Definitive Tests on the Heating and Cooling of Hub Motors

[Punx0r]

Apart from the red trace (what cooling method is that?) it seems aircooling is only marginally effecive.

 

[teklektik]

Hmmm - since these are temp plots along the same route, the shortened plots would seem to indicate higher speeds so the wattage expended/dissipated is likely higher. Shifting the x-axis to distance instead of time might better compare temperatures at identical terrain loading (hills, etc) - given that the throttle may still be somewhat different.

 

[bronz]

I am anxiously awaiting the oil cooling data!

 

[GCinDC]

Apart from the red trace (what cooling method is that?) it seems aircooling is only marginally effecive.

more than marginal.. sorry for any confusion. the chart is from the thread: Oil cooling your hub- NOT snake oil!

allow me to retell the story for the record, in colors that match the lines on the chart.

test 1. i laced a new ebikessf hs3540 in an MC wheel and rode a familiar but imperfect test route, and w/ sidecovers on it overheated alarmingly fast (9 min to reach 140C when i quit the test)

test 2. then i sealed the sidecovers, wire hole etc with permatex and poured the motor half full of dextron, 300ml or possibly more, and was delighted to run the same route and see the motor stay as cool as cucumber. i thought no more testing was necessary as world peace was practically at had. shortly thereafter, the oil leaked, and despite all my attempts to reseal it, it would always leak again...

test 3. then i dragged out an OLD rusty hs3540, drilled out the sidecovers pretty thoroughly, and was astonished to see that it cooled on more or less the same line as the oiled filled motor! at this point, i wished that i had tested the oiled motor longer to see how it heated from minutes 9-16 but i'd never know... at this point, i was convinced, at least for the effort, that vented motors were as good as oiled motors, maybe not slowly under heavy load, but for general commuting purposes at least... eventually, however, the old rusty hs3540 developed a phase wire short, perhaps from being abused in sand and creeks and junk. who knows.

test 4. with that motor down, i slapped its vented sidecovers on the motor stator had been oil filled, thinking it'd be just as cool as the old rusty vented one, but it was NOT much cooler than the stock motor! ????? i was flabberghasted, disappointed and kinda embarrassed, cause i knew it confused my tests here...

test 5. needing a backup, i got a brand new motor, this time laced in a 26" rim, drilled out the sidecovers and was disappointed to see poor cooling performance.

test 6. got tired of bicycle wheel and slapped the stator in the rotor with the MC wheel, and bigger holes, but was confounded as to why i didn't have the same cooling as i'd originally had w/ the vented motor (red line).

i concluded that even tho the motors must have been different at least in quality, venting cooled the motor adequately, manageably, at least at 3kw - except up steep hills, and that oil cooling, with a slotted sidecover at least, was no longer worth my time and mess.

hope that helps.

can't wait till justin makes the point defininitely, whatever it is!

 

[Teh Stork]

GCinDC, you have definitely documented the different motor cooling setups well. In the process I think you overlooked other variables. Did your bike change between runs? Was tire pressure equal between the tests? (It looks as if it is the same kind of tire). Ride style? You only have single runs - what can you say about repeatabillity?

Basically, I don't think saturating the windings with T9 leads to the observed temperature difference. As there are too many external variables not controlled, drawing a definite conclusion is impossible.

 

[pendragon8000]

Just wondering. What would be a good practical way to test different cooling methods on the bike.
I'm thinking a flat oval/park/field limit power to say 1kw and go wot around and around. Or more power if you want make it fun.
"Must stay WOT, for science sake!"

 

[Merlin]

... Two Same motors... One drilled one closed... Two controller, one throttle, one bike...

 

[kdog]

Are there oils that should or should not be used for hub cooling? apparently some mineral oils can attack plastics paints etc. that would be bad for insulation seals etc. im leaning towards adding some oil to my 9c- dont know if this is an issue...

 

[liveforphysics]

ATF.

Mineral oil is a bad choice for many reasons.

 

[madmaxNZ]

I use INOX MX3 in my Crystalyte H4080. It is very low viscosity and hasn't eaten anything. I've used it for about a year and do some very extended hill climbs up forestry roads...temp usually sits around 60degC but can climb to my cutoff point of 85-90degC on one steep-extended section.
See link below to website.
You only need about 100ml - 120ml, so the lower quarter of motor has the windings covered. It will greatly improve cooling.
You need a small breather hole: I drilled a small breather hole (2mm) on the freewheel side-cover near the axel. Also a fill hole on the other side (about 8mm) which you can thread and plug with a bolt. I use a syringe to fill.
That's it... sure beats drilling heaps of large holes in your side covers to let dirt and water in.

http://www.inoxmx.com/inox/mx-3-lubricant/

 

[Cowardlyduck]

Hey Madmax, how did you seal the motor?
I'm thinking of doing this with my HS4080, but was worried about the mess. I've got some household silicone sealant...would that work ok?
Did you have to do anything special to the axle area, or does it not matter since the oil sits below this level?

Cheers

 

[madmaxNZ]

There should be seals on the axel shafts already, so no worries there. I put some extra silicon around the cable exit groove.
I also lifted the side covers and put a bead of silicon under them to form a gasket. Household silicon should work...give it a crack, nothing to loose really. I used "Red RTV" I think it is RTV 162 from memory, but as long as a gasket it formed, it should do the trick. It is easy to get apart again and clean off and reapply. I've replaced hall sensors recently and a broken freewheel mount on the sidecover (Crystalyte have redesigned the freewheel mount with a metal insert)
As you say, the level is quite low, so the axel seals only get splashed...remember that a breather is needed or the oil will be forced out when motor heats up...I didn't have one at first and the oil was forced up the inside of the cables to the Anderson connectors!
I have a page which I need to update, but it shows a pic of my fill plug with a breather drilled thru it; the position of the breather there (at fill plug level) was too low and it seeped a bit, so I replaced that bolt and drilled the hole on the freewheel side, much closer to the axel (higher) any small seepage is on the freewheel side and not the brake disc.

 

[kdog]

Thanks Madmax- the INOX sounds good (1yr nothing eaten) I def like the simplicity of oil and keepIng the motor sealed. I commute on dirt roads and a fine spray of dirt gets every where so air vents= ingress. Also for me the benefit is in reducing the peaks as Im not doing long extended climbs more short steep ones spaced out.

 

[Cowardlyduck]

Thanks Madmax. Yeah I will be doing it on a Stealth modified HS4080 which already has 2 holes for applying WD40 or CRC to the windings periodically to prevent rust. I've been applying insane amounts of WD40 in an attempt to see if a small pool of it would help cooling, but it's too hard to tell.
I might drill a new breather closer to the axle as you suggest though as I think I will need to seal the existing holes since they are already at winding level.
Do you think WD40 would work just as well as MX3?

Kdog, I had that experience when I vented my HS4060. I think the build up of dust/dirt probably contributed to it's ultimate winding short to the stator.
If enough dust/dirt builds up on the windings, it acts as an insulator and traps all that heat. Having it open to the air at that point ceases to help cool it any further as dust/dirt is a very poor heat conductor.

Cheers

 

[kdog]

Irrespective of the test results, I think your point (cd) is the killer for air vs oil for me. Not too keen on dirt crudding up my stator and shorting it. Two holes one bolt a splash of lube and I'm done. Well you know what I mean!

 

[friendly1uk]

WD40 is gay spray. A light oil in thinners to lighten it further. This aids penetration, after which the thinners that carried it evaporates and the remaining oil becomes more viscous. Put some on a mirror and see. You would need many cans and I'm not sure what sort of residue it would leave. Hence putting it on a mirror to see this particular brands usefulness

 

[pendragon8000]

http://www.nulon.com.au/products/Automatic_Transmission_Fluids/Full_Synthetic_Multi_Vehicle_Automatic_Transmission_Fluid/#.Uy7B28IsXFo
A synthetic transmission oil would be ideal right? Its non reactive and 100% pure. Not from crude oil.

 

[Punx0r]

John in CR made an interesting point in another thread: Oil cooling increases the thermal transfer from the windings to the coils (good) but also increases transfer to the magnets (bad). With the magnets normally much less tolerant of high temperatures than the windings what is the overall outcome?

Are we seeing a genuine increase in the motor's ability to shed heat from the case to the surrounding air, bearing in mind its fixed and limited surface area? Or is the oil just adding thermal mass and thermal equilibrium temperature of the motor doesn't change?

I do keep going back and forth trying to decide whether to go for air or oil cooling. Both options seem heavily compromised for all terrain/weather riding.

 

[sardini]

I use INOX MX3 in my Crystalyte H4080. It is very low viscosity and hasn't eaten anything. I've used it for about a year and do some very extended hill climbs up forestry roads...temp usually sits around 60degC but can climb to my cutoff point of 85-90degC on one steep-extended section.
See link below to website.
You only need about 100ml - 120ml, so the lower quarter of motor has the windings covered. It will greatly improve cooling.
You need a small breather hole: I drilled a small breather hole (2mm) on the freewheel side-cover near the axel. Also a fill hole on the other side (about 8mm) which you can thread and plug with a bolt. I use a syringe to fill.
That's it... sure beats drilling heaps of large holes in your side covers to let dirt and water in.

http://www.inoxmx.com/inox/mx-3-lubricant/

Whats is your face amps?
Please try 210 phase amps and dead stops on trails so i can have a comparison.
My H4080 now has a tube which supplies water to the internals of motor.

 

[madmaxNZ]

Guys, the up-shot here is to give it a go.
It's only two small holes and some oil.

Check the viscosity of the oil you intend to use and what it claims not to dissolve. I got a spray can of the MX3 from Supercheap Auto (pretty sure its an Aussie franchise which landed here in NZ a while back, so should be no worries getting it for our Aussie brothers...Pendragon8000 has suggested another option too; I'm sure there are others; not sure about WD40 it may be too thick). I spray it into a measuring container and squirt in with syringe (piece of weez).

Punx0r : the idea is that the motor does not get hot enough to degrade the magnets by using the oil, so overall, the transferred heat is lower than with no oil (it's better than nothing). I set my temp monitor to start foldback at 85degC and cutoff at 90degC. At that point I've gone a fair way up the hill so it's time to stop and enjoy the view anyway.

Sardini: I don't know my phase amps, but I run 24s LiPo and pull about 30 to 45Adc on my extended climb of about 11 Km. I try to keep the speed at around 30Km/Hr without stopping. I'm just reaching the 90DecC point at the top of the hill on a warm day (just in time to head down and cool off). Obviously stopping and starting a lot, will heat things up a lot quicker. I do a total 42Km loop of ups and downs on shingle forestry roads and use about 13AH of my 15AH pack (first 11Km is the longest climb and uses most of the packs power) .

We all have different requirements, oil is a simple option to give a go and see if it suits you.
Gents: Don't over think it...time to "suck it and see" :lol:

 

[parabellum]

Not very much numbers here but I have few Mxus 500w motors to mess around. I would push about 1000w continuous on flat ground to balance at 90-100C on wingdings (30C outside). With 300ml ATF inside it stabilizes on 82C pushing 3000W continuous on flat ground. It is seeing about 7000w peaks while climbing. Thermal mass of the oil gives me opportunity to have much longer bursts.
Setup:
Mxus 36v 20" 500w hub
17" moped rim and tire
Lyens 18fet 4110
24s 25Ah LiPo
Honda CB 125C chasis

 

[liveforphysics]

When thinking about cooling a sealed hubmotor, remember that regardless of what temperatures may happen to be at various places inside, only the temperature of the outside surface of the motor actually gets to transfer thermal energy from the system into the environment (heating air that touches it, radiated energy etc).

Anything you can do to get the outside surface of the case of the motor HOTTER will directly increase the ability for that motor to shed thermal energy.

Many of us have had experienced smoking a hubmotor's windings (>250degC) while the outside shell was barely above room temp. Oil cooling can drastically reduce the chances of that type of failure due to the improved thermal coupling of the system that the oil enables.

The data I want to see, is the exact no-load vs speed changes incurred by doing the partial ATF fill. I know the fluid drag losses should start out like almost nothing, and then increase with RPM, and at some speed it has to become a substantial loss (and all of this fluid drag loss would ultimately end up as heat in the motor). I'm just curious to see that at ebike hubmotor speeds it never reaches the steep part of the exponential fluid-drag curve. My hunch is that even in a hubmotor doing full-speed RPM, that the fluid drag will still be in a very low place in the fluid drag curve.

 

[cal3thousand]

Sounds to me like finned motor cases and ATF are the way to go

 

[liveforphysics]

Sounds to me like finned motor cases and ATF are the way to go

That seems likely.

Something very fortunate about that, is of all the parts of the hubmotor one may to need to make custom, the side covers are fairly simple parts to CNC.

Bearing pocket in one side, along with small surface fins to improve heat transfer with the oil at a minimum additional drag expense, and large fins on the outside to assist in heat transfer to the air passing by them as you ride.

 

[Punx0r]

If you had the patience you could turn a series of small grooves in the face of the covers, say 1mm wide x 1mm deep spaced 1mm apart. That would double the surface area.