I've heard so many things about the proper pinion angle on a car and I want to know the truth. I would love to see some equations and numbers to explain it all, but any info would be helpful.
Pinion Angle
Started by
red300zx99
, Jan 09 2004 22:00
13 replies to this topic
Advertisement
#2
Greg Locock
-
- 6,494 posts
- Joined: March 03
Member
Posted 10 January 2004 - 09:15
Production or racing?
#3
red300zx99
-
- 328 posts
- Joined: February 03
Member
Posted 10 January 2004 - 10:38
Racing, max torque to the wheels. But just out of curiousity what would you say for production cars?
#4
McGuire
-
- 9,218 posts
- Joined: October 03
Member
Posted 10 January 2004 - 12:11
The short, simple answer: For a Hotchkiss drive (live axle - open drive, single Hooke joint on each end of the prop shaft) the optimum pinion angle is zero...or more precisely the pinion should be exactly parallel with the transmission output shaft in all planes. (Or better yet, dead in line.) However, since the axle may wind up under power, some static pre-angle is often desired to keep things straight when max torque is going through the drive. The actual angle required is entirely dependent upon the compliance built into the rear suspension...but typically in the range of 2 to 6 degrees (the latter for leaf springs). A rear suspension purpose-built for racing should require little or no static pre-angle.
#5
AdamLarnachJr
-
- 525 posts
- Joined: June 01
Member
Posted 10 January 2004 - 19:44
Just to reinforce it, McGuire is dead on. I had a problem on my car where the mounts had moved forward, tilting the transmission output shaft up. I fixed this with new mounts and "leg pressing" the engine back into position. 
Oh joy!
Oh joy!
#6
Greg Locock
-
- 6,494 posts
- Joined: March 03
Member
Posted 10 January 2004 - 22:39
1)For a production car the optimum is to have the pinion shaft of the diff parallel to the output shaft of the trans, for all driving conditions.
2)There is also an 'absolute' limit, of about four degrees, on the articulation of each Hookes joint.
3) Finally you need to make sure that the pinion flange does not hit the floor - typically that piece of floor is in the back seat, and you can't afford to put a lump there.
4) Finally there is typically a judder type issue in reverse if the pinion points down too far.
1) is not achievable due to the compliant mounting of the diff, as McG said, typically 2-4 degrees at full throttle in first and reverse, so in practice you pick the area where you have trouble and set that at parallel
2) we normally keep to
3) we achieve using a 4 arm rear suspension with different length upper and lower arms, although on our latest Hotchkiss we have introduced a Crane-Simplex linkage to get the same effect.
4) limits the reverse travel to about 4 degrees, it may even be why (2) is there.
So if 60 mph slight accel was your problem area then you'd set the system up in alignment at say 40% torque in 4th gear. Than your pitch compliance lets you work out where it is at rest. Then you check the other load cases and suspension travel (for a live axle). Then you start adding snubbers etc to suit.
Note that a 2 piece propshaft actually makes this worse, the important length is smaller, so the angles get bigger.
2)There is also an 'absolute' limit, of about four degrees, on the articulation of each Hookes joint.
3) Finally you need to make sure that the pinion flange does not hit the floor - typically that piece of floor is in the back seat, and you can't afford to put a lump there.
4) Finally there is typically a judder type issue in reverse if the pinion points down too far.
1) is not achievable due to the compliant mounting of the diff, as McG said, typically 2-4 degrees at full throttle in first and reverse, so in practice you pick the area where you have trouble and set that at parallel
2) we normally keep to
3) we achieve using a 4 arm rear suspension with different length upper and lower arms, although on our latest Hotchkiss we have introduced a Crane-Simplex linkage to get the same effect.
4) limits the reverse travel to about 4 degrees, it may even be why (2) is there.
So if 60 mph slight accel was your problem area then you'd set the system up in alignment at say 40% torque in 4th gear. Than your pitch compliance lets you work out where it is at rest. Then you check the other load cases and suspension travel (for a live axle). Then you start adding snubbers etc to suit.
Note that a 2 piece propshaft actually makes this worse, the important length is smaller, so the angles get bigger.
#7
red300zx99
-
- 328 posts
- Joined: February 03
Member
Posted 12 January 2004 - 21:38
So if optimal pinion angle is parallel or 0d then how would one calculate the losses associated with a pinion angle of say 1d or something other then 0d? Sorry, too lazy to look for the degree symbol
#8
Greg Locock
-
- 6,494 posts
- Joined: March 03
Member
Posted 12 January 2004 - 21:49
The losses vary proportional to articulation angle, at a rough guess. You'd need to work out the velocity of the needle rollers and the loads on them.
However the answer is very small, so I wouldn't bother. You keep the angles small to limit vibration, not losses.
However the answer is very small, so I wouldn't bother. You keep the angles small to limit vibration, not losses.
#9
BRIAN GLOVER
-
- 465 posts
- Joined: November 01
Member
Posted 13 January 2004 - 02:46
I have some 'back yard' experience in this matter. If you take a look at the car on the left, it has a LS1 Chevy engine with 400 hp and 400 ft/lbs torque and a T56 6 speed. The live axle is a Camaro 'Australian' unit with 3.71 gears.
(I also have an IRS version.)
The car weighs 1780lbs with a GRP body over a mild steel tubular space frame and no doors. It has DOT crumple zones for front and side impact.
It has 52% weight at the back. It has U/L A arms in the front with pull rod coil /overs with zero Ackerman and zero scrub.(No ABS and power steer or power brakes) It has a dash mounted Willwood adjustable brake bias valve to the back.
The drive shaft is 17" long including a 5 inch yoke and Hooke joints are 180' out of phase. (Important.)
Some home grown 'rule of thumb' guide lines discovered by thousands of miles of testing and development are as follows:.
Best way to locate this axle is 3 trailing links with minimum squat geometry and 2 1/2' roll understeer geometry, mounted with 16mm Heims.
A Watts link takes up lateral motion and is mounted at the bottom of the diff backing plate for lowest roll center.. The bellcrank bushing is rubber and the frame mount bushings are polyurethane and all 3 bushings are in the same horizontal plane.
All geometry is determined at full gross weight..IE: Driver and fuel.
All other joints are Heim including coil /overs mounted vertically above lower trailing arm axle housing mounts.
Battery in trunk on left and the 12 gallon 'usable' gas tank is on the right and the same shape as the bottom of the pax seat.
.
The lower trailing arms must be parallel to the bottom of the frame with this load and mounted in such a way as to give you your steer geometry and droop limits..
I got this idea from an old Renault 12 rear axle except the Renault was mounted with roll oversteer.(FWD). The top link points toward the front axle center line and all 3 trailing links have Heim joints. The crankshaft, drive shaft and pinion must be a straight line horizontally..16" wheels on 35 series tires keeps this line as low as possible.
With a live axle the front roll center must be higher on a live axle than an IRS. The rule of thumb here is:
What ever the speed that this car is going to be driven at most of the time the front roll center will be determined.
It is always lower than the rear r/c and the higher the speed the less the difference.. The angle is the same angle as the driver eyeball to the apex for the chosen speed.
The engine has Corvette hydraulic mounts in the center of the engine. The big rubber gearbox mount is a stock Camaro RPO. The gearbox mount and the Watts bell crank rubber bushing are the key to all this working properly..
The set up:
You will always get a vibration if the drive shaft is on the same plane on all axis, contrary to popular belief.. Even when I use the BMW M3 rear IRS where the diff is mounted solid to the frame, the drive shaft is one degree off in plan, but that is another story..The roll centers go underground in this case, but the live axle will beat it in a low speed autocross.
( The BMW IRS gets a fabricated adjustable upper link for camber adjustment and push rod coil/overs transversely mounted inside the trunk.)
I do the DWG 1:1 scale on a chalk board (old school) after a clever employee does it on CAD first. Try it. The only time it all lines up is at straight line constant speed. No vibes.
No production car including a Caterham will touch it in any performance category up to 130 mph. The 1/4 mile comes in 11.2 seconds and it stops in 90ft from 60 with BFG R1s
(I also have an IRS version.)
The car weighs 1780lbs with a GRP body over a mild steel tubular space frame and no doors. It has DOT crumple zones for front and side impact.
It has 52% weight at the back. It has U/L A arms in the front with pull rod coil /overs with zero Ackerman and zero scrub.(No ABS and power steer or power brakes) It has a dash mounted Willwood adjustable brake bias valve to the back.
The drive shaft is 17" long including a 5 inch yoke and Hooke joints are 180' out of phase. (Important.)
Some home grown 'rule of thumb' guide lines discovered by thousands of miles of testing and development are as follows:.
Best way to locate this axle is 3 trailing links with minimum squat geometry and 2 1/2' roll understeer geometry, mounted with 16mm Heims.
A Watts link takes up lateral motion and is mounted at the bottom of the diff backing plate for lowest roll center.. The bellcrank bushing is rubber and the frame mount bushings are polyurethane and all 3 bushings are in the same horizontal plane.
All geometry is determined at full gross weight..IE: Driver and fuel.
All other joints are Heim including coil /overs mounted vertically above lower trailing arm axle housing mounts.
Battery in trunk on left and the 12 gallon 'usable' gas tank is on the right and the same shape as the bottom of the pax seat.
.
The lower trailing arms must be parallel to the bottom of the frame with this load and mounted in such a way as to give you your steer geometry and droop limits..
I got this idea from an old Renault 12 rear axle except the Renault was mounted with roll oversteer.(FWD). The top link points toward the front axle center line and all 3 trailing links have Heim joints. The crankshaft, drive shaft and pinion must be a straight line horizontally..16" wheels on 35 series tires keeps this line as low as possible.
With a live axle the front roll center must be higher on a live axle than an IRS. The rule of thumb here is:
What ever the speed that this car is going to be driven at most of the time the front roll center will be determined.
It is always lower than the rear r/c and the higher the speed the less the difference.. The angle is the same angle as the driver eyeball to the apex for the chosen speed.
The engine has Corvette hydraulic mounts in the center of the engine. The big rubber gearbox mount is a stock Camaro RPO. The gearbox mount and the Watts bell crank rubber bushing are the key to all this working properly..
The set up:
You will always get a vibration if the drive shaft is on the same plane on all axis, contrary to popular belief.. Even when I use the BMW M3 rear IRS where the diff is mounted solid to the frame, the drive shaft is one degree off in plan, but that is another story..The roll centers go underground in this case, but the live axle will beat it in a low speed autocross.
( The BMW IRS gets a fabricated adjustable upper link for camber adjustment and push rod coil/overs transversely mounted inside the trunk.)
I do the DWG 1:1 scale on a chalk board (old school) after a clever employee does it on CAD first. Try it. The only time it all lines up is at straight line constant speed. No vibes.
No production car including a Caterham will touch it in any performance category up to 130 mph. The 1/4 mile comes in 11.2 seconds and it stops in 90ft from 60 with BFG R1s
Originally posted by Greg Locock
1)For a production car the optimum is to have the pinion shaft of the diff parallel to the output shaft of the trans, for all driving conditions.
2)There is also an 'absolute' limit, of about four degrees, on the articulation of each Hookes joint.
3) Finally you need to make sure that the pinion flange does not hit the floor - typically that piece of floor is in the back seat, and you can't afford to put a lump there.
4) Finally there is typically a judder type issue in reverse if the pinion points down too far.
1) is not achievable due to the compliant mounting of the diff, as McG said, typically 2-4 degrees at full throttle in first and reverse, so in practice you pick the area where you have trouble and set that at parallel
2) we normally keep to
3) we achieve using a 4 arm rear suspension with different length upper and lower arms, although on our latest Hotchkiss we have introduced a Crane-Simplex linkage to get the same effect.
4) limits the reverse travel to about 4 degrees, it may even be why (2) is there.
So if 60 mph slight accel was your problem area then you'd set the system up in alignment at say 40% torque in 4th gear. Than your pitch compliance lets you work out where it is at rest. Then you check the other load cases and suspension travel (for a live axle). Then you start adding snubbers etc to suit.
Note that a 2 piece propshaft actually makes this worse, the important length is smaller, so the angles get bigger.
#10
Greg Locock
-
- 6,494 posts
- Joined: March 03
Member
Posted 13 January 2004 - 09:15
Nice to see some numbers. Oddly enough we use a 1 degree plan view offset on our live rear axle. I thought it was there for packaging. Maybe we should move the diff in the IRS over! (no, they'd kill me)
So far as I know the reason to avoid zero/zero setups is that the needle rollers tend to vibrate all the lube out from under them, and then fail. AKA false Brinelling.
BTW your cars are invisible. (grins)
So far as I know the reason to avoid zero/zero setups is that the needle rollers tend to vibrate all the lube out from under them, and then fail. AKA false Brinelling.
BTW your cars are invisible. (grins)
#11
BRIAN GLOVER
-
- 465 posts
- Joined: November 01
Member
Posted 14 January 2004 - 00:57
Greg,
let me aks you a question, but first:
I discovered how important a differential is, recently. Something I had not given much thought to before.
I know F1 cars have markers to set up their differentials before they enter a turn and more and more production cars have electronic diffs. Alanso spun out a few times in practice because the Renault's electronics conked out and the diff was wrong for that turn, amongst other things, TC not withstanding.
I have a 70% limited slip diff(what ever that means exactly), which worked satisfactorily until recently, or so I thought. It just needed more additive.
The inside wheel began to spin as I applied power at the apex, which indicated a clutch pack failure. The slip varied with temperature and a few times around my favorite corner(115mph in 5th), I was able to floor it at the apex and exit the turn with 200rpm more. I was getting more torque to the inside wheel. When it is cold, the inside wheel would spin and I would have to go to 4th, after losing much speed.. Normally, if I floored it at the apex(more like squeeze it to wide open as I am carrying a lot of power anyway) I would induce a serious oversteer(powerslide), which I'm not adverse to doing, but it also kills exit speed. I would normally alter the line and open the radius in order to apply full power.
Now if only I could set the amount of slip electronically, that would be cool.
I see a company in Kit Car Builders magazine called Auburn Gear www.aurburngear.com with a switchable diff. It goes from lock to L/S. Have you any experience with this type of mechanism?
I tried to talk to the local dealer, but couldn't give me the info that I need. This has got to be the hottest thing in performance cars right now and may be as important as tire selection. How does a Porsche 996 electronic diff work? I know it locks up on coast. Since F1 cars have it, it obviously doesn't need ABS circuitry. Can variable slip be achieved mechanically? What would be required to achieve it electronically with a stock GM ECM and Aurburn Gear's ECTED type diff?
I have messed with different planet gear sets for in and out ramps and multiple clutch packs with a local specialty gearbox and axle shop, but neither of us know what we are doing, yet I can't achieve what I did with the slipping diff.. I cant find anything on the net or in any books that I have.
Thanks man.
let me aks you a question, but first:
I discovered how important a differential is, recently. Something I had not given much thought to before.
I know F1 cars have markers to set up their differentials before they enter a turn and more and more production cars have electronic diffs. Alanso spun out a few times in practice because the Renault's electronics conked out and the diff was wrong for that turn, amongst other things, TC not withstanding.
I have a 70% limited slip diff(what ever that means exactly), which worked satisfactorily until recently, or so I thought. It just needed more additive.
The inside wheel began to spin as I applied power at the apex, which indicated a clutch pack failure. The slip varied with temperature and a few times around my favorite corner(115mph in 5th), I was able to floor it at the apex and exit the turn with 200rpm more. I was getting more torque to the inside wheel. When it is cold, the inside wheel would spin and I would have to go to 4th, after losing much speed.. Normally, if I floored it at the apex(more like squeeze it to wide open as I am carrying a lot of power anyway) I would induce a serious oversteer(powerslide), which I'm not adverse to doing, but it also kills exit speed. I would normally alter the line and open the radius in order to apply full power.
Now if only I could set the amount of slip electronically, that would be cool.
I see a company in Kit Car Builders magazine called Auburn Gear www.aurburngear.com with a switchable diff. It goes from lock to L/S. Have you any experience with this type of mechanism?
I tried to talk to the local dealer, but couldn't give me the info that I need. This has got to be the hottest thing in performance cars right now and may be as important as tire selection. How does a Porsche 996 electronic diff work? I know it locks up on coast. Since F1 cars have it, it obviously doesn't need ABS circuitry. Can variable slip be achieved mechanically? What would be required to achieve it electronically with a stock GM ECM and Aurburn Gear's ECTED type diff?
I have messed with different planet gear sets for in and out ramps and multiple clutch packs with a local specialty gearbox and axle shop, but neither of us know what we are doing, yet I can't achieve what I did with the slipping diff.. I cant find anything on the net or in any books that I have.
Thanks man.
Originally posted by Greg Locock
Nice to see some numbers. Oddly enough we use a 1 degree plan view offset on our live rear axle. I thought it was there for packaging. Maybe we should move the diff in the IRS over! (no, they'd kill me)
So far as I know the reason to avoid zero/zero setups is that the needle rollers tend to vibrate all the lube out from under them, and then fail. AKA false Brinelling.
BTW your cars are invisible. (grins)
#12
Greg Locock
-
- 6,494 posts
- Joined: March 03
Member
Posted 14 January 2004 - 21:32
Sorry, I'm going to be no help at all on the internal gubbins of diffs. We use a rather simple LSD primarily for marketing purposes, but not on the variants I was or am involved in.
However, I have some experience with transfer cases, and some of those use multiplate clutches, electronically controlled to provide an active torque split. The clutches are MUCH larger than a diff's, and the relative speed differences are smaller, yet friction material life is still an issue.
However, I have some experience with transfer cases, and some of those use multiplate clutches, electronically controlled to provide an active torque split. The clutches are MUCH larger than a diff's, and the relative speed differences are smaller, yet friction material life is still an issue.
#13
BRIAN GLOVER
-
- 465 posts
- Joined: November 01
Member
Posted 15 January 2004 - 17:05
Bill, what does your new book say about active differentials? What is the state of them now? Does any road car use this technology? what work have you done on LS diffs?
I'm waiting for my copy. Dont you think you aught to get royalties?
I'm waiting for my copy. Dont you think you aught to get royalties?
Originally posted by McGuire
The short, simple answer: For a Hotchkiss drive (live axle - open drive, single Hooke joint on each end of the prop shaft) the optimum pinion angle is zero...or more precisely the pinion should be exactly parallel with the transmission output shaft in all planes. (Or better yet, dead in line.) However, since the axle may wind up under power, some static pre-angle is often desired to keep things straight when max torque is going through the drive. The actual angle required is entirely dependent upon the compliance built into the rear suspension...but typically in the range of 2 to 6 degrees (the latter for leaf springs). A rear suspension purpose-built for racing should require little or no static pre-angle.
#14
McGuire
-
- 9,218 posts
- Joined: October 03
Member
Posted 16 January 2004 - 09:29
Originally posted by BRIAN GLOVER
Bill, what does your new book say about active differentials? What is the state of them now? Does any road car use this technology? what work have you done on LS diffs?
I'm waiting for my copy. Dont you think you aught to get royalties?
Brian,
Hey, nice car. I have identified your problem. This vehicle has an excessive power/weight ratio. Once you correct that the wheelspin trouble will be totally cleared up.
You're welcome. I haven't gotten to the chapter on final drive yet. I am trying not to read the whole book at once but savor it a chapter or two at a time in the evening. I do know the Ferrari axle uses hydraulic control (annular hydraulic throwout bearing-like thingie clamping a clutch pack).
The ELSDs in the OEM pipe employ an electromagnetic coil actuated by a pulse-width modulated DC current. So the circuitry is very simple but the command logic to operate it could get complicated. I don't know of an output driver on your PCM that will work for this purpose (not saying there isn't one, only that I don't know it). I don't really know this for a fact and you didn't hear it from me but I have to believe that in the next +/-18 months GM will have something out with ELSD using a PCM that will support your LS1.
