Publishing my Autocross Secrets

[DG-FNR]

OK gang,

Given my ever-increasing responsibilities at the Regiment, it has become clear to me that my racing days are over. I have not the time, inclination, nor finances to mount a racing effort as intense as I have in the past. And given that nothing is worth doing if it isn't done to the best of one's abilities, that means I'm pretty much retired from racing.

Therefore, there is no further need for me to sit on all my technical secrets, and I am publishing *EVERYTHING*. Every single thing I know is going to go on the farnorthracing.com website.

This will take some time, as I have to squeeze it into my "spare" time, but the intent is to brain-dump all I learned in 8 years of National-level SCCA ProSolo. I had wins in 4 of those 8 years, won three divisional championships, a Street Modified ProSolo championship, and a Canadian National championship in those years, and I busted my ass and spent a ton of money doing it. Well, now you can be the beneficiaries of my blood, sweat, and tears.

Consider it an early Festivus present from me.

The meat of the info is on http://farnorthracing.com/autocross_secrets.html but there will be updates to other websites as well. I'm on a tear to get this done before I no longer remember any of it and don't have the time to complete it.

Be sure to check in regularly for updates. They will be plentiful. There is a lot of information to convey.

Enjoy!

DG

 

[asian312]

brb...this could take a while.

Thanks DG. Your efforts and knowledge is greatly appreciated.

 

[Caithness]

Can't wait to see the finished product. The whole Shocks section looks like excellent reading (comparisons of manufacturers, shock tuning, interpreting shock dynos) and I'm interested in reading more about picking spring rates based on natural frequency. These are concepts I've seen mentioned but never developed a good understanding of.

 

[Elvenhome21]

Very Good info so far

 

[jtmcinder]

Setting springrates by natural frequency is more often used on passenger cars, where the goal is to reduce pitching when the car hits a small rise or swale. (Pitching is the major cause of car-sickness.) It is rarely used on racing cars, since it's much more important to set the handling balance, as in understeer vs oversteer.

Even if you do set up a car in terms of the frequencies, you really don't want to use the simple rule of having the rear frequency about .2 Hz above the front on all cars, because that only produces a nearly-neutral car when the nose is the heavier end. You'd have to use some crazy swaybars to balance an MR2, for example, that had a higher rear frequency. With that said, on DSMs, the rule might provide a good place to start, since our cars are very nose-heavy.

The above brings up the last point with regard to frequency: it ignore swaybars. If you spend most of your time turning, instead of going straight, then it's much more typical to plan things in terms of front and rear roll stiffness, instead, since this measure includes the springs and the swaybars.

- Jtoby

 

[DG-FNR]

And yet, oddly enough, the technique of setting spring rates by natural frequency was taught to me by a former Formula One engineer during one of his race engineering classes ( http://www.optimumg.com )

Even more oddly, it was confirmed by a couple of my fellow students, who included a Trans Am engineer and an ALMS engineer.

Natural frequency, unlike spring rates or wheel rates, is comparable across platforms and provides an empirical measure of "stiffness". It is an incredibly useful tool for getting your springs sorted out.

It is used by professional race teams all the time.

DG

 

[jtmcinder]

Of course. Good engineers don't ignore useful information. It's more of a question of how to weight (as it were) the various types of information that are available when designing a suspension for a particular application.

Road racing cars, especially formula cars, are much more sensitive to pitching, so the engineers in those fields undoubtably weigh the relative frequencies much higher than, say, a person setting up a car for autocross.

In other words, while every decent engineer knows how to calculate frequency and how it affects the behavior of the car, not all of them stress this variable. And, since your new website appears to be aimed at people setting up cars for autocrossing, it surprised me quite a bit that it is so focused on frequency. How much time do autocrossers spend going straight ahead? Why be so concerned with a variable that only concerns pitching when the amounts of front vs rear (lateral) weight transfer are so much more important and is not covered by this variable at all?

Alternatively, if you prefer: how much time do you think that Tunnell or Sias spends on trying to get the rear frequency a few points above the front? Why are ACM and Tevenor so fast when their rear frequencies are more than .5 Hz above their fronts?

There's a bunch of really good information on your new site -- the velocity plots from Koni Sports are very useful as it makes it so clear why one needs to avoid turning them all the way to full stiff -- but the extreme focus on frequency for choosing springrates is not a strength.

- Jtoby

 

[DG-FNR]

In an ideal case, front and rear natural frequencies would be the same.

But it has been found that a slightly higher NF in the rear avoids upsetting the car over bumps by ensuring that the rear reacts to the bump faster than the rear, avoiding dramatic pitch changes.

But - as the article states - there is more to it than that. A slight rearward NF bias promotes a touch of oversteer, which is good for transitional response. It gets the tail moving; helps the car to "dance" which is good for fast slaloms and offsets.

But MORE than a couple of tenths of a Hz rearward NF bias hurts more than it helps. It starts promoting entry "twichiness", compromises power down characteristics (especially on RWD cars) and overall slows the car down.

This has been verified on dozens of other top-flite autocross cars (people were getting me to work on their setups) and was one of my most closely-held secrets. It makes spring rate selection absolutely trivial.

And interestingly enough, since that portion of the article was published, I've had a slew of letters from some upper-level people who calculated the numbers for their cars, whose setup had been determined through exhaustive trial and error, and reported that their numbers matched mine within a couple of percent. I'm not going to name names (that's their business, not yours) but there are several National Champions in my client list.

It works:

You pick a front NF based on the surface you spend them most time running on and how sticky your tires are.

You set the rear NF to be two tenths of a Hz higher. You work out your maximum roll angle at maximum lateral G, and it it is too high, you add bars to bring it back in line with what you need.

You work out the static weight distribution, and you adjust the bars front/back to provide the total roll resistance you need to get the roll angle you need, and you adjust the distribution between the bars such that the roll resistance distribution matches the static weight distribution or a little bit rearward biased.

You set the initial shock valving to be 65% critical, and then you tune it using the suspension speed histograms you get from your suspension position sensors.

And you have a car that is blindingly fast, just like that.

FWD is a special case, but this formula absolutely works and has been tested on many, many cars. If a given setup deviates too far from this formula, that is an indication that the setup may be suboptimal.

DG

 

[jtmcinder]

I agree, of course, that a slightly higher rear frequency is used to avoid pitching over bumps; the whole point of this in passenger cars is to reduce car-sickness.

Where we seem to be at odds is over using natural frequency to balance the car (or, even, to discuss the balance of the car). You can't do this, because natural frequency (as it is calculated by auto engineers) ignores the swaybars. So you cannot say that a car with a slightly higher rear frequency will have a touch of oversteer, since, for all you know, the car has a massive front bar and no rear bar, and actually plows like a pig.

Do you see the problem? Balance is about corner-weights and relative roll-stiffness (all else, such as camber curves, being equal, as we both seem to be assuming). Frequency is directly dependent on corner-weight, but is only one of the inputs to the calculation of roll-stiffness. That it takes corner-weight into account is good, but the fact that it totally ignores other factors which determine roll-stiffness makes it impossible to use to balance the car.

I'm not saying that it's unrelated. I'm saying that it's only part of the story, so writing a page that is so focused on this one issue is not good. You can start with frequency, since getting car-sick on the way to an event wouldn't help your day, but you have to move on to roll-stiffness to be complete.

- Jtoby

 

[DG-FNR]

Carsickness has nothing to do with anything.

As a historical point, yes, working with NF came out of attempts to improve ride quality. One of the amazing thing about talking to the Milikens is about how little was understood about handling and suspension performance up until the late 60s, because all the R&D up to that point had been focussed on ride quality in passenger cars.

But once maximizing performance started to be a goal, and the engineers started learning about how to maximize performance, NF turned out to be a very useful tool from a performance perspective as well.

It is THE only way to objectively determine "stiffness". And it is needed to work out damping.

Big pitching moments don't just upset drivers; they upset cars as well. A car that keeps the sprung mass more or less level over bumpy sections can be driven faster through those bumps than a car that is pitching all over the place. A slight rearward NF bias keeps the car flat, and it goes faster for it.

Believe it or not; I don't care. Those that do believe it will go faster than those who do not.

DG

 

[jtmcinder]

When talking about race cars, I agree: car-sickness isn't relevant. But it's a huge issue for passenger cars and the solution was the small difference in front vs rear natural frequency. (In case you didn't know: I have studied and published articles on this. It's a fascinating topic, since the current explanation for the effect of pitching on car-sickness is that it is due to a system that is built into humans [and other mammals] to prevent death by poisoning.) In any event, the math involved is very simple, so, once you know about it, it's very easy to avoid. The only factors we haven't mentioned that are involved are wheelbase and target cruising speed.

As to the suggestion that the only objective way to measure stiffness is by use of the natural frequency, that's a bit of an over-simplication. There is, of course, a simple relationship between stiffness and natural frequency, but it includes another factor: the sprung mass. The idea that engineers ignore the sprung mass when designing suspensions is, well, a tad insulting to said engineers.

- Jtoby

 

[Jon Lane]

A question from the amateur street side of things: Dennis, Can you work backwards, and if so, how?

In other words, assuming RM anti-roll bars, Koni Sports, adj. perches, and an inch and a half drop -- all ubiquitous parts and methods for, say, 2G cars -- what would be suitable characteristics for neutral street behavior and a reasonably comfortable ride?

 

[steel_3d]

Along the same lines, are you planning on publishing all your measurements and calculations? This would be useful for us slackers in that we wouldn't have to go and measure our motion ratios, camber curves, unsprung mass, and all that stuff ourselves. It's a big waste of time for everyone to be measuring all the suspension parameters of the same car (2g) over and over. Having all the calculations laid out would make it even easier, just plug in the numbers that are different for our setup and get reasonable results. I realize that not everyone's car weighs the same, weight distribution, fwd vs awd, etc etc, but it would be a great starting point. Of course the 1g guys would still be outta luck

Great articles, by the way, really appreciate what you're doing here! Great timing too for me personally, since I haven't started working on suspension yet. Might eventually save me from a couple of missteps.

 

[90AWDTalon]

First I want to say thanks for publishing this.

I have a few questions:

Do these methods apply to all cars or just 2g's? I autoX a 1g awd and a miata.

Is natural frequecy the same as bounce frequecy?

You say to start at 65% critical on shock valving. How do you determin this number? Does it mean if you have 16 adjustment positions on your shock that you would start at 10?

 

[DG-FNR]

I'll answer all these questions later - some in the text of the article, some here on the forum.

But right now, I have to go XMas shopping for Niner.

DG

 

[mavisky]

Good info, glad to see we can have an intelligent discussion on this board without the typical mudslinging that you'd find on most boards.

 

[jtmcinder]

I agree. The idea of setting the front suspension to something between 2 and 2.5 Hz as a starting point is very interesting. It's definitely worth a complete discussion.

- Jtoby

 

[mavisky]

I agree. The idea of setting the front suspension to something between 2 and 2.5 Hz as a starting point is very interesting. It's definitely worth a complete discussion.

- Jtoby

So you're saying my Sportline + agx combo is hopeless

 

[RipperXX]

So are you saying that KYB AGX shocks are crap? It seems so. This makes me curious as to what shocks you WOULD recomend for us 1G guys, seeing as how our parts list is a little shorter.

 

[mavisky]

KYB's are better than stock, but not nearly as good as a set of Koni's.

 

[Rob10_99]

this would be good with illustrations as well, for those of us who cant read

 

[Caithness]

As it says in his guide, Konis are often the best OTS shock for cars with few options because they make them to fit almost everything. It seems for max effort stuff he really likes the Bilsteins rebuilt to custom specs to match the suspension but that'd be a custom job on a 1g.

 

[steel_3d]

Well, this might be an interesting choice for 1g's: http://cgi.ebay.com/ebaymotors/BILS...009QQitemZ190050265879QQrdZ1QQsspagenameZWD1V

Unfortunately, they don't have a 2g application...

By the way, I've seen the argument that twin-tube shocks are inferior compared to mono-tube. Some of the aruments are worse heat dissipation, and smaller piston area presumably leading to less control. Are all Bilstein shocks twin-tube?

 

[pneumo]

That ebay link shows Bilsteins touring class shock, which is about the same as KYB's GR2.

 

[underradar92]

That ebay link shows Bilsteins touring class shock, which is about the same as KYB's GR2.

I wonder if they can be rebuilt/revalved. That may make them a viable option.