10 replies to this topic

[mariner]

Having finished my " much modified corvette" I would like to check the torsional rigidity, partially to see if it is stiff enough to get the actual spring rates the same as the designed ones and partially to see how good a design/construction job I have done!

I know you can do it with software but I can't afford it/use it. I know you can do it with an external frame and lots of dial guages but I dont have time/space. So my idea is as follows.

1) mount two tall vertical posts on diagonal corners of the chassis next to the suspension and spring attachment points.

2)mount my hardware store laser on one vertical post and a graduated target on the other post on the diagonal corner and align the laser spot with one of the graduations

3) Jack up the corner with the laser to a pre determined suspension compression ( say 75% of total bump) and measure the movement of the laser spot on the graduated target.

4) As an alternative, and going right back to old fashoined ways, tension a piano wire between the two posts and then fit two little rollers on each post above it with weights hanging down to pre tension this second wire.

Then as the chassis flexes the fixed wire will droop and the gap to the pre tensioned wire on the rollers will grow. Hopefully calipers and then a micrometer will be able to measure the change in gap and hence deflection.

Would the real engineers here tell me if they think this crude method will give reasonably accurate and repeatable results.

[Engineguy]

If you're trying to measure actual frame flex (as opposed to suspension movement) You have to get the suspension and tires completely out of the picture or you'll not know where movement is coming from.

[Greg Locock]

look around for the sae paper on the winston cup racer, or any fsae report. There are at least three relevant links in the library.

Your method may work but it is a bit odd, you aren't quite measuring what you think you are, I think you'll also be measuring some vertical bending.

Goran's forthcoming book shows how he did it. (Plug)

I'd rather fix one end to the ground at the shock or spring attachment points, and then build a rocking arm at the other end with a fulcrum at a tbd height on the centreline, picking up from the spring towers. This has the great advantage that there is only one mechanical bit to get right, and you are working at one end of the car.

I'm not being funny about TBD, you will get different results depending what height you select. I hope this is a small effect.

Logical values for TBD - ground level, wheel centre height, or level with the spring towers, or the RCH, or the CG. The lower locations are perhaps more realistic, the higher one is more pure torsion.

[Engineguy]

Originally posted by Greg Locock
Logical values for TBD - ground level, wheel centre height, or level with the spring towers, or the RCH, or the CG. The lower locations are perhaps more realistic, the higher one is more pure torsion.

I'd be afraid that ground level or RC (assuming it's pretty low) would result in just bending the chassis sideways as well as/instead of twisting it... I guess that could depend on how you constrain the rear and or opposite front side. Intuitively axle height or CG height seems about right... I think you would want to picture sorta the average frame centroid height... easy to visualize for a backbone or ladder frame, not so easy once you get a roll cage and/or unibody shell in the picture.

[mariner]

Thank you for the advice and help. I do recognise that Greg's sugested method of constraining one end and applying a load at the other is the proper way but I am trying to take a short cut ( probably unwisely but..).

I also forgot to add that I do realise have to constrain the chassis at one point or I will just lift up one side!

I am not primarily interested in the torsional rigidity as a numeric value but a practical question of whether the chassis is stiff enough to make the springs and dampers work as designed. So my idea was to measure only the distortion between the spring/damper mounting points in comparison to the rising load on the springs. This is why I would propose to jack up the hub carrier to 75% or 80% of full bump. If there is no measurable deflection between my two vertical posts then the chassis is stiff enough. If there is significant movement then I must go to lower spring rates and add more ride height etc. since the springs/dampers aren't working as intended.

I know this may seem very crude and basic but since the car is a one off and largely built the only things I can do are to recognise that the chassis is not stiff enough to allow high spring rates or try to see where across the path from spring to spring ( and A/R mount to A/R mount)the chassis deflection is occourring in case it is localised deflection which can be solved with a modification. To determine that I plan to bolt the chassis to the floor in the classic way at several discrete distances from each corner and re - measure the deflection. So if the deflection is , say halved when I constrain the chassis 30cm from the front but stays the same when I constain 30cm from the rear then my torsional weakness problem is with front spring mounting area and not elsewhere.

I know this is not the engineers way but does it make any sense?

[Greg Locock]

Actually it is a perfectly sensible test - in fact I'd go further and secure the body at three of the four corners and just measure the vertical stiffness at the fourth corner. If you look on my website you'll see some test results where we did exactly that on a ladder frame chassis. This eliminates all problems with fancy pants mechanisms. Word of warning - ideally you'd use a spherical joint between the supports and the body.

This will give you an effective local stiffness of the body for vertical inputs. You can then compare this with your spring, and sta bar rates to see if they are working eficiently.

In theory you might be able to work out how much damper force you are losing in this flexibility of the body - but not from just a static measurement.

[RDV]

Another suggestion for a quick and dirty way is to use 1mm shims under setup wheels(rigid), with the whole assembly on scales.

Measure the shims accurately (.05 difference in thickness will give you a 5% drift...), and check the scales also...most scales give tenths of a kilo, so try to build in the equation a max/min from measures.(eg corner weight delta/atan(deflection/track(load application span)).

I have used this in all sorts of cars, and is reasonably accurate, also has the advantage of not constraining anything, so the beam stiffness does not enter into calcs, plus the only thing you need is a reasonably flat floor, rigid setup wheels and a set of corner scales, no rig required.Any other method has the downside that the rig itself could stiffen up what you are measuring.(If no setup wheels a set of empty rims will do, as long as you set it up on the dip in the middle to avoid rim deflection...)

Procedure is to shim front left, noting corner weights and deflection, then shim rear right...these values should be very close as effectively you are applying a clockwise twist(seen from rear)...repeat for front right and rear left.

Note values as you add shims and remove, noting your residual values when back to zero shims,this will give you your hysteresis in measuring.

The stiffer the car the less deflection to be measured before you lift a wheel, at which point cease to measure.
A stiff carbon chassied single seater weighting 650kg will give you about 8 or 9 mil of deflection, around 30Kg/m per degree of stiffness, plus or minus 2%...should be enough, and a heavier/softer car will give you a bigger range.

This measures total deflection including suspension, but excluding tires, so if you want chassi only modify the the system by having struts on the damper/spring mounts onto the scales.

Dont forget to measure all shims together, as sometimes 5 times 1mm shims is slightly bigger than one 5mm shim....;)

Repeat at least 3 times, as you WILL get drift, and the residual will show you micro movements in whole system...PM me and I can send you an example speadsheet with values and procedure.

[Greg Locock]

That is a terrific way forward. As an enhancement, then replace your shocks by a steel bar with sph joints at each end.

If you then compare the results you will either be confused as hell or educated. There is not much difference, except time.

[RDV]

..er yes...dummy dampers also....

[johnny yuma]

Assuming the car is fully assembled with both front springs the same rate,ditto rear springs,tyres pumped HIGH same pressure,very level floor,I reckon for a ROUGH ANSWER as to "workable" or "hopeless" just place a floor jack under a rear CHASSIS corner and jack in tiny measured movements until the corresponding opposite rear chassis point point starts to rise.If its 100% stiff it will move straight away.If you jack 30cm before the other side moves-not real stiff.

[Goran Malmberg]

Its okej trying a fast way of measuring torsional rigidity. However, with all my testings in mind the problem is
to prevent bindings, an other flex that create false readings. Some sort of "floating" restrain points for both the chassis and force application points is good, to make sure we got twisting only forces comming in to the chassis over the mesuring points. Not to loose movements or forces in to bindings.
Depending on how hard we dare to load the chassis the ovarall movements (and twisting) of the chassis will become accordinly large. So, in short, we got a balance act between bindings (solid mount) and floating (result in mounting flex both ends) chassis mount. For an amateur rig I recomend the "floating way" and the use of one very exact scale blub lever guage each end and measure the difference between the two.
Goran.