Ernie's diff....iculties

True hey?? Maybe you are correct there buddy. They have to be setup like that from the factory. Highly un-likely all 5 of our cars are worn,unless its a quick wearing item?? Normal LSD clutch pack/cone setups lasts for the life of the car. My ford is 23yrs old,all original and it has a nice tight LSD still. Just to let you guys know,my 348 has 23,000miles on it. And the wheel spun very freely when it was off the ground......

 

hmmm.. ok, can anyone find a 355 that has no static preload?

and can anyone find a 328 with no ststic preload?

The occurance of no 348 static preload is too many to be a fault, but it's wierd that there are some....

Maybe this would help explain some (but by no means all) of the loosesness and twitchyness of 348's compared to 355.

 

OK-I'm not so sure that with the car in neutral and one wheel off the ground is a valid test. Won't turning one wheel simply turn the diff as well, whereas a true test would involve only the axles?

 

That discussion is irrelevant unless we know the ramp angles of the cup.

If they are very shallow, the locking is higher on drive/over run, to compensate for the lack of static preload.

We know that 348's have a good locker... because they spin both wheels when told to, not just one.

 

Ernie is about as right on in his thinking here as anyone can get, he even has the flow of power nailed down. In a non lsd differential, the spider gears are pined to the carrier and the side or sun gears are simply free to rotate. In these diff's, and this is a basic differential, the pins for the spider gears are not pinned, but are loosly fitted with a sliding ramp on thier ends that corresponds with ramps in the carrier, (the round "case" all the parts fit inside of).

My understanding of having static force on the clutch pack, is that it would be attempting to ramp the spider pins as soon as any differential movement occured. Also, if even one 348 has load, 5 cars without load signals potential wear, if 308, 328, and 355 all have load, and even one 348 have load. So its obvious there should be some amount of load present. Just so everyone understand however, even without load, the lsd funtion would still occur under power, as the torque would still ramp out the spider gears into the sun gears and lock up the carrier.

As to the stacking of the clutch plates, in the previous diagram that was presented, its exactly simular to the other cars (308,328,etc..) in that the two friction plates are fitted between the two static plates (or cross plates as some have called them). I would assume great wear could occur if you were to assemble the carrier with a friction plate on the outside of a static plate, to wear against a shim washer or the carrier itself. If you are not getting the friction load you want, shim it outside of the static plates. I also believe each "stack" of plates should be equal thickness on either side, to keep the spider gears centered in the carrier and not force them right or left more than neccesary.

 

Seems to good to be true FBB Its got me?? I have a feeling that they came like that from the factory. No static pre-load at all. Maybe the 348 FF checked,as i said before,has a sticky brake caliper or over adjusted handbrake shoes on that wheel he spun?? You never know? 5 cars?? May all be worn...............im still waiting for a garage queen to chime in and let us know how they went with theirs

 

Static friction is requied to make the diff load in a wheel free spin condition. If one wheel is hydroplaning, bounced in the air, on ice or gravel, ect., the preload in the diff is what allows power transfer to wheel that has traction to continue. The clutch pack needs to be under load to BEFORE it can transfer any load, pre-load ensures there is adequate power transfer in all conditions. Under most conditions, no preload is required or desired, but when you need it, you need it and most people feel it's better to live with a little drag all the time than not have any traction some of the time.

 

Okay I just sent Rob (rexrcr) a pm asking him what the deal is, since he's the only guy I can think of right now that has been on a 348 challenge racing team.

 

ernie, you have it correct, strong work.

The 348 (355, F40...) LSD is a good one and tunable as you've discovered.

As stated by ferrarifixer, ramp angle is an order of magnitude more influential on LSD behavior than pre-load of the clutch discs.

For track use, I do recommend putting all the clutch disc surfaces into play by alternating them rather than having two of the same configuration in direct contact with each other.

Pre-load per previous recommendations. In fact, I've set up TransAm cars with zero pre-load (though virtually zero clearance in the clutch plate stack, but no pre-load) and tuned with ramp angles.

IIRC, the 348 is a 50/50 ramp angle, in that the on-throttle vs. off-throttle the forces transmitted via the ramps into the clutch stack (hence the force on the clutch stack, hence the amount of "lock" in the differential) is the same.

Downside of 50/50 ramps: potentially greater corner entry understeer. Upside of 50/50 ramps... hmmm, some cars will respond favorably to a little entry push.

For a hard-core racer it would be worth it to build a set of 90/10 ramps for reduction of inside rear tire drag inducing corner entry understeer.

Excellent thread.

Best,

Rob

 

Thanks for your response. Very interesting information. You would have to be a hardcore racer to bother building 90/10 cups wouldnt you?. For the street though,should be fine leaving the car as is.

 

Great thread here cause I had no idea how diffs worked. I just knew if you followed the reverse steps of taking it apart it went back together as it came apart.

Mine for the most part looked perfect according to a friend that does more hardcore rebuilds for local P-car racers so I trusted him and left it alone. Glad your learning Erine incase I ever have to take mine apart again.

 

ROCK ON Rob!!!!! Thanks a bunch, I knew it you would know.

Now it really makes a bunch of sence. I'm convinced that the diff's were set up loose at the factory to create push in the street cars.

I've been thinking about this some more, I know, I need to watch out how much I think could be dangerous, LOL. So if I'm understanding this correctly, With a greater ramp angle that reduces the distance, the retainers for the spider gears need to travel up the V-groove of the cup, in order to get the lock, and the speed at which it locks. With a 90/10 angle that would give you almost insant lock, as compaired to a 50/50. The retainers would be traveling up the grove less. At the same time it would also unlock slower, no? Because of the angle it would be more prone to stay in the V longer, thus maintaining the lock on the clutches a little bit longer, but then once it did unlock it would unlock quickly. It will take alot more force to go up the steaper ramp angle.

I was thinking of it like going down a play ground slide. You are going to go down a 90/deg slide alot faster than you would a slide at a 50/deg angle. Also it will be harder to climed back up a 90/deg slide compaire to a 50/deg slide.

So I guess shimming the plates, is an "easier", and "cheaper" way to keep the clutches engaged, as compaired to having new cups machined with the different ramp angles. With the shimming the spider gears travel less distance, but the force needed to unlock is still the same. Instead of the steaper angle of the V-groove in the cup doing the work to keep the clutches engaged, it's the clutches that are taking the abuse. Then by alternating the the clutches & plates, so that you are making contact with four surfaces instead of just two, and increasing the preload you are doubling the holding force of the clutches. You are acomplishing keeping the lock engaged longer, but at the expense of wear on the clutches.


So in the end I think the ramp angles are the best way to go if you want to make the diff lock up faster and longer. Because you are using mechanical force instead of frictional force to keep things locked.

Okay I think I need a beer now.

 

You got it right there i reckon ernie. Im enjoying a beer with ya buddy
Image Unavailable, Please Login

 

Thanks Rob, quite well explained, but perhaps I could throw another one in the mix...

You stated an alternative to the 348's 50/50 might be 90/10...

Now, When I deal with modified diffs, the numbers I use are the ramp angles, and they can be anything from about 30 to about 130, although they are extremes cases in either direction.

There is no requirement for ramp angles to be expressed as a factor of 100... ie, (without giving too much away ), perhaps an 80/50 or 40/70 is a good diff... and these can be configured either way around for your preference.

Frank Capo from Modena engineering here in Melbourne makes good diffs.

He makes a cup set that has a spare option of angles, and also uses round tipped spider shafts so that they can be flipped around any way you like.

So, one set of cups can give 4 different set ups, and he makes 3 different cups, so if you carry the other options you have 12 options of diff set up.

None of this clears up the query about 348 and preload... Rob, can you remember from your 348 racing whether 348's had static preload (ie, rotate same way when raised off the ground) or whether they were open while static (ie, wheels counter rotated when raised off the ground).

 

yes, preloaded.

 

Wouldn't have a spec would ya?

Was it just "factory", or was it an area for private development for you guys?

 

Definately stick with the OE specification and accept a higher value when placing additional clutch plate surface area into service (by alternating them as previously discussed).

One really only needs enough preload to overcome the "open diff" internal frictional tendency to counter-rotate one drive wheel vs. the other (hence the 40 ft-lb OE spec.).

BTW, the ZF design is very common in the racing world; Hewland calls it Power Flow, also known as a Salsbury differential. It's about the best of the best for a road car and racecar design with the exception of anything electronically controlled. Better than torque sensing helical gear designs (Torsen, etc.), better than a viscious coupling, better than an on/off locker like a Detroit Locker and way better than a spool.

Best,
Rob

 

Rob, would it be a fair test to raise the rear of a 348, hand turn one rear wheel while tranny is in neutral, and then to say that you have 40 ft/lbs or more of static "pre-load" if both rear wheels turn together, and likewise fair to say that you have less than 40 ft/lbs of static pre-load if the rear wheels counter-rotate?

 

nice photos ---nice work, bearing looks like its been spinning on the shaft.friction welding not nice .good timing .you would have kicked yourself if you hadnt checked out the rest and something went wrong that could have been avoided.or has it.put some new plates in if they are worn.set bearing preload and enjoy driving it.could you tell something else was wrong apart from the obvious bearing noise it must have had throttling on and off .or was it one of those annoying sounds that even that new stereo couldnt drown out .did you think you had an LSD problem at the time,i think not.and you thought that bearing was going to be a pain in the butt.can you give me a contact for the fingerless gloves. PS. put the diff in the freezer and then fit those bearings you will be surprised at how little effort it will take to fit.dont get oil on the turkey.IF YOU UNDO THOSE BOLTS AGAIN PLEASE REPLACE THEM.it will save you waking up in cold sweats--------Tony

 

Or oil on the lamb roast....

 

I'm rebuilding a 308 GT4 diff right now, and today I did some tests.

I was surprised at the difference in the static with this... (built with 50/50 grease/oil mix)

With all the plates stacked staggered, the static slip was 75 lb/ft, and with them stacked with abrasive plates touching it was 36 lb/ft. I'll try a half and half set up tomorrow, but I'd guess it'll be about 50-55 lb/ft.

 

Well that's interesting. I wonder what it will be with the half/half?