Interpreting and Comparing Shock Dyno Results

[gt6racer]

There is often confusion on shock measurements, mainly due to the way they are presented.
When we measure the forces from the damper over a single sine wave input ( one turn of the crank ) the most common output we look at is the "football curve" plot of load Vs displacement. We can also look at the force/velocity plot which will give the output you saw, and which riderate has accurately explained. However, these days, neither of these is the way we measure and present damping force.
What is typically done is that we run a series of sine waves at various frequencies, and the software records the peak compression and rebound forces at each frequency. These discrete points are then potted as a single force velocity line as the "characteristic" that we use to specify damping levels. For your purpose, this is the output you will want to use.
Damping models give output similar to this latter line, simply because they don't (typically) include the more complex real world factors that make the real damper act as it does.
The raw load/displacement and load/velocity data is however very useful if you really want to understand what the damper is doing.

[Jersey Tom]

...I still feel like step 1 should be figuring out what they're trying to really accomplish.

To figure out baseline settings for the shocks to go into the initial test with, using simple calculations with critical damping and damping ratios. I am aware that these settings will be adjusted after gathering test data.

So you are trying then to find the damper setting which most closely resembles a constant slope force velocity curve of a specified rate (critical damping). That, is a more specific goal than the very vague, "I'm currently trying to understand how to properly compare different settings on a shock."

Could go about that two ways. One, you could just leave the data as it is. No need for post-processing. Come up with a side data set which is the F/v curve for your critical damping, and see which damper dyno test most closely matches that (by say a R^2 number).

Or the better (more legit) route would be to make a spring-mass-damper model in Simulink or something, and capture the response characteristic of a critically damped system - maybe in the time domain, maybe in the frequency domain. Then run through your damper options replacing the theoretical, linear gain on velocity with something more elaborate. See which damper option best matches the dynamic response of the simple critically-damped system.

How you would choose to represent your rig data is a judgment call by you. That's how it is. Maybe you collapse down the hysteresis, maybe you take away the force offset. In real world engineering you often have to make these decisions - for better or worse - with no reference on what's "best." They don't pay engineers to solve problems which already have a clear answer

Then there's hardware in the loop if you wanted to think about diving off the deep end... but I wouldn't recommend it. I think the above would be how I'd attack it.

[Crazy Bored]

...

Your comments make sense to me and are certainly methods that are within my current capability.

JT, what do you think of the above suggestion:

What is typically done is that we run a series of sine waves at various frequencies, and the software records the peak compression and rebound forces at each frequency. These discrete points are then potted as a single force velocity line as the "characteristic" that we use to specify damping levels. For your purpose, this is the output you will want to use.

Does this "characteristic" have a name? What does it mean, and how would I use it for my purpose? Would/should I plot each of these peak forces against frequency?

[Jersey Tom]

That's one way of coming up with a force/velocity chart, sure.

If in the end you're just looking at this to get a ballpark damper setting for a car, no? Something to have as a baseline for when you get to the track and then adjust from there? If that's the case and you'll probably end up making substantial changes at the track anyway.. what level of fidelity do you really need here?

Might think about just getting a quick and dirty read on it, set your dampers to what should be at least in the ballpark, then get to the track and do some sweeps - get feedback for the real car handling. A/B/A some stuff. Especially if track time is cheap. When track tests cost tens of thousands of dollars...well that's a different story.

Moral of the story being sometimes quick predictive work, at the expense of ultimate fidelity, is more valuable than deliberating over more subtle points.

[Greg Locock]

The f-v diagram is useful, but the potato plots often reveal useful detail missed in the f-v plots. That is to say, two shocks can be valved to achieve a given f-v plot, yet will perform differently on the road, and the potato plot is a more robust comparator. The roehrig software is fine for what it does, but I won't pay pay the extra bucks for the add ons, so I have ended up exporting all the data and picking it up in matlab where I can make my own mistakes.

[Crazy Bored]

Especially if track time is cheap. When track tests cost tens of thousands of dollars...well that's a different story.

It's not expensive, but it's very limited now due to some past events.

[Jersey Tom]

The f-v diagram is useful, but the potato plots often reveal useful detail missed in the f-v plots. That is to say, two shocks can be valved to achieve a given f-v plot, yet will perform differently on the road, and the potato plot is a more robust comparator. The roehrig software is fine for what it does, but I won't pay pay the extra bucks for the add ons, so I have ended up exporting all the data and picking it up in matlab where I can make my own mistakes.

For sure. I'm just sayin let's keep it real here... guy's lookin to rough in some shock settings pretty coarsely with a critical damping target, get to the track and go from there.

I say go for the simplest, fastest, straightforward approximation.. set the things.. load up the car, and get to it.

[RideRate]

Just to post an alternative viewpoint. I am much bigger on force vs velocity plots than the force vs displacement ("football" or "potato") or peak velocity plots (pvp). I don't understand the fascination with either the displacement plots or pvp's since I think they both have major flaws when compared to a few continuous force vs velocity plots. Not to say they are worthless, they are not, but as a shock engineer I think they visually mask the most characteristics of interest.

The football plots are very sensitive to peak velocity. What I mean is plenty of dyno's do not hit the same peak velocity when given the same command and if you look only at the displacement plot (the bulk of which occurs around the peak velocity) you can be easily misled. You may think two shocks are different or yours has vastly changed, but it could very well be a function of the test that was run, but the displacement plots are not going to make that obvious. And because it shows most of the low rate part of the sine wave test, it hides some transient behavior that may be very important.

The pvp plots are just an attempt to make a simple summary of shock forces. What you will find if you run enough tests to tighten the resolution of the pvp plot is it just approaches the low acceleration, no hysteresis response of the shock. You can easily see this by running force vs velocity plots at multiple speeds. So if I run a lot of tests at different speeds to make a pvp plot, I feel I might as well look at all the curve from each test instead of pulling off one discrete point from each.

Just my take.

[Crazy Bored]

I really appreciate the input guys. As I mentioned, this post was all started just from a lab in a course I'm in. I'll be back once we are done designing the chassis and have the car built. The car will be competing in 2015, in order to get some solid testing in before competition.

[Abelma]

The f-v diagram is useful, but the potato plots often reveal useful detail missed in the f-v plots. That is to say, two shocks can be valved to achieve a given f-v plot, yet will perform differently on the road, and the potato plot is a more robust comparator. The roehrig software is fine for what it does, but I won't pay pay the extra bucks for the add ons, so I have ended up exporting all the data and picking it up in matlab where I can make my own mistakes.

Hello Greg....I am looking to a Matlab file where to import my dyno shock data saved in excel files...I can read that you have made something like that? Thank's

[Greg Locock]

Sorry, work paid for me to write it, they own it.

[PlatinumZealot]

I have a personal project having to do with how potholes and bumps impacts the life of shocks.... I would seriously like to get some expert advice on it. It may have to do with a shock dyno too.

[gambler]

Hi Platinum, I haven't been on in a while. I have played with a dyno on an amature level, and they all have huge stroke runs that seem to give poor real life data. I then set up a go-pro on the real car and watched the shock travel on a smooth track, and the movements were very minimal( 1" to 1.5") . Im really thinking the adjustables that only open a bleed hole for" lesser" dampening may not be the ultimate way to go, but rather a good way to find something the tires and driver like, then bringing it back and replicating it on the dyno with out so much bleed. Also we ran some digressives on a dirt track with huge "potholes"and it bent some of the main disks in a crease at the shaft permantly . There weren't enough stacks decending in diameter or thickness to stand that kind of abuse. I never really seen any dynos that replicate a violent pothole hit. Ive given thought to an air cannon or a very high speed rotating wheel with a bump knot to help replicate holes and curbing. There just aren't enough hours in the day for one person. later....

[Greg Locock]

It'd be fairly easy to instrument a shock and drive on the track and build up some knowledge of your real life velocities. I've always wanted to feed those time histories into a shock dyno and measure the resulting forces. As a next best thing the guys strain gauged the mounting clevis on the shock so we could measure the actual forces we saw on the road, and string pots and accelerometers to measure the velocities.

Incidentally the advantage of a potato plot is that you can easily see if a valve is cavitating or choking, which is not easy to spot on an FvsV plot.