Wednesday, September 2, 2020

A Comparison of 12 Greases for Bicycle Use

This post started when  I realized that I needed another ebike for friends to borrow, and I found an old Marin Bear Country MTB for the project.  It was in a scrap metal pile, and consisted of the frame, forks, and handlebars held together by the cables that were still attached, but everything else was missing.  This would require a fair amount of work, but the forks intrigued me.  I have a Very Low Budget, thus all my bikes so far have had either rigid frames or very old front suspension forks. I've realized that a small amount of suspension will increase the speed and efficiency of my bikes, and have wanted to experiment.  These forks actually had a nice sliding surface and said Manitou on them!

There seems to be a lot of debate about the perfect grease for bicycle bearings and forks.  Normally I simply use the better quality lubricants that can be found at any autoparts store and would be hard pressed to use a gourmet grease.  I have a lot of car mechanic experience and after years of using dozens of greases, I do not think a special bicycle grease is necessary for bicycle bearings.  But I wanted to lubricate the sliding surfaces on my first Real Forks the right way, so I started looking into fork greases.  Evidently I needed to get Manitou Prep M grease made by Motorex, which was No Longer Available, and it was time to find a substitute.  The bike forums were useful for identifying what was available and which ones most people were using, but they were pretty short on tech details.  Unfortunately the manufacturers were even worse, in a triumph of marketing over datasheets that might actually help you select a grease for an application, most don't release even MSDS reports.  And the bicycle greases evidently are rated by the name of the manufacturer.  To fill this vacuum it looked like I needed to run my own tests.    I'm putting up this post to save others some trouble.

There are several standard tests that evaluate grease properties such as thickness or viscosity (NLGI rating), heat resistance (for disc and drum brakes), pressure resistance (EP additives), film strength, and water resistance.  These are sort of helpful in a bicycle application, but I've seen enough bicycles (which people were still riding) with bottom bracket or wheel bearings rusted right out to the point of rotating directly on the axle to know that bicycles operate with at least an order of magnitude lighter demands on grease than cars do.  (When these bikes were soaked from rain, the water inside the axles made them pedal easier.)  I also wasn't as concerned with testing the Extreme Pressure rating (spot loads from ball bearings) as I was with sliding (lubricity) and stiction (starting friction), because if a grease can handle a 3000 pound car wheel bearing then it can handle a bike bearing.  Lastly, I was more concerned about grease viscosity making it harder to pedal in winter than I was about liquifying the grease after braking said car repeatedly from 70 mph, and I knew from years of using different greases that some NLGI 2 greases I've used are softer than others and would be better for the cold.  A more sensitive test was needed, that included an evaluation of texture.  Also something with numbers that could be compared instead of forum comments.

The Testing Setup
Looking inside the bottom stanchions on my Real Forks, I could see there were two bearings in each leg.  I could reach a ruler in and measure the top bearings, which were 1 inch in diameter and 5/16 inch tall, which gives almost exactly 1 square inch of surface area.  I decided to drag a 1 square inch piece of metal over another piece of metal with different greases in between them.

The testing jig:  A 1 inch square stainless steel sled with a series of 4 different test weights placed upon it is pulled across a changeable baseplate by a string attached to a cup that is gradually filled with sand.  I mounted the jig on an old tablesaw with a cast iron top to keep it stable and minimize the effects of vibrations.

A closeup of the 1" by 1" stainless steel sled set up with marine grease on a 1/8" thick sheet of  UHMW plastic.  The sled's edges were radiused, then it was lapped with 600 grit wet sandpaper on a sheet of glass, and then polished with a extra fine whetstone.  The strong but flexible upholstery thread pulls the sled to the left, and passes through a low hole in the front sled lip to tie onto a 1" diameter key ring which also holds the weights up off the sliding surface.  The wooden ledger strip on the left is glued onto the plywood base and keeps the test baseplate from sliding.

A well worn but not wobbly computer printer pulley with a ball bearing center spins very, very freely and changes pulling force into a vertical direction.  I bent a couple short pieces of stovepipe sheet metal into ells to hold the pulley in level alignment with the string hole in the sled.

The cup was slowly filled with sand until the sled started moving.  The sled had to stay moving, if it moved only a slight bit and stopped then more sand was added.

After the cup was filled with just enough sand to keep the sled moving, it was removed from the hook on the end of the thread and weighed on a beam balance scale to find the force in grams for the pull.

The four test weights used for loading the sled:
-left, a small aluminum block, 229.7 grams
-center front (on the sled), a washing machine counterweight, 2110.5g
-center back, a long piece of computer dot matrix wire printer anvil, 847.3g
-right, a short piece of computer dot matrix wire printer anvil, 505.9g

The Lubricants in the Test
The lubricants tested, left to right:
-Back row: Shell Rotella oil, Water, (the Acetone was used for cleaning between tests)
-Third row: AGS Lith Ease, Coastal Moly EP, Kendall Super Blu, Castrol wheel bearing
-Second row: Super Slick, Vaseline, Sta Lube marine grease
-Front row: Slickoleum, graphite powder, Protek MPL-1, Slick Honey, Super Lube

MP- Protek MPL-1 synthetic fluoropolymer grease
WA- water
OL- Shell Rotella oil 15W-40
CA- Castrol multipurpose wheel bearing grease 715C
WL- AGS Lith Ease white lithium general purpose WL-15
MG- CRC Sta Lube aluminum complex marine grease SL3184
CO- Coastal Industrial moly EP grease 52353 with polyethelene
GR- 325 mesh graphite powder 99.9% pure
VA- Vaseline (with smudges from graphite that blew into it)
SL- Super Lube synthetic grease with Teflon 21030
SO- Slickoleum multi purpose light calcium base grease
SH-Slick Honey grease LU2005
SS- Rock N Roll Super Slick grease
SB- Kendall Super Blu high temp EP grease L-427

The Baseplates
The baseplate was changeable, it rested on the plywood base and was kept from sliding by a wood ledger strip glued on the base.  The metal plates were 1/16 inch thick, and the plastic sheet was 1/8 inch thick, with no noticeable flex during testing.  I started out with a copper plate because copper is often a component of plain bearings.  The amount of galling during these tests was amazing, and I had to look it up to be sure it was galling and not dirt.  It turns out that soft metals gall very easily, with copper having a large problem, and stainless steel not far behind.  I switched to a stainless steel plate next, which had reduced but still significant galling.  Last I used a Ultra High Molecular Weight (UHMW) plastic base plate, as it was the closest I had to a thick piece of teflon or custom molded bushings.

The copper (CU) sheet started off tarnished but smooth, and I buffed it with 0000 steel wool, then worn out fine nonmetallic Scotchbrite until it was slightly reflective, and then used solvent to finish clean it.  The scratches are from galling that happened during the tests.

The stainless steel plate finish was smooth enough to be reflective with minimal cleaning, and there are less scratches from galling during the tests.

The UHMW plastic sheet had the least galling.

The Test Procedure
The test started with cleaning the base plate and sled with acetone, (I found acetone worked better than lacquer thinner or mineral spirits) and it was easy to tell if all the residue was removed because the shine of the metals and whiteness of the plastic showed any film left on them.  Then a small dab of test lubricant was put on the plate, and the sled was placed on it and rubbed in for a minute to dilute any possible traces of contaminants, leaving them in the globs around the edge and only a thin film in the test area.  I then loaded the test weight on the sled and let it sit for 15 to 20 seconds to squeeze out excess grease.  The test consisted of 4 pulls, starting with the lightest weight and using the next largest weight each time until the heaviest.  All tests were done between 70 to 80 degrees F.  Several lubricants were run through the test procedure only once (4 pulls from the light to heavy weights) because they were known not to be suitable lubricants for bike bearings, (such as water or vaseline), but were done as a reference check.  Others that were more serious contenders were tested several times and their results averaged at the end.  A total of 81 tests (324 pulls) were made.

Then the force in grams for the 4 pulls in each test were added together and the test totals used to rank the greases.  After this basic ordering for functionality was established, the notes from the tests on galling, speed that the sled moved (a proxy for viscosity), and ease of pressing through the film and then sticking while rubbing in were factored into a final evaluation.  The test results were most affected by stiction, but there was also a significant influence of regular friction (lubricity) because the sled had to keep moving and for many of the tests the sled stopped and required the addition of more sand to the cup to get it moving again.  There are a few interwoven requirements in this test, but all lubricants were tested using the same protocol and it's valid to cross compare them.  I tried to be very clean to keep the tests even, and used new rags for preparing each test, but did not worry about random dust in the air (basically none) because I thought that any dust would bring the tests closer to real world conditions.

At the end of the tests, I found I had to make one correction by adding the weight of the hook (11 grams) to every measurement because some of the tests required only the hook without the cup to move the sled and this was not included in weighing just the cup of sand. I then color coded each spreadsheet entry by the amount of sticking (galling) that happened during the pull- clear for none, up to dark red for severe sticking.  A few of the lubricants performed well in one test only to get stuck on a repeat test, and I considered this erratic behavior to indicate the film strength, (for the most part this also corresponded to how easily I was able to press through the film while rubbing in the sled at the start of a test).

I'd like to present the basic data for anyone who wishes to look for themselves at the trends in it, but 324 measurements don't fit well in a blog format.  As a compromise I'll put in 3 screenshots (one each for the copper, stainless steel, and plastic baseplates) of the data that you can zoom in on, followed by a few comments for each baseplate about notable tests.  At the end will be a more readable summary of the greases.

The Copper Baseplate Test Results
I started this test process using a copper baseplate thinking that most plain bearings are soft, easily scratched, and have some copper in them.  Here is a screenshot of the final summary rows of the spreadsheet:
Copper base plate test results, click on image to enlarge
There were 32 tests (128 pulls) on copper 
-the top row is the rank. The order is based on the total grams of pulling force listed in the 7th row
-the second row identifies the grease, (along with it's sequence number in my tests)
-rows 3 through 6 are the pull measurements in grams, row 7 is their total
-at the bottom are notes taken during the test
General comments:
-I started these tests with a bare sled (no lubricant) to get a baseline before any grease was put on, this is the blue column in the spreadsheet, which unfortunately is in the middle of the results (in other words better than half of the greases).  The second test was Slickoleum (SO, tests 2a, 2b, 2c), and I was astounded at how bad the results were (19, 27, 30 out of 32).  (I hadn't yet noticed the warning on the label that Slickoleum was not for use on copper.)  The addition of SO caused the sled and copper plate to suck together- it was similar to two magnets pulling together and dragging.  This happened off and on over the next several greases, and at this point I did some reading about galling because the scratches were becoming apparent on the copper and I had to figure out if I needed to redesign the test.  The procedure seemed reasonable though, so I simply added other the baseplate materials to the tests as a further check and continued on to find out how other greases reacted.
-The thicker greases did better (Super Blu SB ranked 1 and 2), thinner was worse (oil OL, 22 and 24).  The trend was that all the automotive greases were better than the bare test (no lubricant), and almost all the bike greases were worse.  The sled did move much slower during the thicker grease tests though, and I know from experience that Super Blu is very thick and dragging during winter.
-The surprising lubricant was Vaseline (VA) which for some reason unknown to me did well, because in other respects it was thin, easy to push through the film (similar to  oil) and it stuck when rubbed in.
-The addition of graphite to Vaseline made it's performance worse, and it felt like a fine sand when rubbing in the sled.  My experience has been that graphite only works when it is applied in a layer just a few molecules thick and burnished in, and while it's possible to get around this restriction with mixtures (such as DAG) that contain special colloids or binders, graphite is not effective as a grease additive.
-The automotive greases were also more consistent, for example Castrol (CA), ranged from 781, 801, to 814 grams, versus Slick Honey (SH 789, 841, 1025 g) or Super Slick (SS, 819, 909, 995, 1023, 1243 g)
-I saved the Moly greases until last to avoid any molecular contamination of the copper affecting the tests of the regular formula greases.  The Coastal and Super Slick both did well (CO ranked 3, SS 10 and 13) but there was still some galling.  I am wondering if it is better to have galling with a pulling total of 605 grams (CO at 3) then no galling with a pulling total of 1243 grams (SS 31).

The Stainless Steel Baseplate Test Results
The next baseplate was stainless steel, chosen to be similar in composition to many bearings and fork sliding surfaces that don't rust.
Stainless Steel base plate test results, click to enlarge
Another 31 tests (124 pulls) were done to compare galling and the effect of stainless steel on the lubricants versus the copper baseplate.
General comments:
-At the end there were about one third as many galling scratches on the baseplate as there were on the copper baseplate.  The color coding illustrates that there was almost no sticking and galling with the better performing lubricants (again automotive- Castrol, Coastal, Super Blu), and the galling was mainly from the lower ranked greases (oil, Vaseline, Slick Honey, Super Lube, Super Slick)
-Fortunately this time around all of the lubricants were better than the bare (no lubrication) test results
-Super Slick led this round, however it was again erratic (166, 205, 238, 350 grams).  Slickoleum was third but was also erratic (208, 301, 398 grams).  Fourth ranked Castrol was more consistent while still having a lighter pulling force (215, 219,  254 grams)
-A very interesting thing happened with the Marine Grease- on every test it wiped the baseplate clean leaving no lubricant behind.  The grease appears to have caked up during the 15 second sitting period before each test, and then after moving a little ways it started flowing again.  In the bare spots there was no lubrication because it was shiny dry, compared to a faint iridescence (like an oil film on water) around the edges of the bare spot.
After a test and with the weight removed, a 3/16 inch long bare spot with no marine grease is visible on the stainless steel base plate (1/4 way in from right).  There is also a visible crack in the film to the right of it.  Then as the sled continues to move to the left the grease started to flow again and leave a normal film.  The Marine Grease did not do this on the copper and plastic baseplates.
-Slick Honey showed up in spots 18, 19, and 25 out of 31 with a consistent light sticking that raised it's pulling force totals

The UHMW Plastic Baseplate Test Results
The final baseplate was a piece of 1/8 inch thick Ultra High Molecular Weight plastic, which is often used in construction for sanitary washdown situations, as well as for sliding surfaces such as conveyors and chutes.  The plastic used in bearings can range from teflon, to a urethane, or to a harder custom formula with additives.  I don't have a sheet of urethane, but I think the UHMW sheet is a reasonable compromise between teflon and custom formula molded bearings.  I've had plastic bearings degrade more often than not after I lubricated them, usually because the plastic either turns brittle, cracks, crumbles, or swells, so I usually don't lubricate plastic bearings unless there is a special need.  This is an example of the lubricant manufacturers falling down on the job- a system already exists for categorizing plastics recycling, and manufacturers could easily test their products for compatibility with the categories.  A system also exists for oring compatibility.  But if I don't know what's in the grease then looking up compatibility is not possible.  At this moment if a plastic bearing needs lubricant, I use a silicone based grease that was meant for ski bindings which seems to be relatively benign.  It would probably be OK for bicycle pivots if there was a need, but I wouldn't expect it to serve in extreme pressure situations such as racing car suspension bushings.   I thought it would be useful to test these bearing greases on plastic though not only because of teflon coated bearings and molded bushings, but also the seals and shims used in assemblies.
UHMW plastic baseplate test results, click to enlarge
Only 18 tests (72 pulls) were done on the plastic because I felt it was lower priority than the metal baseplates and I didn't do repeat testing except for these 2 lubricants:
1. By this time I had the impression that the Castrol wheel bearing grease was a pretty good choice and wanted to double check it's performance
2. I couldn't believe the price of the Slick Honey so it was also given an extra chance to prove itself
General comments:
-The best lubricant was Protek MPL, which had a super lubricated feeling on the plastic
-Second was water, third was Marine grease
-Fourth was Slick Honey, and while it was slick, it also stuck, and for some reason seemed thicker on the plastic than on the other baseplates
-Castrol was middle of the pack, slightly worse than bare (no lube)
-I thought Super Lube with teflon would work well but it actually stuck twice
-The White Lithium felt gritty on plastic, like fine sandpaper 
-Super Slick was last, it did not agree with the plastic at all

Summary of greases
The results from the 3 baseplates were all different.  I tried to combine the results into one rating through 4 methods:
1. by simply adding the 3 ranking numbers for each grease and ordering by the totals
2. by adding the pulling forces of the 3 baseplates for each grease
3. by adding the pulling forces scaled to adjust for frequency of bearing materials, (the stainless steel results were multiplied by 4, the copper results by 2, and the plastic by 1)
4. by comparing the erraticness (a grease's highest test results divided by it's lowest test results)
Screenshot of 4 evaluative methods, with trends.  In general the automotive greases (the green entries) consistently scored better than the bicycle greases (yellow). click to enlarge

These results do show trends, however they should be screened for a few last considerations that show up more in the test notes than in the pulling forces:

Greases to be removed from consideration
Besides obvious removals such as water and graphite, these factors show up in the notes:
-Vaseline scored well, but it does not have the additives to help it last (service life) and was thin and easier to press through the film while rubbing in than other greases
-Oil had a very low pulling force with a very fast sled speed, but was easy to squish out and then stick.  Oil would be super slippery for a system where it is constantly being replenished, (such as with a pump inside a car motor), and can obviously be made to work in the total loss systems of old 3 speed hubs, but seems to require very frequent replenishment.
-While Marine Grease was in the middle of the results, I would not use it because of the caking that left bare spots in every test on stainless steel
-The Protek MPL is an exquisite grease, it felt very silky and slippery, but had lower results in the tests.  It seemed like it was meant for equipment with lighter loading such as musical instruments and firearms
-The general purpose White Lithium was about as thin as hot soft margarine, and had a gritty feeling on plastic. It was not very robust during rubbing in, and seemed lighter duty, I wouldn't use it on bearings.  (The other greases that have lithium in a compound work better.)
-The Super Lube with Teflon performed as well as or better than the bicycle greases but there is no statement on the label that it is meant for heavier duty use in bearings, only for food grade bearings.  Aside from the Teflon additive it closely resembles Vaseline, and I question it's service life on a bike.

The top greases were automotive
-The Castrol and Super Blu are about equal in effectiveness, however the Castrol was a nice consistency, like soft butter that's been in the sun, that stuck and stayed there but was easy to push around.  This seems like a good fit for fork sliding bearings, and with the wheel bearing pressure additives good for bicycle ball bearings too. 
-The Super Blu is thicker- more like butter in a cool room.  I will continue to use SB on bicycle ball bearings, but would hesitate to use it on forks.
-The Coastal was also soft and consistently better performing than the bicycle greases, but because of the Moly additive I put it in a special category, see the comments on Super Slick below.

The 3 bicycle greases were middle of the pack
Despite being slightly thinner than the automotive greases, the bicycle greases generally took more force to move the sled.  On copper Slick Honey and Super Slick were about the same, and Slickoleum was noticeably lower performance.  However on stainless steel Super Slick was definitely at the top, with Slickoleum mid pack of all the tests, and Slick Honey in the lower third.  On plastic Slick Honey was in the top third, and Slickoleum was at the bottom, with Super Slick last of all 15 lubricants tested.
-I'd rate Super Slick as the best bicycle grease with a big qualification- it is the best between two hard metals, and is not good if there is any plastic in the system.  This matches up with my previous experiences with lubricants that have a Moly additive, I've found that Moly and plastic don't agree.  The plastic eventually seems to turn hard and is deeply stained, suggesting a chemical change.  Because my Manitou forks have urethane bumpers and plastic struts inside them, I won't use either Super Slick or the Coastal Moly EP on them.  There is no statement saying that Super Slick is compatible with plastic other than an unhelpful "best for fork overhauls" on the Super Slick package.  It also is darker than the Coastal Moly EP, and reminds me of CV joint grease which often has some lead in it.  While I like this grease, I'll use it carefully in select applications.
-Slick Honey tested well on plastic and good on copper, but was less than middle of the pack on stainless steel which is probably the most common installation.  Slick Honey is slightly whiter than Slickoleum and has a light yellow brown tinge.  Considering the price and the availability of better choices, I'm not planning on getting more after this little 2 ounce tube is gone.
-Slickoleum was poor on copper, erratic on stainless steel, and poor on plastic.  Slickoleum resembles Vaseline in texture and has a faint dark brown tinge that is similar to the color of oil.  While it doesn't meet my needs, I am impressed that the label does have some helpful information, including the statements that it is not for use on copper or with EPDM (such as certain o rings).  
Both Slick Honey and Slickoleum do not have Moly in them and I would use them with less concern about general compatibility than Super Slick.

What are the Manitou forks being assembled with?
The forks are going together with the Castrol multipurpose wheel bearing grease.  It consistently performed well, often with less drag than the thinner bicycle greases. It had a nice smooth film that was hard to push through but was still supple.  There is much less concern about compatibility with the plastic inside the forks than with the Moly greases.  It has a standard lithium base, probably similar to Prep M (I haven't seen Prep M), and I'm not introducing unknown chemical reactions with metals, plastic, or rubber parts.

The Whole Enchilada
I was not familiar with bicycle greases and they had a mythical reputation on the forums that I had read.  I found out automotive greases often work better.

Another consideration is that a 1 pound tub of wheel bearing grease can be bought for around $8.  I paid $17.99 for 2 ounces of Buzzy's Slick Honey.  There is something wrong here, that says marketing all over it.  Maybe a circumstance will come up that proves SH is valuable, but for now I can do without it.

My one regret is that I did not test one of the newer automotive synthetic greases.  Karl Gesslein seems to love Mobilgrease 28 enough to eat it.  (That could be part of the reason for his unique writing style.)   Maybe I'll test it after I've used up some of the grease around here.

I now think an automotive wheel bearing grease is as good as you can get.  However there is a catch, as not all NLGI 2 greases are the same.  Some are softer, some are harder. Without opening all the tubs on your local auto parts store shelf and sticking your finger into each one, at this moment I'd look for a grease that says multipurpose wheel bearing grease, and stay away from either a grease that is labeled for high temperature (as that is likely to be thicker), or a grease that is labeled only multipurpose without also including wheel bearings (as that is likely not to have the EP additives).

Love, love is a verb
Riding is a doing word