View Full Version : Formula Car floor pan

03-22-2006, 05:12 PM
A friend of mine has asked me to replace the current .064" aluminum floor pan on his Formula Ford race car with a welded steel pan as the aluminum one keeps coming loose. It loosens up over time and the handling of the car slowly drifts off. The car can be fully maintained and dissassembled without removing the floor, so this isn't a concern.

The pan is just under 8' long and is 22" wide at it's widest point, near the center and is completely flat. You can see in the pic that there are cross braces every ~18" between the two side frame rails. My plan, unless someone has a better idea is to plug weld the pan in place every couple inches around the perimiter as well as to the cross braces. I'll punch as many holes as I can with a hand held pneumatic punch and drill the others.

I'm currently thinking 19 gage steel will be adequate as there are no large open areas and it never gets stepped on directly. This is a nationally competitive car so I'd like to keep the weight as low as possible, but without having to risk changing it in a year or two.

If it were my own vehicle I'd just throw 19 or 20 gage on there and hope for the best, but as it's somone elses I want to make sure to get it right the first time.

Any thoughts, ideas or suggestions?



gator 1
03-22-2006, 05:35 PM
Why do the floor pans keep coming loose. I think i would solve this problem first. But if you are going to change to steel and welded in check the rule books as some racing classes state they have to be tiged and the welds have to be certified. Not sure on this class but most cars have the alum so i would try to fix this first. Also you may tweek the frame with the heat of welding in the steel floors, This could effect the set up of the car. As far as thickness goes 19 gage should be fine but this may be speced in the rule book.

Im no expert at racing but i do like to look at things from outside the box, Just my thoughts...Gator

03-22-2006, 07:10 PM
Is the aluminum pan pop riveted in? I'm wondering if it's working loose, as the chassis "works", could you drill, and tap several places, to provide a more solid attachment for the aluminum pan. I think a couple dozen 5/32 or 6mm screws would hold it pretty well. Good luck, Eric

03-22-2006, 07:45 PM
Yes, the current pan is riveted in place. I wasn't involved with the installation of the rivets and haven't seen the floor in it's loose state, so I'm not certain what's happening. I do know that the chassis design in the rear of the car where the floor is coming loose, isn't especially good. ie. not well triangulated. I suspect that this is increasing the loads on the floor and rivets, causing things to come loose. Ideally the chassis design in the rear would get reworked but there isn't much room under the existing fiberglass body work, and I want nothing to do with modifying it. Also, the frame likely isn't thick enough to be threaded and would need inserts installed.

Gator, I hadn't thought of checking the GCR(rulebook) to see if the floor thickness or welding process is controlled. I suspect it isn't but I'll check for sure. I know it isn't for the class I race in, but I should definately make sure.

Thanks for all the suggestions so far...


03-22-2006, 08:34 PM
I'm with Gator on this. I can't see steel as a cure. The problem is probably with the chassis flex. You might try flush Dzus buttons with welded tabs to cure this . Several years ago I helped a friend put a factory stiffening package in his formula 2000 Lola. He used the factory supplied mastic and pop rivets. It was still going strong 2 years later when he sold it. Check the rules before commiting to any change.

03-22-2006, 09:14 PM
By welding a steel floorpan in, you are just moving the flex to an other area. I would look at installing rivnut's into the tube, then a sealer between the frame and pan. This way the pan can be removed, but will still provide some structure, without adding much weight. On the Lotus 7, the steel pan is welded in, and prone to rusting out near the welds.

Just my 2c

Jeff Parker
03-23-2006, 05:59 AM
Does this car have cracked welds, or a sprung frame?? The aluminum is acting as a "Shear plate", and thats why handling diminishes as it loosens. There is a definate problem in the frame's rigidity. The stiffer the frame, the better in those cars. That way the suspension can be tuned without having the "unknown chassis flex" variable thrown in.

I'd check the frame over REALLLY good, before changing the design by introducing steel in lieu of the aluminum.

03-23-2006, 06:12 AM
Do you know if the pan gets replaced very time, or do they just replace the rivets? If the holes have elongated or become sloppy, the rivets won't provide any real shear strength. They'll probably just work loose again.
There is a difference in the rivet types too. Stainless steel would probably be the stongest pop rivet, or you could even go with aviation structural pull type rivets such as cherrymax, pullmax etc.
I like the bonding idea, with tight rivet holes.



03-23-2006, 07:35 AM

From my limited racing experience I can tell you that the floor on a FF car should not be adding any stiffness to the chassis at all. If it is, then you have an improperly braced frame. Going to a welded floor will stiffen up the frame but only in two directions not three. It my cure his problem, but the proper way to do it is to address the frame problem, It's possible that the frame problem could be cured with less work (and weight) than replacing the floor.

03-23-2006, 12:35 PM
I agree that the pan should not be acting as a strenth member. I would concider adding braces or corner gussets to the rear part of the frame. if you are not going to do that I would do the avation rivits and add a few more of them. not sure the steel pan is the best way


03-24-2006, 07:14 AM
The pan has been replaced, but not nearly as often as the rivets. I'll be pulling the pan on Saturday afternoon or Sunday and will then have a better idea of what's happeneing

The floor of a car has a large affect on the torsional rigidity of the chassis. The stiffness in this plane can come from the chassis tubes, or a combination of the tubes and pan. We can easily see in the pic that I posted that there is NO triangulation in the floor. Is is clear then that the engineer(this was a professionally designed chassis) intended for the pan itself to play a key role in the stiffness of this plane.

While there are many other places for improvement in the chassis as well as other parts of the car, they aren't currently in the plan due to time and/or money constraints. The car needs to be back together for the first race in under a month. The car is already mostly apart for other work and now is the time to fix the floor permanently. We'll likely look at re-working the chassis in the future but it isn't as simple as adding a few tubes here and there. We have very tight fitting body work to deal with and the motor still needs to be accessible/removable when we're done. I suspect we'll end up with a bolt in structure, which will take time to get right.

Thanks again for all the input


03-24-2006, 11:20 AM
I think you ought to plug weld in the floor pan.
Think of it as a huge gusset that ties the bottom together. I'm trying to imagine the forces that would tear loose the pop rivits. I'm guessing it's a roll couple distribution problem causing a torque through the driver area. Plug welds have no clearance or "give" as rivits do. A welded shear plane should stiffen this chassis up a bunch.

03-25-2006, 05:17 PM

Just be aware that the floor adds very little torsional rigidity to the chassis. It adds to x and y but only a small amount to z which is the torsional part.

Hope you get it back on the track quick.

Peter Miles
03-25-2006, 07:08 PM
Is the seat bolted to or sitting on the floor?

Can some beads be rolled into the pan to stiffen it?

03-26-2006, 03:30 AM
The floor of a car has a large affect on the torsional rigidity of the chassis.

sorry, but as PaulG says the floorpan in a chassis lends nothing to torsional rigidity. The only place it really helps is in beam. This is pretty important, as the lower rails are the most heavily loaded in the chassis, and actually want to spread under suspension loads.
I've proved this through having an older Van Diemen FF chassis (a copy of your Swift) FEA'd ;)
I've also replaced the floor pans on a number of cars, and made chassis/suspension mods on quite a few.

The cars generally come from the factory with around an 0.080" 7075 aluminium sheet sometimes bonded, but mostly only rivetted to the chassis. Steel can easily be used as an undertray, and the Aussie made Spectrum comes from the factory with a welded steel floor, and most cars here were run with a rivetted stainless floor, mainly as the front would hit the track under heavy braking and the rest of the undertray was bouncing off ripple strips (this is what generally loosens up the undertray) so using a thick stainless floor with no skid plates were used and the car could be run that little lower at the minimum ride height here of 40mm.

I've always used aluminuim as,
1. I was trying to get the maximum amount of weight out of the car as.....
2. I could then ballast them to fine tune the handling. I was prepared to sacrifice a couple of mm of lost ride height by using front skid plates.
Generally, (and this depends to a large extent what tyres you are running in your respective series) this was trying to get as much weight forward as possible.
I mostly used 0.125" 2024 T4 bonded and rivetted to the chassis. At the time, 7075 of any description was virtually impossible to obtain here. Also, bonding and rivetting complement each other. Bonding is extremely strong in shear, but poor in peel, hence the need for rivetts.
We used 3M's Scotchweld 2216B/A adhesive (this is a standard aircraft industry adhesive) and as structural rivets are hard to come by here, and I only posess standard pop riveters, I used Avdel rivets (in monel, and generally one drill size up, eg 5/32") that retain the mandrel and rupture at the work piece first.
The prep for bonding was a pain, with the alclad finish having to be etched (disc sander usually) the steel absolutely clean and lots and lots of methylethylketone and clean rags used on both surfaces until no oxides were left (white rag clean and break water test).
While all this prep was time consuming, it took several seasons of hard racing to wreck these trays, whereas most other teams/ competitors were making running repairs mid season.

I haven't been involved in racing for quite a number of years now, and If I was going to use a steel floor, I'd probably braze it, mainly as I don't have access to a tig these days. ;) If you look at most English Formula Fords, the chassis are mostly brazed. It's an excellent process, and doesn't introduce too much heat into the steel. Surprisingly, Slicone Bronze was used in the UK as the torches were run on 'town gas', which I think was coal gas (someone correct me ?) I always used Nickel Bronze for repairs. Obviously Swifts, including the UK Swift was tigged.

Peter Miles
03-26-2006, 09:16 AM
TDI-Rick - what is the "break water test" to which you referred?

Ah - I decided to use Google to find my answer. The best Link is at:


That is Caswell - Electroplating in Miniature and includes a video. They also describe it as being recognised as ASTM-F-22.

The video link is at: http://www.caswellplating.com/movies/waterbreak.mov

However, there was a Quicktime problem with that video link when I tried it.

To quote from that text link:
To pass the test water will sheet off the part rather than bead off.

Take a cleaned and dried part and set it in a vertical position.
Use a spray bottle containing distilled water.
Spray the part two to three times from at least 6" away.
If the part is clean and free of oily residue, the water spray should sheet off.
If some oily residue remains, the water will tend to bead on the part
Repeat the cleaning process until the part passes the test.

Alternatively, apply several drops of distilled water to the cleaned surfaces.

If the surface is inadequately cleaned, the spherical form of the drop is largely retained, and the surface must be cleaned once more.

If the water runs on the treated surface, then wetting has been satisfactory and the part is ready for plating.

03-26-2006, 02:48 PM
thanks Peter, that says it far better than I ever could. I learned that little trick from one of my mentors, an ex QANTAS aircraft mechanic/engineer.

One thing I forgot to mention in my post above is that Carroll Smith's series of books, Build to Win, Prepare to Win, Engineer to Win, etc are brilliant references for race car mechanics.

twoodbury, I'm curious as to the model year of the Swift you're working on.
I've never seen a US Swift in the flesh, but have seen a few photo's, and the chassis looks little different from the original. A bloke called Dave Bruns designed the Swift chassis, and it revolutionised Formula Ford worldwide when introduced in '83. Van Diemen in the UK basically copied it in '85 (not as nice as he Swift) and refined it in '86 and Swifts were built under licence in the UK from, IIRC, '87 on by (I think) Frank Broadley. The UK Swift stated to diverge a little from the original Swift in '90 (longer wheel base, side pod mounted radiators ala Reynard) and in '91 a new, stiffer chassis and new rear bellhousing/suspension geometry (much higher rear roll centre) and totally changed in '92 with a totally different chassis design and pushrod suspension all round.

03-29-2006, 06:54 AM
Peter, there really isn't much of a seat. There is a sheet of aluminum that rivets to the floor and to the bulkhead behind the driver. On top of that is a removable foam insert that is shaped to the passenger. Think of filling a trash bag with spray foam and sitting in it... Yes, I plan to roll some beads in the pan. Just need to get my HF bead roller reinforced.

tdi-rick, your FEA experiences are interesting, but I'm still a bit skeptical of your results. I've come up with an idea for building a simple model that can be tested for torsional rigidity in two configurations. In the first configuration the 'floor' only maintains the distance between the sides. In the second, the 'floor' also acts as a ridgid plane. The question is, will I find time to build it? It's a '85 Reynard BTW.

I've decided to go ahead and install the steel floor as the owner had asked. While it wasn't really intended for public consumption, I did throw a bunch of pictures on a web page for the owner to see. Feel free to take a look



03-29-2006, 03:46 PM
Travis, I was told by a very good engineer I used to work with that the floor contributed very little torsionally, and our modelling bore that out.
Like you, I was convinced otherwise, until the chassis ('85 Van Diemen) was modelled.

BTW, went and looked at the picture of the car and saw the pushrods :oops: So that's what an '85 Reynard looks like. Actually never seen one before. Later Reynard's (designed by Mal Oastler) went back to Rocker suspension all 'round and were a totally different design. IIRC, the '85 car was designed by Wiet Hu???????er (who later formed Vector race cars) and was a pretty difficult thing to set up/drive.

03-29-2006, 03:50 PM
the '85 had me intrigued, so googled it and came up with a thread on another board you might be interested in for later chassis mods.

03-29-2006, 05:46 PM

It's interesting that you pointed to that thread. The person who asked the first two questions in the thread is the brother of the guy I'm doing this work for. Given the spring rates that are being run on this car, the suspension geometry issues are very much reduced:)


03-30-2006, 01:05 PM
Travis, have the roll centre heights, particularly at the rear ever been plotted ?
If Reynard faithfully copied the DB1 geometry, that could be the unforgiving rear ends problem. I've worked with a number of Euro Swifts, and up until the '90 model, they faithfully used the original US Swift geometry.
IMHO, the rear roll centre is just wrong and way too low. (It's lower than the front) It results in a rear end that has unpredictable snap oversteer in high speed corners, at least on the various treaded tyres used here over the years. Raising the rear roll centre transformed the '90 I worked with/drove. Interestingly, the '91 and later models had a rear roll centre height much high er than the earlier models (new bell housing) and almost identical to '90 on Van Diemens. We also revised the front heights on the '93 and '94 Euro Swifts.

04-12-2006, 01:19 PM
I would try rolling some beads into the pan to make it more rigid and also consider getting rid of some of the flex with more triangulation of tubing.