1.  I'm guessing a couple degrees angular and maybe a mm radial.
2.  Probably on rubber.
3.  No; the resulting moments would break the shafts.
4.  Temperature is a concern.  Maybe a similar industrial coupling would serve better, maybe not.

 

Mike Halloran
Pembroke Pines, FL, USA

thanks for your help Mike.

q2 answer you give as 'probably on rubber'. is the relative softness of the elastomeric and the rubber mounts a factor in this?

do cars typically flex say 1 or 2mm or less or more?, i really have no grasp as to if additional triangulation is likely to bring the chassis twist within acceptable limits.

is life span of these couplings hugely shortened if subjected to 110c?   

Sure.

Cars typically flex more than 2mm.  Your frame appears to be more fully triangulated than a typical car.  It's possible to analyze this sort of stuff, but it's not a whole lot of fun.

Any rubber product suffers at elevated temperature.
 

Mike Halloran
Pembroke Pines, FL, USA

In this instance, I suggest that you think of the rubber elements as fuses.

If you bolted everything together without them, you'd crack some expensive castings.

 

Mike Halloran
Pembroke Pines, FL, USA

I agree with everything Mike says.

I would add.

It is really impossible to tell from one photo of a part of the chassis how much it will flex and twist, but I never saw a race car where the aluminium panels did not fret away at their fasteners, often to the degree that aluminium rivets needed replacing from time to time. 2mm is well within the movement I would guess at. I have seen road cars flex as much as 3/4"with hard use and bonnet hinges bend a bit and door hinges and latches fatigue crack and pull from the pillars.

In drag cars I have worked on, they use a Powerglide gear box. This has a very rigid mounting to the block at the bell housing, so flex there is not an issue, but the extension housing where the output flange is located is quite a bit out behind the box.

The engine is very firmly mounted on thick aluminium plates that bolt to the front and rear of the block and are bolted solidly to the chassis so they in fact form 2 bulkheads with the block as a stiffener for the chassis.

The extension housing may be left just sitting out there with all the whip of the tail shaft being transmitted through it to the transmission hosing. Believe it or not, this seems to work.

The nicer alternative is to put a cross member under the extension housing and let it sit on a rubber pad, like a piece of old tyre tread or piece of old conveyor belt cover or the like.

This then needs to be losely held in place. The best method looks real rough, but a piece of rope tied over the housing and around the cross member works great. Safety scrutineers usually hate it though. I am not suggesting that a rope suits your set up, but a platform with flexible attachment of limited travel over the top, fore and aft and to the sides might work.

Looking at your set up, it very much appears to have IRS so the diff will have half shafts with CV joints going to each hub.

I can't see why you can't mount the diff a few inches further back and allow for more flex in the shaft or you could run back to back flexible couplings.

The temperature resistance of the coupling will depend very much on the rubber compound used and I very much doubt you will find any information on that. Thermoset rubbers which it very likely is tend to harden with age, especially when also exposed to heat.
 

Regards
Pat
See FAQ731-376: Eng-Tips.com Forum Policies for tips on use of eng-tips by professional engineers &
http://eng-tips.com/market.cfm
for site rules
 

Another coupling option might be available; a roller chain coupling.

I.e., a sprocket with the same pitch and number of teeth as the original final drive sprocket, and the pair of sprockets wrapped with a length of double row roller chain.

If you are lucky, an industrial roller chain coupling half exists with the right pitch and number of teeth.

If you are double-lucky, its hub will be large enough to be bored out to clear the final drive sprocket's retaining bolt head, and still maintain enough meat to transmit power.

If you are not that lucky, you'll have to make the sprocket from billet and harden it a bit.  You can get a good enough tooth surface by plunging through the blank and milling radially out.  The approach clearance bevels outside the pitch circle are not necessary because the chain is never unwrapped in service.

 

Mike Halloran
Pembroke Pines, FL, USA

And a slight change in tooth count on the relativity cheap and simple sprockets gives a quick rear end ratio change.  

Regards
Pat
See FAQ731-376: Eng-Tips.com Forum Policies for tips on use of eng-tips by professional engineers &
http://eng-tips.com/market.cfm
for site rules
 

thanks for all your help fellas.

on the subject of twist, do i have a 'fortunate' case in so far as the likelyhood of measurable chassis twist is at the sides of the car and the place that we are concerned with is close to the front/rear centreline, thus less likely to be a problem? even if the sides of the car twist it may be that ensuring the diff flange and drive flange stay in the same axis can be achieved to get within say 1mm?

i have considered a double chain coupling which could be achieved quite easily but figured that setup would be less tolerant than the donut for missalignment? would you suggest flex mounting the diff in this scenario?

another option is a design offered by dax kit cars in Uk. they use the original sprocket on the gearbox and fabricate a flange which has hardened steel pins screwed into it which 'mesh' with the gearbox sprocket. this results in a coupling which takes similar length to the donut but is heat resistant and will allow for small missalignment . the sprocket is considered consumable but at $10 this is fine and catastrophic failure is also avoided since limp home would still possible.    

moving the diff backwards introduces other problems: its not clear on the photo but the top rear wishbones have a diagonal chassis tube supporting them and the further back the diff sits, the greater is the likelyhood of collision between the half shaft and diagonal at upwards suspension travel. it would be possible to move the engine forwards 20mm and the diff back 20mm resulting in absolute max of 80mm (still with 20mm drive flange nut occupying part of this volume). this will require lots of additional work so is a last resort option. it does however mean there is enough space to fit a cross groove cv joint with the diff flex mounted. the flex mounts/cv would ensure that the gearbox and diff axis always met and ensure perfect alignment. does it sound workable? i know cv joints are not designed to be used in a situation which does not give rise to ball movement/lubrication but failure mode would be gradual with 'limp home' reliability when compared to potential heat-induced elastomeric melting.  

  

• http://files.engineering.com/getfile.aspx?folder=77fa9267-facd-41a6-9a77-b2

When those rubber donuts are used for driveshaft service they >>sometimes<< use a "centering element" to keep the driveshaft running on the intended axis.  That feature would make the radial  misalignment tolerance zero.

The only u-joint I can think of that can accomodate radial misalignment was used on GM cars with torque tube drive. It used open ended bushings in place of needle bearings.  A large formed sheet metal spherical housing defined the position of the torque tube, so the real drive shaft inside could be "off center" a bit, and the "u-joint" needed to be able to slide radially continuously to avoid a fight.

one other question about donuts:

when the plastic hardens (accelerated due to heat) what failure mode results. for example, if used in the application described above, the coupling would be subjected to around 100c (although no motion) for several hours after engine shutdown. is the heat without force on the coupling still causing damage. will its life be measured in units of hours?     

The damage is by oxidation and results in extra cross linking and increased hardness and decreased elongation until it eventually cracks. A wild ass guess based on nowhere near enough info would suggest hundreds of hours life at least, maybe a thousand or two.

Regards
Pat
See FAQ731-376: Eng-Tips.com Forum Policies for tips on use of eng-tips by professional engineers &
http://eng-tips.com/market.cfm
for site rules
 

In a D Sports/Racer class in SCCA racing, this motorcycle hybrid is not at all uncommon.  Most have a simple chain/sprocket drive to a live axle.  However, I have seen one a number of years ago that mounted the power plant similar to what you are doing.  In ALL cases the drive train and engine were solidly mounted.  I solid mount my race engine but use a polymer "shim" as it seems to damp out the high frequency vibrations.  In a race car this is not a problem in most cases.  Some solid mts like the original Formula Fords with the early pre Kent Cortina engines suffered block failures at the mounting flanges, a problem that was corrected with the Kent series of Ford engines.

I'm getting a bit off course here, sorry.   The Rotoflex couplings that I have used will stand a good deal of misalignment and are not particularly sensitive to heat deterioration.  I would never directly connect the drive output with the diff pinion...sure failure would result.  I have used this coupling in commercial applications when the standard Rotoflex couplings were failing in a consistent manner (Douglas Aircraft DC-9 paint bldg.)

http://www.bellgossett.com/productPages/VSX-Alignment-Friendly-Coupling.asp

I see no reason to worry about using Rotoflex in your application, Fiat, BMW, Lotus and, a host of others have used them successfully. As to worrying about the temp because the engine/gearbox share the lubricant...From a personal view, from the view of a Mini Cooper racer.."Man, you ain't got no problems"!

Rod  

Rotoflex couplings are used in lots of RWD OEM applications in the propeller shaft/driveshaft.  Ford, BMW, GM, Fiat (in time past).

Also, in certain 60s/70s Triumph they were used in the driveshafts between differential and hubs and in that application had to absorb 12mm of plunge.  The OE Metalastik ones could do this fine and gave good life.  The copy that available now doesn't do too well.

Nick

http://files.engineering.com/getfile.aspx?folder=d610f9ba-4e31-40cd-8882-23a0c7185999&file=coupling10_002.JPG


thanks for all the great help.

just to clarify the coupling type, its not one of the original rotoflex as used in the Triumphs, Lotuses etc (and also still used in farm equipment). they were very stiff and came with a steel band around which you only removed after fitting thus ensuring that the coupling was kept under compression whilst in use.

the coupling in question here is stock fitment in the prop shaft for modern BMW, range rovers and many other vehicles and is very flexible (the photo attached to this post shows me bending it against the floor).

patprimer: sorry for not being clear on coupling type. did your wild guess for life span relate to this type coupling? a couple of hundred hours would be a years worth of use and be great. based upon your description failure mode would likely commence with clunks giving limp home potential before disintegration?

  

• http://files.engineering.com/getfile.aspx?folder=d610f9ba-4e31-40cd-8882-23

Looks like the Guibo allegedly on this 3 series BMW uses a centering ball/device.
http://www.pelicanparts.com/bmw/techarticles/E36-Guibo/pic11.jpg

I'd expect something similar as necessary to provide shaft centering when used as a u-joint replacement on any two jointed shaft handling significant torque and operating at significant rpm.

The centering hardware would have to be left out of a joint expected to operate with radial misalignment.

To predict life we need temperature and load data, otherwise it's a guess.

My guess it will be fine for your application as I doubt you are running much hotter than under a road car in extreme conditions and you will be seeing a lot less use and I expect be more prepared to do maintenance.
 

Regards
Pat
See FAQ731-376: Eng-Tips.com Forum Policies for tips on use of eng-tips by professional engineers &
http://eng-tips.com/market.cfm
for site rules
 

  I am about to encounter the same problem as Dexion with my own little car.  I am intending to solidly mount the engine and final drive solidly together with aluminium plates or a small frame and then mount the whole "transaxle" unit flexibly in the car.

 PP - I think what Mike Halloran is describing is strictly using the chain as a coupling in a direct axial line from one sprocket to the next - not using the chain in a conventional sense.

  Morris Minis used rubber couplings on their driveshafts for many years. They were reliable but were rotted by diff and engine oil leaks.

  Lotus Elans had rubber couplings but these gave the car a terrible springy, surging effect which really spoiled the car.
 I think most existings cars have now replaced the original couplings with all-metal shafts and constant velocity joints.
      

thanks again for a fantastic response.

i'm going to run with the donut!

HEHTex, which Mini used a Rotoflex?  

As to the 'springiness' of the pair of Rotoflex couplings in the axles of the Lotus Elan...well, I guess if you are on and off the power a lot I can see where the rubber band impression comes from.  Me, personally?  I had no problems with them.  I did drive a "U" joint conversion and it was "different" and not in a good sense.  There always seemed to be a 'vibration/noise' that I could not put my finger on.  The center Rotoflex in my Fiat 124 was a problem only because of the rest of the car it was mounted on...Fix It Again, Tony!

Very early Minis used a rubber cupped "U" joint supposedly to reduce 'harshness'  http://www.minimania.com/web/Item/QL5000/InvDetail.cfm   quickly replaced by standard Hookes type U joints. The last bunch used both inboard and outboard CV's.
The "rubber deterioration" spoken of about early Minis occurred after many, many thousands of miles with poor/no maintenance. In no way were they anything like a Rotoflex coupling.

I'm not a great fan of rubber mounted anything in a performance or race car.  However, Rotoflex couplings have been used in such a wide variety of standard and performance cars...race cars from F-1 to Indy that if some inherent problem in the design were present it surely would have made itself known before now.  Proper mounting procedure and maintenance is the key here as in all things racing.

Rod

That type of coupler is used in the C5 and C6 Corvettes. The coupler for a 318i manual was used in the LS1 driveshafts and one a little bigger in diameter and thickness with M12 bolts is used in the newer models. It has big centering pins on the driveshaft. Those couplers actually have hard bushings for each bolt and a lot of cord in them for strength. They are a soft enough rubber they should give in all directions enough for the car you are proposing. I'd think you'd be fine with a similar coupler without a centering device. Remember to mount it with the arrows facing the flanges, not the bolt heads.
 

there are no arrows on the genuine bmw donut or on the genuine (larger) range rover ones i have.

incidentaly, the bmw one (which is the one i will use) has poor manufacturing tolerance for use in my application. the distance across the steel inserts varied by up to .5mm and the faces of some of the inserts were not square to the bores.

this would presumably not be relevant in its normal role on a propshaft where it drives or is driven by a shaft which has two universal joints. in that event the two flanges of the coupling may work ok if not parallel.

my flanges MUST be parallel and to avoid the donut being forced to a permanent state of (probably acceptable) missalignment, i ground the faces of the inserts to be square and the length of each insert to be the same within a couple of thou.

i wont use a centering device but, whist not clear in the photos, the nose of the diff is accurately located relative to the gearbox flange. the coupling was supplied with a spherical spigot bearing as the usual instal for this coupling has a prop with two universals on.

the help given on this thread has given me the confidence to use the coupling and be happy with its ability to cope with the potential missalignment that may be present (more triangulation will help). i'm now lookinq at ways of limiting the heat transfered to it from the drive flange. all other uses discussed here have the coupling in a low thermal stress role
    

"the bmw one (which is the one i will use) has poor manufacturing tolerance for use in my application."

Every rubber coupling I've seen has been "bent" to some degree.  An early ferrari restoration had a vibration at 1X driveshaft rpm. The oem style rubber coupling had a fairly typical amount of runout, but the trans mount was so soft that the relatively stiff coupling easily "had its way", flinging the transmission around at will.  Similar to your machining, careful shimming at each bolt made it better, if not acceptable. Enthusiastic MIT students had designed and built a test rig for something or other. The drive motor was on isolation mounts, and "rubber tire" type couplings were used as u-joints at each end of a short driveshaft.  The rpm was toward the upper range of what the coupling manufacturer would allow without centering devices, so the driveshaft could move off center and whirl if it felt like it, and the tires are very vulnerable to being assembled all skew-geed.  The vibration of the flexibly mounted motor at higher speeds was impressive.

Both malfeasances were measureable as many thousandths of "runout" with a dial indicator when rotating components by hand.  If a quasi static motion/runout check reveals 0.010 inch (10 MILS) TIR/pk-pk motion, there is going to be at least that much at speed.  Balance weights are not likely to make the situation any better.

I don't follow all your coupler alignment issues. You should be using two flanges with 3 bolt tabs or 3 bolt ears on each flange. 3 bolts should face one direction and pass through the coupler and then thread into the engine flange. The other 3 bolts should face the other direction and pass through the coupler before threading into the differential flange.

I didn't mention it before but I would hope you are not actually expecting to install the coupler as shown in your first picture. I had assumed you knew how these are to be installed and had just done that for a quick mock-up.

The length of each bolt bushing in the coupler will have no effect on the installation, within reason. Since there is only a flange on one side of each bushing the rubber can easily give the small amount required for each bushing to pull-up to the flange. Similarily, I would image that the face being exactly square to the bolt hole really would matter either as long as the bushing is not so bad it's putting a bending force into the bolt.

Also, every coupler I have seen has a little triangle or arrow on the outside diameter corresponding to each bolt hole. The triangle points towards the flange, not the bolt head.

 

Star for Lionel; good catch!

 

Mike Halloran
Pembroke Pines, FL, USA

"I didn't mention it before but I would hope you are not actually expecting to install the coupler as shown in your first picture. I had assumed you knew how these are to be installed and had just done that for a quick mock-up.The length of each bolt bushing in the coupler will have no effect on the installation, within reason. Since there is only a flange on one side of each bushing the rubber can easily give the small amount required for each bushing to pull-up to the flange. Similarily, I would image that the face being exactly square to the bolt hole really would matter either as long as the bushing is not so bad it's putting a bending force into the bolt.2

well spotted eagle eyes! i'm well aware of this but the coupling is installed that way at the time of the photo for mock up purposes.

there is a spigot shaft which is an m8 bolt screwed in the gearbox shaft and the unthreaded part of the bolt into an 8mm hole (actually 7.9mm) machined into the diff pinion shaft. the differential has been removed from the diff housing saving 8kg and this means i can hang the diff housing on the spigot bolt. the diff is then perfectly aligned to facilitate mounting fabrication but to relieve the weight on the 8mm bolt i also put the coupling bolts in too, but 'direct' rather than offset as you point out they should be fitted. it was then i realised that when the assembly was turned, the diff rear end made a cyclic rotation and this was due to the variation in the length of the inserts (presumably causing a stress on the spigot).

thus to use the spigot without stressing it for fabrication requires additional support and the additional support was introducing errors which was corrected by machining the dount inserts!

its likley that this small variance of the inserts would not matter in use but for fabriaction it is significant    

 Evel - I had a 1965 998cc Cooper (not an "S"). This had the rubber couplings and I think this type of coupling persisted for quite few years after 1965 (in Oz at least). So it just wasn't very early Minis that had the rubber joints. I don't think the "S" types ever had rubber joints.

 The Elan I drove would surge badly even at relatively constant speed - it really was terrible.  I recently saw an episode of "Wheeler Dealers" where the Elan they were restoring had a similar problem. It turned out that the rubber couplings were starting to break up - so maybe the Elan I drove had the same problem.      

My first Mini was a '63 Austin Cooper and it had standard "U" joints...as a matter of fact, they are still in use in the latest powertrain of my '67 "S".  As far as I know, at least in the cars modified by John Cooper Works,  none have rubber joints.  I have never owned a standard, non Cooper Mini or even driven one as they are not very commonly found on my side of the pond.  Now for OZ......I dunno!  You chaps do tend to do things somewhat differently from time to time.  Sadly, my original Mini Cooper had a rather abrupt end, see attachment.

Oh, and your correct in the evaluation of the Elan's couplings...However, the problem, once it surfaced, should have been repaired with due diligence.  I've been lax on occasion and let one fail...Not pretty and NOT cheap to repair.  I did learn a valuable lesson as it relates to Rotoflex couplings...I also learned to not like them and to substitute other coupling types whenever I can economically/reliably do so.

Rod

• http://files.engineering.com/getfile.aspx?folder=be01a8d8-ef3d-45c7-8de6-97

The rubber band couplings are completely different in almost every respect to the Jurid type the OP was talking about.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376: Eng-Tips.com Forum Policies  http://eng-tips.com/market.cfm?