Project Z06: Polishing the Internals of Our Transmission

J.G.
Update by J.G. Pasterjak to the Chevrolet Corvette Z06 project car
Jan 21, 2020

Sponsored by

Since we had our 2004 Corvette Z06 transmission apart to crack check everything before its rebuild, we figured that everything deserved a proper shine. Actually, GearFX Driveline figured that, because they know better, and we’re products of the Florida public school system and wouldn’t know better anyway.

The key concept here is REM polishing, a very mildly abrasive polishing process that imparts a silky, almost impossibly smooth surface to mechanical bits.

In our case, those mechanical bits were the rotating, sliding and meshing parts of our transmission and differential. GearFX’s Jeff Horton described the polishing process as far from simply a cosmetic operation. “Don’t think of it so much as a power adder,” he said, “but moe as a power loss reducer.” Ultimately, the arrow points in the same direction for both of those terms, though.

So how does this all work? How does shining up a transmission guts result in measurable gains in output? According to the SAE-published research paper “Horsepower retention by ISF (Isotropic Superfinishing) of Automotive racing Components,” the answer lies in reducing internal frictions. Think about it this way: All of the rotational energy that enters the transmission must leave somehow, according to laws of thermodynamics and conservation of energy. We want as much of the rotational energy that enters one end of the transmission to exit as rotational energy driving our wheels as possible. But certain amounts of that energy will be turned into other forms of energy, most notably heat. Reducing friction by smoothing mating and sliding surfaces reduces heat, which translates into more power leaving the transmission as rotational force than thermal force. The whole process for a five- or six-speed transmission costs around $500.

At least that’s the theory. But the theory is confirmed in the paper by actual dyno testing with an engine hooked to a NASCAR-style T101 transmission. In the tests, the baseline, unmodified transmission consistently produced about 5 horsepower fewer than one with REM-polished internals—from 383 horsepower stock to 388 with the shiny guts. That’s a gain we’d be more than happy with if it was from a cold air intake or cat-back exhaust system.

But the polishing process has several orders of benefits inside a transmission. The reduced thermal load is better for bearings, and the reduced sliding friction of interfacing surfaces makes for smoother shifts and reduced chance of fatigue-related failures of interfacing parts.
Like Magnafluxing, the process is fairly straightforward, but requires some specialized equipment. Parts are loaded into a giant vibratory tumbler with the REM media and given a good spin. After a while, the parts come out shiny and smooth as glass.

While it sounds simple, there’s art among the science. Jeff Horton describes the finer elements of the process as “Knowing what precise media to use and for how long to get the results you’re looking for. That’s where the experience comes in. I have different blends of shapes and sizes of the ceramic-based media that we can use and combine for different durations and at different speeds to get ideal results. It’s a relatively simple operation, but it takes some ‘feel’ as well. You can throw the stuff in a flip the switch and it will shine up nice, but an experienced operator can really produce some amazing results.”

We’ve attached the SAE’s pdf document with the data as well. It’s worth a read:

Join Free Join our community to easily find more project updates.
Comments
A 401 CJ
A 401 CJ GRM+ Memberand Dork
1/21/20 1:06 p.m.

Sorry if I missed something.  I just glanced at the thread.  Seems to me that rather than subjecting an unshiny-gear transmission and then an assumed identical shiny-gear transmission to a chassis dyno (well, I guess it could be a before and after of the same trans), why couldn't the story be told simply by the fluid temp?  If I'm losing 5 hp on the unpolished gear set, seems to me that the only place it could be going is into fluid temp.  I don't know how many quarts are in that thing and the efficacy of the cooler but that seems like a lot.  It almost seems like you could save a bit of weight by ditching the cooler.

JG Pasterjak
JG Pasterjak Production/Art Director
1/21/20 1:14 p.m.
A 401 CJ said:

Sorry if I missed something.  I just glanced at the thread.  Seems to me that rather than subjecting an unshiny-gear transmission and then an assumed identical shiny-gear transmission to a chassis dyno (well, I guess it could be a before and after of the same trans), why couldn't the story be told simply by the fluid temp?  If I'm losing 5 hp on the unpolished gear set, seems to me that the only place it could be going is into fluid temp.  I don't know how many quarts are in that thing and the efficacy of the cooler but that seems like a lot.  It almost seems like you could save a bit of weight by ditching the cooler.

 

Well, the testing was performed as part of the research for an SAE paper, and done in conjunction with a bunch of NASCAR teams. So whenever we can leverage someone else's hard work for a cool story, we're going to grab that chance. Fluid temp tells a big part of the story, though. But reduced friction also pays off in reduced mechanical interfaces between the zillions of imperfections in surfaces, each of which could be an eventual failure point. 

A 401 CJ
A 401 CJ GRM+ Memberand Dork
1/21/20 1:22 p.m.

I get that you use the work that's already been done and I applaud you for it.  And when I said "story" I wasn't referring to the GRM story per se but rather just positing that it might be easier to show whether or not it works with a temp gauge than with a dyno.  That's all.

wearymicrobe
wearymicrobe UberDork
1/21/20 2:25 p.m.

I would imagine that time alone would destroy the micro imperfection though. I have pulled some seriously shiny gearing out of transmission on the small industrial scale that did not go in with that finish. when they were perfectly lashed. 

5HP worth of heat is hard to quantify in my mind. I have never seen a conversion setup to do something like that. 

 

EDIT 

 

5hp is 3750 Jules per second. Which if I remember is one Jule = 1 watt. 

So back of the napkin calculation is 10F increase in 5 gallons of oil every 3 minutes in friction heat in a perfect system with no cooling. 

Mr_Asa
Mr_Asa Reader
1/21/20 2:39 p.m.

In reply to A 401 CJ :

How would you ensure that the trans cooler temps before and after was comparing like to like?  You would have to do a run on a track and compare a before and after.  Outside temps, humidity, a sunny or cloudy day will change the readings of the data you get on the track and sure you could account for all that noise in the system

Or you could do a dyno pull or two where all of those things are already accounted for.

Unless I'm missing something in what you're saying?

A 401 CJ
A 401 CJ GRM+ Memberand Dork
1/21/20 3:21 p.m.

I'd make several runs in each configuration.  Then I'd have the law of big(ger) numbers working for me. 

How would your dyno runs be like to like?  Do you have a button to press that automagically puts the shiny happy gears in?

i'm just thinking aloud.  They did what they did and I'm happy GRM reported it.
 

 

Mr_Asa
Mr_Asa Reader
1/21/20 4:01 p.m.

Ambient conditions inside a shop are going to be much steadier than on a track, there's also no driver error, no track conditions, tire slippage, or other issues to worry about and factor out.  If its absolutely necessary to get as close to similar conditions as possible you run the dyno with an electric engine as the input and put the trans output straight into the brake portion of the dyno.

1.3% is such a small difference that you want to limit as much noise in the signal as possible.  A track is the opposite of that.

tyronejk
tyronejk New Reader
1/21/20 6:16 p.m.

This study is a little fishy.  The author works for a company who sells this polishing surface commercially, so there's an incentive to reach a positive conclusion.

The test results are also a little too cut-and-dry.  The fact that 100% of the non-polished parts reached failure but 0% of the polished parts failed seems a little artificial.  Why did the study look at just two kinds of parts: non-polished and super finely polished?  It would've been illuminating to see the lifespans of different surface finishes, not just the two extremes.  At the very least, one of the fully-polished parts should've been run to failure.

There's also the fact that this test was run dry with no oil.  Typical transmissions would be lubricated with oil and there are some applications where a rougher surface finish produces lower friction by saturating itself with oil, such as cylinder bores.

It's still possible this process does help and the dyno numbers show that, but I'd be interested in seeing how much material's been removed from the gears and the increase in gear lash.

Patientzero
Patientzero Reader
1/21/20 7:13 p.m.

SAE is the Society of Automotive Engineers.  This would have been scrutinized just like a paper on some kind of medical research.  It's no different than GM releasing a paper on the internal combustion engine even though they are in the business of selling them.

A 401 CJ
A 401 CJ GRM+ Memberand Dork
1/22/20 9:43 a.m.

Paging Gale Banks.

You disproved fancy-pants high capacity diff covers.  This should be next.

You'll need to log in to post.

Sponsored by

GRM Ad Dept

Our Preferred Partners
FvgI98qMb35JOE6CzQqMcRt7sd7uO9ergoMt6xvWxM9MgVKTz4MqU1mVp2udLd7C