996/997 GT3 Diff Rebuild: Anyone Done it?
Discussion
The following site gives a good overview of LSDs, ramp angles and crucially the effect of preload (which isn't discussed much in this thread): http://www.blueskymotorsports.com/index.php?option...
My Caterham has carbon covered plates in the diff and although using fairly modest accel/decel ramps, it can be massively tuned just by the use of setting a higher/lower preload on the clutch packs. Simply fitting the "right part number" and re-assembling is only half the picture....
My Caterham has carbon covered plates in the diff and although using fairly modest accel/decel ramps, it can be massively tuned just by the use of setting a higher/lower preload on the clutch packs. Simply fitting the "right part number" and re-assembling is only half the picture....
Ade,
Pure diff pron!!!
Fergus,
I assume that the reason its not been mentioned much is because the mainstream thinking on the GT cars is that for a car driven on street as well as the track, the factory preload setting is about right. That the balance of the diff itself is fine (with the 40/60 diff) but that the ability of the street plates to both provide this lock up and do it for more than a few track days is the main issue.
FWIW, the preload setting for the street diff is 5-15Nm. The preload on the diff in my Cup is 150Nm. So while lock up is more or less the same (40/60 v 45/65) and the plates fitted may be the same (assuming a street upgrade with motorsport plates) the behavior of the diff is quite different.
Pure diff pron!!!
Fergus,
I assume that the reason its not been mentioned much is because the mainstream thinking on the GT cars is that for a car driven on street as well as the track, the factory preload setting is about right. That the balance of the diff itself is fine (with the 40/60 diff) but that the ability of the street plates to both provide this lock up and do it for more than a few track days is the main issue.
FWIW, the preload setting for the street diff is 5-15Nm. The preload on the diff in my Cup is 150Nm. So while lock up is more or less the same (40/60 v 45/65) and the plates fitted may be the same (assuming a street upgrade with motorsport plates) the behavior of the diff is quite different.
fioran0 said:
...balance of the diff itself is fine ... ability of the street plates to both provide this lock up and do it for more than a few track days is the main issue.
FWIW, the preload setting for the street diff is 5-15Nm. The preload on the diff in my Cup is 150Nm. So while lock up is more or less the same (40/60 v 45/65) and the plates fitted may be the same (assuming a street upgrade with motorsport plates) the behavior of the diff is quite different.
NeilFWIW, the preload setting for the street diff is 5-15Nm. The preload on the diff in my Cup is 150Nm. So while lock up is more or less the same (40/60 v 45/65) and the plates fitted may be the same (assuming a street upgrade with motorsport plates) the behavior of the diff is quite different.
Hi. Totally agree with all the points above, esp. re plate longevity. The diff behaviour not being solely a function of accel/decel ramp angles was the point I was alluding to in my post. As an aside, the preload for the "street" setting seems very low? Does it take a good stab of throttle to get the diff locked up? i.e. not much effect on low torque inputs/part throttle? I can imagine this is fairly frustrating when on part throttle on the exit from a corner, when you want the diff to be in the process of locking up and the transition may be all or nothing?
You need as much stability as you can get with a 911 in the braking zone and a well set up diff is the key. I like a lot of lock under braking but the down side is a car that would prefer not to rotate so there is a conflict and a balance needs to be struck. I'll take a car that is lovely and stable under braking that I have to work to the apex than something that is easy to rotate. All of the time in a 911 is made up under braking. The lock on accelleration is usefull but the layout is already on the drivers side anyway.
For me at least it is the braking set up of the diff that is crucial.
For me at least it is the braking set up of the diff that is crucial.
Steve Rance said:
You need as much stability as you can get with a 911 in the braking zone and a well set up diff is the key. I like a lot of lock under braking but the down side is a car that would prefer not to rotate so there is a conflict and a balance needs to be struck. I'll take a car that is lovely and stable under braking that I have to work to the apex than something that is easy to rotate. All of the time in a 911 is made up under braking. The lock on accelleration is usefull but the layout is already on the drivers side anyway.
For me at least it is the braking set up of the diff that is crucial.
same here!For me at least it is the braking set up of the diff that is crucial.
fioran0 said:
sportsandclassic said:
As I have mentioned the ABD system is disabled according to Porsche so that is ruled out of the equation for the 997 issue a member mentioned.
Porsche diffs use a % value like 40/65% or 28/40% I believe these to be the ramp angle.. I hope, if so this is my understanding of things.
Mike
My technical booklet for the ABS and PSM systems is Porsche issued and says that the ABD remains on. The Bosch stuff I have also says the same. I see your PIWIS screen says otherwise though so who knows. Porsche diffs use a % value like 40/65% or 28/40% I believe these to be the ramp angle.. I hope, if so this is my understanding of things.
Mike
As per my previous post, its not particularly important either way as far as I see it. If you drive on the road then systems will be on. If you drive on the track regularly then you need to take the same action over the diff regardless of ABD.
The values Porsche use on the diffs are not ramp angles. They are percentages of lock up or as Porsche calls it, "Percentage of anti slip effectiveness".
The GT gear diff has a monster 50/80 % lock up. Its the biggest locking diff for the GT3 cars I have seen. If it were ramp angles it would have the lowest lock up of them all. The 997 diff is weaker than the 996 diff all things being equal. This is because it has ABD and is designed to allow it to operate. Just looking at the pictures are enough to tell you which diffs have more or less lock.
The actual ramp angles on the 996 diff are approx 50 deg on the 40% and 37.5 deg on the 60%. The GT ramp angles are 45 deg on the 50% and 30 deg on the 80%. The 997 diff has 50 deg and way more than 50 deg.
I've enjoyed this thread and would like to add my 0.5p worth for the LSD in the 997GT2.
The 997GT2 runs the 28/40 set up which Porsche state is designed (with the other systems and chassis configuration particularly at the front axle) to offer more neutral handling for increased driving dynamics (compared with the 996GT2 40/60 chassis set up)
In effect this LSD does squat even when rebuilt with CUP plates, I am told it is because of the low pre load specified.
The 7GT2 doesn't seem to really need an LSD as the electroncs, the TC, PSM and associated systems work well together but seem to interject before the LSD can come into play.
Because my road car has a wide 900NM torque band the factory diff was showing signs of wanting to explode so our friend Matt specced me up his proprietary 40/60 billet unit as seen below which was teamed up with the billet side plate.
The most noticeable thing about having this unit is (as has been said in previous posts) there is a definite tendency to understeer, the front end simply doesn't turn in as easily as before particularly off throttle.
In terms of the LSD providing traction ? Well the electronics seem to cut in pretty fast and whilst the LSD is probably helping as part of the system the electronics dominate proceedings when trying to deploy some serious grunt through two wheels.
I'm sure with the TC switched off the LSD would come into its own but unfortunately/fortunately I have a program glitch which has disabled the TC off switch and I haven't been in a hurry to get it fixed.
I presume the discussion about LSD actions on 997GT3.2s are talking about running with all electronic nannies off ?
The 997GT2 runs the 28/40 set up which Porsche state is designed (with the other systems and chassis configuration particularly at the front axle) to offer more neutral handling for increased driving dynamics (compared with the 996GT2 40/60 chassis set up)
In effect this LSD does squat even when rebuilt with CUP plates, I am told it is because of the low pre load specified.
The 7GT2 doesn't seem to really need an LSD as the electroncs, the TC, PSM and associated systems work well together but seem to interject before the LSD can come into play.
Because my road car has a wide 900NM torque band the factory diff was showing signs of wanting to explode so our friend Matt specced me up his proprietary 40/60 billet unit as seen below which was teamed up with the billet side plate.
The most noticeable thing about having this unit is (as has been said in previous posts) there is a definite tendency to understeer, the front end simply doesn't turn in as easily as before particularly off throttle.
In terms of the LSD providing traction ? Well the electronics seem to cut in pretty fast and whilst the LSD is probably helping as part of the system the electronics dominate proceedings when trying to deploy some serious grunt through two wheels.
I'm sure with the TC switched off the LSD would come into its own but unfortunately/fortunately I have a program glitch which has disabled the TC off switch and I haven't been in a hurry to get it fixed.
I presume the discussion about LSD actions on 997GT3.2s are talking about running with all electronic nannies off ?
Edited by TB993tt on Wednesday 7th January 16:34
TB993tt said:
I presume the discussion about LSD actions on 997GT3.2s are talking about running with all electronic nannies off ?
Which you have to do to prevent the car from utterly obliterating its brakes. With the standard diff and everything off it gets a bit frisky under braking, but under power it isn't _that_ bad IMHO. Edited by DiscoColin on Wednesday 7th January 19:03
This is an excellent thread. I think that it demonstrates the problems that Porsche engineers face and how they try to resolve them. From an ultimate performance perspective the 996 GT3 diff set up under braking is far better than the 997. It's a lot more stable and ultimately from my experience more predictable. But it requires a reasonably difficult driving technique to overcome the understeer that it creates and most drivers make that worse by turning in off of the brakes or worse still, on the throttle. It looks like to counter this on the next model (997) Porsche reduced the diff lock on braking and added some rear braking assistance that was programmed to help stabilise the car under braking but enable it to rotate.
If I'm right, it would explain why when the 997 is driven in the classic 911 trail brake technique it does not reward or perform as the 996 does. In other words, the 996 is set up for ultimate performance but only when driven properly - the same as the cup car, but the 997 set up for the most commonly used driving technique. you could say that it's intentionally dummed down. That would explain why the PASM is so horrible on a trail - it's not designed to be driven on a trail. The primary assumption being that the majority of drivers wont drive it that way.
Neil, I'd be interested to hear your thoughts on this?
If I'm right, it would explain why when the 997 is driven in the classic 911 trail brake technique it does not reward or perform as the 996 does. In other words, the 996 is set up for ultimate performance but only when driven properly - the same as the cup car, but the 997 set up for the most commonly used driving technique. you could say that it's intentionally dummed down. That would explain why the PASM is so horrible on a trail - it's not designed to be driven on a trail. The primary assumption being that the majority of drivers wont drive it that way.
Neil, I'd be interested to hear your thoughts on this?
Info taken from this site :
996/997 GT3 Limited Slip Differentials
A few observations on limited slip differentials
The Porsche 996/997 GT3 street car is a marvelous product for street driving enthusiasts who also enjoy weekend track events and or occasional club races. A great chassis coupled with plenty of power makes for a great driving experience. However, all of that horsepower doesn't matter if you can't get the power down. The power from these cars when accelerating off corners causes the inside, or unloaded tire to loose traction; this slows the rate of acceleration, causes unwanted corner exit oversteer, and kills the rear tires. The problem only got worse when the newer 997 GT3 was made available. An even stiffer chassis platform and more power put more demand on the already limited-limited slip.
In the past, we used torsen or torque-sensing differentials, made by various manufacturers, Gleason and Quaiffe being most familiar to correct the loss of traction. These differentials seemed to be the end-all. After all what could possibly be better than a unit that actually sensed the torque, and applied it opposite to the wheel that slipped? The key with torque-biasing differentials is that the wheels must remain in contact with the pavement; even a minuscule amount of traction is needed to engage the diff. Early 911's flexed enough and kept the rear wheels on the ground maintaining the miniscule amount of traction needed (just look at the inside front tires off the ground in older 911s when racing). When the 944 turbo came around, with its stiffer chassis, cars would lift a rear tire in tight or bumpy corners and the torque sensing diffs were no good. We learned this first hand when we built the first Mini Cooper S's for Grand Am. Same problem with front wheel drive, when one drive wheel looses contact with the track, say over a curb or bump, all the torque went to that wheel.
Moving onto the GT3's and their limited slip differentials. A limited slip is a clutch type differential that, with stiff chassis 996/997 based cars, over curbs and in corners, is designed to apply the same force to both rear wheels. Porsche uses a four clutch disc pack in all GT3's. If you jack up one side of a GT3, put a torque wrench on one rear wheel axle nut, you may find it takes less than 20 pounds feet of torque to turn one rear wheel. That's called the breakaway torque. And, that's on a brand new 997. We have checked the differential breakaway torque on two dozen 996/997 GT3 street cars and after one or more track weekends the preload torque is at or near zero. Keep in mind, there are ramps inside the differential that, when drive, or accelerating force is applied, it forces the differential clutches to bind up and act to lock up the rear. However, little or no lock up occurs on deceleration. So, on hard braking, as the weight goes to the front wheels the rears get light, and the inside rear wheel will temporarily lose traction. This causes a slower entry into the corner and generally you will feel the anti-lock brakes engage when this occurs. But what if we could get the differential, to partially engage on deceleration and engage more aggressively the harder we accelerate off a corner, this would be having your cake and eating it too, wouldn't it? You bet it would!
The expensive solution was discovered by all of the long distance enduro GT3 Cup cars years ago. Install a billet, fully adjustable Guard limited slip differential. This nearly indestructible unit is also available for the street cars. Keep in mind, this solution requires removal and disassembly of the entire transaxle. However, the cost of the unit and labor can exceed $7500. This is a bit of over kill for street, track days and even club racing. The expense of the billet housing is way beyond most club race GT3 budgets.
However, using the same technology and top quality parts we can improve your differential beyond its original ability, and make it last for many track days. We can remove your differential (without removing the entire transaxle) and build the same adjustable ramp diff as the pro's have. The adjustable ramps in this unit allow for 40% lock on acceleration and 60% on deceleration. For more aggressive situations the same diff can be set to 50% acceleration and 80% deceleration. The friction plates in the new diff are far stronger than the Porsche factory parts and won't show near the fade. The break away torque is set to 80 pounds feet new, and after a few track miles they settle in at around 60-80 pounds.
996/997 GT3 Limited Slip Differentials
A few observations on limited slip differentials
The Porsche 996/997 GT3 street car is a marvelous product for street driving enthusiasts who also enjoy weekend track events and or occasional club races. A great chassis coupled with plenty of power makes for a great driving experience. However, all of that horsepower doesn't matter if you can't get the power down. The power from these cars when accelerating off corners causes the inside, or unloaded tire to loose traction; this slows the rate of acceleration, causes unwanted corner exit oversteer, and kills the rear tires. The problem only got worse when the newer 997 GT3 was made available. An even stiffer chassis platform and more power put more demand on the already limited-limited slip.
In the past, we used torsen or torque-sensing differentials, made by various manufacturers, Gleason and Quaiffe being most familiar to correct the loss of traction. These differentials seemed to be the end-all. After all what could possibly be better than a unit that actually sensed the torque, and applied it opposite to the wheel that slipped? The key with torque-biasing differentials is that the wheels must remain in contact with the pavement; even a minuscule amount of traction is needed to engage the diff. Early 911's flexed enough and kept the rear wheels on the ground maintaining the miniscule amount of traction needed (just look at the inside front tires off the ground in older 911s when racing). When the 944 turbo came around, with its stiffer chassis, cars would lift a rear tire in tight or bumpy corners and the torque sensing diffs were no good. We learned this first hand when we built the first Mini Cooper S's for Grand Am. Same problem with front wheel drive, when one drive wheel looses contact with the track, say over a curb or bump, all the torque went to that wheel.
Moving onto the GT3's and their limited slip differentials. A limited slip is a clutch type differential that, with stiff chassis 996/997 based cars, over curbs and in corners, is designed to apply the same force to both rear wheels. Porsche uses a four clutch disc pack in all GT3's. If you jack up one side of a GT3, put a torque wrench on one rear wheel axle nut, you may find it takes less than 20 pounds feet of torque to turn one rear wheel. That's called the breakaway torque. And, that's on a brand new 997. We have checked the differential breakaway torque on two dozen 996/997 GT3 street cars and after one or more track weekends the preload torque is at or near zero. Keep in mind, there are ramps inside the differential that, when drive, or accelerating force is applied, it forces the differential clutches to bind up and act to lock up the rear. However, little or no lock up occurs on deceleration. So, on hard braking, as the weight goes to the front wheels the rears get light, and the inside rear wheel will temporarily lose traction. This causes a slower entry into the corner and generally you will feel the anti-lock brakes engage when this occurs. But what if we could get the differential, to partially engage on deceleration and engage more aggressively the harder we accelerate off a corner, this would be having your cake and eating it too, wouldn't it? You bet it would!
The expensive solution was discovered by all of the long distance enduro GT3 Cup cars years ago. Install a billet, fully adjustable Guard limited slip differential. This nearly indestructible unit is also available for the street cars. Keep in mind, this solution requires removal and disassembly of the entire transaxle. However, the cost of the unit and labor can exceed $7500. This is a bit of over kill for street, track days and even club racing. The expense of the billet housing is way beyond most club race GT3 budgets.
However, using the same technology and top quality parts we can improve your differential beyond its original ability, and make it last for many track days. We can remove your differential (without removing the entire transaxle) and build the same adjustable ramp diff as the pro's have. The adjustable ramps in this unit allow for 40% lock on acceleration and 60% on deceleration. For more aggressive situations the same diff can be set to 50% acceleration and 80% deceleration. The friction plates in the new diff are far stronger than the Porsche factory parts and won't show near the fade. The break away torque is set to 80 pounds feet new, and after a few track miles they settle in at around 60-80 pounds.
Sorry, missed the URL - http://www.bodymotion.com/GT3_Differentials.asp
I have no idea if the information is 'correct' or not, just thought I would post it here for information
I have no idea if the information is 'correct' or not, just thought I would post it here for information
There is some truth in it but as you say, exit traction isn't a huge problem in 911's - old or new. Most LSD applications are mainly relevant under power on exit. But the 911 diff - certainly the water cooled cars - can be a key aid under braking and into the apex. I rely heavily on a diff set up to accommodate my driving style which I believe is quick and effective. Not sure about the multi diff set ups for pro's though. Most that I know that drive 911's are happy with a single set up.
There is an interesting description of a worn diff on a 997 under braking though and it's pretty accurate.
Either way, my view is that the 997 diff needs a plate upgrade and a 60:40 set up with a fair bit of pre load.
There is an interesting description of a worn diff on a 997 under braking though and it's pretty accurate.
Either way, my view is that the 997 diff needs a plate upgrade and a 60:40 set up with a fair bit of pre load.
Steve Rance said:
This is an excellent thread. I think that it demonstrates the problems that Porsche engineers face and how they try to resolve them. From an ultimate performance perspective the 996 GT3 diff set up under braking is far better than the 997. It's a lot more stable and ultimately from my experience more predictable. But it requires a reasonably difficult driving technique to overcome the understeer that it creates and most drivers make that worse by turning in off of the brakes or worse still, on the throttle. It looks like to counter this on the next model (997) Porsche reduced the diff lock on braking and added some rear braking assistance that was programmed to help stabilise the car under braking but enable it to rotate.
If I'm right, it would explain why when the 997 is driven in the classic 911 trail brake technique it does not reward or perform as the 996 does. In other words, the 996 is set up for ultimate performance but only when driven properly - the same as the cup car, but the 997 set up for the most commonly used driving technique. you could say that it's intentionally dummed down. That would explain why the PASM is so horrible on a trail - it's not designed to be driven on a trail. The primary assumption being that the majority of drivers wont drive it that way.
Neil, I'd be interested to hear your thoughts on this?
Steve,If I'm right, it would explain why when the 997 is driven in the classic 911 trail brake technique it does not reward or perform as the 996 does. In other words, the 996 is set up for ultimate performance but only when driven properly - the same as the cup car, but the 997 set up for the most commonly used driving technique. you could say that it's intentionally dummed down. That would explain why the PASM is so horrible on a trail - it's not designed to be driven on a trail. The primary assumption being that the majority of drivers wont drive it that way.
Neil, I'd be interested to hear your thoughts on this?
Apologies for not getting on to this until now. It was a great post and I hope I didn't cause things to fall away too much by being tardy?
Leading on from your first point a little, its also interesting in the context of what was happening with the cars as the appeal was widened. There has been a continuous and definite move with each iteration of GT3 to broaden its usability. This approach, coupled to the increase in ultimate performance that each iteration carries often does a good job of covering up some of what is actually lost by the process.
On one level its hard to argue against it, clipping some of the peaks in exchange for access to more of whats left, more of the time is a positive in many respects.
However you then get into something like a 997 cup and feel just what is left on the table with respect to the platform and its hard not to lament this approach. The 997.1 has it slightly tough in that regard as the performance jump from the previous iteration was actually quite small and as a result there is less camouflage there.
With specific reference to the diff. Its an interesting one also.
A diff with a high deceleration lock up % acts as a yaw damper on corner entry promoting understeer HOWEVER If you can get that diff lock up to really bite then you can use that same understeer inducing, yaw dampening diff to actually promote entry oversteer. The rear remains locked and the inside rear wheel breaks away.
The 997 therefore responds less well to aggressive trail braking IMHO simply because there is less benefit to be obtained from it. The 40% deceleration side isn't really enough to provide that rotation that you experience with the 60% diff going in hot with the axle locked up. Trail braking is still something that benefits, especially in 2nd gear corners and some 3rd gear ones where it makes all the difference to rotation, but there isn't that extra level of reward available from hitting it really hard (as I am sure you do).
The 996 by contrast equally benefits being trail braked in the same corners as with the 997 but can also offer that extra bit of rotation when worked aggressively as a result of its deceleration ramp offering 60% lock.
This situation is magnified of course on the Cup cars because the preload is so high. Driving one of the guard 50/80 diffs combined with a large preload one needs to really lean on the diff lock rotation effect to get the car turned at all.
The whole diff decision in the 997.1 was poorly judged by Porsche IMHO but thats for another post. I found your PASM comment interesting. I sometimes felt that it modulated in an odd way when worked heavily (and this would include heavy trail brake wrt to the front dampers) but I know many don't feel it. I am certainly not the most sensitive nor the fastest thats for sure so its certainly not that playing in. Its impossible to say whether its a limitation of system, design/testing, or a figment of imagination all together.
Edited by fioran0 on Wednesday 14th January 00:18
Steve Rance said:
There is some truth in it but as you say, exit traction isn't a huge problem in 911's - old or new. Most LSD applications are mainly relevant under power on exit. But the 911 diff - certainly the water cooled cars - can be a key aid under braking and into the apex. .
This +1The diff working to create rear end stability during braking is crucial on the 911. If you don't have this it gets incredibly twitchy back there as the rear tries to swap ends with the front. In fact the first indication that you diff has had enough is this sensation under braking. You almost never notice it under acceleration due to the weight advantage (and in my experience it will be throttle steering where you can tell rather than wheel spinning stuff).
I always use the rear wing on the 997 cup at maximum angle for rear end braking stability too (amongst other reasons). With the wing flat I can feel the reduction in stability (even with the diff) when braking really heavily from high speed. My experiment in increasing Vmax was a short lived one
Theres a second conversation in here somewhere about the 997.1 given its 40% deceleration lock up.
Edited by fioran0 on Wednesday 14th January 01:07
Just to pop back in to this.
I was rummaging around and found some hand over notes from one of the first 996 Cups I bought. In the notes for setup from the team that had previously owned it it listed:
For general oversteer, soften rear bar.
For general understeer, trail brake more.
Hardly ground breaking stuff in terms of secret sharing but it was amusing seeing it after all these years in the context of this thread.
I was rummaging around and found some hand over notes from one of the first 996 Cups I bought. In the notes for setup from the team that had previously owned it it listed:
For general oversteer, soften rear bar.
For general understeer, trail brake more.
Hardly ground breaking stuff in terms of secret sharing but it was amusing seeing it after all these years in the context of this thread.
Gassing Station | 911/Carrera GT | Top of Page | What's New | My Stuff