Bicycle bearing grease lab test help

Joined
Mar 18, 2017
Messages
12
Location
Serbia
Hi all,

My name is Relja Novovic, I'm from Serbia (Europe/former Yugoslavia). This is my first post, though I've found this forum (and the bobistheoilguy.com site) a great source of good information. That's why I'm posting here, hoping to get more opinions, ideas and corrections. Apologies in advance for my Tarzan English and probably incorrect English technical terms. All suggestions and corrections are more than welcome.



If all goes well, before the end of this month, I'll get access to a lab testing machine of a local mechanics faculty. My goal is to test how various types of lubricants (greases, might do a test with oil, just for comparison) fare when used in bicycle bearings.

The presumptions, based on my previous experience and knowledge, but haven't found any tests (in books or on line) to scientifically confirm it, is that bicycle bearings don't take much load (compared to motorcycles, cars, not to mention heavier machinery) and that paying for any premium grease is not justified. That is what I want to test and compare if more expensive, higher quality greases make any measurable difference. With and without EP and other special additives. I also plan to test if filling the bearings 100% with grease, not leaving any gaps as is generally recommended, prolongs the life of bearings exposed to water and dirt, or harms them - or makes no measurable difference.


What I'll have at my disposal is a machine with controlled number of revolutions per minute and static load on the bearings. The machine has an axle, onto which a bearing can be screwed on. I will also be able to perform controlled impacts on the bearings (Charpy pendulum in English?).

My plan so far:
1. Using cartridge bearings, for easier mounting and lower price than cup and cone hubs. 6903 model.
2. Assuming an average wheel circumference is 2.1 meters (26" are smaller, even with wide tyres, 28" are a bit larger, but that's about middle ground).
3. Simulating 30 km/h speed for 90% of the test, with 10% at 100 km/h. With 2.1 meter wheel circumference, that's about 240 to 800 revolutions per minute.
4. Assuming one wheel carries 60 kg of weight, 30 kg per bearing. That load will be simulated, as constant.
5. Since rolling bearings that roll while loaded don't suffer measurable damage from road buzz, only impacts, I will disregard road buzz simulation. This will make the test less useful for the headset bearings, but good for all the others.
6. For impacts, simulating drop off a 20cm high kerb, with 50 kg load, 25 kg per bearing, without any damping from rim and tyre deflection, I'd test with 5 kilopond-meter impacts after every 12 hours of testing (about 50 Joules). My estimate is that's reasonably high impact, since off road bikes have wider tyres and suspension, and riders will usually use hands and elbows to absorb the shock. While road bikes don't get those impacts often (and they still have tyres, rims and riders to dampen the impact). 10 impacts after roughly every 12 hours of testing.
7. Assuming bearings are never submerged into water, but often sprayed with rain and dirt. A mixture of 1dl water, 1 spoon of sand and 1 spoon of soil will be poured over bearings after every 12 hours of testing.
8. Will degrease bearings in diesel, then "medical petrol" (is that the correct English term), then add tested lubricants.
9. Will do a test with 30% less lubricant than 100% full bearings, and compare that with 100% grease packed bearing. See which option fares better for this purpose, since bike bearings roll rather slowly.
10. Will use a fan to simulate air cooling, if there's measurable heat build up over 50 degrees Celsius. Will also use more stops for cooling down if required. Reduced speed as well (to 20 km/h average speed).
11. Wear measurement will be done by completely cleaning/degreasing a bearing and measuring it on a precise scale. Also, visual inspection for ball and race wear will be done at the end of testing.
12. I will do a first test on a 3000 km service interval. Then 5000 km. Then 8000 km. With 30 km/h estimated average speed, 5000 km will take about a week. If there's not much heat build up, I'm considering increasing the test speed to 60 km/h, and max speed to 120 km/h.
13. Assuming side loads are not relevant for lubricant performance comparison and the application, so will not bother with them (would have to improvise to achieve lateral loads).


Questions:
a) Are the estimates of loads and shocks realistic (strict) enough? Any better ideas on rain/dust/mud simulation?
b) Did I miss something?

Thanks in advance for any help.
 
Last edited:
You've got an interesting test regimen planned, hopefully it goes as planned. What is your goal from this testing? I've been a cyclist, on & off road for 30+ years and confess that the cost of a small quantity of bearing grease was and is insignificant compared to the overall cost of riding decent quality bikes. People with cheap bikes probably don't ever have the bearing greased, people with good bikes don't have a problem paying $10 instead of $5 for grease.

Who will benefit from your findings?
 
Originally Posted By: AZjeff
You've got an interesting test regimen planned, hopefully it goes as planned. What is your goal from this testing? I've been a cyclist, on & off road for 30+ years and confess that the cost of a small quantity of bearing grease was and is insignificant compared to the overall cost of riding decent quality bikes. People with cheap bikes probably don't ever have the bearing greased, people with good bikes don't have a problem paying $10 instead of $5 for grease.

Who will benefit from your findings?


Good question.
smile.gif


It's mostly my own curiosity for start. Afterwards, depending on results, we'll see.
 
It would seem to me that changing out the grease in the wheel bearings/pedal crank bearings, etc. would be much more labor dependent than anything else. I'd buy the most expensive grease one time if I knew it would last the life of the bike.

When I found out the cost of tools and labor to take apart the cranks on my back up 1977 Raleigh it was quite apparent it would be cheaper to get a new bike or at least a quality used one. Same thing when I briefly considered rebuilding my 2006 Raleigh sport bike after 10,000 miles. Fwiw, I put those 10,000 miles on over 8-10 years and never once did anything to any of the bearings on the bike. Chain lube and new tires was all the maintenance it ever got. For $450 it was a good deal over 8-10 years. Didn't even have a single tune up done. Went that distance on one chain. Bearing grease is a non-issue for a bike like that. I bought its replacement last year...an almost never used 2010 GT 16 speed for $275...everything on it basically new. In this price range...ride 'em until they drop. Though this time around I will change out chains more often so that I don't destroy the cassettes.

But, good luck on your test to find just what kind of grease/lubricants hold up the best per cost.
 
Where I live, most bikes (about 99% from my rough estimate - I service bikes) use cup and cone wheel and steerer bearings, that require annual service in order to last long.
Not sure about the most of the world.

As for the other parts, Shimano pedals are serviceable. Not sure about other makes.
 
On my boy's BMX race bike, we use Kluber Isoflex LDS18 Special A on all the bearings. The hubs and bottom bracket use ceramic balls with titanium races.

 
Last edited:
Originally Posted By: Phishin
On my boy's BMX race bike, we use Kluber Isoflex LDS18 Special A on all the bearings. The hubs and bottom bracket use ceramic balls with titanium races.


Have you measured bearing drag? Ceramic balls, being a lot harder than steel, even more so than titanium, so are more likely to cause pitting in the races, causing even more drag than regular, all steel bearings after a while.
 
Originally Posted By: CycleGremlin

Have you measured bearing drag? Ceramic balls, being a lot harder than steel, even more so than titanium, so are more likely to cause pitting in the races, causing even more drag than regular, all steel bearings after a while.


No, I have no way to measure it. BMX races last 40 seconds or so. It's an all out 1/4 mile sprint. If the wheels spin ultra smooth and free, make no noise, or show any signs of "hiccups", that's really good enough for our application. As soon as any sign of a issue or problems presents itself, the bearings and races are replaced.
 
Originally Posted By: AZjeff
You've got an interesting test regimen planned, hopefully it goes as planned. What is your goal from this testing? I've been a cyclist, on & off road for 30+ years and confess that the cost of a small quantity of bearing grease was and is insignificant compared to the overall cost of riding decent quality bikes. People with cheap bikes probably don't ever have the bearing greased, people with good bikes don't have a problem paying $10 instead of $5 for grease.

Who will benefit from your findings?


Its true that many cyclists spend $5K on bicycles and costumes today as a form of status; however, there are still a few of us who ride older bicycles...gasp...and practice a strong economy.

We are the ones not wearing spandex outfits into the local SBUX Forcing all the customers to see our 60-year-old glutes.

And I can think of a few old bike shop owners who would prefer to save overhead costs by purchasing a quality grease at half the price.

Cheers

To the OP:

Welcome to the forum and your English is better than 90% of us Native speakers.
 
Originally Posted By: CycleGremlin
Where I live, most bikes (about 99% from my rough estimate - I service bikes) use cup and cone wheel and steerer bearings, that require annual service in order to last long.
Not sure about the most of the world.

As for the other parts, Shimano pedals are serviceable. Not sure about other makes.

Speedplays are servicable and the bike bearing suppliers(Enduro mostly) have bearing kits for Look and Time.

In the US, good bikes used sealed headset and bottom bracket bearings and it's a tossup for wheels, cheaper ones with Chinese/Taiwanese hubs as well as Shimano use cup & cone, the rest are cartridge.

Chris King and Hope are weird in the fact that while they do use sealed cartridge bearings, they do encourage periodic cleaning/relubrication with "synthetic" grease. Chris King being tight-lipped about their recommendation of lube, they did have Finish Line lubes make their hub lube but they seem to like XOM lubricants. I have a King headset and it hasn't been adjusted or relubricated. It still works like new.

As for my hubs, I'm mostly Shimano. I currently have M1 grease in my Ultegra 6800 hubs and I notice they spin quicker. Before that, it was Nemco purple synthetic sold as Pedro's Syn Grease and it was heavy grease(ISO460 base oil vs. M1's ISO220 base). I'll do my 'cross/commuter bike with M1 as well if I plan on doing a race or two this season. That bike has Deore XT M8000 QR hubs, it's a Specialized Tricross. The bike always had shifting problems, wide 135mm mountain bike hub spacing + road gearing(SRAM PG1070 11-36 cassette with a Ultegra 6700 road compact crank) = fun in good and bad ways.
 
Last edited:
Originally Posted By: CycleGremlin
Have you measured bearing drag? Ceramic balls, being a lot harder than steel, even more so than titanium, so are more likely to cause pitting in the races, causing even more drag than regular, all steel bearings after a while.

Not all true. Ceramic balls have less deformation at the point of contact, resulting in less drag and less heat generated. They're popular for racing applications.

I've never heard of titanium being used for raceways... not even in aircraft application. What's the engineering behind that?
 
Jobst Brandt described the loads typically seen in use in his book 'The Bicycle Wheel' It may offer you some useful information, it describes the engineering, as well as the process of building a wheel. He does include a finite element analysis of a wheel.

He suggests that wheels do see radial loads of 400 kg or more when striking a bump. He also points out that hubs are not efficient when passing the torque load to the left side of a driven hub, which I would see as unbalancing the load to the bearings during pedaling.

I question the reasoning behind the speeds you chose to test at. 100 kph isn't seen often, likely only seen on a severe descent in optimum conditions.

Good luck with your testing !
 
Originally Posted By: CycleGremlin
... Questions:
a) Are the estimates of loads and shocks realistic (strict) enough? Any better ideas on rain/dust/mud simulation?
b) Did I miss something?

Thanks in advance for any help.
That looks like an interesting experiment. Good luck, and keep us posted on results. A few points:

1. Although absolute load on bicycle wheel bearings is obviously a lot less than that of a small car, EP qualities of the grease do matter, because specific loading (stress within the much smaller bearing races) is often higher. That's one reason why life of bike bearings is generally shorter.

2. Friction of any bike wheel bearing in good condition and adjustment is insignificant, compared to tire rolling resistance or aero drag.

3. Chain tension can add greatly to overall load on right rear wheel bearings, especially when climbing hills. Do the vector math, using reasonable assumptions, and you'll see. On steep hills, it can exceed load due weight of rider+bike.

4. To better simulate riding in rain, you need to subject your bearing to periods of continuous spray, not just dump mixture over them intermittently. The life of bike bearings tends to be highly dependent on the water resistance qualities of the grease.

5. I agree you needn't bother with lateral loading.

6. Front wheel bearings (at least typical conventional cup-and-cone ones) do not necessarily outlast rear ones, despite lower loading, because the front bearings are smaller.
 
Thanks for all the replies. Here's my thinking out loud - don't take it as arguing and please correct me if you find my reasoning flawed:

Have read The Bicycle Wheel. From the time it was written, hubs have somewhat changed. From my experience - material quality and hardness has reduced (even in high end Shimano hubs), while rigidity and load transfer has improved - more load transfer to the left side of the rear hub. Not 50/50, but not much worse. Will double check that data (and the source).

As for bump loads - 400 kg seems reasonable, but since my test totally disregards the flex of the tyre and the rim (even the little flex hub material has), my guess (if I manage to do a measurement, or at least a calculation I will) is that 25 kg hit directly to the race of a single bearing is quite severe. Like said, if I manage, I'll double check it.

100 km/h speed will be used for shorter intervals, to simulate fast downhill riding. If there's much heat build up, I'll reduce the overall average speed, but will do a few tests at that top speed. Having said all that, if it's possible, I'm considering making a pilot test with a higher hardness grease (NLGI 3), just to see if the higher than realistic test speed is biased towards the "lighter" greases. If the heavier grease shows measurably more wear, I'll reduce the testing average speed (still going with some 5% at 100 km/h to simulate high speed descending).
 
Last edited:
Originally Posted By: Kestas

Not all true. Ceramic balls have less deformation at the point of contact, resulting in less drag and less heat generated.


True - while all is new. My thinking is that a lot higher hardness and lack of elasticity is more likely to cause (metal) race pitting sooner. Fully ceramic ones (both races and balls) make somewhat more sense IMO, though I don't believe they would provide any (stop watch) measurable advantage - except if one believes they make them faster - that does help and work.

Originally Posted By: Kestas

They're popular for racing applications.


Pros have used patents far from optimal - since sponsors need to market and sell stuff. People often buy what pro racers use believing it is the best - since the pros use it.

Originally Posted By: Kestas

I've never heard of titanium being used for raceways... not even in aircraft application. What's the engineering behind that?


+1
But I'm taking the poster's word for it - he's using it, guess he knows what it is. Didn't do much research on "exotic" cartridge bearings.
 
Last edited:
Cycle gremlin, your "thinking" is contrary to all data that is commonly known among bearing designers. Raceways are redesigned to accommodate ceramic balls so that raceway loading is within acceptable design parameters so spalling won't be an issue.

Making ceramic raceways has its own challenges, which make it impractical to manufacture and install.

Racers don't use ceramic rollers because of advertising pressure. They use it because it works. They see the test data, which time and again prove ceramic rollers perform better on the track. Plus they are willing to pay the premium for this performance.

By the way, I don't take people's word for it when it comes to bearings. I rely on data from testing.

I think you need to study a bit more in depth about bearing design, ncluding ceramic rollers, and learn about bearing failure analysis to understand all the different reasons for raceway spalling.

Also, when bearing manufacturers perform kerb crash testing, they include the tire in the assembly. The test is chiefly to look at brinelling of the raceway. Grease is not a variable in this test; it will not affect results.
 
Originally Posted By: Kestas
Cycle gremlin, your "thinking" is contrary to all data that is commonly known among bearing designers. Raceways are redesigned to accommodate ceramic balls so that raceway loading is within acceptable design parameters so spalling won't be an issue.

Making ceramic raceways has its own challenges, which make it impractical to manufacture and install.

Racers don't use ceramic rollers because of advertising pressure. They use it because it works. They see the test data, which time and again prove ceramic rollers perform better on the track. Plus they are willing to pay the premium for this performance.

By the way, I don't take people's word for it when it comes to bearings. I rely on data from testing.

I think you need to study a bit more in depth about bearing design, ncluding ceramic rollers, and learn about bearing failure analysis to understand all the different reasons for raceway spalling.

Also, when bearing manufacturers perform kerb crash testing, they include the tire in the assembly. The test is chiefly to look at brinelling of the raceway. Grease is not a variable in this test; it will not affect results.



For shock tests - do you think I should just exclude the shock tests?
If not - without a tyre, with a reduced impact, wouldn't results be similar? The tyre would probably reduce the maximal impact and spread the impact force over a slightly longer time period, but I don't think that would change the results - what do you think?

Do you have any data and educational links I could use?
 
Originally Posted By: CycleGremlin

But I'm taking the poster's word for it - he's using it, guess he knows what it is. Didn't do much research on "exotic" cartridge bearings.


I'm sorry. I misspoke. The races are NOT titanium. They are 52100 steel. My mistake.
 
I think you should exclude the kerb crash test. This test is used to test the raceways for dent resistance. As I said, grease is not a variable in this test. You could run it without grease and have the same results. The tire is included because this is how the test was developed. The test has very specific parameters (i.e., drop height, load, angle) that I am mot intimately familiar with.

There are books in the library of educational institutions that cover bearing design and theory. Common bearing failure modes can be found on the internet.

Your water and dirt tests have merit.

I wouldn't measure wear by weighing the bearings. It's not practical and it won't yield anything useful. Wear is typically assessed by examination of the raceways. It's not really quantified.

Another test that may be useful is a raceway fretting test. This tests the effectiveness of grease during micromotion of raceways, such as the steer tube bearings. Different greases have markedly different results with this test. I believe there already is a standardized test for this in the industry.
 
Back
Top