Ceramic Bearings vs Steel

John in CR said:

In normal use they will commonly see higher than the 64°C, and much higher for those running moderate to high performance. Plus common ebike hubbies don't even have axle seals.

Is the 64°C limit a good number? I recommend against opening hubmotors any more than absolutely necessary, so can we re-grease from the outside? Is there a grease that will raise the temp limit significantly?

Click to expand...

http://www.machinerylubrication.com/Read/798/grease-selection

There are a great many number of options for grease, to summarize, there are various greases that can handle many different conditions and temperature ranges. This includes the temperatures you might anticipate on an electric motor.

Bearing losses on a hubmotor are negligble, well at least if they're an actual bearing. On 9C I purchased some years ago, I'd hesitate to even call those bearings they were so poor new and unused.

At least bowlofsalad gets the real issue here. Can anyone comment based on experience with the different types of grease?

speedmd said:

Motor torque certainly loads the motor bearings. What else transfers the drive force the motor generates? It is only connected to the frame through the bearings.

Click to expand...

I suppose there is some truth to this, in that the forwards motor thrust on the bike from the motor is coupled through via the bearings. But thrust forces are just a fraction of the downwards load forces from mass vehicle mass, and much less even than the load forces from a pedal chain, as a pedal chain tension is like 5-10 times higher than the vehicle thrust it produces. So ultimately it's fair to say as Chalo did that motor torque doesn't in any noticeably way increase the loads that the bearing sees.

Anyways the reason that I'm piping in here is because we run into this misperception at our shop all the time, with people thinking that they notice extra drag in their hub motor and assuming that this must be the result of a shot ball bearing. But even if you DO have a shot ball bearing, the extra resistance that it causes doesn't even remotely register when compared to the cogging drag, tire rolling resistance, or anything else impeding the movement of the vehicle, and to think that you could actually see a difference in your measured power consumption with a new vs. damaged ball bearing even with precision instrumentation is extremely unlikely.

That said, we've seen lots of worn out ball bearings in need of replacement, and water/corrosion damage is one of the prime reasons for that, (poor hub machining tolerances resulting in too tight / too loose a fit is another). Most hub motors now have decent rubber shaft seals on them to keep the bearings protected from the elements, but not all of them. If other materials like Ceramic would intrinsically hold up better with environmental exposure, then that could have some merit and justify the extra price.

So to repeat what others on this thread have stated, there can be valid reasons for changing your motor ball bearings, but reducing wheel drag / increasing motor efficiency is not one of them.

speedmd said:

Motor torque certainly loads the motor bearings. What else transfers the drive force the motor generates? It is only connected to the frame through the bearings.

Click to expand...

Tire>rim>spokes>hub shell>magnets>stator>hub spider>axle>dropout. At no point does the motor torque have to pass through the axle bearing. The bearing constitutes a free axis as regards motor torque.

Chalo said:

speedmd said:

Motor torque certainly loads the motor bearings. What else transfers the drive force the motor generates? It is only connected to the frame through the bearings.

Click to expand...

Tire>rim>spokes>hub shell>magnets>stator>hub spider>axle>dropout. At no point does the motor torque have to pass through the axle bearing. The bearing constitutes a free axis as regards motor torque.

Click to expand...

That was my first reaction too when I saw Speedmd's post and was scratching my head thinking "he's smarter than this!", But then I noticed that he carefully phrased it as transferring "drive force" and not "drive torque". So the motor generates torque, and this torque is converted to a forwards thrust at the wheel diameter which then pushes forwards on the bike via the bearings to the axle to the dropout. So indirectly, the motor torque does load the motor bearings.

But unless your setup is powerful enough to accelerate on the order of ~1G, it's a pretty negligible amount.

Thanks for sharing your real world experience with this Justin. There really isn't any proper unbiased hard numbers or facts backed up with decent evidence except in the in the hardcore world of ' wheelchair racing' .

So am I assuming correctly, that opposite to the sales pitch 'HQ, high performance' ceramic bearings, that when calculated into rolling resistance a wheel would be more greatly effected by minimising the tyre contact patch than allowing the wheel greater freedom of rotation, any marginal benefits seen would be countered by the cogging of the freewheel?

Also, that any HQ steel bearing properly maintained with a 'High Temp' grease suitable for our -15 to 150+C would minimize the risk of a bearing failure and it's ceramic counter part is more for people who don't want to have to open their hub/pop the bearings every 2000km to regrease and check for signs of wear?

I only use 2000km as a guideline from my current HS3540 SKF6203ZZ failing at 1600km with all weather usage @ 1500W. Maybe it should be a check every 1500km?

Agree, drive forces are relatively low but significant- directly related to torque, and that corrosion is what kills most all of them, but looseness- hammering kills / degrades a bulk of them in this type of application also. Chalo is correct in that the harder balls will certainly be harder on the steel races. Critical to use quality units. They still will dent up when you overload the bearing hitting pot holes and the like. The hard ultra smooth SiN balls however never seem to get that egg shell surface you see in steel balls that kills their smoothness. You can hear it when they get tired. Several bearing manufacturers I know of sort - grade bearings also by listening to them. All kinds of other issues can happen, but they are still worth a try just to experience for your self if your looking for the best you can get in a particular app.

Where I think the small but relatively important drive forces matter is the constant loading and unloading (forward -rearward) that may cause the balls in the bearing to be away from the races in the main load directions. This would increase the slapping action caused by the extra movements. In a hub application, I would just stick to a good grease filled bearing with the best seal I could find. Roller bearing even possibly. For something like a astro or joby..... I know what I would be looking for. :lol:

When a bearing fails, the majority of the time it's the bearing 'race'? So a Hybrid bearing with ceramic balls and a steel bearing wouldn't last longer than a full steel as the 'race' is the weakest link? This is unusual, isn't it the part that moves the most and takes the 'load' normally the failure? Guess ceramics are just a hype.. This is good stuff!!

Both the balls and the races see various forms of degradation from use. Pitting from corrosion is a killer. Dents are another. Balls roll around so they somewhat self heal by peening them selves back into a good looking shape (with the naked eye) easier than the races can as they are stationary. Don't let this fool you. They ( the balls ) get dented up badly in normal use. They end up looking much like a disco ball under a scope when well used. If the races can not rotate slight amounts over time, they definitely will have a wear side that goes much quicker than the rest. One small engine Mfg I know of had bearing pocket tolerances so well dialed in that it allowed stock caged bearings to rotate slowly over time minimizing this wear side on the race. In a hub You can just loosen / rotate the axle 180 degrees - let the inner race change positions every so often.

This article on bearings is massive, but quite informative. Although they do claim a 4% decrease in rolling resistance the ceramic bearing mounted in a wheel. Which, we've already had explained is very far fetched, but even if it was 4% the gains are around 10W using the motor sim.