Cyclone 650w-1680W up to 60v @2100W without overheating

[richj8990]

When you put more voltage on this engine you need to consider:
- The weakness that came from a cheap motor kit
- The problem that came when you put a lot of power on a bike not build for that (overheating and transmission problems)


Weakness came from cheap motor kit
Cyclone 1680w engine came with mounting brackets in ergal, very light but not strong enough!

I know what you mean, on a different chain drive kit I'm really needing to shore up the non-drive side. I think it will finally work but yes the motor/sprocket will flex about 10mm over to the drive side and not only can the chain pop off, the sprocket can hit the pedal, or should I say the pedal hits the sprocket.

If you put enough power on the engine, the brackets flex under the engine torque causing the engine chain spill out from the cog on crankset (terrible situation).
To make things worse, the only way to fix engine at the opposite side of the cog (see the picture below) is with steel clamp holding the left bracket …..no way that it works with the desired power!

 
Overheating problem
After many tries, I figured that you have overheating problem if you run the engine using the incorrect gear (maybe you haven’t the correct one!).

The only way to avoid overheating, that is a very common problem for this engine (even without power upgrade), is to use the engine at elevated speeds. Thanks to Endless Sphere I learned that more fast engine spin less heat is made (it something related to the current impulse), it works!

Make some calculation:
Engine specification from Cyclone website are “Electric Motor kit 1680 Watt 48Vup to 60Nm, speed 4500rpm” and they realized this graph (handworking?)

If my readings are correct, to have maximum torque the engine must run at 1000 rpm and to have the maximum efficiency (EFF) is @ 3000 rpm.

The gear reduction is about 1:30. Then, if the engine spin at 3000rpm, the pedaling cadence is about 100rpm (3000rpm /9.33 primary gear reduction/3.15 secondary reduction). For humans is quite hard keep 100 pedaling rate per minute. At upper speed it became

Bottom Line: if you want to run your engine at the optimal speed forget to follow the engine by pedaling!

How I can manage that?
It is quite simple, going on flat I love pedaling. I use engine at not-best performance (cadence at the crank set about 60 rpm)
On mountain, I would like to use it near to maximum speed 4500+ rpm (good torque and good EFF), simply pedaling is impossible.

Configure the correct gear to avoid overheating
The easy way: put throttle to the maximum, if the engine speed increase slowly you must change gear to the light one.

To avoid overheating you need to change your riding stile, the throttle must be always near to the maximum (full gas) and you can regulate the bike speed you want by using the bike gears.


The overheating problem came on long slopes, after a while engine stop spinning and became very hot.
It is because engine strive al low rpm unable to reach optimum speed. The only way to avoid overheating is to reduce gear to reduce engine strive and reach the correct engine spin.

Unfortunately neither with 32t front and 36t rear gear I can reach optimum engine spin, so can we do?

40t 42t sprocket/cassette/chainring adaptor/expander MTB 7/8/9 speed. I put on the adaptor the 44t 104mm cog (my deore long cage derailleur liked it)
To avoid cog slip I also fixed my sprocket cassette with the sprocket adaptor with two bolts to distribute torque generated by the engine on the entire hub freewheel.
The result is: no more overheating and incredible climbing hill ability. My fatbike wheel loose traction on rocks before enging stop spinning.

I think that's true for a lot of these motors in general; mine seems also to be around 30:1 total reduction, but that calculates to 50 rpm pedal cadence, not 100 if it's at the normal efficiency of 1500 rpm. Are you sure the best efficiency is 3000 rpm and not 1500 rpm. The other issue is that if you are in a lower gear, even uphill the cranks will spin out and you can't go any faster. That's really annoying. On an e-bike, going 3-4 mph uphill even on singletrack seems ridiculous; I can go 2.5-3.0 rpm purely on human power up a 7-12% grade with a 25-30 lb normal bike anyway, so what's the point of having electric power? You want "and" need to go faster on an e-bike than that uphill, so a really low gear is not going to cut it for the entire purpose of having the bike to begin with, efficiency be damned. If you can't go more than 5 mph uphill, what is the point of having an e-bike offroad?

Right now I'm using 46T drive chainwheel, 36T cassette chainring, and 11-52T rear cassette. Those three pieces of metal are large enough that somewhere in-between there is a decent efficency for the motor uphill. I like either 28 or 34T in back; that gives fairly close to 1:1 for the cassette chain, and then 46/16T (sprocket) gives close to 3:1 reduction. You can't make it work with 32 front / 36 rear? For me, going to 46 or 52T in back again seems ridiculous and my chainline won't allow it anyway. I'm using the rear wheel for other non-bikes so that's why I'm still keeping an 11-52T cassette on it; but it's simply not needed for the mid-drive.

 

[richj8990]

Ooookey.
If you want to know the speed, you need to identify the cadence, that is the number of rotation of the crankset (rpm = round per minute).

Cadence = <Engine Speed under load> x <Total reduction Ratio>
Cadence = <70% of Engine Speed declared> x <Gearbox Reduction x Engine Freewheel/engine cog>

For example in my case:
“Electric Motor kit 1680 Watt 48V up to 60Nm, speed 4500rpm”

The values:
- Engine Speed declared= 4500rpm
- Engine Speed under load = 3150rpm
- Gearbox = 1/10 (from Cyclone datasheet, for Cyclone 3k I thing 1/6)
- Engine Freewheel = 13t (Standard Cyclone is 14t)
- Engine Cog = 48t (Standard Cyclone is 44t)

Cadence = 3150rpm x (1/10) x (13/48) = 85 rpm

For 1800-3000W
"Cyclone 1800-3000W 36-72V 100Nm 100Km/h 600-1000rpm"

The values:
Here is the messy point....MY GUESS is that 1000rpm is not the Engine Speed but the Engine Freewheel Speed
IF it is true
Cadence (@72Volt) = 1000rpm x 70% x (14/44) = 198 rpm WTF!!!!!
Cadence (@48Volt) = 700rpm x 70% x (14/44) = 155 rpm

Note:
Normal people have pedaling cadence from 60 to 90 rpm

Once you had the Candence evaulate speed is quite easy, we can use this calculator Bicycle Bike Gear Ratio Speed and Cadence Calculator.

In my case (fat bike has 30inc wheel)

At full speed yesterday I reach 54km/h it is because the engine overvolted speen faster then 4500rpm.

Note about the torque:
there is NO WAY that my engine put 60N/m @4500, my gess is that engine put 6N/m, considering the reduction 10:1 in the gearbox, at the Engine Freewheel We have 60N/m (like a Ford Fiesta).

For Cyclone 3k 100N/m is at the Engine FreeWheel.

I'm guessing my internal reduction is 10:1 too. You are using a 44T chainwheel and the sprocket is 13-14T. That's at least 3:1.

Under load 3150 rpm..., 30:1 reduction is yes a bit over 100 rpm, but that's assuming a 1:1 pedal chainring to cassette, as in 36T front 36T rear. Correct me if I'm wrong but the Bafang BSSxx models only have around a 22:1 reduction; they come with a 44-52T chainring and most people use 8-speed, so the max gear is usually 34 or 40T. That's far worse of a scenario than this one with 30:1 reduction. And their drivetrains work. Something is wrong in the calculations...I think it's the engine efficiency that is not 3100 rpm. It really should be 1500 rpm, no?

 

[richj8990]

One other thing: I think you are assuming full throttle with the cadence calculations. Half-throttle may not be proportional, eg. 50% throttle may not really be 2250 RPM when the max rpm of the motor is 4500, but for the sake of argument let's say it is.

50% throttle = 2250 rpm = may not be exactly half of 1680 watts, but will provide a ballpark of 840 watts, which is a LOT for climbing with a mid-drive

10:1 internal reduction = 250 rpm

44T chainwheel / 14T sprocket = 3.14 reduction = 79 rpm

cadence normally 50-90 rpm, so if you had a 32T in the front, then you can use between 28T - 40T in the rear


This makes sense because on a flat surface, if you are doing 32/11 you just negated the drive chain reduction and are back to 250 rpm IF the throttle was 100%. But it's not. Throttle on a flat surface is maybe 15-30% which means rpm is 15-30% of 250 = 40-60 rpm cadence for a slow, cruising pedal speed. This makes more sense...you have to estimate the watt output first with how much throttle you are using for what gradient.

 

[DanGT86]

Old thread here but figured its worth a shot.

I have one of these old cyclone 48v 1680watt motors. I just ran it without the gearbox attached on 75v.

With no load it pulls 3.7amps at 9400rpm and it heats up to 50 degC in seconds. After a few min its still around 60deg C.

Does this seem normal? Seems too hot. Maybe the rpm is just too high and there are eddy current issues making all this heat?

I was going to use it for a project but dont want to invest the time in it is going to overheat instantly.

 

[amberwolf]

No load you should have very very little heat generated. (assuming it's brushless)

Either it's a false-positive phase/hall combo, or as you say the RPM is too high and it's eddy current heating. The latter would take quite a bit of "overspeed" to cause, AFAIK.

Other possible causes include bearing issues or lack of lubrication, but those would also probably cause quite a racket, and might even just sieze once cooling down.

 

[DanGT86]

No load you should have very very little heat generated. (assuming it's brushless)

Either it's a false-positive phase/hall combo, or as you say the RPM is too high and it's eddy current heating. The latter would take quite a bit of "overspeed" to cause, AFAIK.

Other possible causes include bearing issues or lack of lubrication, but those would also probably cause quite a racket, and might even just sieze once cooling down.

Its a nucular controller so it did its own hall/phase learn and motor setup.

Tried it in FOC, square, sensored, and sensorless. Had similar result. It sounds nice and smooth but the heat was instant.

I was hoping someone with one of these cyclones would know if the instant run up to 50 deg was normal.

Im going to run it with a couple other controllers and see how it reacts.

It cogs really bad so its hard to figure out if its a bearing. It looks like the board would need to be desoldered to get the rotor out to check the bottom bearing.

3-4 amps didnt sound too out of line to me.

 

[amberwolf]

If it has severe cogging even spinning it by hand without being connected to the controller, maybe it has a partially shorted winding (to the laminations).

If it's a hard short you can test this with an ohmmeter. If it's a soft short it requires some voltage to cause it so the ohmmeter will probably read normal open circuit between any phase and the laminations.

You cant test for phase-phase shorts with a regular ohmmeter, but it's likely this kind of short would cause controller errors or spin problems anyway.

How does it behave at it's nominal-rated voltage range?

 

[DanGT86]

got about the same readings with my multimeter.

Tried limiting the motor to 54v and at full speed unloaded it gets to 60deg C in 2 min.

 

[amberwolf]

got about the same readings with my multimeter.

I don't know what you mean? Readings of what? tested how? compared to what?

Tried limiting the motor to 54v and at full speed unloaded it gets to 60deg C in 2 min.

FWIW, the image shows 84v for "max motor U"? Is max motor U actually a voltage?

Note that the full unloaded speed for 54v should be proportionally lower than at the 84v. Was it?
If not, it's not doing what you'd need to do to see if it makes a difference, since the point of lowering the voltage is to lower the rpm at full throttle to see if it's rpm-related.

Also, if it's limiting the voltage by limiting the PWM duty cycle, it may have different results than using a power source for the controller that is actually the lower voltage. (I don't know what specific difference, as that depends on the motor and how the controller does whatever it does).

 

[DanGT86]

Sorry for my careless vague post. I was posting from a phone.

I was trying to say that my cheap multimeter showed 0.1ohms between any 2 phases which is roughly the same phase resistance the controller was displaying.

I'll check for a partial short to the laminations when I open the motor again.

The 84v motor setting in the pic was before I set it to 54v and the max speed was proportionally lower at 54v as expected. It was mid 6000rpm. This is more in line with the 48v nominal this kit was sold as.

So it still gets hotter than I would have expected with no load. I cant imagine it will stay cool under load unless for some reason free spinning is worse than loaded for creating heat.

 

[amberwolf]

I was trying to say that my cheap multimeter showed 0.1ohms between any 2 phases which is roughly the same phase resistance the controller was displaying.

Unfortunately, as I'd noted before, you can't use a regular meter to check phase to phase resistance; your meter leads probably have as much resistance as the windings, or more.

You would need a meter at least as sensitive as the DE5000 I use here to do a phase to phase check, assuming it's a hard short. To check for a soft short you'd need something that can create high voltage to do the test, like an isolation tester. (this is how I found the problem with an MXUS some years back).

But if it did have a phase to phase short, it would probably not run correctly anyway, or the controller would indicate some sort of problem with a current imbalance (assuming it's smart enough to check for that).

I'll check for a partial short to the laminations when I open the motor again.

You can check this without opening it, by measuring between the phases and whichever part of the motor is fixed to the stator (either shaft, axle, or case, depending on the motor design).

The 84v motor setting in the pic was before I set it to 54v and the max speed was proportionally lower at 54v as expected. It was mid 6000rpm. This is more in line with the 48v nominal this kit was sold as.

If the speed is lower and it still gets that hot, then my money is on a fault with the motor, or with the way the controller is driving it (incorrect detection of motor properties, etc--autodetect/autolearn functions don't always work correctly).

So it still gets hotter than I would have expected with no load. I cant imagine it will stay cool under load unless for some reason free spinning is worse than loaded for creating heat.

I'd guess that under load will probably heat up beyond what the motor could deal with.