Scale levelling process

[Rick Kirchner]

Does anyone have a scale leveling. process that they would share? I've searched but there's nothing particularly specific on the forum.

Seems like it should be simple, but I've been fooling around all day and can't get it to converge. Should have taken surveying in college....

Setup is our scales on four separate longacre levelers.

I pick one pad, make that the reference, then move from there to the others. I use 123 blocks, a square steel tube, and a machinist's level good to .0005/ft. Measure how far off it is by placing feeler gauges under the beam, then calculating how many 1/4ths of a turn of the adjusters, re-check with the level.

When I get them all level from the reference, I check the cross and the side-to-side across the end that doesn't have the reference, and I'm at least .040 off on one of them.

I've tried twice around, three times around, going up halfway on one end and down halfway on the other, and finally picked a point in the center o the pads, put up a stack of reference blocks, and leveled each pad from the center. Basically the same result.

Driving me freaking crazy....

[Purple Frog]

.040 could be as much as grit under the feet of a scale... just saying.

I level a scale, then stand on it. Check it again. My 200 pounds will crush the small stuff.

In a jam, its mostly important that they are level left to right. Front to back, not so much.

Three tries... too soon. LOL

Also, that steel beam... are you overhanging it the same over each 1-2-3 block? If it sticks out 2 feet over one, and only 3" over the other, that could be your .040.

[R. Pare]

Also, mark your bar both for where the level will sit every time, as well as the orientation of the bar itself - eg - which face of the bar faces forward each time.

Perfect levelness is not as important as the elimination of changes in that error from one pad setup to another. As long as the levelness ( or out-of-levelness) is exactly the same every time you set the scales up, you will be better than "close enough". It is when you get variances in the pads from one setup to another that will end up screwing you.

Also, for the majority of cars, you really need only scales under the front wheels - it is that end of the car that needs to be set square ( or off-square for certain tracks), and the evenness of that end will determine where the back end ends up.

[Rick Kirchner]

I'm chasing some kind of stiction/crack/loose part problem where it takes more rear pushrod to move weight than I think it should - and then it's not consistent. My hillbilly platen has a bit of give in it, so I wanted to take the variables out. Borrowed these steel Longacre's from Art Smith. Must have been designed for a stock car because they are hell for stout.

But slightly bigger than my usual setup, nice machine feet instead of plain old bolts. I figured it would level right up.

To y'all's points - beam is marked in the center for the level. I try to keep the ends fairly equal - I have a 60" run, a 97" run, and a 144" run. the level is in the middle so I just kind of eyeball the center off the space between the pads.

since it's cross weight where this shows up I'm trying to get as flat as possible.

[Bob L.]

Not sure you can reply on sq. tube and level not to introduce error due to sag, moving around, and not being "machined".

I've done it 2 other methods...

First and oldest was to do with tygon tubing and water.. make a manometer/water level gauge.... leg up at each pad to get a level, fill system, get all bubbles out (bigger tubing at base helps, could use hose or pipe to connect up vertical legs at corners..) Work off meniscus. Think method was in Fred Puhn's old book on set up.

Other was to use laser level(s) and shoot down length, across. They make rotating laser levels to put in suspended ceilings, think you can rent those at Home Depot/Lowes, seen in some pawn shops. Not tried, but sure seems like a good way.

Believe, you're trying to do more than those parts will allow...doesn't take much to be out.

(FYI. I used to do machinery alignment in the field.. first thing learned was sag in setup is a killer..next was temperature on turbo machinery makes a good stable reference tough...had to build water cooled towers to mount proximity probes so as, most error was dialed out of ref. points.)

Bob L.

[Rick Kirchner]

I know droop is there. The level is about 4 lbs I guess, and it's sitting on a tube, not a I beam. I'm just betting on being in the center of the (very slight) caternary.

I'd think with a manometer you are limited to your visual resolution, somewhere between a 16th and 32nd.

Lasers - That was my business for many years in the Navy. The problem there is that any laser you can afford has a spot size larger than 1/32 at 6' or so. And getting the laser level itself is extremely hard to do. Certainly possible with lab grade equipment, but not Home Depot contractor grade.

Years ago I built an ART - like laser leveler, tossed all the parts after first use due to the issues described above.

Last time I did this I made these same measurements, drew out a little diagram, looked at the numbers and AHA! A solution just came to me. Should have written that down.....

Here's what's driving me crazy:

B is the reference.
A-B is level
B-D is level
A-D is level so with these we have the proverbial 3-legged stool

Then leveled A-C

After this C is .025 higher than D, and C is .038 higher than A. ARRRRGH.

Being picky of course because with 1000 lb/in springs it doesn't take a lot to put you well off...

[jtaphorn]

I use a Bosch Laser level. Relatively quick exercise.

https://www.homedepot.com/p/Bosch-Se...B&gclsrc=aw.ds

[alangbaker]

Being picky of course because with 1000 lb/in springs it doesn't take a lot to put you well off...

Rick...

The spring rate doesn't matter.

What is the WHEEL rate?

[mark elder]

Put a thick washer (or coin) in the exact middle of each pad (same thickness...), try again.

Whenever I have had this, the actual pads were not perfectly level. It does not take much and the error gets amplified by the distance to the next pad.

I have also had a level that was not centred (bubble sits in a different place when turned around). Again, the error gets amplified as you work your way around.

I also agree with Mr Frog, the front pair just need to be parrellel to the rear pair. Slope between front and back is too small to be important to us (I am assuming your garage floor is approximately flat). Using this approach, it does not matter if your level is not centered, as long as the left end is always on the left (ie any error neutralises itself).

Let us know the outcome.

[EricP]

...I just want to say that this makes me feel much better... I think I spent two full weekends last summer trying to get my setup pad level and pulled out the little hair I have...

[Vne165]

Water level (5/16"ID clear flexible tube), Mk1 calibrated eyeball and a ream of A3 paper, sheets under each scale as required.
Seems to work ok for me. I'm always careful of overthinking such matters.

[problemchild]

Whenever I have had this, the actual pads were not perfectly level. It does not take much and the error gets amplified by the distance to the next pad.
.

I have a fancy homemade jig with machinist level that I put across the pads but use a digital level to level each pad individually. With my pad setup, I don't worry about each scale laterally, but am very diligent in getting each pad level longitudally. This is more important if your car's wheelbase does not match the pad length. Ideally, you should get the same values, if the car is 3" forward, or rearward, as when centered on each scale.

I cannot eliminate the scale reading differently as car placement creeps out of my target patch, but have found, as Richard says, that using the same process, with the same faults, but being diligently consistent, I can align a fleet of cars with complete consistency. It is not like the driver will ever tell the difference between 50.7% and 50.8% CW. I doubt most drivers at the club level can tell the difference between 50.4% and 51.0% CW.

Even on my winged cars, I am not worried about end to end being under 1/8" in variation, if each end was level side-to-side with level scale pads.

I have no idea if my pad setup is under your 1mm variation. I do know that I set it up 15-20 times per year at various racetracks (in less than an hour) and get complete consistency in my alignments, and my cars.

[S Lathrop]

I use a transit level and a stand that has a machinist scale that is used as the target. The magnification is enough to use the edges etched lines on the scale to line the cross hairs of the scope. The resolution of the system is less than 1/64 of an inch. I set the transit level up 10 to 20 feet behind scales and then shim all the scales to the same dimension on the scale. This was common practice among the Indy car teams decades ago.

This is the beat system I have seen but I don't know if the additional accuracy it allows is worth the cost of the equipment for club racing.

[Ralph Z.]

Is .040" relevant? If so, mine is always off.

[Purple Frog]

Like Ralph said and Greg expressed; I don't think my ass-o-meter would feel .040" going into turn 1. Just saying.

edit: Steve might have just proved me wrong in post below.

[S Lathrop]

Is .040" relevant? If so, mine is always off.

You can answer the question your self. Say you spring rate at the wheel is 200 pounds per inch. It will take 8 pounds to move the wheel .040". How close do you try to balance the weights on your car?

[stonebridge20]

Been using these for 35 years with some success.

[Ralph Z.]

You can answer the question your self. Say you spring rate at the wheel is 200 pounds per inch. It will take 8 pounds to move the wheel .040". How close do you try to balance the weights on your car?

OK. I get it.

[LFFR]

Where is the track surface within .040" of level?

John Mihalich, Jr.

[chrisw52]

Been using these for 35 years with some success.

I have been using this for a couple years now. I will use this to get the pads level across all four pads, then verify with a 6' beam bubble level gauge.

[RussMcB]

86 cents per year. I guess that's not too bad.

Been using these for 35 years with some success.

[Rick Kirchner]

In the past I used a couple of cheap bubble levels to level each pad, and then the machinist's level to go after the whole setup. This time I did what Greg Rice does - I used a Accu-Level electronic device on each pad and then the machinists level setup.

Then Art Smith dropped by with his 18" Starrett machinists level, and we spent about an hour and a half on one pad and found out how far of it was despite the Accu-Level. And here's where it get's interesting: We could get it level from front to back - and the diagonals would be off, and vice versa. Essentially if we could get two of the directions level, the other two would be off. We had to be diligent to watch for the "teeter-totter" where three of the pads would have weight on them and one little adjustment later one pad is hanging and dorking up the results.

If you think about it, the single pad problem is just a microcosm of the larger platform problem, but it should be easier because all the corners are actually connected.

We stuck a 123 block in the center of the pad, but that didn't make a difference. I do believe we are looking at the problem that Mark Elder describes.

The bottom surfaces of these scales are machined, but the tops are as-cast, and have a deep texture pattern in them - except for the Reb-co logo in the center, which is where we are trying to make the measurements from - but the level of course spans the entire pad.

I will try the washer method today. Also going to call longacre and see what they recommend.

In my prior professional life I ran a group of engineers that did a lot of photogrammetry - which is not just adjusting from a sort-of-round earth to flat screens or projecting sensor bearings into targeting parameters but also things like calibrating a four-post camera based missile detection system where the corners are 7 miles apart. The cameras have to be able to look up into the sky and see the same point in space with a relative accuracy of just a few feet (while correcting for the individual spherical aberrations in the lenses). That was all done with GPS locations and a transit and then a bunch of optical processing. I still think this problem is one where if you can set up a reference and measure the corner of each scale WRT that reference, it's just a matter of putting numbers into an algorithm and having it spit out the adjustment for each corner.

[APEowner]

I use a Bosch Laser level. Relatively quick exercise.

https://www.homedepot.com/p/Bosch-Se...B&gclsrc=aw.ds

This is exactly what I use. I leave the level setup when I'm scaling and use it as my ground reference for ride height as well.

[Daryl DeArman]

If you start with a level platform and put 350# on the right rear scale and 225# on the right front scale how far out of level is your platform now?

[DaveW]

I'm chasing some kind of stiction/crack/loose part problem where it takes more rear pushrod to move weight than I think it should - and then it's not consistent....

IMO, one major caveat in all this is what Rick mentioned above. I routinely check suspension friction in a very simple way.

Step 1. Push down on one end of the car as far as it will go and then gently release the load. Then measure the shock length.
Step 2. Pull up on the same end of the car as far as it will go and then gently release the load. Measure the shock again.

The difference in shock length between 1 & 2 times the spring rate divided by 2 [(L1-L2)x(rate/2)] (for 1 wheel) gives you the suspension friction force at the shock. Then multiply by the motion ratio. This gives you the friction force at the wheel. If this is more than ~5-lb, not only will you have issues getting a good cross-weight measurement, you will be giving up grip. Excessive friction kills grip.

By doing this with the opposite end of the car jacked up in the middle you can also see if one side has more friction than the other - the stickier side will rebound less when you release the force you push down or pull up with.

I have no issues getting the front cross-weights equal on my Citation within a couple of lbs.

[stonebridge20]

86 cents per year. I guess that's not too bad.

My first set was given to me by Joe Stimola. I had to replace them after a really nice lady ran my scale pad and levels over at Lime Rock about 15 years ago with her Suburban.

[BeerBudgetRacing]

IMO, one major caveat in all this is what Rick mentioned above. I routinely check suspension friction in a very simple way.

Step 1. Push down on one end of the car as far as it will go and then gently release the load. Then measure the shock length.
Step 2. Pull up on the same end of the car as far as it will go and then gently release the load. Measure the shock again.

The difference in shock length between 1 & 2 times the spring rate divided by 2 [(L1-L2)x(rate/2)] (for 1 wheel) gives you the suspension friction force at the shock. Then multiply by the motion ratio. This gives you the friction force at the wheel. If this is more than ~5-lb, not only will you have issues getting a good cross-weight measurement, you will be giving up grip. Excessive friction kills grip.

By doing this with the opposite end of the car jacked up in the middle you can also see if one side has more friction than the other - the stickier side will rebound less when you release the force you push down or pull up with.

I have no issues getting the front cross-weights equal on my Citation within a couple of lbs.

For those of us always learning this stuff: Wheel/tire on the ground when you do this or a frictionless surface.

[R. Pare]

Where is the track surface within .040" of level?

Not even relevant to setting the car up - it is the CAR that has to be set accurately and as close to exactly the same every time in order to not introduce unwanted handling inconsistencies.

As a quick "for instance" in what sort of repeatable accuracy you want to achieve. quite a few years ago with an FF using bias tires, we found that at MO, if the corner front corner weights were set the same, the LF tire would grow more than the RF, throwing the handling off starting about half way through the session or race. We then switched to setting the car up even, and then reducing the LF weight via the pushrod by 12 pounds. The RF tire then grew larger than the LF, making the handling even worse. Reduced to 8-10 pounds, and the tires grew evenly, and the car "came to the driver" more and more throughout the race.

In other words, just a few pounds of inconsistency will screw the handling up,and if you cannot figure out why because the scales are never the same, you are basically screwed.

[Rick Kirchner]

Made a couple of calls to some other Indy Car buds - couldn't get an actual written procedure out of them, but their practice follow what many have said here. I was surprised that cross measurements were't done in practice, but it seems that might be a bridge too far.

Called Longacre, nothing written, but a good conversation - screw all the pad levelers in, find the highest pad with a level and start from there using it as the reference. they thought that with the pads I have, there will be issues because the measuring capabilities of the levels exceeds the manufacturing accuracy of the pads.

Back out in the garage, trying to level the reference pad. I had to get a washer larger than the v-groove in the bottom of the level. This produced considerable rocking. Went to the 123 block on the 2x3 plane - this reduced the rocking considerably but still there. Eventually I took some large files and started going after the surface and filed down a bunch of high spots. Reduced the rocking probably 90%. In the end I can get it level within .015 over two diagonals and either the front to back or the side to side, but not all. Just going to have to call that good enough.

After this my scales are going to the machinist....

By the way - check your scales for buggered edges. I had to take a file to all of mine yesterday as there were a lot of dings and burrs from use.

[R. Pare]

I
The bottom surfaces of these scales are machined, but the tops are as-cast, and have a deep texture pattern in them - except for the Reb-co logo in the center, which is where we are trying to make the measurements from - but the level of course spans the entire pad.

I've never understood why some of the manufacturers make to top surface of their scale anything but perfectly flat and smooth - all that the fancy texturing does is make it impossible to be sure that the tire has the same amount of surface area touching the texturing, which in turn will change just how far into that texturing the tire sinks, in turn screwing up the suspension positions and the actual corner weight.

[EACIII]

The tops of mine are slightly convex. I have little shims I use to level the laser level I use to go between scales.

Ed

[Steve Demeter]

Water level (5/16"ID clear flexible tube), Mk1 calibrated eyeball and a ream of A3 paper, sheets under each scale as required.
Seems to work ok for me. I'm always careful of overthinking such matters.

I use more or less the same technique except measured difference in height of water in each tube. Found the highest one then made shim blocks to raise the other 3 up to the same height.

Marked the spots on the floor. And each time I use the scales, put the blocks down and shoot across them with a cheap laser lever to verify nothing has moved.

[Daryl DeArman]

Sounds like you've solved your issue, at least for now.

I was going to suggest you start with "C" as your reference pad since it is clearly the high spot and bring everything up to it.

There is simply no way you can have:

LF to RF level
LF to LR level
RF to RR level

and not have LR to RR and both diagonals level. If you do your "level" standards are off more than your measurement standards.

Don't make it more complicated than it needs to be. Sometimes best to just walk away and come back with a clear mind and fresh eyes.

[Robert J. Alder]

My shop floor has drains all over the place. A nightmare ot set up four Longacre scales on the floor. Not a level spot anywhere. Built this rig with adjustable feet on the four ends of the rails (2"x2"x1/4" wall alum.). Just used a SmartLevel (Already had that camber tool) to level the rails (end to end). Then sent the two trays (3"x3"x 1/4" alum angles with 1/8" alum plate between.) on the rails. Used the SmartLevel to get both trays level side to side while on the rails. All easily done with the adjustable feet. DONE. Tape measures with a sticky side (God love Amazon) were taped to one rail for easily locating the trays for correct wheel base, and then one tape each put on the trays for easily setting the Longacre scales on the tray for various front & rear tracks. The adj. feet on the scale pads need not ever be touched again after a one-time leveling them all the same height and level front-back and side-side. Then added some gizmos on each end to hold a string. $400 vs. a $4,000 store bought rig. Oh, Harbor Freight ramps.

[Rick Kirchner]

Never got the platform as level as I wanted. Yesterday, I leveled LF (ref) to RF to RR and LF to RR diagonal (because that's the measurement that was giving me fits). Then RR to LR and called it a night.

This morning, I removed the rear pushrods and made them equal. Took the top of the shock off and ran thru whole range of motion, no stiction.

Took the front completely apart (and those of you that have had a 90-95 VD know what an absolute PITA that is) re-set the front pushrods, went through the entire range of motion, no stiction.

Put it all back together and on the scales. Front was off 50 lbs, just as before, dialed in a whole turn of RR pushrod and saw not enough change, just like before. Lucky I keep my gun in a safe, or something would have got shot - maybe the car, maybe the scales, maybe me.

Just for ****s and giggles, rolled the car back and stood on each scale and I almost missed it - LF was light by 20 lbs! Swapped LF and LR pads, rolled the car back on, and it scaled almost perfectly.

Figured it was either the connector or the channel in the head. At first, because it repeated almost exactly between when they were assembled on my rig and then again on Art's, I thought it was the head, but after the swap, the LF is now just two pounds off compared to the rest - I believe it's corrosion or bad contact on the connector pins. What are the odds???? I would have expected work/no work, not an offset.

On these units, Reb-co used a mil-spec connector but there isn't enough room to twist the lock ring, so they just omitted it and turned it into a push-on connector, and there's not a lot of retention force without the lock ring.

Thanks to all that provided advice. Now I'll reference weight the scales after assy instead of just trusting the zero. But there's still the matter of leveling - that remains a mystery.

[Robert J. Alder]

Where is the track surface within .040" of level? John Mihalich, Jr.

Kinda' what I was wondering. How many angels are on the head of this pin? Of course, John and I race vintage FFs (aka "Flexible Flyers.") So.......... .040 becomes a bit moot.

OK contemporary formula cars with mega stiff chassis and springs are undoubtedly another world. But..... what is an allowable difference in corner weights (say at the front) before it has a measurable negative impact?

And what if the track has a combined net camber on "the line" for the the combination of all corners. Maybe that would warrant "tuning" the corner weight to that track's overall net camber at the corners?????? Just trying to stir up the curiosity juices. It's more of a rhetorical question rather than a serious need to know question.

[Rick Kirchner]

With my car at lest, a clue to corner weights being way off is instability under heavy braking.

[Robert J. Alder]

With my car at lest, a clue to corner weights being way off is instability under heavy braking.

What was the magnitude of "way off." You recall?

[Rick Kirchner]

What was the magnitude of "way off." You recall?

Based on my approximate 20 lb offset in the scale, I'm guessing 30-40 lb

[Robert J. Alder]

Based on my approximate 20 lb offset in the scale, I'm guessing 30-40 lb

Yup. That's bad. Way more than any repeatability error range. With my new set up, I can usually get repeatable readings within 3 or 4 pounds on corners. Perfectly side-to-side equal weights (at both front and rear) isn't always achievable due to the car being heavier on one side so minimizing difference at the front is the best compromise.