Increase Torque up to 30%...Ferrofluid fills magnetic gap

[John in CR]

Fechter mentioned in Justin's motor cooling thread that he heard about benefits of ferrofluid in the air gap. I found this paper https://www.wpi.edu/Pubs/ETD/Available/etd-041811-170656/unrestricted/ajudge.pdf , which talks about a 30% reduction in Kv by eliminating the air gap losses using ferrofluid to fill in the gap of permanent magnet motors. I immediately think wow we can increase torque/amp by 30% without increasing copper losses.

A cursory review of the paper shows that it's most effective with low speed motors (under 500rpm), but it's hubmotors that have the biggest heat limitations. Plus enough guys have used oil fill in hubbies to demonstrate that the negatives aren't too big to overcome. Maybe the rotor magnets will hold enough ferrofluid in place that less fill is needed.

My main concern is electrical conductivity. Can we easily obtain a fine enough iron powder and mix it with ATF to create a good ferrofluid?

I've got an unused pair of hubbies that would benefit from a lower Kv that I'd be willing to sacrifice. If one works then I'd use the 2nd to demonstrate the vital importance of added exterior surface area when using oil fill cooling.

 

[heathyoung]

Iron powder will just get pulled out of suspension and stick to the magnets I would think.

https://en.wikipedia.org/wiki/Ferrofluid - States they are nano-sided particles coated with a surfactant to stop clumping.

Although someone here - http://www.instructables.com/id/Make-your-own-ferrofluid-in-5-minutes/ made it with MICR (magnetic ink character recognition - aka cheque printer ink).

 

[John in CR]

You're probably right that DIY may not be good enough, but in our use the spinning rotor my provide enough mixing action to eliminate clumping while running. From what I read the surfactant ends up about twice the content of the magnetic material, so eliminating the surfactant in favor of more working material for the same viscosity may end up a better ferrofluid for our use.

I can collect all the magnetite that I want with a trip to the beach. In fact I probably have enough with what the kids had fun collecting a few years ago. It shouldn't be too difficult to find a local lab with a ball mill to grind it fine enough. I've never seen a ferrofluid in person, so I should make some with the kids even if it never makes it into a motor.

 

[Punx0r]

Interesting

How about testing the fluid first in a pot with a spare neo magnet and maybe a stirrer? If the iron ends up coating the magnet you know it's no good.

 

[John in CR]

The DIY aspect is only about making it cheap and using a proven oil as a the liquid. Making it part of a cooling approach is also secondary. It's the possibility of a 30% increase increase in torque with the same current without increasing the copper losses at hubmotor type rpms that has my full attention.

 

[wb9k]

Tweeters for speaker systems often use ferrofluid in the magnetic gap, but it's mostly for cooling and damping purposes. It's hard for me to envision ferrofluid inside a motor, but it's an interesting concept.

 

[Punx0r]

The toner powder with a few additives sounds like the best bet if you can't obtain "proper" fluid. It will be very interesting to see what results you get in a motor 8)

 

[John in CR]

Toner powder is only 40% magnetic stuff from what I read.

I went through the paper more closely and the 30% gain was based on modeling with a ridiculously large air gap. He did achieve 10% or so improvement in our rpm range with a small real world motor that had an appropriate gap. Oh well, alarms should have gone off that a 30% gain was too good to be true.

If it's done in conjunction with an oil fill for cooling purposes, then I wonder if it could actually be better if the magnetic stuff is pulled out of solution and held in place by the permanent magnets. As long as the clumping still retains some liquid properties so there isn't a big increase of frictional/viscous losses then it could be interesting as something that to a large decree eliminates the magnetic gap.

 

[Punx0r]

10% still sounds worthwhile, especially if someone is already considering liquid cooling?

 

[speedmd]

Ten percent torque increase is what in current and heat transfer? Looks impressive to say the least. Reading the link a bit makes me wonder if it can not be dialed in so that it works best (moves into the gap) at higher current levels and spins mostly out of the way as the revs increase via centrifugal force and current also drops . Mildly magnetic may keep it from wanting to find a way out also. Looks like something worth exploring a bit.

 

[John in CR]

Ten percent torque increase is what in current and heat transfer? Looks impressive to say the least. Reading the link a bit makes me wonder if it can not be dialed in so that it works best (moves into the gap) at higher current levels and spins mostly out of the way as the revs increase via centrifugal force and current also drops . Mildly magnetic may keep it from wanting to find a way out also. Looks like something worth exploring a bit.

Speed doctor, the paper is written that rpm/volt decreases. Whatever % Kv decreases is the same as saying Kt (torque per amp) by that same %, because Kt in Nm/amp = 9.549/Kv in rpm/volt. This simple and direct relationship isn't commonly understood, because it means motors with identical Kv's make the same torque per amp regardless of size. It's just that for the same Kv a larger motor will have thicker copper, so it can handle more current, and a bigger stator for more torque making magnetic force before reaching saturation. I apologize if you already understand it, but I repeat this stuff as often as possible, because less than 1% of members understand the implications of these relatively simple facts.

Ten percent torque increase is what in current...?

It's an increase in torque with no increase in current. The only "cost" for the increased torque is the added frictional losses called windage, because the ferrofluid has higher viscosity than air. I'm thinking that I'd want to fill in the gaps between the stator teeth with a non-magnetic material to make the stator perimeter as smooth as possible to minimize these losses.

While the author mentions a 500rpm limit of good benefits, he used a small low pole count motor, and I wonder how that would apply to our hubmotors. Is rpm a good guide, or should we look at operating frequency or velocity in the air gap instead?

In terms of heat transfer, I don't think we'd want to use the minimum liquid required to obtain the torque benefit, because that would create a low resistance to heat transfer only between the stator and magnets. I for one don't want to pass the bulk of the heat loss through the magnets. That means more oil fill to obtain a good heat pathway to the side covers, where I'd add more outside surface area to move heat faster to the environment for lower temperatures. That's were every oil fill I've seen falls down, because getting heat to the side covers quicker is mostly meaningless as the system reaches a stabilized temperature, since to remove heat faster either the outside surface needs to be hotter or surface area needs to be greater.

 

[speedmd]

speedmd wrote:
Ten percent torque increase is what in current...?


It's an increase in torque with no increase in current. The only "cost" for the increased torque is the added frictional losses called windage, because the ferrofluid has higher viscosity than air.

Claim is they are making a better magnet with it's addition? That alone is worth a try. Would think there is a certain velocity where the shear losses become a issue, unless this fluid acts differently than most fluids. Compared to mineral oil fill, I would expect similar losses.

The magnet ring is prime for some fins as on old drum brakes on most hub motors. Your wheel spoke plates could double as fins. Thinking a direct heat path across gap through mags to the outer rotating ring would be a good addition at heat loss and help prevent mags from overheating.

 

[John in CR]

Claim is they are making a better magnet with it's addition?

Exactly. Magnetic attraction/repulsion are greatest with the magnets against each other, and it gets weaker the further apart they are. To the extent the ferrofluid is magnetic, it completes the magnetic circuit similar to having them touch but with much lower frictional losses.

As I understand the primary limits are how magnetic the fluid is, it's viscosity (the related losses increase with rpm), and the activation/relaxation time as the stator teeth are energized.

 

[Punx0r]

I imagine the 500rpm limit stated probably relates to what was considered tolerable windage losses. Since hubmotor users often run oil anyway, that consideration can probably be discounted. You're probably running hub motors at higher rotational velocities than almost anyone else on this forum, John.

 

[speedmd]

If this stuff is a great addition for low RPM's, I am sure simple methods can be used to help get it mostly out of the way when motor is spinning fast if that is the only significant issue. Cost may be the killer.

 

[John in CR]

...I am sure simple methods can be used to help get it mostly out of the way when motor is spinning fast if that is the only significant issue...

I take it that you've never had to get iron shavings off of a hubmotor rotor. Once we put this stuff in a motor it will be all but impossible to get out, and the magnets and centrifugal force will hold it in place.

I obviously wouldn't try it in my high speed motors, but my pair of Wilderness Energy BL36's would give me the ability to do a side-by-side comparison of modded vs stock. Then if it works, I could proceed to compare oil fill with stock covers vs covers modified to increase surface area. I have a number slower wind lower power hubbies collecting dust that would make great low power emoto's for kids, so it's worth the risk and the effort.

 

[speedmd]

The ferrofluid I have played with in speakers has no metal grit in it at all. Totally colloidal and only attracted to magnets ever so slightly. Just enough to stay in the gap and not leak all over the place. They use it in most high end high frequency drivers so how much drag can it give without totally screwing up the sound performance. This sounds like a great idea if you can get enough of it at a reasonable cost to play with some of your dust collector pile. The audio stuff is Super expensive. $220 a liter for the classroom stuff. http://www.teachersource.com/product/bulk-ferrofluid-100ml-1000ml/electricity-magnetism?gclid=CJPTjv2thscCFYGRHwod_qQOuw
http://www.magnet4sale.com/1-litter-magnetic-ferrofluid-for-fun-400-gauss/?gclid=CIj4veOuhscCFYMWHwodmCYIfw

 

[Skalabala]

What will happen if you use finely crushed to powder ferrite cores for this use?

 

[speedmd]

This stuff has been around for quite some time. Trying to reinvent it will most likely be a long hard road IMO. Best to focus on what grades (oil /ester or some other base) would be worth trying out and how much/ little you would need to give the results your after. The direct coupling of the stator to the rotor via a fluid like this would possibly be a game changer in heat transfer. Maybe it does not yield as high a peak efficiency, but yields a motor that is much more difficult if not impossible to burn up without really trying to do so.

 

[Toorbough ULL-Zeveigh]

why no FF/Gatorade sticky?

I would not recommend DIY ferrofluid. The APG stuff will last much longer in service and it has a proven track record. The size distribution of the particles, the carrier fluid and the surfactant that keeps the particles from settling out are critical to long term performance.

for anyone else that wants to try;
"Better than store bought"
[youtube]6L8yUY-doNc[/youtube]