Downforce Data Acquisition

[jtaphorn]

I am considering running a vintage event this summer at High Plains raceway and thought I'd watch a track video to get a lay of the land. What surprised me in the video was the downforce telemetry this F1000 was revealing. I realize the depth of knowledge of many on this forum is very deep and I am treading in shallow water by comparison. However, I thought this was so cool and appears to bring forth some real transparency as to the effectiveness of one's wing set up. What type or application of a sensor is acquiring this downforce data?

F1000 lapping at High Plains Raceway -- May 29, 2020 - YouTube

[Rick Kirchner]

I'll take a guess. You put shock pots on all corners. You know what the static positions are, and you know the motion ratios, etc so that you know the displacement vs force curves.

Then you write a math channel that sums all the forces. Weight transfer takes weight off one corner vs another, but downforce will add to all.

[romoman]

I'd be very curious to see how the person in that video was coming up with those numbers.

The ways I have seen are either a strain gauge mounted in a suspension point (shock body, mount, etc) or wing mounts, or by measuring shock movement on a calculating additional loads. I've always seen it referenced as straight line measurements and not in corners.

If you have a third spring, it's much easier to measure.

[jtaphorn]

My initial thought also was a sensor measuring shock loading. However, that seems inconsistent with the data as the car is visibly rocking front to back at the end of the straight, due to undulations in the track, and hitting curbs on several apexes. I would expect that to be revealed in the data, were it shock related, and I don't see that it is. It seems a bit sexier, perhaps a calculation of differences in air pressure at two points?

It is more downforce than I expected: 60mph - 120lbs; 75mph - 200lbs; 85mph - 300lbs; 100mph - 400lbs; 110mph - 500lbs; 130mph - 800+lbs.

[99sh]

There are no shock pots, I've seen this car in person. Let me point the owner (Jeff) to this thread. You could also ask Gary Hickman, it was his old car. I think he's using a strain gauge.

Sent from my Pixel 4 using Tapatalk

[Rick Kirchner]

I'd think the DF numbers are believable. Seems like most of these cars are capable of ~2.5G. On a 1000 lb car that would require nearly 1500 lb of DF.

[romoman]

I'd think the DF numbers are believable. Seems like most of these cars are capable of ~2.5G. On a 1000 lb car that would require nearly 1500 lb of DF.

I'm not sure that is correct. In some fender cars we see around 2G with minimal downforce. My old 914 does over 1.5 with no aero.

[Teuobk]

This is my video, so let me pull back the curtain

Rick is correct: shock pots are on all four corners, the spring rates are known, the motion ratios are known, and the static positions are known. From there, it's a matter of making math channels to put the data together and spit out a downforce number.

Having said that, there are some caveats that decrease accuracy.

One is that the tire sidewalls will also deflect as load increases. This obviously can't be seen by the shock pots, so this source of error will cause the downforce to be slightly under-reported. (EDIT: This turns out to be incorrect. See DaveW's response later in this thread)

Another is that track camber in corners will increase or decrease the load. In an on-camber corner, the figure from the downforce math channel will overstate the actual downforce, and in an off-camber corner it will be understated. Similar issues exist at bottoms and tops of hills.

I put the downforce number in this video (and others) mostly because I thought it was neat. It's nowhere near 100% accurate, but I'm pretty confident it's at least in the ballpark.

What was unexpected is how the downforce numbers in those videos have caught the eye of non-racers -- you know, the types who would ask "how fast does it go?". It's been a great icebreaker to talk about the various things that make formula cars so amazing.

Jeff

[phantomjock]

Very nice and thorough description of the D/F acquisition devices you are using. Quick question - what data acquisition "tool" are you using (AIM, RT, other)? and, also please the software you are using to post the info on your video.

Thanks in advance.

Cheers - Jim

[S Lathrop]

Teuobk,

I have done some coast down testing with a F2000. Your numbers seem quite reasonable.

Coast down testing is here you bring the car up to some speed, say 130 mph and then put the car in neutral and let it cost until the speed drops to say 60 mph. We were able to use an airport taxiway that was on an inactive runway for a day.

On the car I did the testing I used rotary post that were gear driven on the bell cranks. The big advantage of this system is that the rotary post are relatively inexpensive, around $50 when we rigged the system. The gearing allow the pot to rotate over 200 degrees for the shock motion we were measuring. This improved the resolution of the data immensely over any other system I had seen at that time.

The advantage of doing coast down testing is that you can measure down force over a speed range. In out case we were particularly interested in not only the total amount of down force but how that changed with speed, especially how the balance changed with speed.

I found coast down testing infinitely more valuable than wind tunnel tests. That is because the data has a direct correlation to the car as it is on the track and the data can be confirmed on the track as well.

[DaveW]

...Having said that, there are some caveats that decrease accuracy.

One is that the tire sidewalls will also deflect as load increases. This obviously can't be seen by the shock pots, so this source of error will cause the downforce to be slightly under-reported....Jeff

I don't totally agree with that statement. Yes, the tires deflect and that deflection is not measured. However, if you are measuring deflection through the shock pots, and using wheel rates to determine downforce, nowhere in that equation do you need to know how much the tires deflect or what their spring rate is.

If you were measuring ride height using an ultrasonic or other tool to determine height off the track, then you would need to include the tire spring rate in the force calculation.

[jtaphorn]

Jeff, Thanks for clarifying.

While the numbers may not be error free, they certainly appear to provide constructive data feedback on the relative effect of wing adjustments. Wish that I had that tool on my FC. I don't get enough track time to get sensitive enough at "seat of the pants" adjustments. I read about the talents of some drivers on Apex who recognize the subtle effects on handling, when things occur such as when their shocks warm up; but alas, that is not me. I rely on the best of intention guidance of others and hypotheses. I'd love empirical feedback on wings to optimize them.

Well done.

[Teuobk]

Quick question - what data acquisition "tool" are you using (AIM, RT, other)? and, also please the software you are using to post the info on your video.

The data is collected using an AiM EVO4. I was very fortunate in that Gary Hickman, who I got the car from, put together a very comprehensive datalogging system.

I exported the data using Race Studio from AiM and then used RaceRender to overlay the data onto my GoPro video.

Jeff

[Teuobk]

Coast down testing is here you bring the car up to some speed, say 130 mph and then put the car in neutral and let it cost until the speed drops to say 60 mph. We were able to use an airport taxiway that was on an inactive runway for a day.

...

The advantage of doing coast down testing is that you can measure down force over a speed range. In out case we were particularly interested in not only the total amount of down force but how that changed with speed, especially how the balance changed with speed.

That's a great idea. The runway (or taxiway) would eliminate most of the issues related to camber and elevation changes, and the ability to run it in both directions would allow the effect of wind to be removed. Seems like it would make for really clean data.

Jeff

[Teuobk]

I don't totally agree with that statement. Yes, the tires deflect and that deflection is not measured. However, if you are measuring deflection through the shock pots, and using wheel rates to determine downforce, nowhere in that equation do you need to know how much the tires deflect or what their spring rate is.

An excellent point; I was wrong. The same force would still be seen by the spring regardless of how much the sidewall was deflecting.

Always nice when an error term can be eliminated!

Jeff

[Rick Kirchner]

I'm not sure that is correct. In some fender cars we see around 2G with minimal downforce. My old 914 does over 1.5 with no aero.

Not sure I'm buying that. You get some addition grip due to the physics of racing tires and soft compounds, but a full G? nah

I could see the front of your 914 building downforce with a good air dam and splitter and a little rake - but I'd think the low pressure area behind the backlight would create lift.

[lowside67]

Not sure I'm buying that. You get some addition grip due to the physics of racing tires and soft compounds, but a full G? nah

I could see the front of your 914 building downforce with a good air dam and splitter and a little rake - but I'd think the low pressure area behind the backlight would create lift.

I assume he means peak, not continuous, G load; I am with you that there's no way a sedan is doing 2.0G without downforce.

My basic Stohr sports racer which has original 2004 aero (aka not much but still a decent splitter and wing relative to a street car) but has around 2x the tire to weight of a sedan was seeing peak loads of around 2.1G and sustained in the 1.7-1.8 range. This was on brand new sticker ultra-soft hillclimb tires on a well rubbered in track.

-Mark

[lowside67]

Rick is correct: shock pots are on all four corners, the spring rates are known, the motion ratios are known, and the static positions are known. From there, it's a matter of making math channels to put the data together and spit out a downforce number.

Hi Jeff,

Can you share the formula for your math channel? It's super interesting stuff and I'd love to better understand how this all works. (I would think that to improve accuracy you would need to subtract the unsprung weight on each corner?)

Cheers
Mark

[sccadsr31]

This data is from an AMAC P2 car at ThunderHill, data collected the same way, linearly shock pot, known motion ratios, static ride height and etc.
This is no longer my car and I still do the data but stripped off the numbers, however the numbers are very similar in total magnitude to the video. We have been a bit surprised at the magnitude being as high as it is and I'm crawling through the calculations for possible errors, have not found any. This is also an AiM EVO4 recording at200hz. What it revels in the dynamic balance is very interesting like the F1000 data. We need more front DF. What I want to know is all the steps to sinc the GoPro, that is way cool!

[romoman]

Not sure I'm buying that. You get some addition grip due to the physics of racing tires and soft compounds, but a full G? nah

I could see the front of your 914 building downforce with a good air dam and splitter and a little rake - but I'd think the low pressure area behind the backlight would create lift.

My car is setup like the original 914-6 GT with flares, oil cooler duct, and GT valence. I have about .5" of rake. Graph from the back straight at Watkins Glen. I would have to dig it out, but I have data from the Pocono tri-oval that shows that shows it just as much.

Production cars make over 1G on street tires now https://fastestlaps.com/lists/top-grip-kings

[Rick Kirchner]

Production cars make over 1G on street tires now https://fastestlaps.com/lists/top-grip-kings

And have since the Vette did it in the 90s.

You have to correct your G-numbers for the banking........

[romoman]

And have since the Vette did it in the 90s.

You have to correct your G-numbers for the banking........

I'm familiar with the data and corrections.