Help for a complete novice

[DAGONITWIT]

Hello and thanks in advance for any help or guidance.

I have a need for a device to control wheelstands on a drag race car. I've done some limited research and have found a component that I think may work toward that goal but I don't know how to utilize the technology the device holds. It's an Inertial Rate Gyro Sensor. See link.

**broken link removed**

What I really need is a dc voltage signal that I will use to trigger my ignition system to reduce ignition timing to the engine as long as the signal is present. The reduction in ignition timing will reduce available horsepower enough to keep the car from rising higher without decreasing throttle position. When the car returns to the position that the signal is lost, the ignition timing would return to normal.

I know nothing about what it takes to make this device sense that there is a change in pitch. I can give the sensor a "start" point that I would want to start getting the signal from it by jacking up the front of the car to the point that I don't want it to exceed but I'm not sure if that's the way to approach it.

I know that more information is probably needed but this is a start.

Any thoughts? Suggestions? Am I looking in the right direction?

Karl

 

[Boncuk]

How about using a tilt sensor?

Google for acceleration circuits.

Boncuk

 

[stevez]

You'll likely have to look at all three axes if measuring acceleration - lots of processing to do but probably can be done.

I recall that some cars used to have a pair of smaller wheels on a spring bar trailing the rear axle (wheelie bars?) to actually limit the rise. Maybe a similar setup except that the assembly is in constant contact with pavement and you measure the relative position as an indication of how much rise there is in the front end.

Another thought - use ultrasonic measurement of distance from pavement.

 

[DAGONITWIT]

Thanks

You'll likely have to look at all three axes if measuring acceleration - lots of processing to do but probably can be done.

I recall that some cars used to have a pair of smaller wheels on a spring bar trailing the rear axle (wheelie bars?) to actually limit the rise. Maybe a similar setup except that the assembly is in constant contact with pavement and you measure the relative position as an indication of how much rise there is in the front end.

Another thought - use ultrasonic measurement of distance from pavement.

Thanks for the responses.

Here's the problem as I see it. I do not want to measure acceleration. In fact, I want to disregard acceleration and only measure tilt or pitch. Using any type of acceleration measurement device won't control what I'm attempting to control. The wheel stand. In drag racing we want to achieve maximum acceleration without losing traction and preferably staying on all 4 tires. High horsepower cars with good traction tend to wheelie.

In the type of racing that I'm doing we aren't allowed to use wheelie bars on the car to mechanically prevent or lessen the amount of wheelie that the car makes. This rule tends to even the playing field for all competitors.

The device has to be able to measure tilt without being influenced by acceleration or g-forces. The device that I showed a link to seems to be able to measure tilt without being influenced by acceleration.

The other part that I need for my uses and that this device seems to provide, is an output trigger signal when the device senses a predetermined angle of the car as compared to the ground.

Please, keep the thoughts coming. I know this can be done.

 

[Willbe]

The device has to be able to measure tilt without being influenced by acceleration or g-forces.

You could use an accelerometer to correct the tilt sensor output.

Next question: do you want to anticipate, and correct for, an impending wheelie or use a bang-bang control system?

 

[DAGONITWIT]

You could use an accelerometer to correct the tilt sensor output.

Next question: do you want to anticipate, and correct for, an impending wheelie or use a bang-bang control system?

I'm very interested in a system that would anticipate a wheelie but figured it would be more than I could take on. Just the bang-bang system is stumping me.

Please forgive the fact that I have no experience in this area and started here on this site because I know when I need help and I'm not afraid to ask. It would be great if I could partner up with someone who could help me devise this piece. I know what I need it to do for me and how I will apply the signal to my existing systems. The system that I use for Engine Management is very comprehensive and software driven with lots of user options that can be adapted to control horsepower. We're using a built in component to the system that's monitoring driveshaft speed rate of increase to predict loss of traction. I just need to be able to keep the front wheels on the ground so my guy can steer the car and see the race track in front of him. It's OK if the wheels come up a foot or two for while but not the entire length of the race track and not past a foot or two.

Do you have a thought on predicting and correcting for the potential for wheelie?

Please keep in mind I've never worked with these type systems so bare with me if I ask some questions that seem obvious. I'm looking for major guidance on how to put this thing together.

Any help is greatly appreciated!

Karl

 

[Willbe]

If you can monitor the torque in the spinning rear driveshafts [maybe this parameter is already available in your vehicle], and the vehicle acceleration [avg. accel = avg. speed/time], I think I have a way that you will be able to use a controller to adjust the front wheel weight to whatever you want, including zero.

No tilt monitor for this concept.

My idea is that whatever torque is not being used to accelerate the vehicle is being used to lift the front wheels.
If you calculate the moments around the rear axle (including the driver weight and position) you can maintain the rear wheel torque at exactly the right value.
For this you need the vehicle dims, the vehicle center of gravity location and the radius of the rear wheels.

For both methods you need to measure or calculate acceleration, so it comes down to the tilt meter vs. the torque monitor.

At this point we can probably post at least three system design diagrams for this project and then weed them out. Diagrams = no ambiguity.

You go first!

 

[Pommie]

Why can't you measure the front wheel suspension length as this is a direct indication of front wheel weight.

Mike.

 

[Mickster]

Could some form of range-finder beneath the front axle, pointing at the track, be made suitable to use on a moving object?

 

[Diver300]

An accelerometer would work, as long as it has one of its sensitive axes vertical, and you put it at the front of the car.

You might also want to ignore it until the front suspension tops out, or it would notice the front suspension movement. Better would be to mount the accelerometer on the front suspension.

Accelerometers have up to 3 axes. All the axes are at right angles. In a drag car, the accelerations that you see are:-

Forwards. Lots of acceleration, hopefully more than the opponent sees.
Sideways. Not much, just small steering corrections.
Vertical. Gravity only.

Those are all independent, and each can be measured without the others affecting it. In normal running the vertical acceleration of the front suspension is near 1G all the time. If you lift the front wheels, you will see sustained vertical acceleration larger than 1G.

There may be practical problems. You have to make sure that the large forward acceleration is not measured. You have to mount the sensitive axis of the accelerometer very close to vertical, and you might have what you mount it on change angle as the suspension rises. If you set it with the suspension topped out, that should be close.

There will be noise from irregular road surfaces. I know drag strips are quite smooth, but you will have to do some damping.

 

[nickelflippr]

Would an IR distance sensor work in this situation? A max distance of a Sharp GP2Y0A02YK would go to about 60 inches. Put one on each front corner and average because of torque.

 

[Mickster]

Would the emitter/detector LED angles be required to be different at stand-still / max speed?

 

[nickelflippr]

Hard to say about the angle, maybe only talking a few tenths of a foot per a-d conversion at speed, so could be insignificant? Reflectivity could be a problem tho for the IR, as you may be going from a concrete launch/start area to blacktop a little further down the track.

 

[microtexan]

Maybe I'm missing something but, wouldn't a simple microswitch arrangement which senses when the front wheels have reached the top of the their suspension point give you the signal you need?

 

[Pommie]

A more complex approach would be a 2 axis accelerometer arranged as a tilt sensor, one axis pointing along the car and the other vertical. By doing a bit of maths the two readings can be combined to give an acceleration and an angle of acceleration. This would be by far the most accurate way to do this but would require some rather good programming and maths skills.

An alternative approach would be to just have the vertical sensor. At rest it would just return gravity, when accelerating horizontally it would return gravity but when wheelieing (sp?) it would return a value (a lot) less than gravity.

Mike.

 

[nickelflippr]

The big question is what type of inputs are allowed, and how do you affect your timing with that input. Is there a data sheet or manual for your management system? Getting the 12V signal is not a problem using a 2n7000 npn transistor to level shift the 5V output of what might be a small programmable microcontroller (some sort of comparator circuit would work here also).

 

[nickelflippr]

Simple enough for a microcontroller with an analog to digital function along with the digital outputs for your 3 or 4 inputs to the engine mgmt. You would control the digital outputs using conditional statements, based on the a-d value, to fit your wheelie height criteria. The window of criteria is absolute, so would work both on the way up, or coming back down. Using a Basic like syntax your code would look sorta like:

...
...
#define height2 410    ;10 bit digital value of a 2V output (5V reference)
#define height3 350    ;would need to correlate height to a digital value
dim wheelie as word
...
Set PortC Off
Main:
wheelie =  ReadAD10(AN0)
...
...
If wheelie > height2 and wheelie <= height3 Then
    Set PortC b'00000011'    ;Set two outputs to 12V+
End if
...
...
goto Main

 

[Willbe]

The only input to the Engine Management System that I can utilize is 12-16vdc. I have recently learned that I can use either 3 or 4 separate 12-16vdc inputs to the EMS. They all act identically but independently. Each of those inputs can adjust a preset amount of ignition timing. So if I can get the system to generate 3 separate 12-16vdc signals, say 1 at 4" of rise, another at 8" of rise and finally another at 12" of rise, I could potentially have 3 stages of progressive ignition timing retard to kill horsepower in 3 stages if the car doesn't respond to the initial retard. It seems like these additional stages would help to anticipate and prevent the unwanted very high wheelie as well as smooth out the transition.

Of course it's important to be able to put the ignition timing back in in reverse order as the car comes back down in order to maintain acceleration and to keep the car from slamming down. All of this without the driver ever having to change throttle position.

I don't have a way to provide any info. from the system manual. Honestly, this part of the system is pretty basic. Give the system a 12-16vdc signal on a specific wire in the harness and it will retard a predetermined amount of ignition timing.

Thanks, Karl

So you do/can have a step-wise proportional control system.

With a system this complex, adding this distance-sensing component could cause unforeseen and dangerous system instabilities, so testing of whatever we all come up with should proceed very cautiously.

I imagine that an experienced driver corrects for a lot of system problems, maybe without even knowing it. It becomes second nature to him/her.
But now we are adding an unknown to a known and familar system.

The worst outcome is that engine power goes into flipping the car or having it bounce up and down rather than propelling it forward, just like an engine without a harmonic balancer will destroy itself, and a helicopter flying in a forbidden horizontal-speed/vertical-height region may or will destroy itself.
In both case the considerable engine power is diverted into destruction rather than propulsion. And once you're in these regions no one can recover.
This situation could probably be called a widowmaker.

So, you may want to add a sensor that detects this undesirable condition, and responds much faster than any human can, in shutting down the engine. You may even want this sensor to apply brakes.

Toto, we are not in Kansas anymore!

 

[Willbe]

Assuming the spectators are a safe distance from the track. . .

what's a convenient height for the distance sensor, that works with one or two candidate sensors in the hyperlink?
The sensor example graph had the same voltage output for two different distances, so we need to have the sensor working in the graph area to the right of the peak.

Decide
the longest vertical distance from the road to the sensor mount that is "normal" and
the shortest vertical distance from the road to the sensor mount that means a wheelie is in progress.
If there is overlap here, that may be a problem.

Then the task is to find or design a field-adjustable comparator circuit to output the limiting signal to the engine control, that mates with the sensor output and the engine control input.
This circuit may need a metal box to shield it from radiated ignition noise.
Including the box and shipping costs, the total mat'ls cost may come to $15.

Then testing.

Sometimes, what is built works exactly as designed, but once it's built and operating, it becomes clear that something else is needed.
I guess we cross that bridge when we come to it.

 

[rwahrens]

I have used ultrasonic sensors in dirty/noisy environments with good success. Try an industrial automation supplier, like AutomationDirect.com. Here is a link to a sensor on their site that might work:

https://www.electro-tech-online.com/custompdfs/2008/09/prox30mmultratu.pdf

They have others as well.

This would be simple to install and calibrate, no need to write any code. Some of these sensors are pretty smart. More expensive that $15.00, but a lot less developement time.