61 replies to this topic

[27neil]

Please axcuse my ignorance but could someone please tell me what exactly is a De Dion rear axle/suspension.

[Ray Bell]

Devised by de Dion, this was a form of axle that disassociated the mechanical elements from the axle beam.

Typically called a de Dion 'tube', the axle that bore the wheel bearings and controlled the wheels usually formed a kind of arc around the differential and the drive went to the wheels via shafts that were universal-jointed at both ends.

The springing, damping and locating media all attached to the de Dion tube...

[xflow7]

de Dion is non-independent like a conventional live axle. However it offers an unsprung weight advantage by fixing the diff to the chassis, although I believe total weight is often slightly higher than an equivalent live axle due to the U-joints (or CV's) etc.

[robert dick]

http://www.stoneleis...rs/Steaming.htm

Original idea behind the De Dion axle - from Pomeroy's "Grand Prix Car"/page 265 :
"This drawing... from the 1894 patent application made by Count de Dion... shows his arrangement of a dead axle beam with separate exposed half-shafts, each having two pot-type universal joints. The ends of each half-shaft pass through the hubs and connect with spokes engaging with the periphery of the wheel. The spokes of the wheel proper were thus relieved from driving stresses."

Strictly speaking the De Dion axle was designed by Charles-Armand Trépardoux, Albert de Dion's and Georges Bouton's associate - Trépardoux axle would be the better name.

In principle, the chain driven cars of the pre-WWI period had a De Dion rear axle whereby the task of the driving shafts was taken over by the chains.
Difficult to understand why so many historians maintain that the shaft driven cars were the modern alternative...

[2F-001]

"xFlow7"?? You sound like a kindred spirit...

One current application of a de Dion rear end (which xFlow7 will, for one, be aware of) is the Caterham 7 - the evolved Lotus 7. One other factor that leads these cars to be typically heavier than a live-axled variant is the reliance on proprietary diff from a large, heavy, mass-produced vehicle.

[Ray Bell]

True enough... it's hard to work out why they ignore the lightweight proprietry components from mass produced cars that they could be using.

[bradbury west]

If someone has a copy of Hugh Haskell's book "Colin Chapman Lotus Engineering" and the technology to scan and send the picture, try page 36 for a shot of the original de Dion configuration. It says it all

keep up the interest, the threads and the replies


Roger Lund.

[RaymondMays]

I have been trying to get my head around the de Dion axle for a while, and reading some new books I got for Christmas, including Mike Lawrence's "Grand Prix Cars - 1945-1965", I thought I'd take a look on TNF for more information.

It seems like most early F1 cars used the de Dion axle, but how does it work, what does it do, and what were it's benefits?

To me, it doen't look like the neatest or lightest solution, but was independent rear suspension not an option in those days?

What prevented designers from using other solutions?

[David Birchall]

Remember that the primary purposes of the deDion axle was to a/ Keep the unsprung weight as low as possible and b/ Keep the wheels parallel to each other to avoid rear wheel steering. The independent suspensions, until the late fifties even, frequently failed on the second point due, apparently, to the designers not requiring the loaded wheels to stay upright as the vehicle leaned in curves. Unequal length A arms were the answer but it seems to have taken a long time for the requrement to be realised. Chapman made some beautifully light deDion axles but when the Lola Mk1 came out with it's independent rear suspension using essentially unequal length wishbones they became redundant overnight. Chapman had designed sophisticated rear suspensions of course but the Lola was a step ahead.

[fester82]

The early designs for independent rear suspensions were swing axles used by Auto Unions in the '30s and many other vehicles like Beetles, Spitfires, and Corvairs for instance. The problem with swing axles are keeping the tires perpendicular to the road surface and can be quite easily seen by looking at one of those cars mentioned above with a load. The camber ( top of the wheels point to each other) is quite noticable. I had a rather large friend that had a Spit with a permenant list to port. Cornering the outside wheel stays pretty much perpendicular, but the inside has a visible camber problem reducing the contact patch with the road surface.

de Dion rear ends did a better job of keeping the wheels 90 degrees to the road than swing axles. Think of it as a live axle without the weight of the differential since it is fixed to the chassis. Reducing the unsprung weight has many advantages, especially back in the days when shocks were friction type and not the gas/hydraulic tube types we have today. Less weight means less inertia and the suspension is much more responsive to road irregularities and control with less inertia. It is not necessarily lighter overall as mentioned by others, but the "axle" (a bent beam around the diff) and halfshafts weigh less than an axle with the differential.

[VAR1016]

Lancia as usual excelled: in 1937 the superb Aprilia was introduced: independent all round with Lancia's traditional sliding pillars at the front, the car had rear suspension that comprised semi-trailing arms with torsion bars plus a transverse leaf spring. Complex but very effective.

One has to drive an Aprilia (as I was priveleged to do recently) to realise what a fantastic achievement it is. Imagine in 1937 that for £395 (say £20,000 today) you could buy an Aprilia in England. Then imagine further that this 1370c.c. four seater saloon, with no B pillars had a maximum speed of nearly 80 mph (i.e. 20% faster than most other 1300s and faster than other larger cars) and cornering powers that put rather a lot of expensive metal to shame in the 1950s - never mind the 'thirties - it was well known that a Jaguar couldn't stay with an Aprilia on a country road - no chance.

A system of semi-trailing arm independent suspension was used on Lancia's Aurelia introduced in 1949. From this cam the famous B20GT so popular with the G.P. drivers of the day. The first three series of these retained the independent rear end, but the latter three series were fitted with de Dion axles.

The old boys used to say that the de-Dion cars had better traction, but others have more recently said that, given more modern tyres, the early series cars handle better. I have followed a seasoned old racer around Goodwood whilst he was driving his 2nd Series Aurelia; looks excellent from the back I have to say.

As usual, Lancia were ahead of their time.

PdeRL

[David Birchall]

You must understand the meaning of the terms "Sprung" and "Unsprung" weights to appreciate the benefit of the deDion axle and, to a certain extent, independent suspensions. If you are unsure of the meaning of those terms I would suggest learning about them first.

[Bonde]

Another advantage of the DeDion axle relative to the live beam axle is that the DeDion enables providing the rear wheels with static camber and toe angles as required; on the live axle the wheels are obviously square-on. This, IIRC, was actually one of the rationales behind the deDion as originally conceived as the highly cambered roads (hence the term 'camber' for wheel 'lean') then prevalent seemed to some to call for positive camber on the rear wheels as well so that the tyres wouldn't be running on-edge on their shoulders.

A bit of static toe-in also typically improves stability.

In addition, by keeping the differential from moving about with suspension travel, the chassis could be built lower - even more so if the driveshafts were permitted to have some upwards slant towards the road wheels, and clever axle location design could provide greater freedom in location of roll center.

The DeDion also allows designing-in some intentional flexural compliance. IIRC, this was used by Rover, among others.

Efficient independent suspension wasn't that easy to implement on the then-prevalent ladder chassis either.

Re. swing axles versus DeDion: Auto Union did realize the less favourable features of the combination of a road-level front roll center (trailing link IFS) with a high rear roll center (swing axle), particularly lethal if the rear wheels are allowed to go into positive camber, their later models having DeDion rear. Swing axles are the simplest IRS to implement on any chassis configuration with RWD, and with the rear/central engine location, a live axle wasn't really an option for Auto Union. Dr. Porsche was apparently deeply inspired by Hans Ledwinka, as everything the pair of them designed in those days seemed to be branded with swing axle IRS!

[VAR1016]

Originally posted by Bonde

Re. swing axles versus DeDion: Auto Union did realize the less favourable features of the combination of a road-level front roll center (trailing link IFS) with a high rear roll center (swing axle), particularly lethal if the rear wheels are allowed to go into positive camber, their later models having DeDion rear. Swing axles are the simplest IRS to implement on any chassis configuration with RWD, and with the rear/central engine location, a live axle wasn't really an option for Auto Union. Dr. Porsche was apparently deeply inspired by Hans Ledwinka, as everything the pair of them designed in those days seemed to be branded with swing axle IRS!

Yes but when Uhlenhaut drew up the W196 Mercedes-Benz, he deliberately opted for swing axles (albeit low pivot ones) because he had seen the breaks in the black lines from the de-Dion-equipped cars in the 1930s.

PdeRL

[Ian McKean]

Originally posted by VAR1016


Yes but when Uhlenhaut drew up the W196 Mercedes-Benz, he deliberately opted for swing axles (albeit low pivot ones) because he had seen the breaks in the black lines from the de-Dion-equipped cars in the 1930s.

PdeRL

What do you mean?

[VAR1016]

That the trend at the end of the 'thirties was for de Dion axles.

Rudolf Uhlenhaut, who was around at the time obviously turned his back on the idea.

PdeRL

[Ian McKean]

I thought you were implying that the charecteristics of the de Dion axle left different rubber lines on the road suggesting different traction or roadholding.

[mat1]

Re unsprung weight of De Dion axle vs live axle:

With a live axle the weight of the cardan axle (from the engine to the rear axle) is to be taken into account as well. If you do that, it is obvious a De Dion axle can be much lighter.

Another advantage is the possibility of positioning the engine in the back of the car. With e alive axle this is almost impossible.

A reason the independent rear suspension came rather late to blossoming: if you use IRS and the rear wheels are driven, you need a kind of sliding arrangement in the axle (because the radius from wheel to differenatial is not constant). (An even more complicated alternative is a layout with a couple of CV joints.)

This is difficult to manufacture, and prone to wear and breaking down. Of course this is also true for the DD axle.

Anyway, the advantages of the DD axle led Forghieri c.s. in the 70s to try it on a F1 Ferrari (312T2 I think.)

mat1

[VAR1016]

Originally posted by Ian McKean
I thought you were implying that the charecteristics of the de Dion axle left different rubber lines on the road suggesting different traction or roadholding.

Yes that's right and is what Uhlenhaut had observed which prompted him to take the course he chose.

Please see "Rivals" by Chris Nixon

PdeRL

[Hippo58]

Hi folks!

My first post on the board!

I thought I can provide a photo so that those who don't know at all how a deDion axle looks like, to get an idea.

http://www.ifrance.c...gtv6/4/b/18.htm

This is from an Alfa Romeo GTV6. From the 70's Alfettas to the 75 and SZ in the early ninety's alfa was using this layout on some of its cars.

Note the inboard rear disk brake layout contributing to less unsprung mass. Also as you can see this setup uses a transaxle (clutch, gearbox, diff) in the rear for better weight distribution (50/50 on these cars).

Take care!

Petros.

[Roger Clark]

Originally posted by David Birchall
Chapman made some beautifully light deDion axles but when the Lola Mk1 came out with it's independent rear suspension using essentially unequal length wishbones they became redundant overnight. Chapman had designed sophisticated rear suspensions of course but the Lola was a step ahead.

Didn't the Lotus 12 and 15 preceed the Lola with strut-type independent suspension?

[David Birchall]

Treading very gently here....
The 12 did I am sure but the 15?
According to Ian Smith in "The Story of Lotus" and David Pratley in "Lola's First: the Mark 1", both cars arrived in early 1958.
David B

PS Just reread David's post and realised that the Lotus rear suspension and the Lola's were quite different

[Roger Clark]

The Lotus 15 first appeared at Oulton Park on April 12th, the Lola at Crystal Palace on 5 July.

[David Birchall]

Roger, the Lola's rear suspension is not strut. The way to tell the difference is to imagine removing the shock absorber assembly. On the Lotus the wheel would simply fall down, while on the Lola the suspension would remain intact. NO I Know that is not the purpose of the suspension design before somebody corrects me but it shows there is a basic difference. The Lotus used the "Chapman Strut" where the shock assembly is rigid part of the suspension. The Lola uses upper and lower A arms (Actually the upper is the driveshaft but that doesn't matter). The Jaguar independent rear suspension is similiar to the Lola in many respects but Jag mounted the shocks on the lower part of the suspension instead of the upper like the Lola. Similiar principle.
David B

[Ray Bell]

Originally posted by Hippo58
Hi folks!

My first post on the board!

I thought I can provide a photo so that those who don't know at all how a deDion axle looks like, to get an idea.



This is from an Alfa Romeo GTV6. From the 70's Alfettas to the 75 and SZ in the early ninety's alfa was using this layout on some of its cars.

Note the inboard rear disk brake layout contributing to less unsprung mass. Also as you can see this setup uses a transaxle (clutch, gearbox, diff) in the rear for better weight distribution (50/50 on these cars).

And a good contribution... thank you...

I'm wondering too about Uhlenhaut's statement... is he implying that the de Dion cars had some kind of axle windup?

[Roger Clark]

Originally posted by David Birchall
Roger, the Lola's rear suspension is not strut. The way to tell the difference is to imagine removing the shock absorber assembly. On the Lotus the wheel would simply fall down, while on the Lola the suspension would remain intact. NO I Know that is not the purpose of the suspension design before somebody corrects me but it shows there is a basic difference. The Lotus used the "Chapman Strut" where the shock assembly is rigid part of the suspension. The Lola uses upper and lower A arms (Actually the upper is the driveshaft but that doesn't matter). The Jaguar independent rear suspension is similiar to the Lola in many respects but Jag mounted the shocks on the lower part of the suspension instead of the upper like the Lola. Similiar principle.
David B

I appreciate the difference, although I can see that my wording may have indicated otherwise. I thought that your earlier post suggested that Chapman had used de Dion rear suspension until the Lola appeared.

On the wider subject, can anyone tell me why some designers went to great lengths to achieve articulation of their de Dion tube, while others appear not to have bothered?

[VAR1016]

Originally posted by Ray Bell


And a good contribution... thank you...

I'm wondering too about Uhlenhaut's statement... is he implying that the de Dion cars had some kind of axle windup?

I cannot put my hands on Nixon's book at the moment (flat is always a shambles )

I recall him being quoted as having said that he had noticed breaks in the black marks left by the de Dion-equipped cars, whilst the swing-axled cars left continuous black lines.

The Mercedes-Benz low pivot approach apparently addressed some of the other "characteristics" of swing axles. I do not profess to understand this - geometry is horrible stuff!

PdeRL

[willga]

Does anyone have any photos of the deDion on Foghieri's 312T2?

[xkssFrankOpalka]

My experience with De Dion rear axle cars is with the Elva Mk 2 that I raced, the De Dion is very stable and predictable. My Iso has a DeDion also, works very well. Maserati tried a Birdcage style tube in their rear engined car but found it flexed under trailing throttle, not very stable. lotus 11s had De Dion. my Elva Mk 4 had a form of independent which was poorly designed and the rear wheel steer was awful. The sliding joint in the tube is used by Rover, maybe other cars, as the axle moves up and down there is a change in length which must be delt with. Most cars have universals in the half shafts, or rubber donuts, but Rover chose to make a sliding joint in the tube itself!!! Very clever.

[David Beard]

Originally posted by Roger Clark
On the wider subject, can anyone tell me why some designers went to great lengths to achieve articulation of their de Dion tube, while others appear not to have bothered?

If my understanding of articulation within a de Dion is correct, this was a feature I only recall in the Rover 2000 arrangement. What others were there?

(The purpose in the Rover was that it aimed to have the most over-engineered suspension system ever seen in a four door saloon car. The complication of de Dion tube at the rear with its extra twiddly bits was matched by the horizontal coil springs at the front, thrusting against the passenger compartment bulkhead)

[Ray Bell]

Yes, that Rover suspension is a real bundle of complication and confusion... the only dampers ever required to have more resistance on compression than rebound, for instance, leading to the fact that most manufacturers wouldn't make them.

The joint in the middle of their de Dion tube did more than save them the trouble of having sliding yokes in the axles... it prevented the de Dion tube from doing what it was always intended to do... instead of keeping the wheels upright, it enabled them to roll with the car!

[Roger Clark]

Originally posted by David Beard


If my understanding of articulation within a de Dion is correct, this was a feature I only recall in the Rover 2000 arrangement. What others were there?

(The purpose in the Rover was that it aimed to have the most over-engineered suspension system ever seen in a four door saloon car. The complication of de Dion tube at the rear with its extra twiddly bits was matched by the horizontal coil springs at the front, thrusting against the passenger compartment bulkhead)

The Mercedes-Banz W125 was one. Pomeroy wrote:

From this arrangement it follws that as the rear wheels rise and fall they swing about a 36 in. radius imposed by the arms, and also that if only one wheel rises this wheel traverses an arc and has an angular relationship to the other rear wheel, This would only be possible by twisting the de Dion cross tube which would then act as an excessively strong torsion anti-roll bar whereby (apart altogether from the severe stresses which would thus be set up) the handling qualities of the car would be impaired.

To relieve the De Dion tube from torsion it was necessary to divide it so that one half could turn in relation to the other, ....

I believe the 3-litre auto-Unions had a similar arrangement.

Elsewhere, Pomeroy says of the 4.5-litre Ferrari:

The rear wheels are maintained parallel with each other and vertical with the road by a de Dion tube which is made in three parts, but does not require any provision for oscillating movement of one side against the other, such as was provided on the pre-war German racing cars. This follows from the fact that the radius arms which drive the car and contain engine torque and brake reaction are formed in pairs and lie parallel one above the other in the same plane as the wheels.

I must admit that I can't see how either the radius arms, or sliding yokes in the drive shafts would relieve the de Dion tube of the torsional forces that the Germans went to such lengths to avoid.

With regard to the front suspension of the Rover 2000, I thought it was to allow space for the gas turbine engine that the car was originally designed for. It sould also be that the designers wanted to feed suspension loads into the most rigid part of the car.

[Ray Bell]

I don't think any of that specifically says the de Dion tubes on those cars were like that on the Rover, Roger...

The M-B one specifically says about the tube having to twist the bar to achieve this, and that it would introduce stresses that were undesirable to do so.

[Roger Clark]

Originally posted by Ray Bell
I don't think any of that specifically says the de Dion tubes on those cars were like that on the Rover, Roger...

The M-B one specifically says about the tube having to twist the bar to achieve this, and that it would introduce stresses that were undesirable to do so.

I didn't intend to imply that it did. It was the German cars I was thinking of when I asked the original question.

[xkssFrankOpalka]

I think the 3 pc tube refered to was the construction of the tube instead of the curved tube some builders used.

[Catalina Park]

Does anyone have a picture of Jim Halls 2H de Dion?

[mat1]

Originally posted by Roger Clark





I must admit that I can't see how either the radius arms, or sliding yokes in the drive shafts would relieve the de Dion tube of the torsional forces that the Germans went to such lengths to avoid.

I think the "trick" int he Ferrari is using two parallel radius arms (on each side, presumably), instead of on. If you use one, you must make the connection of radius arm and axle rigid, and then, when the car rolls, you get a torsion in the axle. If you use parallel arms, you can use joints at every connection, and the torsional effect is gone. There will be a steering effect, though.

Can this be the explanation?

mat1

[Roger Clark]

Originally posted by mat1


I think the "trick" int he Ferrari is using two parallel radius arms (on each side, presumably), instead of on. If you use one, you must make the connection of radius arm and axle rigid, and then, when the car rolls, you get a torsion in the axle. If you use parallel arms, you can use joints at every connection, and the torsional effect is gone. There will be a steering effect, though.

Can this be the explanation?

mat1

You may be right. Certainly the majority of Italian post-war racing cars with de Dion suspension had twin radius arms. However, I can't see why a single radius arm requires a rigid connection, nor why having flexible joints removes the torsional effect.

Perhaps these things are obvious to everybody else.

[Ray Bell]

The single arms would be required to maintain the stability of the de Dion tube... keep it 'on the level' so to speak. Like a radius arm on the front of a Ford 10.

[Kvadrat]

Picture from Cimarosti book:

[Roger Clark]

Some newer members may find parts of this old thread to be of passing interest.

http://forums.atlasf...=&threadid=5596

[Roger Clark]

In case there is anybody else interested in the question of de Dion articulation, I post this diagram in "The Grand Pris Car".



The upper two diagrams show the single triangulated arm as used on the pe-war German cars. As one wheel rises there was an angular movement at the end of the axle tube, which, therefore, hoad to be split to allow rotation. The lower diagrams show the arangement introducrd by Ferrari and used on most post-war cars. THe twin radius arms prevent rotation of the tube and allow it to be made in one piece with a saving in cost and weight at the price of a variation in wheelbase.

[Roger Clark]

Somefurther investigation has revealed that the Delahaye monoplace was first to use a single piece de Dion located by twin radius arms.

[Roger Clark]

I feel that I'm coducting something of a dialectic on this tread, but the next question is, if twin radius arms prevent a de Dion tube acting as an anti-roll bar, why don't they prevent an anti-roll bar acting as an anti-roll bar?

I would appreciate some help as my answers have to be in by 31 ÿíâàðü. Ooh! What a give-away...

[Ian McKean]

A minor point, but you're still going to get a variation in wheelbase with the single link on each side, unless the link is infinitely long.

The Saab 900 had an interesting location of the back (dead) axle using Watts linkages on either side. The compliance in the bushes was sufficient to allow body roll without winding up the dead axle but inhibited roll sufficiently to make a rear anti-roll bar superfluous. Of course it wouldn't work on a racing car, but this setup avoided any variation in wheelbase.

Even more off topic, I used to think my Saab 900 T16 oversteered more on left handers than right handers, possibly due to the Panhard rod location. I was investigating this phenomenon after mulling it over while consuming a couple of Guinesses, when I spun the car through a farmer's fence. I did a lot of damage (£4,000 IIRC) and also damaged a lot of hedge and fence, but the car was still driveable (just, after changing a wheel) so I got the hell out of there.

But when I got up the next morning to survey the damage in the cold light of day, I realised thet my front number plate had fallen off at the scene of the crime...

[mat1]

Originally posted by Ian McKean
A minor point, but you're still going to get a variation in wheelbase with the single link on each side, unless the link is infinitely long.

The Saab 900 had an interesting location of the back (dead) axle using Watts linkages on either side. The compliance in the bushes was sufficient to allow body roll without winding up the dead axle but inhibited roll sufficiently to make a rear anti-roll bar superfluous. Of course it wouldn't work on a racing car, but this setup avoided any variation in wheelbase.

I believe the original Alfasud had the same arrangement. I seem to remember this layout was abandoned on the Alfa 33, presumably because the 33 had a less rigid chassis, and this layout only would work in a rigid chassis.

Originally posted by Ian McKean

Even more off topic, I used to think my Saab 900 T16 oversteered more on left handers than right handers, possibly due to the Panhard rod location.

Interesting. I always have asked myself whether the (small) differences in bump-behaviour between the left and right side on a car with a Pabhard-rod. I can imagine it makes a difference.

mat1

[David Birchall]

Originally posted by mat1



Interesting. I always have asked myself whether the (small) differences in bump-behaviour between the left and right side on a car with a Pabhard-rod. I can imagine it makes a difference.

mat1

I am no engineer, but wouldn't the rear roll centre change with a Panhard rod depending on the side the vehicle is "Leaning" on? Usually the Panhard rod is replaced on a racing car with a Watts linkage or track system.

When I was a service manager we would occasionally get someone come in complaining that when they followed their own car they could see the front and rear tracks were out of line. We had to explain that this was due to the Panhard rod but they didn't always accept that. They thought a train had hit it or something!

Roger, was that a serious question about the rear roll bar? (It is separate to the trailing links is the answer)
David B

[mat1]

Originally posted by David Birchall


I am no engineer, but wouldn't the rear roll centre change with a Panhard rod depending on the side the vehicle is "Leaning" on? Usually the Panhard rod is replaced on a racing car with a Watts linkage or track system.

When the loading of the car is such that in rest the panhard rod and the axle itself are exactly in parallel (which usually is not the case), and if it is properly designed, and if the vehicle is just leaning, than the roll centre will be in the same place, whether the car is leaning left or right.

A lot of ifs, and in reality this will not be the case.
And of course, if both wheels bump at the same time, the axle will move a little bit to one side.

So, a simple but crude system.

mat1

[vivafroilan!]

Originally posted by willga
Does anyone have any photos of the deDion on Foghieri's 312T2?

I'm not able to scan, but there's a photo in Niki Lauda's The Art and Science of Grand Prix Driving that might be what you seek, if that's any help (p.157)... The caption for this says:

"In mid 1975 Ferrari began tests with the De-Dion [hyphen in original] rear suspension. Gone is the massive De-Dion tube of the good old days, but this design still has the desired effects: the rear wheels have the same track and camber whatever the car's attitude. The solution shown here is a purely strut-design which puts the individual components only under tension and compression. It has the added advantage that the amount of toe-in can be easily adjusted -- by means of two universal couplings of variable length, attached to the rear extension of the upright. The design works in such a way that a racing car can be converted from the hitherto usual wishbone suspension to the De-Dion system in only a few hours."

[Ray Bell]

Is this one a Lotus 11 or similar?


Robert Britton photo