Relative Viscosity Retention Comparisons Among Five Brands of Automotive and Motorcycle Oils
by John C. Woolum/ Ph.D.
Professor of Physics
California State University, Los Angeles
The central dogma of motorcycle oil manufacturers and distributors has always been that motorcycles put different demands on their lubricants than do automobiles. In particular, they point to the facts that motorcycles run at higher temperatures and use the same oil in their transmissions as in their engines. The transmission gears supposedly put extreme pressures on the oil molecules, thus causing the long oil polymers to break down. High temperatures can have the same basic effect, as well as additional effects such as the increase in oxidation products.
When the size of the oil polymers decreases ("cut up by the transmission gears," as at least one manufacturer claims), the oil thins. In other words, its viscosity decreases, as well as its ability to lubricate properly. For example, what started out as a 40-weight oil could effectively become a 30-weight oil, or even a 20-weight, after prolonged use. What this means, effectively, is that if the claims of the motorcycle oil producers are valid, they can easily be verified through measurement of viscosity changes on various oils as they are used in different applications.
Measuring the viscosity drop in oils did not seem like too difficult a task, especially since measuring viscosity of solutions of large molecules is a common practice in many biophysics laboratories - mine included. My lab had all the correct equipment - in fact the viscometers that I normally used for solutions of DNA and proteins were originally designed for oil measurements.
Setting the Stage
Viscosity is a measure of the friction between two layers of a liquid sliding relative to one another. It is usually measured in poise, or grams per centimeter per second (g/cm. sec). The basic principle of many viscometers is to measure the time required for a known amount of a liquid to pass through a capillary tube under gravitational force. The time taken will depend on the viscosity and the density of the liquid. The more viscous or less dense the liquid. the longer the time it will take to flow through the capillary.
Therefore in reality, this kind of viscometer does not measure viscosity directly, but rather the ratio of the viscosity to the density of the liquid being tested. This ratio is called the kinematic viscosity. and the common unit for expressing it is in stokes or poise cm^3/gram.
The viscometer used for my measurements was an Ostwald-type, Cannon-Fenske 200, designed to measure kinematic viscosities in the range of 10 to 100 centistokes (a centistoke is one-hundredth of a stoke). The oils being measured had kinematic viscosities between about 10 and 25 centistokes.
For the test samples, I decided to use two types of oils designed specifically for motorcycles and three types of fairly standard automotive oil.
The automotive oils were Castrol GTX 10W40 (petroleum based, $1.24/qt.), Castrol Syntec 10W40 (synthetic, $3.99/qt.) and Mobil 1 15W50 (synthetic, $3.48/qt.). The motorcycle oils were Spectro 4 10W40 (petroleum based, $4.99/qt.) and Honda HP4 10W40 (petroleum/synthetic blend, $5.99/qt.).
Each of these oils was run in the same motorcycles 1984 Honda V65 Sabre-under as near to identical conditions as possible. The oils were sampled for testing at 0, 800 and 1500 miles each.
As temperature has a strong effect on viscosity, I had to make certain it was carefully controlled for the experiments. Using a laboratory temperature control chamber, all measurements were made at 99 degrees Celsius (error factor of plus or minus 0.5 degrees), which is about 210 degrees Fahrenheit. This is the most common temperature used for oil viscosity measurements. It usually took about 15 minutes for each sample to achieve equilibrium within the chamber.
Each oil's kinematic viscosity was compared with its own kinematic viscosity at 0 miles to establish the viscosity ratio. In addition, measurements were made of each oil's density at each state of the tests. The densities were found to change by less than one percent, which is about the limit of the accuracy of the measurements. Therefore, a ratio of the times taken for the oils to pass through the viscometer effectively gives the ratio of their actual viscosities, since the densities cancel out.
What this all means in layman's terms then, is that the ratio established for each oil at the end of each test is a percentage of the amount of original viscosity retained at that point. For example. the Castro] GTX sample at 800 miles showed a relative viscosity of 0.722, meaning it had retained 72.2 percent of its original viscosity. Or, if you want to look at it the other way, the Castrol had lost 27.8 percent of its viscosity after 800 miles of use in the motorcycle.
Just for comparison sake, I also tested the viscosity drop of the Castrol GTX automotive oil after use in a 1987 Honda Accord automobile. At 3600 miles of use, the Castrol GTX showed a relative viscosity of 91.8 percent.
As the Mobil 1 had retained so much of its viscosity after the 1500 mile test, it was the only oil I allowed to run longer in the motorcycle. After 2500 miles, the Mobil 1 recorded a relative viscosity of 79.1 percent.
Also, it is worthy of note that from a testing standpoint, the two most similar oils were the Castrol GTX automotive oil and the Spectro 4 motorcycle oil. By similar, I mean that they tested as having almost the same absolute kinematic viscosity and density right out of the container. So starting out as equals, the Castrol maintained its viscosity several percentage points higher than the Spectro, under the same use in the same motorcycle yet the Spectro costs about four times the price of the Castrol.
The Error Factor
As a scientist, I must always ask myself. Are there possible errors in these measurements that would make them invalid? One possibility here would be that there was more particulate matter (contaminants) in some oil samples than in others, which would increase the viscosity numbers of that oil. Particulates disrupt the streamline flow and so increase the viscosity. (Einstein was the first to derive the quantitive expression for the increase in viscosity due to spherically, shaped particles.)
Large particulates should have been removed by the oil filter, and a new filter was used for each test. Still, to determine the effect of smaller particulates the oil samples were centrifuged at 11,000 g (11,000 times the acceleration of gravity) for a period of 10 minutes. A considerable amount of particulate matter was found and removed in all of the 800 mile and 1500 mile samples. However, the change in viscosity made by eliminating these particulates was found to be negligible.
Another possible source of error would be that the conditions to which the oils were subjected were different. In all cases, the distances were comprised of approximately 70 percent city riding and 30 percent freeway riding. The range of temperatures and the average ambient temperature during which the motorcycle was ridden were approximately the same. If anything, the average ambient temperature was higher during the operation of the motorcycle with the Mobil 1 oil, which should have put it at a disadvantage, yet it scored the highest overall in the viscosity retention tests.
Of course the motorcycle did age somewhat during the testing period, which took place over a year-long span. It registered about 4000 miles at the beginning of these tests and about 14,000 at the end. The order in which the oils were tested was:
1) Castrol, 2) Spectro, 3) Mobil and 4) Honda.
The motorcycle oil producers have suggested that other criteria. such as the amount of wear metals and contaminants, might be unacceptable when using automotive oil in a motorcycle. To test this theory, I sent a sample of the Castrol GTX at 1500 miles to SpectroTech. Inc., for a complete oil analysis. Their findings were that all contaminants (water, dirt, coolant and sludge) were normal.
SpectroTech also reported that all wear elements (antimony, titanium, silver, copper, lead, tin, aluminum, nickel, chromium, cadmium, sodium and boron) were normal except for iron, which was reported as "mildly above normal" at 51 parts per million.
SpectroTech lists acceptable levels for all of the above listed metals except iron, for which they state, "values vary greatly with systems and parts." so it is not clear what exactly is meant by "mildly above normal." Perhaps it was in comparison to cars with 1500 miles on the oil. Also, this could have been due to cam wear, since the early Honda V-4s were known for excessive cam and rocker arm wear.
In any case, again I could find nothing to support the argument that automotive oils were somehow less effective than motorcycle-specific lubricants when used in a motorcycle.
It could appear from this data, then, that there is no validity to the constantly-used argument that motorcycle-specific oils provide superior lubrication to automotive oils when used in a motorcycle. If the viscosity drop is the only criterion, then there is certainly no reason to spend the extra money on oil specifically designed for motorcycles. There does, however, appear to be a legitimate argument for using synthetic and synthetic-blend oils over the petroleum based products.
In speaking to a number of people involved in the production, marketing and distribution of motorcycle-specific oils, we could not find anyone who could present a valid argument for discrediting the testing done by Dr. Woolum. In general, they all tried to turn the conversation another direction by bringing up other possible advantages to using their products, while ignoring the viscosity-retention question. Yet without exception it is their own advertising that consistently brings the subject up, touting the special shear-stable polymers as the primary reason motorcyclists should purchase their products.
It is this practice to which we take exception, as we have been unable to find evidence to support these claims. In short, it seems to be nothing more than a clever marketing ploy designed to enhance their products' image and separate motorcyclists from their money.
MCN is ready to print any research or test results provided by the oil companies to support their claims of superior viscosity retention, with this one proviso: The comparisons must be against actual, SG-rated oil products that can be purchased off the shelf at the average auto parts store. Tests against generic, basic-stock mineral oil or against the lower-rated SE and SF oils would lack any credibility in a real-world context.
Despite more than six months of research, reading all the claims and counter-claims printed by dozens of industry experts and lubrication experts, MCN cannot and does not purport to know all there is to know about the differences between automotive and motorcycle oils. However, what we do know is that we can find no substantive evidence that using a high-quality, name-brand automotive oil in an average street motorcycle is in any way harmful or less effective in providing proper lubrication and protection than using the more expensive, motorcycle-specific oils.
Figure I Petroleum Based, Multiple Viscosity, SG-Rated, Oils
Best Retail Prices Found
Maxum 4 Premium
Torco MPZ 2.95
Average Price/qt. 3.80
Automotive Oils <
Average Price/qt. 1.19
Average Price Differential: 319.5%
Synthetic Based and Petroleum/Synthetic Blend
Multiple Viscosity, SG-Rated Oils
Best Retail Prices Found
Golden Spectro 4
Maxum 4 Extra
Torco T4-R 5.99
Average Price/qt. 6.53
Automotive Oils <
Pep Boys Synthetic 3.99
Average Price/qt. 3.53
Average Price Differential: 185.0%
Figure II Relative Viscosity Retention
(as a percentage of initial viscosity retained
after normal use in the same motorcycle)
0 miles 800mi 1500mi
Spectro 4 100%
[ Home | Top of Page ]
[ Search | Site Map | Feedback ]
moi je dit on paye plus cher l'huile de moto ,alors je vais mettre de l'huile automobile