40 replies to this topic

[TD1]

Hi , I looking to buy , rent of go to someone who has a Trosion test rig to test my chassis.
Any ideas please MOST welcome, i`m in Oxford but can travel.

[rachael]

Hi , I looking to buy , rent of go to someone who has a Trosion test rig to test my chassis.
Any ideas please MOST welcome, i`m in Oxford but can travel.

RML in Wellingborough has one - doubt it will be cheap though!

[DaveW]

Hi , I looking to buy , rent of go to someone who has a Trosion test rig to test my chassis.
Any ideas please MOST welcome, i`m in Oxford but can travel.

Try contacting the Multimatic Technical Centre in Thetford, UK (+ 1842 755744).

[cheapracer]

Last one I ever did some time ago was drop the roller door over the front of the chassis with some bolts slotted through the door and a long pipe with one guy hanging off it. Had a 2" pipe under the center of the rear of the chassis.

Kept increasing the length of the pipe until we got the rear corner to move down more than 1/2", was a fair length of pipe for that one too, chassis copied from a Radical with additions.

Knew the weight of the guy (had a set of kitchen scales) and the length of the pipe and the 1" of movement - simple bit of math and we had our result.

And that was all for the CAMS Technical guy standing there who approved the method.

Previously I have dyna bolted the front or rear down, I didn't think the door would hold but it did

[TD1]

haha, Basic but it would work, i`d like to be a little more sophisticated than this as it does have to be approved.
Thanks,
TD

Last one I ever did some time ago was drop the roller door over the front of the chassis with some bolts slotted through the door and a long pipe with one guy hanging off it. Had a 2" pipe under the center of the rear of the chassis.

Kept increasing the length of the pipe until we got the rear corner to move down more than 1/2", was a fair length of pipe for that one too, chassis copied from a Radical with additions.

Knew the weight of the guy (had a set of kitchen scales) and the length of the pipe and the 1" of movement - simple bit of math and we had our result.

And that was all for the CAMS Technical guy standing there who approved the method.

Previously I have dyna bolted the front or rear down, I didn't think the door would hold but it did

[cheapracer]

Ok we have above;

a/ Sylva Riot

b/ Unusual homebuilt Locost

c/ Carleton Uni FSAE (guess how I knew that one!)

d/ MEV Rocket

What do I win?

[Greg Locock]

The yellow one is good enough, and I really like that he's measuring the deflected shape along the rails, essential for finding out where your FEA is wrong.

[mariner]

But they tell us that FEA is never wrong!!

[mariner]

Puting aside my flippant copmment above I do have a serious question on test rigs.

How do you measure the suspension attachment/bracketry stiffness?

I can see the benefits of the dial gauges along the chassis to find out where the flex is happening but I suspect that some of the devil is in the attachment detail so to speak.

These bits are hard to test even off the car because the parts are small and the loads are at their maximum as you feed in the suspension/spring loads. It is a guess but I can see a large proprtion of the deflection coming from areas close to the mounting points. Is this true?

[rachael]

Puting aside my flippant copmment above I do have a serious question on test rigs.

How do you measure the suspension attachment/bracketry stiffness?

I can see the benefits of the dial gauges along the chassis to find out where the flex is happening but I suspect that some of the devil is in the attachment detail so to speak.

These bits are hard to test even off the car because the parts are small and the loads are at their maximum as you feed in the suspension/spring loads. It is a guess but I can see a large proprtion of the deflection coming from areas close to the mounting points. Is this true?

This test can only give you an overall feeling for the torsional stiffness and the extra dial gauges along the chassis only give you a clue as to which area to concentrate your efforts to improve the stiffness. There are definitely some pretty horrible ways of attaching brackets to tubes both for strength and stiffness and only an experienced eye or a good fea analysis is going to be able to sort the wheat from the chaff.

[DaveW]

Puting aside my flippant copmment above I do have a serious question on test rigs.

How do you measure the suspension attachment/bracketry stiffness?

I can see the benefits of the dial gauges along the chassis to find out where the flex is happening but I suspect that some of the devil is in the attachment detail so to speak.

These bits are hard to test even off the car because the parts are small and the loads are at their maximum as you feed in the suspension/spring loads. It is a guess but I can see a large proprtion of the deflection coming from areas close to the mounting points. Is this true?

For a typical open wheeled vehicle, half the hub-hub compliance would be expected to be chassis compliance, & half suspension (including rocker posts, wheel bearings, etc.). A four post rig can be used to estimate the total & the chassis contribution for a complete, assembled, vehicle in one hit....

[cheapracer]

I can see the benefits of the dial gauges along the chassis to find out where the flex is happening but I suspect that some of the devil is in the attachment detail so to speak.

.... also just putting a long pipe in different places all over the place can show up some interesting things.

I believe Frank Matich would tie bits of string in selected places, bolt 3 corners down alternately and jack the 4th corner up and see what the string did.

Oh and if you just run beam axles front and rear then you don't have to be so worried about it all

Edited by cheapracer, 29 November 2010 - 17:18.

[PhilG]

But they tell us that FEA is never wrong!!

Oh dont, please, my ribs are still sore and i cant laugh.

The main reason FEA fails is the inability to establish the load case, or more importantly, the inability to know what is actually needed .

[rachael]

Oh dont, please, my ribs are still sore and i cant laugh.

The main reason FEA fails is the inability to establish the load case, or more importantly, the inability to know what is actually needed .

That's not a reason to not use it. It's just like any other tool, you have to know how to use it properly and understand the implications of any assumptions.

[Greg Locock]

I believe Frank Matich would tie bits of string in selected places, bolt 3 corners down alternately and jack the 4th corner up and see what the string did.

Mazda do that as well.

On my old website I posted an exercise a couple of us did on a 1955 Tbird chassis, 3 pillars and a lot of sandbags, measuring the deflection at many points.

The comparison with the FEA was eventually really good, although it took me a while to identify all the problems with the FEA (mostly joint efficiencies).

BUT a ladder chassis is a lot easier to model than a monocoque or unitary build car. A spaceframe should be easier than a ladder frame because the joint efficiency is generally better - I had to model some of the joints in detail to find out how exactly a C section feeds loads into a box, etc.

Edited by Greg Locock, 29 November 2010 - 22:28.

[mariner]

Long,long,long ago one of the people who built the first Lotus chassis' suggested that the relative longevity of those early chassis was due to using brazed not welded joints. They suspected that the flexibilty of the brazing allowed just enough movement to spread the loads across multiple tubes and hence reduce local failures.

Could the same argument apply to a modern chassis i.e if you make one joint so rigid that it transmits all the load down one critical path then there might actually be more overall deflection than allowing some more flex at the original joint so the load is more evenly split. For example putting some of it up thru the roof as opposed to feeding everything into the side rails from the front suspension towers.

Is this one of the ways FEA is used to minimise overall weight versus stiffness?


If so how do you degrade a particular joint to see if it helps overall , is there some target seeking tool like meet XXX deflection between wheel lines in the programs?

[cheapracer]

Long,long,long ago one of the people who built the first Lotus chassis' suggested that the relative longevity of those early chassis was due to using brazed not welded joints. They suspected that the flexibilty of the brazing allowed just enough movement to spread the loads across multiple tubes and hence reduce local failures.

My business partner in the late 80's entered the first year of the Diahatsu Charade Rally Challenge in Australia and rather than seam weld it all up we prepped it by filling up the sills and pillars and any section we could with 2 pack expanding foam - most of the shells developed cracks to the point of requiring reshelling at the end of that year but ours and others who didn't seam weld had no issues at all (and they all used the same spec supplied cage).

[rachael]

Long,long,long ago one of the people who built the first Lotus chassis' suggested that the relative longevity of those early chassis was due to using brazed not welded joints. They suspected that the flexibilty of the brazing allowed just enough movement to spread the loads across multiple tubes and hence reduce local failures.

Could the same argument apply to a modern chassis i.e if you make one joint so rigid that it transmits all the load down one critical path then there might actually be more overall deflection than allowing some more flex at the original joint so the load is more evenly split. For example putting some of it up thru the roof as opposed to feeding everything into the side rails from the front suspension towers.

Is this one of the ways FEA is used to minimise overall weight versus stiffness?


If so how do you degrade a particular joint to see if it helps overall , is there some target seeking tool like meet XXX deflection between wheel lines in the programs?

In welding the Heat Affected Zone is likely to be much worse that brazing. The material becomes more brittle in the HAZ than the parent material and is thus much more likely to crack.

If you 'make one joint so rigid that it transmits all the load' then you have most likely made a stiffer structure than one where you have multiple load paths. In good design you try to minimise discontinuities in the load path i.e. severe changes in cross section or right angle bends as these induce stress concentrations that start fatigue cracks, occassionally you might weaken a load path in order to relieve an over stressed feature but it usually means the overall design is poor.

In a structure like a spaceframe FEA is good at showing which tubes are working to support the load and which are less useful. If you increase the gauge of the tubes carrying the load and minimise the gauge of the less useful ones you get a stiffer and lighter space frame. FEA is less good for directly showing where adding additional tubes would increase the stiffness - you need to view the deflected shape and look for large relative deflections.

There are fea optimisation programs where for a given package space and suitable design constraints they will produce a structure that will meet a target stiffness for minimum weight - google Optistruct for one such program.

[Tony Matthews]

In welding the Heat Affected Zone is likely to be much worse that brazing. The material becomes more brittle in the HAZ than the parent material and is thus much more likely to crack.

If you 'make one joint so rigid that it transmits all the load' then you have most likely made a stiffer structure than one where you have multiple load paths. In good design you try to minimise discontinuities in the load path i.e. severe changes in cross section or right angle bends as these induce stress concentrations that start fatigue cracks, occassionally you might weaken a load path in order to relieve an over stressed feature but it usually means the overall design is poor.

In a structure like a spaceframe FEA is good at showing which tubes are working to support the load and which are less useful. If you increase the gauge of the tubes carrying the load and minimise the gauge of the less useful ones you get a stiffer and lighter space frame. FEA is less good for directly showing where adding additional tubes would increase the stiffness - you need to view the deflected shape and look for large relative deflections.

There are fea optimisation programs where for a given package space and suitable design constraints they will produce a structure that will meet a target stiffness for minimum weight - google Optistruct for one such program.

[cheapracer]

Rachael Welch?

[Tony Matthews]

Don't lower the tone, Cheapie (Tone - geddit), just let me dream...

[Greg Locock]

Long,long,long ago one of the people who built the first Lotus chassis' suggested that the relative longevity of those early chassis was due to using brazed not welded joints. They suspected that the flexibilty of the brazing allowed just enough movement to spread the loads across multiple tubes and hence reduce local failures.

Could the same argument apply to a modern chassis i.e if you make one joint so rigid that it transmits all the load down one critical path then there might actually be more overall deflection than allowing some more flex at the original joint so the load is more evenly split. For example putting some of it up thru the roof as opposed to feeding everything into the side rails from the front suspension towers.

Is this one of the ways FEA is used to minimise overall weight versus stiffness?


If so how do you degrade a particular joint to see if it helps overall , is there some target seeking tool like meet XXX deflection between wheel lines in the programs?

You are confusing stiffness and strength(durability). I think that it would be impossible to design structure that would be stiffened by reducing the stiffenss of a joint, but I agree that structure's durability can be improved by that. For example the original Polo's twistbeam had a little hole in the gusset plate. Using a solid plate instead, and carrying the weld all along the gusset resulted in broken twist beams.

FEA optimisation tools do exist, in fact for the Tbird exercise I wrote one in Matlab because the commercial ones didn't do what I want, as my model used beam elements, not plates.

Here's a pretty typical commercial optimiser http://www.altairhyp...OptiStruct.aspx

Um, above written before I read Rachael's reply. What she said.

Edited by Greg Locock, 05 December 2010 - 21:23.

[gruntguru]

Don't lower the tone, Cheapie (Tone - geddit), just let me dream...

The "tone" has risen a lot since Rachael appeared - interesting.

Nice photo Rachael - you're not really that attractive are you?

[rachael]

The "tone" has risen a lot since Rachael appeared - interesting.

Nice photo Rachael - you're not really that attractive are you?

This being the technical forum I thought you guys might be more interested in what I had to say than what I looked like - apprently not. I will admit that it is quite an old photo - please forgive my vanity.

[gruntguru]

This being the technical forum I thought you guys might be more interested in what I had to say than what I looked like - apprently not. I will admit that it is quite an old photo - please forgive my vanity.

Interested and impressed by what you had to say.
Are you saying you were more attractive when you were younger?
What's a tiny bit of vanity in the face of such honesty?

[Tony Matthews]

This being the technical forum I thought you guys might be more interested in what I had to say than what I looked like - apprently not.

Not true for me, as you know, but the picture is a bonus. The one of you, too! :

[cheapracer]

This being the technical forum I thought you guys might be more interested in what I had to say than what I looked like - apprently not. I will admit that it is quite an old photo - please forgive my vanity.

My gosh, we're human! But I am interested in anything you have to write that I agree with Rachael.

Actually all I can see after just getting back from a short walk in the mountains (2 days all either up or down stone steps) is an old red decrepit race car, darn it.

[rachael]

My gosh, we're human! But I am interested in anything you have to write that I agree with Rachael.

Actually all I can see after just getting back from a short walk in the mountains (2 days all either up or down stone steps) is an old red decrepit race car, darn it.

The old race car is taken from a painting of a Ferrari 330P4 by Tim Layzell.

I get the feeling from you lot that female + engineer + does not look like Clare Balding equals something near impossible to comprehend, which is typical but never the less annoying! Perhaps this will ease the comprehension; I can't be a proper engineer because I sit in front of a computer all day where there is no risk of getting oily or breaking a finger nail.

Could we get back on topic now.

[cheapracer]

Could we get back on topic now.

No.

[Tony Matthews]

no risk of getting oily or breaking a finger nail.

That's a mechanic. Many years ago I heard that girls were being put off engineering as a career because the image they had was of a man wearing a boiler suit, carrying a grease gun.

[Canuck]

I'm willing to admit that "rachael" and "she" didn't connect until Greg made his statement. Totally escaped me, what with the general lack of actual proper names (Tony and Grunt being the obvious exceptions). I will further admit to profile checking after that conclusion. Either way, I was already well intrigued by your posts - I'm always curious and envious of those doing (or appearing to be) what I wish I'd done .

[Magoo]

Could we get back on topic now.

Indeed. It has been rather uncomfortable to watch.

http://www.theonion....-by-2018,14400/

[cheapracer]

Indeed. It has been rather uncomfortable to watch.

http://www.theonion....-by-2018,14400/

As usual and in full NASA tradition, the estimated costs of that project are out of this world.

[Magoo]

As usual and in full NASA tradition, the estimated costs of that project are out of this world.

Sure. The Chinese space agency could do it for one-third that amount.

[Fat Boy]

In welding the Heat Affected Zone is likely to be much worse that brazing. The material becomes more brittle in the HAZ than the parent material and is thus much more likely to crack.

Some of this is going to be a product of the materials being used. Low carbon steel doesn't tend to have much of a HAZ because there isn't enough carbon to make one. A 4130 chrome-moly frame might have HAZ issues to deal with that are fairly easy to remedy with a torch afterwards. Brazing can be a very good answer to frame building, though. You can use very high quality steels if you want because since you aren't actually melting the parent materials, you don't get the HAZ problems you do with welding. Also, generally, you can use thinner wall tubing (for a given fabricator skill level) and it has the convenience of a nice fillet at the tube junction with any sound joining. I'd have no problem running a brazed chassis as opposed to a welded one.

[Tony Matthews]

As usual and in full NASA tradition, the estimated costs of that project are out of this world.

Not much compared to a some flowers, a meal and a lifetime of servitude.

[ralt12]

These guys are the mothers of all stress testing...check out the wings....
Now that's a test rig!

Edited by ralt12, 12 December 2010 - 05:25.

[cheapracer]

Ralt, thanks for that, simply incredible!

I had known for the bulk of my life that Jumbo wings were stressed with a deflection of 40 feet (40 feet ffs!!) but until today I had never seen a picture, made my day - cheers!!

(yes I know it's not Jumbo).

[mariner]

What is even more impressive is that the fuel tanks are in the wing sealed only with "rubber" and they don't leak.

Each time I sit in a jumbo ( or any big long haul plane) I stil get the childish thrill from watching the wings come up as the thing accelerates down the runway. Leaving Hong Kong's Kia Tak made it even more exciting as apparently nothing but harbour was left on each side.

[gruntguru]

When turbulence seems bad, I always look at a wing tip and take great comfort from the knowledge that wings are designed for several times the deflection I am seeing - without failure.

[gruntguru]

I get the feeling from you lot that female + engineer + does not look like Clare Balding equals something near impossible to comprehend

No - but a rare beast and a very desirable combination. (I do mean that in the nicest possible way.)

Could we get back on topic now.

OK (without prejudicing my right to have the last word).