In the last installment we deduced that there was likely a mechanical problem with T-Rex. The GPS data acquired at the track indicated reduced grip in many of the medium- and high-speed corners. Also, on the odd occasion during the initial phase of extreme braking at the track the car became very unstable in the rear, providing the driver an unwelcome wake-up call. Fortunately, this provided an additional clue to the mystery of lackluster lap times.
With the suspicion that the OEM limited-slip differential (LSD) had worn out on my ’10 GT3 with some 5,500 miles on the clock, I started the physical investigation of the problem in my garage. By jacking a rear wheel off the ground with the other wheels blocked I was able to perform what is affectionately know as the Spin Test. With the e-brake released and the transmission in Neutral, you try to spin the wheel that’s off the ground. If the preload on the LSD is reasonable, the wheel will not spin at all, indicating that the diff is likely in good working order. However, if the wheel spins more or less freely (taking in to account the fact that the brake pads will be dragging slightly on the rotor), then the LSD could be worn out, or the preload very weak. Both rear wheels spun rather freely, indicating that this may be the problem.
I’ve been told by a couple of sources (my Porsche dealer being one) that the Spin Test is not accurate because the factory static load breakaway (preload) torque is very low, some 9 to 11 ft-lb on a new 997 GT3. A more precise estimate of the preload can be obtained with the rear wheels off and using a range-proper torque wrench, but there are still problems. Regardless of the measurement or how it’s done, since the LSD is not actually under load at the time when such test measurements are taken, any measurement could be irrelevant. Also note that the Porsche shop manual procedure to check the LSD preload is done with the unit removed from the transaxle (i.e., a bench test). So what does a poor guy do?
I took it to the dealer to get an opinion, providing them with my experiences and findings. Their testing was done with the LSD in the car and resulted in a preload torque measurement of 8 ft-lb; I was told that Porsche would not replace a unit unless it was below their 2 ft-lb minimum specification. I was also told that I’d need to pay up for the labor costs if I wanted a bench test. If the LSD unit measured below the 2 ft-lb minimum, Porsche would replace the unit and cover the labor cost. Otherwise I was on the hook for all labor regardless, and parts should I want a new diff installed. I’m not a gambling man, so I declined their kind offer.
After bit of web research, I decided that a fresh, OEM LSD was not up to the task of enthusiastic tracking because its friction discs are made from brassgood for longevity, but rather poor for friction and bite (it would need to be replaced again in 5,500 miles, or less). A stronger and more reliable solution was definitely in order. There are many ways to go if one desires a completely new aftermarket LSD unit, but installation requires the transmission to be removed for proper installation and setup, a very expensive proposition. The solution I found to be the simplest, at a reasonable price, was having the OEM unit rebuilt and upgraded to a stronger specification. This method saves quite a bit of shop time because the LSD can be removed with the transmission in the car, and no special setup is needed when reinstalling because the OEM diff will fit perfectly. Even better, if you’re so inclined, you can remove and replace the LSD in the comfort of your own garage and save a little more.
I sent my unit to Matt Monson at Guard Transmission in Boulder, Colo., for a complete rebuild and upgrade. He offers a few rebuild options, all with improved plain plates and friction discs (clutch packs), and the choice of retaining the OEM ramps 28/40, Guard 40/60 ramps, or Guard 50/80 ramps (the most aggressive setup). I opted for the rebuild with Guard 40/60 ramps as Monson recommended this as an excellent street and track solution. FYI: The cost for the Guard LSD clutch pack rebuild with upgraded moly-coated/heat-treated friction disks was $950, plus a further $650 to upgrade to Guard ramps, and $100 ring gear removal and installation if you aren’t willing to mess with this yourself. Note, when the first Porsche GT3 hit North America by storm in 2004/5, marking the climax of production for the 996 series, it came equipped with a 40/60 limited-slip differential, so interesting that this specification was not used on the 997 GT3. I did learn that the steeper 28/40 ramps were used so the traction and stability control systems can interact more transparently and to reduce the car’s inherent understeer.
Prior to tearing down my OEM LSD Monson told me that my preload was not even measurable, but the plot thickened once the unit opened up. Your car has obviously been driven the way Ferry intended it to be driven, was his comment, and noted that my plates were blue due to heat, something he has only seen on GT2s. He also observed that some of the plates and friction discs showed little or no wear on the inside edges, but noticeable wear on the outer edges, indicating that the factory plates were not flat and true when new. Again, interesting.
With the freshly rebuilt Guard LSD installed and the transaxle filled with fresh Mobil 1 Delvac synthetic gear oil this project is again ready for action. Note that this gear oil contains no friction modifiers, and that after 1,000 miles on the street or a full day at the track you should change the oil, as this will flush the transaxle of any particulate matter created while the LSD friction discs get settled in. The life of the LSD will depend on how much track time the vehicle endures, but a combination of street and DE track days will very likely see the LSD lasting until the transaxle is rebuilt. My first impression after driving the car for the first time with the rebuilt diff was enlightening. I was able to feel the LSD truly working in the corners; a more than subtle feeling of stiffness and grip in the rear. Clearly, my factory LSD was completely shot. The new unit works quietly too, but I did notice that the GT3 exhibited a touch more understeer in low-speed corners with the more aggressive LSD. Perhaps some sway bar adjustments will be in order to tune the handling when we get back to the track in the spring.
In the next installment we’ll attempt to lose some weight, add some power, and improve our handling, all in one giant leap. Any guesses?
One additional LSD wear test
I performed one final test before making my decision. This was blocking the front wheels and lifting both rear wheels off the ground. Then, with the car in Neutral and the e-brake released, I hand-turned one of the rear wheels. The other rear wheel stayed stationary, likely indicating a very weak preload. After all of this, I was pretty sure my LSD was pooched. A strong, working LSD allows the other wheel to turn in the same direction (a regular open diff will turn the wheel in the opposite direction).
R&R GT3 LSD
With the GT3 safely on a lift or axle stands, remove the rear-most undertray to access the transaxle and driveshafts. Drain the transaxle gear oil, and remove the T-50 bolts from the inner driveshaft to the differential output flanges. Loosen and remove the central flange bolts and carefully pull out the flanges. Next, remove all 12 bolts from the left side of the transaxle differential casing and then gently, with a wooden peg or hammer end, tap out the differential cover. With the left-hand driveshaft out of the way (and being mindful not to damage the coolant hoses above), gently and carefully, with a firm, clean grip (fresh and dry latex gloves), remove the 30-pound LSD unit from the transaxle. You will have to be very patient and try several orientations to get it out. Reinstallation is the reverse, but lubricate the end bearings first and pay special attention when meshing the ring and pinion gears and inserting the output flangeseverything should rotate smoothly upon testing by hand with the transmission in Neutral. Torque up as per the manufacturer specs. Work time is approximately 2.5 hours for removal and about the same for install. Take your time and have fun!
Properly packing your LSD for shipping
Be sure to protect your valuable and very heavy LSD when shipping. Believe me when I say you don’t want to have to buy another unit if yours gets damaged en route. Pack it twice: first into a small box with plenty of bubble wrap and paper, and then into a second, larger box stuffed with paper on all sides. This will prevent it from settling and rolling about in transit. The sucker weighs 30 pounds and will move around if you don’t pack it correctly. I speak from experience; fortunately, I was lucky. Now, read this paragraph again!