Circle Track Sway Bar

[JMortensen]

One of the main things that I wanted out of a circle track style hollow bar with splined ends was lighter weight. Turns out the arms that you can buy off the shelf are 3/4" thick steel and are REALLY heavy, so this ended up being a lb or two heavier than the modded MSA bar that I took off.  If you're looking at this for a light weight bar, look elsewhere.

This is probably revision 457 of my sway bar setup. Some of the highlights of my previous attempts: Started with an MSA front and rear bar set. I had originally bought MSA front and rear bars. Then I had changed the rear end links and they ended up hitting the frame rails. Then I got an ST rear bar and modded it by drilling out the bushing saddles and adding zerks. Made double shear control arm attachments for modified stock arms. I tried to get a guy who custom bends hollow bars to make me one, can't remember the name now (Simonson???), but finally gave up after repeated attempts to contact him, including ones like: "I've got cash in hand. Want to buy a bar. You have made the sale. All you need to do is answer this email." No reply, so I finally decided to do the same thing a friend of mine had done successfully: I cut the ends off of aftermarket MSA front and rear bars and welded on 3/8" x 1" steel bar and drilled the new ends for adjustment. Then I mounted the bars on heims joints to reduce friction (more on this later). Then I mounted the front bar on top of the frame rail for splitter clearance. Then I installed an oil cooler, and the bar interfered, so I moved it back underneath, and that is what led to these latest revisions.

The previous bar in its last iteration had the 3/8 x 1" bar stock welded on at an angle so that the end links leaned in towards the fenderwell significantly. I had found that this would give me the best clearance around the suspension parts. Before I ever ran it, Richard (260DET) pointed out that this would give me a falling rate sway bar, so I revised it by bending the ends vertical. There were several adjustment holes on the bar, but after fixing the end link angle, the end links bottomed out on the bar or the control arm mount that I built, so I could only use one hole per side, so I basically spent a lot of time cutting up a sway bar and welding crap to it and ended up with a non-adjustable sway bar with a bunch of useless shit welded to it. I probably should have left it the way it was, but even in that state the whole thing was pretty janky, and the only saving grace was that my spring rates were so stiff that it probably kept the bar from being used enough to bend the ends or snap the welds off. People have a big problem with welding to a spring, and I get that, but having done it and knowing others who have done it on race cars with stiff rates (meaning not a lot of sway bar twist) it works. It's not ideal, and I'm not saying that it is, but it works. It did hold up for 5 or 6 autoxes with 2 drivers but I wouldn't put this up as my best fabrication project. I wanted the new one to be functional, mount on the bottom of the frame, fit under the splitter, and clear the power steering ram on the rack which sticks out quite a ways on mine. 

Mounting on heims is controversial. One of the main things I was after with my suspension was reducing friction, so when tube80z suggested this as an alternative to poly, I was immediately interested. The idea is to have ears on the sway bar that connect to a rod end that is installed into the frame. This allows the sway bar to move up and down much more freely, but it also allows the sway bar to twist and doesn't restrict its shape while it twists. The bar should make a (slight) S shape in the center, and allowing the movement that the heims joints do should make it just a little bit easier for the bar to do its thing. I don't know that this is a huge deal, I think the bigger thing is allowing the bar to move freely, but there are IMSA and GT cars running this kind of setup, which is what tube80z showed me, and that's what convinced me to try it. I already had reinforced the stock sway bar mounting areas with a plate on top and bottom of the frame rail, and tubes connecting top to bottom in the front and back of the plates. I had used a pretty thin walled tap tube, but later decided it was probably too flimsy, so changed to a much thicker walled tube welded nuts on top and bottom. I'm still using those tubes. This still produces a lot of "WTF are you doing?" kinds of comments. Back in the day Cameron (Heavy85) was pretty outspoken as a critic and Coffey seemed to like the idea IIRC. You decide whether it's great or a waste of time and effort or doesn't make any difference at all. I was convinced enough to do it. 

The stock bar has the dip in the middle for oil pan clearance, and with the ears I had on the stock bar the center part of the bar would hit the splitter as it moved. I was hoping that the straight center bar would improve things, but no, just created a different problem: it hit the power steering ram. I figured I could use longer mounting ears to drop the sway bar down away from the ram. It was tough to see how everything would line up when it was all done, so I made them longer than necessary. I spent a lot of time making sure that the end links wouldn't bind and that the bar could move all the way through the suspension travel without hitting the ram or anything else. Had to trim sections off of the arms as well for sway bar clearance. 

The arms are very thick and strong, and I was only able to put about 1/16" bend in one with my 12T press. Was thinking of cutting or grinding them thinner so that I could bend them, or just giving up, but found a local metal fab shop with a huge press. Took it down there and explained what I wanted and watched/helped as the guy bent them to fit. Also wanted double shear so I welded an extension to the arms as well, and used high misalignment bushings on both ends to keep it from binding. Found that 3/8 male and female rod ends weren't short enough, so had to trim the threaded part of both to get them short enough, but there is no binding. The front adjusting hole on the arms is for the droop limiters, not a sway bar adjustment. 

Cleaned it all up, welded, painted, stuck it in the car and the ends of the arms where the bolt holds the splines tight hits the splitter. Now that I had it all on the car though, I could see that the ears were longer than necessary. Cut an inch off of the ears, redrilled and they're on there now. I haven't put the splitter back on, but I'm sure everything clears.  Couple pics of older iterations in here as well.

Off topic: the brake duct doesn't clear the arm, so one of my projects for the winter is to point the brake duct inlet forwards for clearance.

[clarkspeed]

That is a lot to digest, but I think I follow your journey. Some things I have run across in my current and previous designs:

1. Weight is a concern. If you end up with a setup considerably heavier than a couple Suspension Technique bars and bolt style end links, you are going in wrong direction.

2. I usually target around 1 degree of body roll at max lateral G.  So if I run relatively stiffer springs, then sway bar rate can be reduced. I once set a car up with 400 lb springs, a stock front sway bar, and a ST bar in the rear. Both made adjustable (so the front was actually stiffer than stock).

3. Running small to medium size bars have other advantages in weight reduction. For argument, I will say medium is less than a 1" solid bar with 14" legs. You can weld on  adjustment brackets with simple angle iron. And you can use smaller rod ends. I think I ran 1/4" with the stock bar, maybe 5/16, I can't remember. But less than 3/8.  And if you are running a hollow circle track style bar, you can go with aluminum arms. They say they flex and move around, but that is only a problem at higher bar spring rates. I also try to keep the arms as short as possible to keep the arm weight down and reduce flex.

4. It looks like you solved the binding problem. When running very short end links you end up giving the control arm adjustment holes also to match up. I have also done this with a simple piece of angle drilled and bolted to the arms.

5. I like the rod end mounting. Kind of makes sense at least for reducing friction. I would put a camera under there to see what the are doing under load. I've seen mounted in roller bearings also.

6. It doesnt take much adjustment to make a big difference. You can do the math, but for me 1/2" is usually difference between understeer and oversteer. So if you have selected the correct bar rates, you really only need one bar to be adjustable.

 

 

[JMortensen]

Interesting stuff Clark, thanks.
1. Same weight as MSA 1", so good there. Haven't run a rear bar in 20 years, so call that a weight savings... I will say that the ST rear bar is a LOT lighter than the MSA rear bar, you don't have to drill holes in the chassis to mount it, and if you space it back a little it doesn't bind. I have dents in my frame rails from the MSA bar end links hitting them, so back in the 90s and early 00s I was definitely doing it wrong.
2. According to my WTW, I'm at .99 degree for 1g, and I have a lot stiffer springs 550/600. I am a lot wider, maybe that makes the difference. So may need to stiffen up later as the car should be capable of more than 1g. Haven't measured g's yet.

4 and 5. Yes, definitely needed the adjustment on the bottom, and the double shear means I can run a clevis pin too. One thing I forgot to mention is that when I welded in the mounts on the previous iteration I had an end link solidly bolted in. This made it difficult to slide the end link into the control arm mount. This time I took a HF blister pack from a tool I had bought and used some of the clear plastic. Just cut out a square and punched out a 3/8" hole in the middle and put that in between the rod end and the arm before I welded. I measured the plastic at something like .006 or .010 thick, can't remember, but now if I want to move the rod end it slides nicely to the next position really easily. Probably would clunk in a street car, but I'm doubting I'll hear it or care...
6. I calculated the spring rate and matched it up. I can go stiffer or softer, so hopefully I'll have a setting that works, but won't know until I get it back out again. I have some spring options, so might need to change those to get what I want, and if it needs stiffer anyway, no biggie.

Edited January 28, 2021 by JMortensen

[jhm]

@JMortensen, thanks much for posting.  Will definitely want to spend some time digesting all this, so the amount of info and details are greatly appreciated.  I've been looking at several options for an adjustable front bar, but have yet to settle on a design.  There are several hollow adjustable bars made for Miatas, whose front bar has similar width and length dimensions as the S30, but has a straight center section.  DP Racing offers some nice hollow sway bars, but they're a bit too pricey for me.

 

 @clarkspeed, your input is very much appreciated as well.  How have you made your stock front bar adjustable?  Sliders?  Any pictures or fabrication details would be greatly appreciated.

 

This is great stuff, and why I like HybridZ so much!

 

 

 

 

[clarkspeed]

John, I anticipated you had already considered everything on my list. Your car has more planning and engineering in it than I could imagine.  I just thought I would list some of the lessons I have learned. I did make a mistake above, I didn't run a rear bar with the stock front bar. Writing about jogged my memory. That is how I ended up with a stock front bar. I was running your standard front and rear aftermarket bars that I had made adjustable. It was  a perfectly balanced setup, but I wanted to drop the rear bar to get the power down better and save weight. After doing the math, I found I didn't need very much spring rate in the front to keep the same F/R roll resistance. For everyone else reading, the roll bar is basically another spring when cornering. So if the spring provides 500 lbs/in resistance and the roll bar has 200 lbs/in then you have 700 lbs/in effective spring rate. It's a little more complicated than that due to motion ratios and roll bar configuration, but you get the idea. So it you have 700lbs in front and say a 450 spring and a 150 lb bar in back (600), and you remove the 150 lb bar, you can imagine the 700 in front combined with the new 450 in the back will roll considerably more in the back and will understeer until you get the total F/R stiffness ratio back to where you started.  So....to make a long story short, if you remove the rear bar, without any other changes, you probably need a smaller front bar. 

 

I will try to find some pictures of my adjustable mods. 

 

[260DET]

Great post Jon but no no no, the supplied speedway arms are only used for the part that clamps onto the bar, the rest is cut off and chucked with suitable light weight arms welded on. The design of the replacement arms is dictated by the load involved not the bulk of the clamping part. The advantages of a splined set up are too good to be discarded.

[JMortensen]

Clark, I was taking it as advice and appreciating it, so thanks.

I got that same feedback elsewhere Richard, will probably be revising. The good news is I won't have to figure anything out next time, just duplicate with thinner metal.

[clarkspeed]

Kind of off topic, but here is a picture of the adjustment I added to a stock arm.  Pretty simple and it was just bolted on to the roll bar and a clevis was bolted to the control arm.  It worked perfectly.  

[tube80z]

On 1/27/2021 at 9:44 PM, clarkspeed said:

 

5. I like the rod end mounting. Kind of makes sense at least for reducing friction. I would put a camera under there to see what the are doing under load. I've seen mounted in roller bearings also.

 

 

 

 

Hi Clark the rod end works better in most cases.  If you have a roller bearing it can still bind when the arm flexes in the middle.  A common problem people run into is they install collars to hold the ARB in place but when it bends these hit the inner ARM mounts and cause bind.  Give it 1/8 to 3/16 space and the binding stops.  The rod ends are also great on older cars that may not be as stiff as a tubeframe chassis.  I stole the idea from a Riley & Scott trans-am car but have seen it all over the place.  The Realtime racing Acuras that raced in IMSA used this method as did a lot of older BTCC cars.

 

All you have to do is mount a normal bar in a poly mount that's all greased up and see how much force it takes to move.  Now jack on the end of a frame rail at the front of the engine bay and try again.  I usually saw very different numbers.  Sometimes 50 or more pounds from a fish pull scale.  With rod ends that have been burnished it will be in the 2 to 3 pound range.  That's free grip and helps make the car more consistent.  

 

Cary

[clarkspeed]

Makes perfect sense to me. I will incorporate in future builds!

[JMortensen]

Got some pics of fabbed arms. Looks pretty easy, basically just hacked the original arm off and welded up some boxed sections with some tube through to stiffen. Will put on list of things to do this winter.. 

[clarkspeed]

A GoPro or similar mounted underneath can give a wealth of info of what is going on. I bought a couple of tiny 720p cameras for 20/ea and ty-wrap those things everywhere. I lay video snap shots in PowerPoint and estimate max distance traveled for each component.  Poor man's data acquisition!

[JMortensen]

What do you have running all the cameras? Or do you just have them plugged into old phones? Was thinking about wiring in some cigarette lighters so I could run phone chargers and then running a couple phones I don't use anymore and either using the phone cameras themselves or plugging into something else, tablet or laptop that would capture the video. I'm not a techie though, so not really sure how to synch them all up and all that.

[clarkspeed]

For quick down and dirty videos I have a couple of rechargeable cameras that record to microSD like these:

 

https://www.ebay.com/itm/Nanny-Cam-Video-Camera-Tiny-Small-Pinhole-DVR-Digital-Spy-Hidden-Secret-Wireless-/381414328453?_trksid=p2349624.m46890.l49292

 

I bought them used a few years ago, but I see you can order from China for 12.95. They record a good 45 minutes or more and are easy to ty-wrap in unusual locations (no jokes please). You have to manual turn on, but they are pretty durable in a metal case even though I don't care if they got destroyed. I've mounted them in fender wells, under the hood, pointed at gauges, and pointed at my feet.

 

Race Render software is really good If you want to sync with other videos and edit.  Easy to use.

 

If you want track videos, I recommend an old GoPro or clone that is hidef. They are just so cheap now and easier to mount than a phone.

 

Edited February 7, 2021 by clarkspeed

[JMortensen]

Interesting. Cheap enough that if they fall off or get hit by a cone I wouldn't really care. Looks like some of them have remotes to turn them on and off or turn on via motion detection. That would probably be better than crawling all over under the car. Very cool. Thanks for the idea. I have a good phone mount that I've been using for a couple years. Works well, but I do feel a bit hesitant about the phone falling out and getting destroyed. I guess the advantage of the phone would be using a piece of software that would measure g forces and speeds and all that. Can't do that with $13 cameras yet...

https://www.ebay.com/itm/Cop-Spy-Nanny-Camera-Wireless-Hidden-Motion-DV-HD-1080P-Mini-Car-Dash-Cam-USA/324272628932

 

[clarkspeed]

I only use those things to answer the question "what's going on under there"?  I run a full data acquisition system and add forward and rear facing cameras when in a race. I got real heavy in the data side of things a few years ago and now do analysis for others too.

 

If you want to dip your toes in data, the phone apps are getting pretty good. Also the Aim Solo. And I think some of the new action cameras collect data. 

[JMortensen]

I need some DA, both for the car and to analyze my driving. My intention was to get some app for the phone and then get some data from 2020, but I ended up not racing at all last year. Problem is that DA isn't all I need.  

Right now I have a little cash together, and it's still looking like my racing schedule is going to be sketchy at best this year. So the question now is; new set of Avons (don't want to buy a set and let them sit a year), data acquisition, wing, shocks (car could really use these), online driving coach (people will look at your gaming DA and coach you on driving) or buy fiberglass supplies and learn to do that better and maybe make a wing. If nothing happens this year, could just sit on the cash and see what's going on next year, since DA stuff seems to be getting cheaper over time. Thinking if I have to sit the majority of the year out, online coaching and fiberglass supplies makes sense because it gives me something to do and will improve results when the racing resumes.

[clarkspeed]

You are mostly autox'n correct? That is a special category.  If HPDE or time trials let me know, it will be a different answer. First of all, for a beginner, DA is easiest applied to driver analysis.  Next concern is analyzing a 60s run you can only look for major issues or make a comparison between run 1 and run 2. If you tried a slalom 2 different ways, you could definitely determine the best choice, but analyzing a single run is doubtful you would see something you dont already know.

 

I am interested in your online coaching statement. I assume you are racing online. I also run iRacing quite a bit. I personally wouldnt invest more in sim driver analysis unless I wanted to improve my sim racing.  I couldn't imagine sim coaching would make me better in a real car. I am not going to list the reasons now.

 

Based on your post, I would say there are 3 definite ways to get faster (WIN). A real school, either autox or track. Lowest cost way to improve the driver. Shocks if you really need them, but that is another $$ catagory.. Tires, always make you faster for the next race, and fiberglass??? I say do that in addition to whatever else you decide. I just started down my fiberglass learning curve and it is fun, and not that expensive. You can definitely master it quickly.

 

 

 

 

[JMortensen]

Plan was to start doing HPDEs last year. Again, this year might or might not happen, but yeah otherwise primarily autoxing. I did do an Evolution 2 school a couple years ago and wasn't that impressed and didn't learn anything that I could point to. Look ahead, no further ahead, no further than that was about it. They wouldn't let me spray my R1Rs down between runs and they wore the shoulder off really bad due to overheating. I left kinda pissed off about the whole thing. Their Evo 3 school would probably suit me better. Basically you and a really good driver both drive your car, back to back, switching drivers. That seems like it would be much more helpful.

I'm already a pretty decent driver, came in 2nd, 5th, and 14th overall at the 3 autoxes I went to in 2019 and that's on used slicks that were 2 years old already. 14th was the day after a demo derby and the track was slick and having gigantic tires just made it like driving on ice. I think a new set of Avons would put me right up at the front. When I jump online I find myself getting creamed though, so I know there's a trick or two I haven't learned yet.

As to sims, I started with a TH300 wheel and pedals and a T3Pa shifter on the PS4  and I had AC, ACC, Dirt 4, and GTS. Found myself sticking with GTS and AC. AC was the more realistic, GTS is more of an arcade game. ACC just wasn't fun, Dirt 4 never really captured my interest. Then I started changing things around, bought Fanatec pedals which are much better, and I switched to PC. I was doing iRacing but found it expensive and the steering didn't feel realistic at all. Also bought RF2 and a crapload of cars and tracks, but it has the same dead wheel problem, so right now Assetto Corsa is my game of choice. I hear that if you use a direct drive wheel iRacing is much better, but I'd rather not spend all that money on a wheel when I can put it towards the real car. AC has H2H, but I've been racing against the AI and just turning laps. Recently I had a break through and picked up a lot of time on a particular car at a particular track, but the annoying thing is that I don't know why I got faster. That's the kind of thing that I think online coaching could teach me. I mean, you have the inputs there, so should be able to use the data just the same way you would out of the race car, in theory.

I do need shocks. I had my Bilsteins revalved and the guy who convinced me to do it that way was... unorthodox. Stock valving was something like 250/75, Katman suggested 300/100, this guy valved them 100/100. Then he retired. LOL. Car porpoises a bit. The 30mm Bilsteins have linear pistons, would really like to move up to 36mm, digressive and have some adjustability. But if I'm not going racing, I think FG would at least be a creative outlet, and I already bought a vacuum pump.

[clarkspeed]

OK then.  Yes a school that puts the coach in the passenger seat or even better, let a national champ drive your car with you riding.  I friend of mine who was a very respected driver let John Thomas drive his car one time and was blown away at the techniques he used.  The one area where sim racing = real racing is thinking out of the box.  That is what separates the good drivers from the mediocre. Being able to try something new and learn from it immediately. Guys like me are doing that to gain seconds, the pros do it to gain tenths.  And no substitute for more seat time.  Anyway, from my race experience, anyone less than professional level drivers can usually gain the most time by driver improvement. And for both racing and Autox, easy to drive, comfortable, confidence in the car makes you faster.

 

Sounds like you are way ahead of me on sim racing.  I run the Thrustmaster stuff on iRacing and don't know any difference. Personally i think AutoX and lapping are just too different to make a direct comparison.  An unfamiliar 60s blast averaging 40mph through cones with slaloms is not like lapping a big track. Personally I suck at Autox for that reason.  I am still not exactly sure my sim racing will directly affect my actual track racing.  Time will tell.  A SCCA Runoffs driver I worked with ran sims for a couple of months, then hired a local driver coach for feedback during practice, then hired me to analyze his data for a new track he went to.  That covered all the bases and his driving was near perfect and he got a podium.  There are also many differences between the sim cars and an actual race car.  The sim cars seem kind of sterile compared to the real thing.  There are so many variables they hold constant. Long story short, IMHO, online coaching in a sim racing format would not equal faster AutoX times.  Learning a new track? Probably. Being a better sim driver, definitely.  Like most things, the question is not if you can get from point A to point B in skill, of course you can with practice. But can you get there faster with coaching?  Personally data analysis is my thing, I am very good at it, but I have yet to look at a single lap from my sim runs.  I am usually about 2-3 sec behind the fastest person and I am confident I could find another second if I had more seat time (time to experiment).  If I ever I feel like I have hit a wall and can't find anymore speed, I will dive in.  Finding speed myself and learning race craft still seems like the fastest way up in sim for me, which is not the case at all when I am racing a real car.

 

So tires are a "no brainer".  They usually are.  Shocks sound like a good investment also.  I had the Bilsteins on my last car.  I called them and gave them my sprung and unsprung weights and motion ratios.  They send me back valving equivalent to 300/100.  Which looked like the 100 compression was linear and the 300 rebound was digressive on the dyno chart.  I raced those for a number of years and they felt pretty good.  If you are 100/100 you are probably jacking down during a run.  Anyway, a revalve can make the car feel much different if the low speed valving changes. Low speed shock valving is probably the secret sauce to a good Autox run due to all the quick transitions.

 

Clark