Air Cooling my x5

[ZapPat]

OK guys, there's one common misconception here on the ES forum and specially on this thread lately that almost everyone has been repeating and that annoys me:

Watts don't heat up your motor, it's really the motor/phase current that heats it up!

The sad news is that your average Watt's up / turnigy watt meter / cycle analist does not display this value to you, it shows *battery* current only, so it is a bit complicated to know just how bad you are really abusing your motor. And the lower your speed is, the greater your motor-heating phase current is compared to your easy-to-read battery current and your displayed watts too of course.

Keep that good testing info coming, El_Steak!!! I just hope you are patient enough to remove and reinstall the wheel and covers for each test to make it to test #5!

Pat

[John in CR]

Pat,

I understand that it's the current through the windings creating most of the losses in a motor, but since they have a torque constant, why is it wrong to look at power out from the battery as long as you consider the motor's efficiency curve? Losses at the controller can't be too significant or variable (mine are near constant warm temps), so other than a small amount of wire losses, the bulk of the losses occur in the motor (ignoring losses to wind friction). Power out of the battery - work done = heat in all its forms.

[ZapPat]

I understand that it's the current through the windings creating most of the losses in a motor, but since they have a torque constant, why is it wrong to look at power out from the battery as long as you consider the motor's efficiency curve?

The biggest source of losses in a hub motor are dependant on motor phase current squared or more (copper + magnetic eddy & hysteresis losses), the RPM dependant losses look to be mostly linear instead of squared, except for eddy losses also going up squared with RPM. Source here.

It seems hard to me to use an efficiency estimate to help you because efficiency changes with both currrent and RPM.

Anyone please rectify anything here I've said that's not exact... I'm not a motor pro for sure, I mostly know controllers pretty well.

Pat

[John in CR]

I don't think the relationship of current and efficiency can be stated that way. eg A motor's efficiency doesn't change a lot between it's points of maximum power and maximum efficiency, but current changes a good bit. Otherwise I think we're saying the same things just from different viewpoints.

I think a confusing part is how controllers do that current multiplication thing at takeoff to generate more torque, but doesn't it do that by reducing the effective voltage for a certain power from the battery? Yes those motor losses are high due to low rpm, but there's also more work getting done due to the higher torque than if the current multiplier was turned lower, making the increased losses have a shorter duration to reach the same speed. Overall losses are the same for a slower lower current acceleration to that speed, because the same work is done over the same efficiency curve related to RPM.

If that's incorrect then I need help with that explanation too.

John

[El_Steak]

Run C is done and numbers are logged !

Its cooler, with a top temp of 117, over 20 degrees less than Run A and Run B. I made the run at dusk though and at 19C is was starting to get chilly. How does ambient temp affect the results? I don't know, I guess I'll have to do another run to test it out

This latest run is with my old covers, with 6 additional 7/16 holes close to the center and 24 x 7/32 holes close to the outer perimiter. As suggested by John, the outer perimiter holes are "angled" so their edges "bite" in the inside air to push it out. Angles are in opposite directions on each side.

Go see the results for yourself in my test post:
viewtopic.php?f=2&t=9791&p=278852#p278852

I've decided I'll squeeze in another run with that same cover. The only difference will be that I'll tape over the 12 holes in the middle of the cover. As John suggested, they might be counter productive as they could be letting the fresh air out before it reaches the outer holes. Hopefully it will be a bit warmer tomorrow night as well.

If you plan on drilling your covers, here's my little jig to drill perfect holes in the curved area around the perimeter:
The "axle" in the lower piece of plywood allows me to rotate the cover and always maintain the exact position and angle for the drill bit.

Then to create the final oriented tapering as suggested by John, using a portable drill, stick the drill bit in the hole and slowly tilt it while it spins to the desired angle. Works great, aluminum is like butter.

Works great for the other holes on the flat surface as well.



OK, got to go give some time to the GF, shes not too impressed with my testing...

[John in CR]

El_Steak,

You're aces man, great follow through. I'm glad you're seeing solid positive results. I think it can do significantly better, so we should talk about interior blades specifically for the 9c. I don't think it will be about getting the air spinning as much as it will be about directing the flow away from the side cover as a more turbulent flow at the stator and maybe even at the windings too. I can't wait to go to town on my 9c when it arrives.

The higher peak amps is interesting. I wonder if that's just because of cooler windings lowering the resistance. Did you notice any difference in how hard the bike pulled?

How about the sonics, did you get any whistling? Did it change that other noise the 9c is known for, or did yours not do it to begin with?

Can I get you to do 1 pic? I'd like to see the cover from the outside but with the view parallel to the motor. I want to see where the hole placement is in relation to the flange where the cover's bolt holes are. I ask because I think you can draw the flow out from even closer to the windings by butting those holes right up to that flange where the cover is near perpendicular.

One other question, how did the covers feel to the touch? That's really all I know, since I've yet to use a temp gauge, which will change shortly.

Again, great job, and great job sharing with everyone.

John

[Arlo1]

So the ambient temp is 9 deg cooler and the last run the motor was in the 10-15 deg cooler range. So if the ambient temp was the same we could expect 1-6 deg better cooling then no holes at all? Looks to me Johns Bulshit didn't do much at all!
Great work el_Steak We are ll thank full for your work and the fact you posted the numbers and mesured them.

[Arlo1]

Arlo1,

That's why I handed you the shovel, just for the entertainment value in seeing you dig the hole deeper. BTW, just like I said before:
1. You owned those holes as soon as you made them bigger to copy your hero Methods.
2. During your wheelie sessions you're not running anywhere near the power you claim.

You kidding me? Have you seen my video? Where is you video racing a truck, car, bike, or what ever, rolling smoke show? Burnout? Dude ask anyone if I wasnt running the power I claim then how did the video happen?
I also chamfered the holes but your selective hearing choses not to hear that. I chamfered them to pull air in the inner holes and out the outer hole and on my wheel the innners are as close to the center as posstible and the outers are beside the windings so not as bad as you are trying to make everone belive.
As for methods he may not be my hero but he is a great shit he works hard for the endlessphere comunity and guys like you have to ruin it WTF?

Edit: Maybe you should read this on page 21 John!
Test #1 With fan on!
Outside temp 9.5 deg C(mesureed from outside wireless temp sensor)
Motor start temp 17 deg C (house temp)
Test point #1 temp 60 deg C
Test point #2 temp 70 degC

Test #2 With fan off!
Outside temp 8 deg C
Motor start temp 18 deg C
Test point #1 temp 59 deg C
Test point #2 temp 122 deg C
Yeh thats right 52 deg hotter without the fans and the ambient temp was 1.5 deg cooler!

[John in CR]

Keep on digging boy. Next I'll send you a rope.

[Hyena]

Great work el steak. I was watching this thread a while back but you've added alot of good info to it lately!

The holes in my GM cover aren't placed ideally - just outside the bearing on one side and about an inch from the outer lip on the other side but they have dropped the motor temperature noticably. I didnt have a temp sensor in there when it was sealed but before I'd burn my hand after a hard ride, now it's hot but I can keep it there. I'm running less extreme power - 18S lipo with a 100a limit but I haven't seen temps much over 90oC after a hard ride(max was 93 I think)

Air cooling mods aside it's good to see you having cooked any motors yet with internal temps well over 100oC.

OK, got to go give some time to the GF, shes not too impressed with my testing...

Don't I know that situation all too well. Wait until one day your GF becomes a wife and then there's a baby or 2 thrown into the equation.

[Toorbough ULL-Zeveigh]

.

Hyena:"Wait until one day your GF becomes a wife and then there's a baby or 2 thrown into the equation."

natural_parenting.png (21.31 KiB) Viewed 432 times

[ZapPat]

So El-Steak, will the next run be with the mid-radius holes plugged up? These extra holes in the middle may be reducing the effectiveness of the other well-placed holes, maybe?

[markcycle]

I'm going to add my experience with air cooling holes

it's my observation that the big problem with these motor is not the winding temperature per say, but the magnet temperature, as above 80C the magnet start to demagnetize. I have found with the motors I sell that the winding can sustain 140C measured and peaks of 160C. But if the magnets get above 80C then a degenerating effect begins because of the magnet getting weaker the motor needs more current hence more heat and this positive feedback continues until the motors finally burns out or the magnet let go and spin.

Air cooling holes in the side covers helps keep the magnet below 80C preventing this degenerating effect. With air cooling I can regularly see 125C on the winding but can keep the magnets below 70C. without air cooling I have to set the thermal shutdown at a winding temp of 100C to prevent the magnet form going over 80C.

I have been running air cooling holes on my test motorcycle for a year now (3000 miles) and have found that with good screens over the holes there is no problem with dirt or rain. I regularly inspect the windings and they look amazingly clean.

Mark

[markcycle]

for those who haven't seen the 600 series motor EnerTrac sells with cooling holes

[El_Steak]

The higher peak amps is interesting. I wonder if that's just because of cooler windings lowering the resistance. Did you notice any difference in how hard the bike pulled?

No, it felt the same. I don't think those "peak" numbers have much value, they last only a split second and vary a lot. I'm still recording them jsut for the sake of completeness (and I have to wait 2 minutes after each segment anyways).

How about the sonics, did you get any whistling? Did it change that other noise the 9c is known for, or did yours not do it to begin with?

My new 9c makes the same growl as the others I could hear in various videos. I didn't notice any difference in terms of sound with the new swiss-cheese cover. I wear a full face helmet though so my hearing is not the best.

Can I get you to do 1 pic? I'd like to see the cover from the outside but with the view parallel to the motor. I want to see where the hole placement is in relation to the flange where the cover's bolt holes are. I ask because I think you can draw the flow out from even closer to the windings by butting those holes right up to that flange where the cover is near perpendicular.

I'll take that pic tonight. I think there's still some space between the flange and the edge of the holes, I left some meat so I could enlarge those holes if need be.

One other question, how did the covers feel to the touch? That's really all I know, since I've yet to use a temp gauge, which will change shortly.

Hummm, I don't have good consistent data on that, I remember touching them a few minutes after I came back last night and they were just warm. I don't remember touching them on the previous runs though.

[El_Steak]

So El-Steak, will the next run be with the mid-radius holes plugged up? These extra holes in the middle may be reducing the effectiveness of the other well-placed holes, maybe?

Yes, that is what I intend on testing tonight.

[BlackArrow]

Hi Guys,

This post is really interesting, amazing work's El_Steak and ZaPpat you really read my mind I have the same thinking about the holes. Mark good information is nice to see the hole on your motor I never see it before thank you.

P.S. Arlo1 please, if you have some grief’s about John idea's can you send him a email, and let’s us try different approach and testing for cooling hub motors thank you!.

Good Day Guys!
Black Arrow

[ZapPat]

I don't think the relationship of current and efficiency can be stated that way. eg A motor's efficiency doesn't change a lot between it's points of maximum power and maximum efficiency, but current changes a good bit. Otherwise I think we're saying the same things just from different viewpoints.

I don't think this is true, and I double checked using Justin's simulator to see some power/efficiency relationship examples. Using some arbitrary settings I found one example to give maximum power at 35kph (4300W), and maximum efficiency for the same settings at 75-80kph (1300-1700W). It's really too bad that the simulator doesn't generate a "motor/phase current" line on the chart, it would be pretty easy to add it seems to me. Which brings us to your next point...

I think a confusing part is how controllers do that current multiplication thing at takeoff to generate more torque, but doesn't it do that by reducing the effective voltage for a certain power from the battery?

Controllers use PWM to actually *reduce* startup torque, not generate more. Having no PWM in the controller would be like using a brushed motor and a On/Off switch... this would give very brutal startup torque. Motor current and voltage in this situation would be equal to battery current. Note that this situation can happen at startup even when using a PWM controller if: your motor's+controller's+wire resistance is high enough, and/or battery voltage is low enough, and/or your current limit is high enough to naturally limit the current flow without needing the controller do PWM to reduce the motor voltage. The controller would just act as a commutation source in this case. However, in most high performance setups this would not be desirable to have happen, as you will be dumping most of your energy into heat while doing this.

Yes those motor losses are high due to low rpm, but there's also more work getting done due to the higher torque than if the current multiplier was turned lower, making the increased losses have a shorter duration to reach the same speed. Overall losses are the same for a slower lower current acceleration to that speed, because the same work is done over the same efficiency curve related to RPM.

I understand what you mean here, but I think there is an ideal acceleration profile to aim for to achieve maximum efficiency (Wh/km) with any given system. We would of course have to do a number of acceleration tests at different rates to find out what is the best acceleration rate to optimize efficiency for any given setup in a given situation. Changing the slope of the hill or the wind speed would change this acceleration rate though. In fact the ideal rate itself (determined by motor torque/current) will likely be non-linear with respect to motor speed, and this brings us to your next point...


ES member Miles said (in another thread) that a motor's best efficiency point is when the copper losses are at par with the other parasitic losses. I might dare to expand this statement to say that a motor's best efficiency is achieved when all the current-based losses are at par with all the RPM-based losses. This implies that as speed changes, the most efficient motor/phase current to use also changes. However, since both the RPM and current based losses are each quite complex with many linear and exponential variables, the ensuing relationship to find the best efficiency point will not be linear. Here's a good post about a motor's theoretical losses.

Pat

[ZapPat]

it's my observation that the big problem with these motor is not the winding temperature per say, but the magnet temperature, as above 80C the magnet start to demagnetize. I have found with the motors I sell that the winding can sustain 140C measured and peaks of 160C. But if the magnets get above 80C then a degenerating effect begins because of the magnet getting weaker the motor needs more current hence more heat and this positive feedback continues until the motors finally burns out or the magnet let go and spin.

Air cooling holes in the side covers helps keep the magnet below 80C preventing this degenerating effect. With air cooling I can regularly see 125C on the winding but can keep the magnets below 70C. without air cooling I have to set the thermal shutdown at a winding temp of 100C to prevent the magnet form going over 80C.

I have been running air cooling holes on my test motorcycle for a year now (3000 miles) and have found that with good screens over the holes there is no problem with dirt or rain. I regularly inspect the windings and they look amazingly clean.

Mark

That's very good info indeed Mark! We had not really considered this aspect very much so far in this thread, but it does make very good sense.

So this means that they use regular low temp "N" grade neo magnets in these motors. Do you know what strength grade they mostly use in these DD hubs BTW, Mark? I hope it's at least "40", but I wouldn't be surprised if they use real cheap weak ones like N35 or N38...

Here are some neodymium magnet specs for anyone interested.

Pat

[John in CR]

Great point Mark, that we need to keep the windings cool so they don't cook the magnets. The possibility of cooking the windings is only during very short term peaks because the varnish on magnet wire handles far higher temperature than the magnets. Getting some air flow through the gap, was the primary reason I decided to go with intake just on one side (exhaust on both sides).

Also, thanks for sharing that a screened intake works well, though we need to keep in mind it will significantly restrict flow, so we need bigger cross-sectional area for the intake holes if we screen them to achieve the same flow. How did you attach your screens? Are they removable for cleaning.

Lot's of mileage and motor still clean...Great, I need to stop worrying so much, screen my intakes, and put my ventilated motor in permanent service.

John

[BlackArrow]

Hi Guys,

I have made 8 holes, like those ones on the picture on one side only. One thing for sure I never drive my ebike at fast speed (60 to 70 Kph)for so long period of time and when a stop the motor is cooling very quickly each 3 or 4 second I can see the motors cooling down on my probe.

I didn’t if another member here at ES has drilling is holes like I did but I’m really happy with my result so far and will try to do the same holes on the other side when I can find a free wheel removal tool.

Good day guys!
Black Arrow

[El_Steak]

Run D is done and numbers are logged !

This run is with the same covers as Run C, but I covered up (taped) the middle row of holes (12 holes per cover). It was suggested that they might be counter productive by allowing cool air to escape before it has a chance to hit the winding. The results are pretty close to Run C. A bit hotter in the last segment, but I think I popped one wheelie too many

Ambient temperature was a bit less chilly than last night at around 21C.

Go see the results for yourself in my test post:
viewtopic.php?f=2&t=9791&p=278852#p278852

Run E will be after adding the cooling fan blades inside the covers. I have a backlog of stuff to do now, so for this next step I will wait to see what designs John and Pat come up with on their covers. Radial would be easy with the existing ridges in the cover, but forward curved might be more efficient at low RPM.

[El_Steak]

Can I get you to do 1 pic? I'd like to see the cover from the outside but with the view parallel to the motor. I want to see where the hole placement is in relation to the flange where the cover's bolt holes are. I ask because I think you can draw the flow out from even closer to the windings by butting those holes right up to that flange where the cover is near perpendicular.

Not sure if that is what you were looking for, but I added the pic in the test thread under "Run C"

[John in CR]

Can I get you to do 1 pic? I'd like to see the cover from the outside but with the view parallel to the motor. I want to see where the hole placement is in relation to the flange where the cover's bolt holes are. I ask because I think you can draw the flow out from even closer to the windings by butting those holes right up to that flange where the cover is near perpendicular.

Not sure if that is what you were looking for, but I added the pic in the test thread under "Run C"

Thanks, yes that's the angle I was looking for. Your holes do reach the part of the cover perpendicular to the sides. I've pretty much decided on slots instead of holes, though I'm not sure how I'll make them. I want to stay on the flat part at the perimeter, up against the flange and not reaching the curve.

I ended up going with a blade angle as a reverse curve. My rationale was that they create the most pressure, and I was concerned about flow restrictions. That's the kind of blade/vane shape used in vacuum cleaners and I believe the pathways in disc brakes on cars. Maybe we can get someone in here who really knows this stuff to help us go the best route with the blades and angles. I remember being up in the air, and it was the greater noise level of forward curve and radial that was the final straw to push me to the reverse.

One thing I have also been thinking about is the intake, and I'm going to stick with 1 sided, so about half of the flow through the motor has to flow through the stator and magnetic gap. Both covers will have exit vents, so it's not pulling in one side and out the other, but instead in one and out both. I have a couple of thick AL disc's, so if I go with in-wheel use I can reinforce the intake side cover and go with very large intake holes close together without concern about structural integrity of the cover. Large intake with some kind of screen that can be removed easily for cleaning, without pulling the motor, is the moving target I have in mind right now.

[icecube57]

Intresting data. I seem to have your B run covers on Steroids. I might drill more holes at the edges to make them more like your C run covers but i dunno if there would be a cooling increase since my existing holes are so big. I think it all ties back into maybe we are over complicating this a bit. Remember we brought up the questions about intake vs output. That the holes needed to be the same size. We also talked about hole surface area. In you D run you coverd up alot of holes. I think it messed up your input output hole surface area ratio so it ran a little hotter.