My E90 M3 Dedicated Track build (with pix!)

[dogbone]

PART 3: Conclusions about the OS Giken differential versus Stock M3 differential

I have learned over the last few years that making a car faster is actually about making it more predictable over a wide variety of challenges. Brakes are a good example: brakes don't make the car go faster, but you can achieve better lap times with a BBK and race pads. Why? Because they brake consistently every time you press the pedal on any lap at any temperature. You learn to trust them and are willing to push harder. The stock brakes in an M3 will last a couple laps but then the fluid boils and you're done----foot goes to the floor. A good suspension effectively widens the track. More lines are available to you with a quality suspension. If you can jump on a berm because the suspension can handle it, then you have more track to work with. Plus, the quality suspension will hold the tire to the ground more effectively than a cheap damper. The more time your tire is in contact with the ground, the more control you have. The more control you have the faster you are willing to go. Same thing with aero, subframe bushings, etc. All these things add more predictability to the situation.

The same thing is true with the differential. If you have a differential that makes the car react in a more predictable way, then the possibility exists to go faster in the long run. The differential itself isn't a go-fast part necessarily. First of all, if you've learned your car with one kind of diff and then stick another one in there, there's a learning curve. We saw in Part 2 in slow corners that I'm guilty of pressing the gas too much with the new diff in there. So, if you learn the new equipment properly, the differential makes the car more predictable, and then we can say we've spent our money well. But the diff was a mysterious piece in the car. Very few people seem to understand them, and then there's even less data. I hope Part 2 changed that.

We set out by asking a number of questions. Let's revisit them and see if we have answers:

-Does the OSG lock on acceleration as promised? Yes, it locks on EVERY single press of the gas pedal. It does so in a very consistent way.

-Does the OSG lock quickly on acceleration? Yes. The wheel-speed spikes are virtually eliminated when the gas pedal is pressed. We can confidently say the power is getting to the weighted wheel quickly.

-Does the OSG lock on initial braking? Yes. I have mentioned before, the car feels like it's being slowed by a parachute. I would say that braking feel is the MOST obvious feel change.

-Is the OSG open when entering a corner? Yes. The data clearly shows the diff is open on every corner entry. This is important.

-Can you actually see the performance characteristic of the OSG looking any different than the Stock unit? Yes. The OSG lines are much smoother than the Stock unit.

-Does the stock unit lock on acceleration? Does the stock unit lock quickly? Someone with more experience than me would have to look at these charts to determine how much the stock unit is locking. It's not obvious and it's not quick. After every corner, there is a very consistent spin up of the unweighted wheel.

Clearly the Stock diff is not "slow" because my current standing PR was done on the Stock unit and I have not been able to eclipse it------yet. But I just chalk that up to the learning curve with the OSG. As I mentioned in Part 1, I have run faster segments throughout the track with the OSG, but I haven't been able to put all the pieces together in one lap. That's ok. I'm patient………I'M REALLLLLLLLLY PATIENT!!!!!!!! OH YEAH-----SUPER DAMN PATIENT!!!!!

I am glad the AIM Solo DL was able to track the wheel spins along with all the other data. I have to say, I hate the whole "Butt Dyno" thing. In my opinion, the Butt Dyno is very much affected by rate of drain on the wallet----it's an inversely proportional relationship: the less money left in your wallet, the more your Butt Dyno tells you your new mod is effective. Perhaps we could find a way for the AIM Solo to somehow track that too……

So, after going through the data, I am comfortable in saying that the money was well spent and that the car is more predictable------more predictable on acceleration, more predictable when it's sliding and more predictable on braking.


Couple other thoughts:

-As usual, tires are the limiting factor. The grip limit of the tire is the boundary for a differential. You can transfer all the power of the world quickly to a tire, but if the tire can't handle it, then I hope you enjoy going sideways. So, a custom setup diff needs to account for the tire. Which leads me to my next thought:

-Lots of engine power can help overcome the delay of an open diff----again the tires are a factor here. My sense in all of this is that a car with a lot power can overcome a diff that remains largely open and still send enough power to the loaded wheel quickly so that the max grip of the tire is being used. Tires only have a certain amount of grip, so if the supercharger, with it's 50% more power than stock, can slam a bunch of power out, some of it will end up at the weighted wheel quickly and that amount of power is probably near the limit of most tires. With the Stock diff, I was able to go sideways whenever I wanted. So, I think having a supercharger had allowed me to get good power to my tires quickly with the Stock diff, but you could see that the situation would always be pretty dynamic when I got on the gas where the OSG could achieve the same results while being more predictable and less dramatic. Again, people would say that driving with the OSG is less "ham fisted".

-A lower powered car probably benefits more from a performance diff than a higher powered car. A low powered car does not want to waste any time or power spinning up an unweighted wheel. If I had to guess, I'd say the diff setup of a lower powered car probably involves having it lock really quickly. You want it to lock immediately so you can squeeze every horsepower possible to the weighted wheel.

-It's worth noting that a diff that remains open is considered to be a "safe" diff. A high performance diff that locks easily can send a car sideways very quickly. This would not be ideal on the street where an inexperienced person with a powerful car is goofing around. By staying open on a monster gas press, the diff delays the onset of the car going sideways, and perhaps the traction control can kick in to save the situation.

-What I am learning from staring at these charts is that braking earlier and getting on the gas earlier seems to net a better result than braking later and waiting to get on the gas later. In my earlier track days I thought having a BBK meant I should wait to brake until the absolute last second and go careening into a corner, but after looking at the data quite a bit, it seems the earlier braking gets better times with less stress and panic. Now I'm not suggesting coasting on the brake from half way down a straight. I'm just saying that waiting to the absolute last second and having the car in a stitch at the apex requires that you wait to get on the gas, and it seems to me that staying at a higher speed longer and braking later is worse for lap times than braking a little sooner and being able to get on the gas quicker so that you're at a lower speed for less time. I guess my realization falls into the saying, "slow in, fast out". It's just interesting seeing it play out in my AIM data.

I want to thank a few people for the knowledge they shared with me on this topic:

-driftflo is a forum member here. He does two things: he has a differential building business and he runs a drift school in Germany. He has been extremely generous with his deep knowledge on this topic. I almost bought his diff instead of the OSG because I'm quite certain his diff builds are awesome. Ultimately, cost due to VAT tax, my proximity to OS Giken in Los Angeles and shipping time were what swayed me in the end. Since I was new to this topic, I wasn't sure that whatever diff I bought would be the right setup, and being so close to the OS Giken shop was comforting that I could quickly deal with changes. Also, I figured if I started with the OSG and didn't like it, I could sell that much more easily than if I didn't like driftflo's diff and tried to sell a diff no one was familiar with. I would love to compare driftflo's diff to what I have someday…….Anyway, I have no hesitation in recommending him.

OS Giken techs - the guys at OS Giken were patient with me and answered all of the ridiculous questions I had as well. And they were lightning quick on their turnaround. I ordered the diff from an authorized OSG vendor, they sent the P.O. to OSG in Los Angeles, I had two conversations with OSG, they built it. It was in my hands two days after ordering it.

diffsonline - they were able to provide me with a couple parts for my OSG that are hard to find and also answered some questions.

minicorsa - Poor Danny had to deal with all this nonsense.

Anyway, I hope these postings on this topic help people to understand diffs better. If nothing else, I am quite sure they can cure insomnia!!

[nholmes]

Good shit here. Too much data for me to digest, but awesome.

[M3 Number 86]

swapping diffs for certain tracks would help as well.

[dogbone]

Originally Posted by nholmes

Good shit here. Too much data for me to digest, but awesome.

Hehe yeah I know, it's a bunch of stuff. But I figured one whole lap was the best way to look at it.

[dogbone]

Originally Posted by M3 Number 86

swapping diffs for certain tracks would help as well.

Ooh, now THAT would be a hardcore solution. Better have a few bucks laying around for all that swapping though.

[M3 Number 86]

Originally Posted by dogbone

Ooh, now THAT would be a hardcore solution. Better have a few bucks laying around for all that swapping though.

yep - i guess it's next level shit because there are a few who do it for short/long tracks/etc.

i just got.....ONE seat haha.

[italyix]

Glad someone actually posted an extremely detailed diff analysis, good read and ty for doing so.

You actually have me really excited now as my car is getting its diff rebuilt today with the driftlo unit I bought last year! I too was looking at the OSG but ultimately decided on driftlo's. Price was much better than OSG as well.

[dogbone]

Originally Posted by italyix

Glad someone actually posted an extremely detailed diff analysis, good read and ty for doing so.

You actually have me really excited now as my car is getting its diff rebuilt today with the driftlo unit I bought last year! I too was looking at the OSG but ultimately decided on driftlo's. Price was much better than OSG as well.

Of course, when writing all this stuff up, I knew that this wasn't the most popular topic in the world. And the length is obviously an issue for a casual reader. The problem with this topic is that you can't just look at only one corner and say you've done your homework. You just have to make your way around every corner of a dynamic track to see what's going on. And you want to see how they compare directly, so they need to be overlaid on top of each other. But then the data overlaps and it's hard to see what's going on, so you want to see them isolated.

This is one of those things that hopefully will pop up in a Google search and be useful to people who are seeking the info.

As far as the driftflo diff, yes I knew you had one from your postings. I never did hear any feedback. I guess you just never installed until now? Anyway, I've spoken to him several times now. He is a very nice guy and is super knowledgable on this topic.

Next on the list will probably be dealing with the DCT Cooler.

Good luck with your rebuild. Report back here about your experiences with the driftflo unit if you don't mind.

[roastbeef]

i don't quite understand all of it, but i sure as hell enjoy reading it. thanks for the read.

[italyix]

Originally Posted by dogbone

As far as the driftflo diff, yes I knew you had one from your postings. I never did hear any feedback. I guess you just never installed until now? Anyway, I've spoken to him several times now. He is a very nice guy and is super knowledgable on this topic.

I bought a house and car stuff got put on hold. I finally sorted out the finances and am able to get back to business! I have seats also going in and full spherical conversation in front as well

[Downshift22]

Nice build! I haven't had a chance to go through every page yet, but I did see talks of oil and dct coolers.

I installed the CSF oil and trans cooler on my car during the offseason. The last two events prior to install, the car only had bbk and nt01s. Oil temp neared the 300 mark via cluster after <15 mins of running hard. DCT would have a lot of lag at this point. Ambient temp was in the 80s.

I did my first event with them a couple of weeks ago, weather was high 60s most of the day. Car now has suspension and aero added, weighs 3675 with driver (3500 curb).

Unfortunately I wasn't able to get my solo DL hooked up in time, but based on the gauges the oil temp stayed very consistent and didn't once overheat, even when pushing the car hard for 35-40 mins in the last session of the day. Very happy with the oil cooler thus far.

The trans still developed a slight lag after about 20 mins. At that point I would do a lap or two with less shifting to cool it off and it would go back to normal. So trans cooling does seem to be slightly improved, but still not solved.

My next event will be the first weekend of April and I will have the DL hooked up, so I will report back with actual numbers

[dogbone]

Originally Posted by Downshift22

Nice build! I haven't had a chance to go through every page yet, but I did see talks of oil and dct coolers.

I installed the CSF oil and trans cooler on my car during the offseason. The last two events prior to install, the car only had bbk and nt01s. Oil temp neared the 300 mark via cluster after <15 mins of running hard. DCT would have a lot of lag at this point. Ambient temp was in the 80s.

I did my first event with them a couple of weeks ago, weather was high 60s most of the day. Car now has suspension and aero added, weighs 3675 with driver (3500 curb).

Unfortunately I wasn't able to get my solo DL hooked up in time, but based on the gauges the oil temp stayed very consistent and didn't once overheat, even when pushing the car hard for 35-40 mins in the last session of the day. Very happy with the oil cooler thus far.

The trans still developed a slight lag after about 20 mins. At that point I would do a lap or two with less shifting to cool it off and it would go back to normal. So trans cooling does seem to be slightly improved, but still not solved.

My next event will be the first weekend of April and I will have the DL hooked up, so I will report back with actual numbers

Thanks for the feedback on your coolers. I would be curious to see an AIM chart showing oil temps from the same track before and after. It would be nice, as you keep tracking into the warmer season, if you could compare days with more similar air temps and overlap them in an AIM chart. A 60s day is quite different than an 80s day for my car.

To me, the CSF DCT cooler seems quite similar to the stock unit. One extra cooling tube and 20% more surface area is not going to help a car that's overheating in two hot lap attempts at a PR. My opinion is that something more formidable is needed. And as I mentioned before, the problem will get worse if I put slicks on some day.

[Downshift22]

Originally Posted by dogbone

Thanks for the feedback on your coolers. I would be curious to see an AIM chart showing oil temps from the same track before and after. It would be nice, as you keep tracking into the warmer season, if you could compare days with more similar air temps and overlap them in an AIM chart. A 60s day is quite different than an 80s day for my car.

To me, the CSF DCT cooler seems quite similar to the stock unit. One extra cooling tube and 20% more surface area is not going to help a car that's overheating in two hot lap attempts at a PR. My opinion is that something more formidable is needed. And as I mentioned before, the problem will get worse if I put slicks on some day.

Of course. I'll work on getting some charts going. I still need to sit down and learn race studio.

I'm going to look into getting a temp sensor on the transmission as well.

[///Mobbin]

dogbone I'm wondering if you've had any more thoughts on the GTS DCT programming and S4 vs S5 vs S6. I haven't tried it at the track yet but I've noticed that my car lunges forward slightly when upshifting in certain gears in the middle or mid-upper end of the RPM range. Perhaps closer to redline this will go away or be minimal but I'm nervous this might upset or unbalance the car on the track. Downshifts are generally smoother which is good but occasionally I feel like the rev-matching is a little off (gear shift complete on early side of throttle blip). Again, I'm not driving on the street like I would at the track so perhaps everything will be smoother in the proper environment with shifts closer to redline. And generally I do enjoy it on the street, just wondering if the OEM programming might be smoother for me at the track (S6 was butter smooth for me before). I also have the Dinan 3.45 diff if that has any effect on your thoughts here.

[dogbone]

Originally Posted by ///Mobbin

dogbone I'm wondering if you've had any more thoughts on the GTS DCT programming and S4 vs S5 vs S6. I haven't tried it at the track yet but I've noticed that my car lunges forward slightly when upshifting in certain gears in the middle or mid-upper end of the RPM range. Perhaps closer to redline this will go away or be minimal but I'm nervous this might upset or unbalance the car on the track. Downshifts are generally smoother which is good but occasionally I feel like the rev-matching is a little off (gear shift complete on early side of throttle blip). Again, I'm not driving on the street like I would at the track so perhaps everything will be smoother in the proper environment with shifts closer to redline. And generally I do enjoy it on the street, just wondering if the OEM programming might be smoother for me at the track (S6 was butter smooth for me before). I also have the Dinan 3.45 diff if that has any effect on your thoughts here.

Regarding the GTS DCT programming, I know what you mean about the car feeling a tad "lungey" on upshifts. I felt it too, but only on the street. On the track, as you pound through the gears at WOT near redline, there's NO lunge---at least not in my experience. The shifts are quick and decisive. With a supercharger, you get a nice sharp kick in the pants.

And yes, I stick by the smoothness of the GTS DCT downshifting. Excellent EXCELLENT downshifting. I don't know why the stock car doesn't downshift like that. Maybe it's less gas efficient or something making the car rev and work harder to make it so smooth. You'll see comments to my videos that people marvel how I seem to be able to downshift---and multiple-downshift---ANYWHERE. Regardless of what the upshifting feels like, to me the downshifting is more important because it's happening at a more critical moment. When you're setting up for a corner, braking and thinking of turning, you don't want your downshift to change the balance of the car. That's the true beauty of the DCT. It can pull that off sweet as pie.

As far as diff ratios, since I have a fair amount of extra torque/power from a supercharger, I stayed with the stock ratio on my diff setup, so I can't offer much insight there. But I guess it's possible the "lunge" may be more accentuated on your setup with that 3.45. Also, I'm pretty satisfied with my shift points at Buttonwillow which I consider to be my "home" track. For me, DCT makes it mostly unnecessary to worry a lot about shift points. I've ridden passenger enough in 6-speeds on track to see how it's WAY more important a topic for those guys.

[///Mobbin]

Originally Posted by dogbone

Regarding the GTS DCT programming, I know what you mean about the car feeling a tad "lungey" on upshifts. I felt it too, but only on the street. On the track, as you pound through the gears at WOT near redline, there's NO lunge---at least not in my experience. The shifts are quick and decisive. With a supercharger, you get a nice sharp kick in the pants.

And yes, I stick by the smoothness of the GTS DCT downshifting. Excellent EXCELLENT downshifting. I don't know why the stock car doesn't downshift like that. Maybe it's less gas efficient or something making the car rev and work harder to make it so smooth. You'll see comments to my videos that people marvel how I seem to be able to downshift---and multiple-downshift---ANYWHERE. Regardless of what the upshifting feels like, to me the downshifting is more important because it's happening at a more critical moment. When you're setting up for a corner, braking and thinking of turning, you don't want your downshift to change the balance of the car. That's the true beauty of the DCT. It can pull that off sweet as pie.

As far as diff ratios, since I have a fair amount of extra torque/power from a supercharger, I stayed with the stock ratio on my diff setup, so I can't offer much insight there. But I guess it's possible the "lunge" may be more accentuated on your setup with that 3.45. Also, I'm pretty satisfied with my shift points at Buttonwillow which I consider to be my "home" track. For me, DCT makes it mostly unnecessary to worry a lot about shift points. I've ridden passenger enough in 6-speeds on track to see how it's WAY more important a topic for those guys.

EXACTLY what I wanted to hear, thank you. Returning to the track in 3 weeks and can't wait to get back out there to try this out!

[Richbot]

Very good analysis

But isn't the stock diff in the e9x m3 a GKN visco-lok? You mention it's a Torsen in your first diff post.

The GKN unit is a riff on clutch diffs with a torque sensing viscous fluid filled "puck" that tells the cluthes when to lock when speed differential is enough

http://www.gkn.com/driveline/our-sol...coLok-engl.pdf

Different from Torsen which is clutch-free. If you could update that first post with this info (which driftflo has also posted) it would be good for those who read it later because it is otherwise very well done and would not want someone taking it as authoritative on the rest to get wrong info and cite you that's no way to win the Internet!

[dogbone]

Originally Posted by Richbot

Very good analysis

But isn't the stock diff in the e9x m3 a GKN visco-lok?

Thanks.

hmmm Regarding the stock diff---thanks for bringing that up. Yes, the GKN Visco loc is the stock unit. People (techs, shops, etc.) have always generically referred to it as a Torsen unit, so that's all I've ever known about it. (One person who built high end custom cars lectured me on the topic for 15 minutes telling me it was a Torsen diff.) On my end, after hearing all the same info for years, I never looked up the particulars of the stock unit because it was on it's way out and I wasn't considering buying it again.

Now that you've brought it up, I've actually looked, and the GKN Visco loc is really a Speed Sensing unit. I cannot tell you how annoying it is to find that out now!

Anyway, interestingly, when you look at the stock unit data, it makes even more sense knowing that it's a speed sensing unit. Because the stock unit doesn't seem to respond until one side is quite spun up. In my original analysis, you can see I commented about how there's always a delay in the stock unit and that it needed to spin up first. I always wondered how a torque sensing unit could respond so slowly to torque, especially considering I have a supercharger that puts out even more torque than stock. But now, I see the GKN Visco lok is responding only after it's spun up on one side----because it's a speed sensing unit. Geez----why is there so much mystery and misinformation surrounding this topic?

Thanks for the heads up!

[dogbone]

Originally Posted by Richbot

Very good analysis

But isn't the stock diff in the e9x m3 a GKN visco-lok? You mention it's a Torsen in your first diff post.

The GKN unit is a riff on clutch diffs with a torque sensing viscous fluid filled "puck" that tells the cluthes when to lock when speed differential is enough

http://www.gkn.com/driveline/our-sol...coLok-engl.pdf

Different from Torsen which is clutch-free. If you could update that first post with this info (which driftflo has also posted) it would be good for those who read it later because it is otherwise very well done and would not want someone taking it as authoritative on the rest to get wrong info and cite you that's no way to win the Internet!

oops---didn't see your edit until now. Yes, I will update the posting.

[dogbone]

Originally Posted by Richbot

If you could update that first post with this info it would be good for those who read it later because it is otherwise very well done and would not want someone taking it as authoritative on the rest to get wrong info and cite you that's no way to win the Internet!

Ok, it's updated. Thanks again for helping to clarify one more piece to this puzzle.

[RickyBobby]

What percentage lockup did you set it to? I had a clutch type differential on a previous car and took it apart 3 times to flip clutch plates before I found my preferred settings.

[dogbone]

Originally Posted by RickyBobby

What percentage lockup did you set it to? I had a clutch type differential on a previous car and took it apart 3 times to flip clutch plates before I found my preferred settings.

I don't know % lockup. OS Giken decided all of that and didn't tell me. They purposefully remained vague about some of the details saying that some of the setup info is proprietary and is based on their database of all the race cars they have setup over the years.

Is % lockup something that can be calculated based on number of active plate sets and ramp angles or some other parameters? The OS Giken unit in particular is a little unique from what I understand because it has different Pressure Ring Tuning Spring options in the 16 holes in the Pressure Rings (mine is pictured below). Those, along with different Cone Spring options and the setup as far as pre-load goes starts getting to be a little "Black Box-like"…..

Here's all I know. Maybe you can tell me the % lockup based on this? I would love to know…..

-1.5 setup
-Ramps: Based on Y-axis---35º on Power, 25º on Decel. (On X-axis: 55º/65º)
-All 10 clutch plate sets are active.
-There's "hardly any preload" is what I was told. (whatever that means)