G28.03 differential set-up
11-30-2021, 06:07 PM
#1
Rennlist Member
Thread Starter
G28.03 differential set-up
I'm setting up the differential bearings as part of a complete rebuild of a G28.03, and am a little surprised by things. I am using all the factory tools as described in the WSM and following the procedure, with one exception. Which is that I am setting S-total using turning torque on the bearings rather than the WSM's measure-the-gap method. I do this, for all transmission types, because the measure-the-gap method is horribly imprecise. I shoot for preload that puts the turning torque in the range recommended for the same bearings back in the type 915 WSM, and with the *** bearings I am using it is 3.9 Nm. *** spec is 3.5 - 6.5 Nm.
So one strange thing with this transmission is the change to S-total required. As received, it was S1 = 0.50 mm, S2 = 0.90 mm, and so S-total = 1.40 mm. The S-total I've arrived at with preload is 0.40 mm, a drop of a whole millimeter! Quite surprising that there might be that much fatigue in the diff carrier - best guess as to why the change. I've seen big changes like this with for example early magnesium 911 cases, but not in a 928 aluminum case. No signs that "someone else was in there" but it's possible.
The other strangeness is that setting S1 to zero (no shim) and S2 to 0.40 mm, I am at the high range of free play for the Hurth ring-and-pinion set, measuring 0.25 to 0.28 mm around the ring gear. Spec for Hurth is 0.20 to 0.25 mm. To get it lower, I'd have to put a shim under the S1-side bearing inner race, quite possible using for example using a 0.25 mm shim from the 911-series boxes. But weird! Or I could increase S2, decreasing bearing preload.
Am I possibly missing anything? Anyone seen this sort of thing? Cheers,
So one strange thing with this transmission is the change to S-total required. As received, it was S1 = 0.50 mm, S2 = 0.90 mm, and so S-total = 1.40 mm. The S-total I've arrived at with preload is 0.40 mm, a drop of a whole millimeter! Quite surprising that there might be that much fatigue in the diff carrier - best guess as to why the change. I've seen big changes like this with for example early magnesium 911 cases, but not in a 928 aluminum case. No signs that "someone else was in there" but it's possible.
The other strangeness is that setting S1 to zero (no shim) and S2 to 0.40 mm, I am at the high range of free play for the Hurth ring-and-pinion set, measuring 0.25 to 0.28 mm around the ring gear. Spec for Hurth is 0.20 to 0.25 mm. To get it lower, I'd have to put a shim under the S1-side bearing inner race, quite possible using for example using a 0.25 mm shim from the 911-series boxes. But weird! Or I could increase S2, decreasing bearing preload.
Am I possibly missing anything? Anyone seen this sort of thing? Cheers,
11-30-2021, 09:37 PM
#2
Addict
Rennlist Member
Rennlist Member
Join Date: Oct 2003
Location: Gone. On the Open Road
Posts: 16,523
Received 1,665 Likes
on
1,080 Posts
Good questions. I can't help you but am interested in any answers you receive.
I'm also very interested in where/how you acquired the WSM tools for the transaxle. Should I assume that they are not recent acquisitions?
I'm also very interested in where/how you acquired the WSM tools for the transaxle. Should I assume that they are not recent acquisitions?
11-30-2021, 11:06 PM
#3
Former Vendor
Dave:
Not sure what tools you are looking for, but I picked up an extra set of late pinion nut loosening and tightening tools.
9218 and 9219.
Not sure what tools you are looking for, but I picked up an extra set of late pinion nut loosening and tightening tools.
9218 and 9219.
12-01-2021, 12:44 AM
#4
Rennlist Member
Thread Starter
There are Porsche-only tools that you have to get through your friendly neighborhood Porsche parts department. Anything beginning with a P or with a 9xxx identifier is likely to fall into this category, eg, the 9218 and 9219 tools Greg mentions. Among these, I picked up the G28 tools a few years ago, first time a customer brought me a G28 to service. I was surprised to learn that a) they were available from Porsche, and b) compared with other Porsche special transmission tools, they were a bargain. Oh, and several of them are painted fire engine red!
I hope this is helpful.
12-01-2021, 01:17 AM
#5
Archive Gatekeeper
Rennlist Member
Rennlist Member
I have exactly zero justification for needing a VW385 set, but that hasn't stopped me from looking at them on vw-snapon for a couple of years. The first time I checked multiple years ago the set was about $3800, now it's $5800.
https://vw.snapon.com/SpecialToolsDe...temId=34100108
https://vw.snapon.com/SpecialToolsDe...temId=34100108
The following users liked this post:
KevinGross (12-01-2021)
12-01-2021, 01:25 PM
#6
Former Vendor
I'm setting up the differential bearings as part of a complete rebuild of a G28.03, and am a little surprised by things. I am using all the factory tools as described in the WSM and following the procedure, with one exception. Which is that I am setting S-total using turning torque on the bearings rather than the WSM's measure-the-gap method. I do this, for all transmission types, because the measure-the-gap method is horribly imprecise. I shoot for preload that puts the turning torque in the range recommended for the same bearings back in the type 915 WSM, and with the *** bearings I am using it is 3.9 Nm. *** spec is 3.5 - 6.5 Nm.
So one strange thing with this transmission is the change to S-total required. As received, it was S1 = 0.50 mm, S2 = 0.90 mm, and so S-total = 1.40 mm. The S-total I've arrived at with preload is 0.40 mm, a drop of a whole millimeter! Quite surprising that there might be that much fatigue in the diff carrier - best guess as to why the change. I've seen big changes like this with for example early magnesium 911 cases, but not in a 928 aluminum case. No signs that "someone else was in there" but it's possible.
The other strangeness is that setting S1 to zero (no shim) and S2 to 0.40 mm, I am at the high range of free play for the Hurth ring-and-pinion set, measuring 0.25 to 0.28 mm around the ring gear. Spec for Hurth is 0.20 to 0.25 mm. To get it lower, I'd have to put a shim under the S1-side bearing inner race, quite possible using for example using a 0.25 mm shim from the 911-series boxes. But weird! Or I could increase S2, decreasing bearing preload.
Am I possibly missing anything? Anyone seen this sort of thing? Cheers,
So one strange thing with this transmission is the change to S-total required. As received, it was S1 = 0.50 mm, S2 = 0.90 mm, and so S-total = 1.40 mm. The S-total I've arrived at with preload is 0.40 mm, a drop of a whole millimeter! Quite surprising that there might be that much fatigue in the diff carrier - best guess as to why the change. I've seen big changes like this with for example early magnesium 911 cases, but not in a 928 aluminum case. No signs that "someone else was in there" but it's possible.
The other strangeness is that setting S1 to zero (no shim) and S2 to 0.40 mm, I am at the high range of free play for the Hurth ring-and-pinion set, measuring 0.25 to 0.28 mm around the ring gear. Spec for Hurth is 0.20 to 0.25 mm. To get it lower, I'd have to put a shim under the S1-side bearing inner race, quite possible using for example using a 0.25 mm shim from the 911-series boxes. But weird! Or I could increase S2, decreasing bearing preload.
Am I possibly missing anything? Anyone seen this sort of thing? Cheers,
It's very common for the bearing race on the ring gear side of the differential to spin, eroding the seating surface. This is especially an issue with the heavier limited slips, but also commonly occurs on non limited slip carriers..
The resulting loss in dimension results in what you are describing.
With all of your apparent transmission experience, I'm a bit shocked at your generic preload technique.
I'd sincerely recommend that you review your dimensional conclusions.
Last edited by GregBBRD; 12-01-2021 at 01:40 PM.
12-01-2021, 03:52 PM
#7
Rennlist Member
Thread Starter
New differential bearings or old bearings?
It's very common for the bearing race on the ring gear side of the differential to spin, eroding the seating surface. This is especially an issue with the heavier limited slips, but also commonly occurs on non limited slip carriers..
The resulting loss in dimension results in what you are describing.
With all of your apparent transmission experience, I'm a bit shocked at your generic preload technique.
I'd sincerely recommend that you review your dimensional conclusions.
It's very common for the bearing race on the ring gear side of the differential to spin, eroding the seating surface. This is especially an issue with the heavier limited slips, but also commonly occurs on non limited slip carriers..
The resulting loss in dimension results in what you are describing.
With all of your apparent transmission experience, I'm a bit shocked at your generic preload technique.
I'd sincerely recommend that you review your dimensional conclusions.
I don't understand what you mean by generic preload technique. The bearing manufacturers quote turning torque spec's in their literature (INA/***, SKF, for example) and Porsche really just shares those figures. I believe an *** 32010 bearing is the same part regardless of the transmission it's installed in, and don't know why its preload should vary from G28 to 915 to 930 to G87 to... you name it. How do you preload bearings, what technique do you use? I'm always pleased to learn something new! Cheers,
Trending Topics
12-01-2021, 04:24 PM
#8
Former Vendor
Greg, thanks for your reply. New bearings. No sign of spinning of the outer races in their carriers but I'll recheck. They came out and went in with proper interference. What you describes makes sense, and I toss a lot of diff carriers which have had races spun, for example, when there's been a pinion gear failure. These bearing carriers appear to be just fine.
I don't understand what you mean by generic preload technique. The bearing manufacturers quote turning torque spec's in their literature (INA/***, SKF, for example) and Porsche really just shares those figures. I believe an *** 32010 bearing is the same part regardless of the transmission it's installed in, and don't know why its preload should vary from G28 to 915 to 930 to G87 to... you name it. How do you preload bearings, what technique do you use? I'm always pleased to learn something new! Cheers,
I don't understand what you mean by generic preload technique. The bearing manufacturers quote turning torque spec's in their literature (INA/***, SKF, for example) and Porsche really just shares those figures. I believe an *** 32010 bearing is the same part regardless of the transmission it's installed in, and don't know why its preload should vary from G28 to 915 to 930 to G87 to... you name it. How do you preload bearings, what technique do you use? I'm always pleased to learn something new! Cheers,
Actually, the inner bearing race spins on the differential.
Loose bearings in the aluminum side carriers are rare, but an easy way to quickly check the carriers is to swap side to side (since the bearings are new, you can do this) and see if your backlash dimensions change.
In terms of preload, it's not the bearing manufacturers that determine the preload (tapered bearings can take insane amount of preload)...it's the transmission engineers.
The specifications are determined by considering the amount the bearings will "wear in" and how much the transmission case will change with heat. (Heat being the primary consideration.)
The preload specifications you are using is from an early magnesium case 915 transmission. These cases did not expand much (but they did distort the side cover, as we all know.) The preload was .15mm, resulting in the quoted spinning resistance you are using....apparently for all your transmissions.
The aluminum case 915 transmissions had a higher preload, because of the material (We've all seen, dozens of times, the "pounded" pinion bearings in the 915 aluminum transmission cases. This is a result of the pinion bearing actually being "loose" in the case, when it gets hot. Getting rid of the steel insert for the pinion and mainshaft bearing, found in the magnesium cases, was not a very good idea!)
The preload on the 928 differential is .30mm...twice that of the early 915. The "turning resistance" will be over double that of the 915 transmission. (And this is simply all about how much the transmission case expands with heat.)
By the time you get to the G-50 transmission, the preload is .40mm!
Now that almost every tidbit of information is on the Internet, I went and looked for a document about taper bearing preload. The first one I could was from Timken...and is quite good. I suggest any professional working with tapered bearings to read this document.
https://www.timken.com/wp-content/up...Brochure-1.pdf
Especially good is their method of determining preload by shimming until spinning resistance is present and then adding in (or subtracting in the 928 transmission) .009" of shim.
Once you have done this once, with the brand of bearing you are using, you can then measure the turning resistance of the differential and forever use that information to set-up your preload. (As long as you do not switch brands of bearings.)
Finally, a word of caution about backlash. The workshop manual (and the gear manufacturer) only provide a "new" specifications. The backlash changes, with wear (and with the removal of the nitrate break-in coating, obviously.) Resetting the backlash to "new' specification will almost always result in rapid pinion wear/failure. (Smaller headed pinions, like the 7 tooth 915 pinions, are very sensitive to this.)
A "good" number that I've always used for used ring and pinions is to add 10 to 20% to the new specifications....dependent upon the side of the pinion diameter.
Last edited by GregBBRD; 12-01-2021 at 04:37 PM.
The following 2 users liked this post by GregBBRD:
Kevin in Atlanta (12-01-2021),
Speedtoys (12-06-2021)
12-01-2021, 08:35 PM
#9
Rennlist Member
Thread Starter
Greg, thanks again for the detailed information. I've read through the Timken document, good information in there. The guidance with respect to end play, preload, and normal value ranges for them is particularly helpful. The discussion of probabilities is quite interesting, but more useful to the manufacturer than the guy servicing a transmission forty years later. Among their methods, I guess I am pretty clearly a Torque-Set guy!
The inner races on this G28 haven't spun. I pulled them using the VAG1582 set, normal interference felt in getting them off and pressing new ones on. (Love that tool, a life saver when dealing with inner races on the ring gear side of a lot of LSDs where they sit flush.)
I went back to this particular transmission, followed the WSM procedure, and at the end of the process ended up with the original S-total, S1, and S2 shims, which is comforting to a degree. I still find the measure-the-gap method imprecise, but I guess precision is a relative concept. The fact that the S-total remained the same as-received is consistent with no material loss due to spinning, also a comfort. Good backlash numbers with the original S1 and S2, to boot.
However, the one thing I am still a little stuck on is the fact that with the 0.30 mm preload, turning torque remains darned small. You mention that "The "turning resistance" will be over double that of the 915 transmission." The 915 WSM calls for 0.40 mm preload, both for the early (mag) and late (alu) cases unless I am overlooking something. Porsche did call for 0.15 mm for the 901 transmissions, which perhaps you're thinking of. Whether 0.30 mm preload gives you 3.0 - 6.5 Nm turning torque in a 915, I have no idea, I never found that technique helpful. I'm curious to know how much turning torque you typically see with 0.30 mm preload in a G28 - do tell! Maybe it doesn't have to be that high. As another data point, VW and Porsche called for 2.0 - 4.0 Nm turning torque with new bearings in the 944-series transmissions, which featured relatively robust aluminum cases not terribly different in material and strength from the G28.
Your insights are much appreciated! Cheers,
The inner races on this G28 haven't spun. I pulled them using the VAG1582 set, normal interference felt in getting them off and pressing new ones on. (Love that tool, a life saver when dealing with inner races on the ring gear side of a lot of LSDs where they sit flush.)
I went back to this particular transmission, followed the WSM procedure, and at the end of the process ended up with the original S-total, S1, and S2 shims, which is comforting to a degree. I still find the measure-the-gap method imprecise, but I guess precision is a relative concept. The fact that the S-total remained the same as-received is consistent with no material loss due to spinning, also a comfort. Good backlash numbers with the original S1 and S2, to boot.
However, the one thing I am still a little stuck on is the fact that with the 0.30 mm preload, turning torque remains darned small. You mention that "The "turning resistance" will be over double that of the 915 transmission." The 915 WSM calls for 0.40 mm preload, both for the early (mag) and late (alu) cases unless I am overlooking something. Porsche did call for 0.15 mm for the 901 transmissions, which perhaps you're thinking of. Whether 0.30 mm preload gives you 3.0 - 6.5 Nm turning torque in a 915, I have no idea, I never found that technique helpful. I'm curious to know how much turning torque you typically see with 0.30 mm preload in a G28 - do tell! Maybe it doesn't have to be that high. As another data point, VW and Porsche called for 2.0 - 4.0 Nm turning torque with new bearings in the 944-series transmissions, which featured relatively robust aluminum cases not terribly different in material and strength from the G28.
Your insights are much appreciated! Cheers,
Last edited by KevinGross; 12-01-2021 at 08:38 PM.
12-01-2021, 11:02 PM
#10
Former Vendor
I'm not sure how you could increase the shim thickness on a 928 transmission and get better preload and better backlash. Seems a bit "opposite, to me. I'll need to think about that.
Since it has been a minute or two since I did a magnesium case 915 transmission, you got me to go look at the WSM.
My 911 WSM, Volume IV, Group 3, Page 5.1-5/5 tells me the desired preload is .15mm. The next page tells me that this results in a turning torque of 40-65cmkp (for *** bearings.)
Regretfully, I'm not a "turning torque" guy. Don't measure it...never have.
However, I'm right in the process of building another 964 Tiptronic for Singer. (Yes, I'm sadistic enough to be able to rebuild and update these things (to hold the increased torque and horsepower of their bigger engines) in my sleep, which seems to be somewhat of a "lost" art, according to them.....I consider this to be a curse, these days, BTW.)
The 964 Tiptronic has a desired "press fit" of .24mm
Just for ****s and giggles, I'll measure the turning torque on that. (Albeit, the differential bearings on a Tip differential are larger than those on a 928 and the data may be useless.)
Since it has been a minute or two since I did a magnesium case 915 transmission, you got me to go look at the WSM.
My 911 WSM, Volume IV, Group 3, Page 5.1-5/5 tells me the desired preload is .15mm. The next page tells me that this results in a turning torque of 40-65cmkp (for *** bearings.)
Regretfully, I'm not a "turning torque" guy. Don't measure it...never have.
However, I'm right in the process of building another 964 Tiptronic for Singer. (Yes, I'm sadistic enough to be able to rebuild and update these things (to hold the increased torque and horsepower of their bigger engines) in my sleep, which seems to be somewhat of a "lost" art, according to them.....I consider this to be a curse, these days, BTW.)
The 964 Tiptronic has a desired "press fit" of .24mm
Just for ****s and giggles, I'll measure the turning torque on that. (Albeit, the differential bearings on a Tip differential are larger than those on a 928 and the data may be useless.)
12-03-2021, 09:33 PM
#11
Addict
Rennlist Member
Rennlist Member
Join Date: Oct 2003
Location: Gone. On the Open Road
Posts: 16,523
Received 1,665 Likes
on
1,080 Posts
Good discussion so far. I need to catch up on the last few days of posts.
I guess I need to make a list. But...
Basically, everything I need to take apart a manual box and put it back together. (Recall a discussion you and I had a couple of years ago.)
I guess I'll need to look again. A couple of years ago, I did look and spent hours with very little result.
Basically, everything I need to take apart a manual box and put it back together. (Recall a discussion you and I had a couple of years ago.)
12-03-2021, 10:28 PM
#12
Former Vendor
Good discussion so far. I need to catch up on the last few days of posts.
I guess I need to make a list. But...
Basically, everything I need to take apart a manual box and put it back together. (Recall a discussion you and I had a couple of years ago.)
I guess I'll need to look again. A couple of years ago, I did look and spent hours with very little result.
I guess I need to make a list. But...
Basically, everything I need to take apart a manual box and put it back together. (Recall a discussion you and I had a couple of years ago.)
I guess I'll need to look again. A couple of years ago, I did look and spent hours with very little result.
You need to be able to loosen and tighten the big nuts on the pinion shaft. Different tools required for early and late model. (The 9218 and the 9219 tools I bought do the '85-'95 transmissions)
You need a little bit more, if you need to set up a ring and pinion.....but not nearly what the people at Porsche claim that you need. (There's some serious "tricks" (only in the 928 transmissions) that one can use to even eliminate the need for a VW385 tool set (99% of the time....although I do rely on mine to do this task, since it is handy.)
I made up a few special tools that Porsche didn't supply in order to install the seals at the perfect depth and perfectly square. This can be faked, although oil seepage from the seals is common, if they are not perfectly square.
A few months ago, I was going to make up (I was going to have a CNC guy make up) a short run of these seal installers, but didn't get enough interest to even warrant a short run.
Last edited by GregBBRD; 12-03-2021 at 10:33 PM.
The following users liked this post:
Ducman82 (12-07-2021)
12-03-2021, 10:44 PM
#13
Archive Gatekeeper
Rennlist Member
Rennlist Member
Perhaps I missed the memo, I'd have gone for a set of trick trans tools- count me in if you change your mind. You don't have much else to be thinking about, right?
12-04-2021, 01:54 AM
#14
Addict
Rennlist Member
Rennlist Member
Join Date: Oct 2003
Location: Gone. On the Open Road
Posts: 16,523
Received 1,665 Likes
on
1,080 Posts
12-06-2021, 07:36 PM
#15
Rennlist Member
Thread Starter
Since it has been a minute or two since I did a magnesium case 915 transmission, you got me to go look at the WSM.
My 911 WSM, Volume IV, Group 3, Page 5.1-5/5 tells me the desired preload is .15mm. The next page tells me that this results in a turning torque of 40-65cmkp (for *** bearings.)
My 911 WSM, Volume IV, Group 3, Page 5.1-5/5 tells me the desired preload is .15mm. The next page tells me that this results in a turning torque of 40-65cmkp (for *** bearings.)
Regretfully, I'm not a "turning torque" guy. Don't measure it...never have.
However, I'm right in the process of building another 964 Tiptronic for Singer. (Yes, I'm sadistic enough to be able to rebuild and update these things (to hold the increased torque and horsepower of their bigger engines) in my sleep, which seems to be somewhat of a "lost" art, according to them.....I consider this to be a curse, these days, BTW.)
The 964 Tiptronic has a desired "press fit" of .24mm
Just for ****s and giggles, I'll measure the turning torque on that. (Albeit, the differential bearings on a Tip differential are larger than those on a 928 and the data may be useless.)
However, I'm right in the process of building another 964 Tiptronic for Singer. (Yes, I'm sadistic enough to be able to rebuild and update these things (to hold the increased torque and horsepower of their bigger engines) in my sleep, which seems to be somewhat of a "lost" art, according to them.....I consider this to be a curse, these days, BTW.)
The 964 Tiptronic has a desired "press fit" of .24mm
Just for ****s and giggles, I'll measure the turning torque on that. (Albeit, the differential bearings on a Tip differential are larger than those on a 928 and the data may be useless.)
In thinking about this over the weekend, I have a suspicion that the big difference between my turning-torque method results and the "dimensional" (feeler gauge) results with this 928 just might have something to do with taking measurements with the end cover removed. Because this is how the procedure is pictured in the WSM. I suspect that with the end cover installed, its two dowel pins and twelve studs/nuts give the case greater rigidity. Next one I do, will measure both methods and compare, with the end cover installed. Cheers,
Last edited by KevinGross; 12-06-2021 at 07:45 PM.