Compression vs. Rebound- Not What You Think

Most people seem to find the effect of compression forces from shock absorbers and what they do fairly easy to understand. How rebound forces from the shocks affect the car seems to be more difficult. This may be a clue that most of us are thinking incorrectly about what the shocks do during a turn.

Time for a thought experiment. I’m not going to help you out with diagrams or pictures. You can find those lots of places. Consider this a challenge. Just THINK about it.

Say you turn the steering wheel left. The car turns left and at nearly the same time weight begins to shift from left to right. The car body, because it is supported by suspension springs, begins to roll to the right. The right-front spring compresses, putting more load onto the right-front tire. The left-front spring extends and load is removed from the left tire. The additional vertical load from the springs is directly proportional to the extra spring compression. Voila! We now understand weight transfer. Right? Are you sure?

Caveat: For the purpose of this thought experiment please forget about the energy dissipation effect of shock absorbers. That amount of energy is small, in any case, and has more to do with how the tires are controlled over bumps, which we are not talking about here.

Let’s go with this idea a little longer even though most of you already know it’s wrong (or at least incomplete) and consider the shocks.

When the right-side spring compresses the shock is also compressed. The shock resists this motion with a compression force. That force pushes down on the tire via the suspension linkages and up on the body. So, the shock adds to the weight transfer and slows the body roll. There’s more downward force on the tire while the shock is moving, which is prior to the steady-state condition, so it sped up weight transfer. Once the car stops rolling the shock force becomes zero but by then it has been traded for spring force and we are steady-state cornering, fat, dumb and happy, as they say.

This is the way I thought about shock forces for a long time. But, then I started thinking about the left-side shock. It’s extending as we enter the turn and the body begins to roll and is therefore resisting that extension with rebound force. So, it pulls up on the tire, reducing the load at the tire patch and at the same time pulls down on the body of the car, slowing the roll. Note: Rebound forces always pull weight/load off the tire just like compression forces add weight/load to the tire.

Now I have a question for you: Where did the load go that got pulled off the left tire?

Well, the only place it can go is over to the right tire. But, how does it get there?

See the problem? How does the load that the left shock picks up from the tire patch get over to the right side and push down on that tire? Do all the free body diagrams you want… I don’t think you’ll find it. We think we understand how the right shock pushes down on the tire with compression force, but how does the left shock, all by its lonesome, send weight over to the right?

Maybe it will help to understand or at least appreciate the dilemma if we remove the right shock entirely. The left shock is over there resisting the roll of the sprung mass of the car by picking load up from the tire patch but how does it speed up weight transfer and create a load on the right tire? Magic?

The way out of this thought-experiment dilemma is found by gradually stiffening up the shocks until they won’t move at all. Just replace them with solid bars. You can leave the springs in there, but they can’t do anything if the shock is locked up, right?

Now we have a kart. The suspension cannot articulate and the springs are just dead weight.

Is there any weight transfer when we turn a kart? Of course there is. Turn the wheel left and it wants to throw you out the right side of the kart. Plenty of weight shift from one side to the other. The tires on the outside of the corner see much higher vertical (and horizontal) loads during a turn even though there is no suspension.

This is called unsprung weight transfer and it happens faaast, i.e. instantaneously, as long as we neglect the elasticity of the tires and the structure itself. If the tires produce a lateral force there is instant, unsprung weight transfer. Can’t get faster than that. (Please don’t bring quantum physics into this!)

All cars with a working suspension have both unsprung (fast) and sprung (slow) weight transfer. Sprung weight transfer is often called “elastic” weight transfer and it’s a bit complicated to calculate (and has been discussed previously here) but we don’t need to worry about that now.

Springs create (or allow) the sprung weight transfer at the expense of unsprung weight transfer and then add a little extra to the total. The softer the spring, the more and slower the sprung weight transfer. The full weight transfer cannot be complete until the sprung weight transfer is complete, i.e. the body stops rolling. In the end the total weight transfer is the unsprung weight transfer that happened instantly plus the sprung weight transfer that took a finite amount of time.

If the shock is replaced with a solid bar and prevents the spring from moving then you get the fastest possible weight transfer. Effectively, you have converted the sprung weight transfer (that would have taken some time to accomplish) into unsprung weight transfer that happened instantly. You have also slightly decreased the total weight transfer by eliminating the extra bit that comes from allowing the body to roll.

Now you know the answer: both shocks, left and right, by acting somewhat like stiff, solid bars and resisting body roll temporarily convert some of the elastic, sprung weight transfer into unsprung weight transfer. They speed up weight transfer by “borrowing” some elastic weight transfer from half a second in the future and bringing it into the NOW. The stiffer the shock the more like a solid bar, the more like a kart, and the more weight transfer gets borrowed from the future, making it all the more easy to dynamically overload the tire.

This is why shock forces that are too high, compression and/or rebound, can contribute equally to causing that right front tire to go over its lateral grip limit during turn-in to a left corner and start to slide, producing understeer. It’s the total that matters.

Yes, the extra compression force from the right side shock helps to counteract the increase in instantaneous weight transfer. But, similarly, by jerking weight off the left front tire with too high rebound force you can start an oversteer sliding event from the inside tire which then cascades to the outside tire.

Even if you have near zero of one type of force, say very little compression, if the rebound forces are too high you can make the same bad things happen. And vice-versus.

An Autocross Season- Part 16: Finally, some tires!

I’ve been on pins and needles hoping that one of my three orders for rear tires at Tirerack would come through. Finally, a set of 305/30-19 Falken 660s arrived. This will allow me to do another event before Nats. The present tires have 84 runs, so I wouldn’t have run them again until the test & tune course in Lincoln.

I still have an order for a pair of the new 315s, but they are tentatively expected in October, not soon enough for Nats.

One of the other orders was for a pair of Yokohamas. Just got an email from Tirerack saying they are now expected Spring of 2023!

An Autocross Season- Part 15: Bristol Pro & Tour

Fig. 1: Early Arrivals At Bristol Motor Speedway

After the last local event we had a good setup. We thought.

Aaaand, we were right!

At Bristol Silver Ghost was great! Grip for miles, transitioned well, no bad habits. I mean, you can provoke understeer if you ask too much, too fast on corner entry, and you can produce oversteer if you give it too many beans on corner exit, but you really have to try hard. You can slither it in the slaloms and 4-wheel drift it around offset cones like nobody’s business.

Fig. 2: Yellow-jacketed Co-driver Spotted At The Yellow Bridge

The drivers were not quite as good as the car. At the Pro-Solo I moved from 4th and out of the trophies to a trophy and 2nd place when I finally combined a good start with a good run in the last session. 2nd, 3rd and 4th were extremely close, but we were all far behind first. This is only my second trophy at a Pro. I was pleased to be 3rd on the Masters index, got into the Super Challenge, won the first round and should have won the second, but a 1+ second reaction time (due to forgetting to switch off traction control until the yellow lights were counting down) put an end to the Challenges for me. I lost that challenge match by 0.04s after giving away ~0.40s at the start.

Fig. 3: Pro-Solo Start At Bristol

The Pro-Solo courses were unusual, even by Bristol standards. Each side, though not symmetrical, had the longest acceleration zone I’ve ever seen at a national event, good for 80mph for the C6Z06 AS cars on the left side. My car maxes out at 70mph in 2nd gear, so I had to sit on the rev limiter for 1.5s on the left side and 1.0s on the right while my auto-box competition continued speeding up. I expect entering a slalom at 70 to 80mph was a new thing for many drivers…. it sure was for me. Such unusually long, high-speed sections damage fairness both within the class index system and within different cars in specific classes and endanger course workers. I saw a worker run to avoid being hit by a car sliding through what would normally have been a safe location given slower speeds. The car slid directly over the spot where he and others, including a person from my region, had been standing earlier.

Then there were the mid-site bumps, which are always a consideration at Bristol and part of the site’s charm and challenge. In the Pro they were located at the 60-65mph point of the long acceleration zone on both sides. The coneage was incredible. If the driver made an input, or otherwise had the car unevenly loaded when crossing the bumps, the car would take a big hop one way or the other, at the least, or spin, at the most. (My work assignment gave me a perfect rear view of the shenanigans.) The Porsche GT3s had a terrible time of it. The back end would violently bounce. (With the exception of the 96/196 SSP car. That one traversed the bumps relatively smoothly. Better shock valving? Less stiff springs?)* The mid-engined cars were next worse. My front-engined Corvette on Penske’s? After the first round someone asked me how my car was handling the bumps. I answered, “What bumps?” Only after that did I really take note of them while driving. But, on one run my co-driver proved that if you didn’t have the wheel straight when hitting the bumps bad things (like an off-course) would happen even to my car.

Fig. 4: Beaver Creek Winds Around Bristol Motor Speedway

The courses for the Tour event were the best I’ve seen on this site. Kudos to course designers Dave Marcus and Charles Krampert! The day 1 course really kicked my butt. I was a full second off the pace and don’t really know why. If I had to guess I’d say I was slow to appreciate how much speed you could carry through certain sections. Plus, I’m consistently not as fast through slaloms as my co-driver. I think I’ve been slow to find the dynamic limits of the car now that it has such good lateral grip. This put me in fourth place at the end of day 1, one spot and half a second out of the trophies. On day 2 I was right up there, however. I turned in the second fastest time in the class on my second run with a good chance to improve on the third. It didn’t happen. I totally blew the first big turn by entering it too fast after doing it well previously. I’ll present a technical discussion of that turn below. I stayed in 4th and my co-driver took 5th.

First Turn On Day 2

The First Turn on Day 2 was similar to ones I’ve seen several times previously at that location on this site. It’s quite complicated to look at and maybe confuses some people, but the fast way to do it, I think, is simple in theory if difficult to accomplish in practice.

Below is the turn as drawn on the course map, except that I think cone 503 had been removed. The exit was not pinched as shown below.

Fig. 5: First Turn As Drawn On The Course Map (courtesy Charles Krampert)

I think most of us, with nothing more to go on than the map, would plan to drive this corner something like the dotted line path as shown below. Enter wide, be on cone 523, cross from 523 to 502 as fast and shallow as possible, exit on 502 at the angle that allows maximum acceleration yet still make the 500/501 gate.

Fig. 6: Correct Path Assuming No Complications

In reality there is a complication, i.e. a complex slope. Take a look at the next three figures to see that we have a downward slope from left to right that flattens out in the corner exit path, plus another slope that rises from front to back in the center of the corner.

Fig. 6: First Turn With Section Views Annotated

Section A-A from Fig. 6 shows a slope from above cone 523 down through 502, extending a distance beyond 502 and then flattening out. Section B-B shows another camber from an axis through cones 523 to 502 sloping upward to the rear of the corner as defined by cone 515 and others creating a back wall.(These figures are not to scale. In reality the Sect. B-B slope is less than the Sect. A-A slope.) You may need to think about the sections for a while to understand the two slopes I’m trying to illustrate if you’re not familiar with section views.

Fig. 7: Section Views

What happens to most drivers when they attempt to drive the path shown in Fig. 6 is that the slope shown in Section A-A makes them understeer wide of cone 502, fall down the slope crossing the chalk line with the left-side tires and land in the flat below 502 where they struggle to get turned and then back up the slope toward cone 449 while not hitting cone 501. Therefore, the vast majority of drivers find themselves on a path similar to, or worse than, what I’ve shown in Fig. 8. It feels terrible to an experienced autocrosser. It feels, and is, very slow. I got sucked into this line plenty of times at previous events held on this site. I vowed not to let it happen again!

Fig. 8: Actual Line Most Often Taken, If Inadvertently

When walking the course my plan was to take advantage of the slope shown in Sect. B-B. Everyone seems to see and worry about the other, more severe slope, the one shown in Sect. A-A, but taking advantage of the B-B slope seemed to me to be the solution. My thought was to go deeper into the turn, effectively using the discredited “late apex” strategy, but using the B-B camber to help slow the car down in the first half of the turn and increase the efficiency of the entire turn to allow an exit close to cone 502 and at an exit angle that allowed the car to 1) stay on the A-A slope, never venturing lower than necessary and especially not all the way down to the flatlands, and 2) achieve an early acceleration into the next, fairly significant acceleration zone. This would also give the advantage of not spending as much horsepower having to climb back up the A-A slope on the way to cone 449. This planned path is shown in Fig. 9 and is basically what I drove on my first run as verified by data and video.

Fig. 9: Path Taken On Run 1

Below is a video clip of this corner as I actually drove it in Run 1.

Fig. 10: Video Clip Of First Turn, Run 1

Well, the first turn felt pretty good in Run 1. The 180 degree turn in the back of the corner seemed efficient and quick. But, I was way off of the entry cone, cone 523, and thought it could be done better, i.e. shorter. (It always Saves Time to go shorter and slower in a smaller arc than longer and faster on a bigger arc.)

On the next run I tried to do essentially the same thing but wanted to be right on cone 523 and slow down more so that I could turn sharper with less distance traveled. The result was a line similar to what’s shown in Fig. 11, though I over-slowed a bit on entry and flattened out the arc in the middle, meaning that the tires fell below their multi-tasking limit for a moment and costing time. (You can see the double move with the steering wheel.) The clip of the actual run is shown in Fig. 12.

Fig. 11: Path Taken Run 2
Fig. 12: Video Clip of First Turn, Run 2

While I think I was faster than most in this corner in Run 1, the path taken in Run 2 was 0.20s faster than Run 1 as measured from the entry braking point to the next braking point.

Let me know in the comments how you think this corner should really have been done! I’d like to hear your thoughts.

One more thing: this is one big reason why the Bristol events are so successful:

Jeff Cox, event master, 8PM on Friday, still working to prepare the courses for Saturday
  • *I later learned that this car had electronic shocks with aftermarket tuning, something only allowed in Super Street class at present.

An Autocross Season- Part 14: A Shocking Mistake!

In planning a test program for Silver Ghost’s next event (which was today) I discovered a problem. I made a mistake when specifying force values for the Penske shocks.

It’s a little hard to explain, but I’ll try. I did the force calculations for 3in/s shaft velocity. Then I instead told Penske to put the knees at 2.5in/s +/- 0.5 in/s. Well, Penske put the knees more or less right where I said, at 2.5in/s, and they hit the force targets for 3in/s accurately with the adjustments at a particular setting. This means I have 20% too much force at 2.5in/s at the nominal adjustment settings. (3/2.5 = 1.2)

So, instead of having 90%Critical damping at 2.5 in/s I actually have 108%Critical. The shocks are stiffer at the nominal setting than anticipated. This is what has caused no end of issues with grip and balance as it turns out that for my car, at least, 108% of Critical is definitely too much damping for best grip on a bumpy site.

Mistake Explanation

Here’s how I screwed up, explained with the chart above.

I calculated a force at 3 in/s, let’s say it was 100lbf. (100lbf is not far from the real number on my front shocks in bump.) That’s point 1 in the chart. (These numbered point are called “knees” in the shock dyno curve.) I asked for as linear an increase as possible from zero to that point. In theory, that produces the line I’ve labeled “90%critical damping.” The force at every velocity along that line represents 90%critical damping. I did exactly the same thing for rebound.

But, when I wrote the spec, I changed that point to be at 2.5in/s, that’s point 2, and I forgot to recalculate the force for 2.5in/s. So, Penske gave me point 2, which is 100lbf at 2.5 in/s. See how the line from zero to point 2 has a steeper slope? Every point on that line turned out to represent about 108%critical damping.

What I really wanted was the line that leads from zero to point 3, which is a lesser amount of force at 2.5in/s and would lie on the original 90%critical damping line.

Also, you may notice that to the right of point 1 the almost horizontal line is significantly above the sloping line to the right of point 3. This means that for sharp bumps, which produce higher shaft velocities, the forces are also higher than they need to be. At those higher shaft velocities we want less force, not more, to give us grip over bumps. That’s why we want highly digressive shock pistons. Beyond the knee the slopes “digress” to much lower values. A horizontal line would be best, but is not achievable.

Before today’s local event I took the dyno charts and figured out what settings, front and rear, would get me back onto the 90% line and also knock down the slope of the line after the knee. I had to put bump at full soft and then reduce rebound until the total got me back on the 90% line. Then I dropped the rebound even more, down until it was equal in magnitude to bump (I didn’t want rebound to go under bump) and found that at that point I was at 83%Critical. All these numbers are approximate, by the way. What I’ve drawn in the figure above is a very simplified, idealized representation of what are really much more complex curves. To get the shocks to be really what I wanted, with some adjustment in either direction, I’ll have to have them revalved. (Maybe next year.)

That’s where we started today, at 83% with bump forces roughly equal to rebound forces, i.e. a 1 to 1 ratio. The car was soooo much better! Grip was high and it didn’t push at all, using exactly the same tire pressures on exactly the same surface that was so difficult to run on two weeks ago and again one week ago at Peru. I was near the top of Pro class, much closer to the two or three really fast guys than I was 2 weeks ago.

During the lunch break I adjusted the shocks to the new 90%critical settings. I got a little push from the front tires and little more bounciness over the bumps. In the end my co-driver and I decided we liked the 83% setting on the front with the 90% setting on the rear, along with some tire pressure changes that we also tested today. That’s what we’ll use at Bristol, which is a bumpy asphalt site like our local site.