What's the diff? (3.64 review)

Helps you get the most out your Traqmate System
Club Racing
Posts: 28
Joined: Wed Jul 19, 2006 6:45 pm

What's the diff? (3.64 review)

Postby Hornswoggler » Thu Jan 10, 2008 5:57 pm

One statement that always intrigued me is when I hear somebody say that a shorter differential is "a poor man's supercharger". A couple years ago when I was shopping for more power, I shrugged this off and figured if I was going to have gains, I want to be able to measure them, not seat of the pants. Being a track junkie, I went the reliable approach (as opposed to FI) and installed a cam kit back in 2004. This wasn't to rule out the differential, as I always kept it in the back of my mind as one of those "what could I ever possibly do" kinda mods.

This past summer my trusty ol' stock diff started leaking so I considered my options. I don't know much about rebuilding differentials and replacing seals, clutch packs, etc. so figured I would leave that part to the experts and get a rebuilt custom diff on exchange. After 120k+ miles on the stock diff, I figured it was about due for clutch pack replacement anyways and made for a great excuse to upgrade. The stock unit that came with my 1997 M3 has ratio of 3.23:1 and 25% lockup on the limited slip. After doing some searching and plugging some numbers into a gearing calculator, i decided on the 3.64 w/ 40% lockup. With my 5-speed transmission, this puts my top speed at 140 mph, should be plenty fast for the tracks I visit (assuming I do not add power without more revs or forward gears).

So I order my diff from Diffs online. Fantastic support and excellent product. Dan @ diffsonline was able to address all my newb questions with ease. While I was going through the effort to install, I decided to also swap rear subframe bushings to PowerFlex solid urethane replacements. Install wasn't too bad, just time consuming (<- not a mechanic) and a lot of work. Thankfully some local friends were able to help with a few parts (thanks again Martin for pressing the diff bushings, and Andy for subframe re-install!).

All back together and broken in, how is it? It definately goes through the rev range quicker, which gives a sensation for speed. Trying not to be tricked by my senses, I tune in closer to actual acceleration but since it had been about a month working with the rear subframe, its hard to tell. To me, it feels pretty good but not enough to blow me away. I notice I am spending more time shifting... not that I mind but time spend shifting is time you are not accelerating. Of all the gears, I probably notice the improvement most in 4th gear. Was it a poor man's supercharger? Eh, that term might have been used in the negative sense. I noticed a MUCH bigger improvement with cams and supporting mods (intake manifold, injectors, etc).

Not being one to settle for seat-of-the-pants assessments, I was anxious to use my traqmate in comparing track data from before and after. My next two HPDE events were at tracks I did not have a baseline. Around the same time I also removed my car stereo and replaced the sunroof with a carbon fiber piece, resulting in about 100-150 lbs weight loss.

First event with the new diff was Road America. Without having a baseline to compare, hard to tell how much difference the diff made, but I will say the 3.64 was the perfect ratio for this track and my 5-speed. At three points on the track I reached speeds between 130-135 mph, which was just about redline in top gear. If I ever added power, I would need either more revs or another gear for Road America. This was the first time I ever felt like I was getting the most out of 5th gear. My best lap, according to my traqmate, was a 2:42.689 (best session available for download at their Share and Compare website). I am very happy with that time. Could probably improve a second or two with some more data analysis and practice. Was only my third day driving Road America (and first one hardly counts as it was my first event... back in 2003!)

Eventually I made it back to my current home track of Autobahn Country Club in Joliet, IL to run their 3.5 mile full course. Having about 10 previous days seat time among the three track configurations (north, south, and full), I am fairly consistent around there and know the track pretty well. Getting straigh to the laptimes from full course:

Previous baseline: 2:54.498
w/ 3.64, 100 lbs lighter: 2.51.748

Driving impressions: Previously, with my 3.23, there were two points on this track that I was "between" gears. Going into South 4 and North 6, I would run out of steam in 3rd gear and would have to make a brief shift to fourth. I am in fourth gear earlier so these two weird moments are no longer an issue. OTOH, there are atleast two points with the new diff where I run out of 4th gear and have to either make a short shift to 5th or a few seconds of coasting before the braking zone. Coasting is bad when going for laptimes, but by the time I am done shifting, its almost time to brake. :shrug:

3 seconds is a nice improvement. The challenge now is to isolate whether these 3 seconds came from acceleration/traction (the diff), reduced weight (should slightly improve all performance metrics), or improved driving technique. Let's dive into some traqmate data!

The best gains from the new diff came between 50 to 80 mph, sometimes seeing a gain of .3 seconds each time. I see this advantage about 4 times per lap around autobahn full course, adding up to 1.2 seconds improvement.

What about longer acceleration runs? Looking at the longest stretch (South to North connector, 50-105 mph), the new diff consistently gains 0.1 seconds and about 1 mph at end of the straight. In the below picture, the red plot is with new (3.64) diff, and blue is stock 3.23 diff:


With the diff gaining about 1.2 seconds between the range of 50-80 mph, where did the other 1.5 second improvement come from? I gained .25 seconds in the North-South connector by maintaining a higher minimum speed. Found .25 seconds with a higher apex speed in South 4, a .5 second improvement with higher min speed at South 5. Getting on the gas a bit sooner in S8 gave me a .3 second improvement. Getting on the gas sooner in N9 was worth .3 seconds, in addition to the 50-80 mph advantage. This adds up to a little more than 1.5 seconds, but I lost a little in S6 as I was not driving that corner as hard as I could have.

Being able to get on the gas sooner was probably a benefit of the higher lockup % on the limited slip. Hard to say the condition of my old clutch packs, I didn't think they were a problem, but the fresh 40% could have been just a bit better.

It is hard to isolate how much of the gains were from the weight reduction and how much was from the differential. If this were a good science experiement, I would have done them separately.

- In the perfect world, you would run gearing specific to each track
- .3 second advantage seen between 50-80 mph (~1.2 second improvement around Autobahn full course)
- Acceleration gains between 3.23 and 3.64 are sometimes negated by time spent shifting
- Higher % lockup (from 25% to 40%) may have helped application of power on corner exit (~0.6 second improvement around full course)
- Wish I had more revs to better exploit the lower gears!

Andre' Surles
Team Traqmate
Posts: 651
Joined: Mon Feb 07, 2005 5:53 pm

Postby Andre' Surles » Fri Jan 11, 2008 11:10 am

Thank you so much for posting! Great read and I really appreciate the info on this. Having helped install the same diff on a friends 95' M3 and subsequently driving it I got similar impressions. Not a night and day change, but it did make the car noticeable more fun through 3rd and 4th. Just felt more on boil.
Great to see the data back it up too! Good luck finding more revs and keep posting.
Andre' Surles
Track Systems
Redline! Redline! Redline!

Club Racing
Posts: 28
Joined: Wed Jul 19, 2006 6:45 pm

Postby Hornswoggler » Fri Jan 11, 2008 1:28 pm

I figured this would make a great "real world" example of how using a traqmate can answer typical car myths like differentials "multiplying torque". Must be a pretty small multiplier, lol.

If I had to choose between a $1,600 differential and a traqmate, I know which one has made me faster. :)

Return to “Data Analysis”

Who is online

Users browsing this forum: No registered users and 1 guest