand...

the one things most people don't consider when either building or rebuilding a project is acceleration times, which really equates to hill climbing ability but these are typically poeple that ride on flat roads.

i've discovered the hard way that it makes a great deal of diffrence between usually safe and having drivers actively try to run you over because you're in the way of a green light. as long as they don't "feel" stuck behind something that's "too slow" getting started from a stop, they don't react with violence. makes a significant difference to other things too, but that's a pretty important one to be able to keep riding.

most drivers are reasonable...but too many are not, and it generally seems that the bigger thing they are sitting in, the worse the problem is (the more invulnerable they feel? dunno).

anyway...it's also more fun to have more accleration available, in occasional moments here and there. (even for the same top speed)

Also he said that they would do basically any custom winding... But; he doesn't know what phase resistance is, and he pointed out that his technician (possibly the same guy that ran that test on the 1000W motor lol) said it wouldn't be important (*cough cough*). Which, he said, is the reason why they aren't disclosing that information...

Click to expand...

pretty sure they aren't disclosing it because they dont' want to shell out for a meter that will test it.

also, *and* then actually be held accountable for it when people get motors that aren't built the same and have different properties, because they went cheaper with the wire after a while, or some other greed-based thing.

fwiw, i wouldn't necessarily hold my breath on anything custom. based on some other forum posts over teh years, you're as likely to just wait for weeks only to be told they won't do it (after they said they would), as to wait and then get it and it not actually be custom, but rather just a regular whatever, as to actually get what you asked for (which you'd have to be able to test all parameters of to verify they actually did it). the same applies to most, probably all, manufacturers, unless you want hundreds of the same thing, and then they'll probably really do it.

To me it seems very much like a sluggishly done test, it just could be much better, and at least some numbers are wrong. If we're lucky it's only the no-load power input, but how can one be sure about that...

Well actually... Even with the current motor, which needs 10-12 seconds to reach 50 km/h, I'm usually far ahead of most other cars, before they start to catch up again.
More acceleration definitely is fun, no doubt about that!

Yep, just got the answer from Nick (quote):"For mid drive motor, different voltages are only corresponding to different speeds, and the windings have not changed." Well whatever, as to be expected...
I've told him that our trade depends on if he can get me phase to phase resistance, and preferably inductance, of the 2000W motor...

Yes, I've got all the necessary equipment to measure out motors, so that wouldn't be a problem, and I'd probably send them back more numbers than they themselves have.

I thought those Perm 132's were supposed to be pretty good motors, so unless it's being replaced with a much higher rpm motor then it's unlikely that a much lighter brushless motor rated at half the power is going to cut the mustard. Plus 48V and 250 Kelly phase amps doesn't seem like enough controller to me for a 180kg vehicle even geared to that low speed. That's far heavier than even my heaviest 2 person scooter.

Yes, the PMG132 are quite good motors (even in optimal conditions with sub-standard brushlife), when run in the right direction (in the Cityel they are running in reverse direction) and at their designed voltage (60V). There are a few problems with them running in reverse;
1. the timing is off and can't be corrected (which produces heat and causes the brushes to wear out even faster than they usually do)
and 2. the brushes and the commutator are also axial-type, and designed for forward direction (=even more premature wear by sparking etc.).

There are a few more things about it, affecting its peak power (e.g. its low heat shredding capabiliy). Biggest problem with it is though that you can't throw high peak currents at it without burning up the brushes + commutator in the long run.

The Cityel is no really quick vehicle (neither with the old Thrige, even converted to SepEx, and even less so with the Perm PMG132), and I don't intend it to be with the brushless motor conversion; what I'm after is the lowest possible consumption on flat roads @ 50km/h.

I've played around with the ebikes.ca calculator quite a bit, @240A phase amps and low gearing (top speed without limiter is about 60 km/h), the acceleration is 6-7km/h-s, so 7-8s until 50, which is fine. The motor just needs to have a low enough phase to phase resistance now