Author Topic: Want To Build Accelerometer to Control My Dangerous Rototiller  (Read 2424 times)

0 Members and 1 Guest are viewing this topic.

Offline wstr75Topic starter

  • Beginner
  • *
  • Posts: 2
  • Helpful? 0
Want To Build Accelerometer to Control My Dangerous Rototiller
« on: October 22, 2012, 03:10:02 PM »
Hello, newby here.  Last weekend my 8 hp BCS rototiller put a serious hurt on my leg when the tines hit a rock and the tiller suddenly vaulted forward.  Unfortunately, I did not let go of the handle bar and safety switch and when it suddenly stopped and began tilling again, my body did not stop and I slammed into the rear of the tiller big time.

I found this site in searching for a 3 axis accelerometer.  The tiller has a Honda engine and its existing safety switch grounds out the electric juice to shut off the motor.  Rather than use a battery (because the tiller rarely gets operated), I would like to pull start the engine and let its magneto output charge a supercapacitor with some sort of added on voltage regulation.  The output from the supercapacitor would run the CPU and sensor.  The CPU would control a relay to ground (shut off) the engine circuit.  I would like to program the CPU to "interrupt" the engine momentarily.  I have found running the clutch intermittently in two second bursts prevents the dreaded "tiller jump" when either in hard clay or hitting embedded rocks.  Essentially, this accelerometer/relay arrangement would momentarily shut off the motor for long enough to prevent the accelerating jump but not long enough to keep the motor from completely spinning down. 

There would need to be a mechanism for determining the G force level for triggering the engine interruption.  It would be great to have a some kind of recorder that tells the maximum voltage spike seen by the X, Y and Z axis accelerometer outputs.

Questions:
1.  What CPU do you recommend?
2.  Is the supercapacitor arrangement described above logical?
3.  How do you see adjusting the G force sensitivity?  Programming?  Using a potentiometer?
4.  Recommendations for means to record/capture the G force voltage output so as to make adjustments to the settings?  This is on a rototiller.  It naturally vibrates and constantly moves in all three axis directions.  It is the severe acceleration move events to be controlled.

Any thoughts or equipment suggestions are welcomed.  I am not concerned as much about cost as I am at having an effective solution that can be implemented without too long of a trial and error period.
Thanks in advance for your suggestions.
Bill

Offline Billy

  • Robot Overlord
  • ****
  • Posts: 132
  • Helpful? 2
Re: Want To Build Accelerometer to Control My Dangerous Rototiller
« Reply #1 on: October 22, 2012, 07:15:05 PM »
Unfortunately, I did not let go of the handle bar...

You see plenty of videos of newbies on motorcycle that can't bring themselves to let go of throttle...so you're not alone.

Questions:
1.  What CPU do you recommend?
2.  Is the supercapacitor arrangement described above logical?
3.  How do you see adjusting the G force sensitivity?  Programming?  Using a potentiometer?
4.  Recommendations for means to record/capture the G force voltage output so as to make adjustments to the settings?  This is on a rototiller.  It naturally vibrates and constantly moves in all three axis directions.  It is the severe acceleration move events to be controlled.

Unless you really want to use this as an excuse to get into programming, I don't see a need for a CPU at all. You can do this with analog electronics easily enough.

Your issue isn't G forces - your issue is the length of time the G forces are present, and that is what you should focus on. The tiller bouncing around is going to make an accelerometer go crazy, but all forces not lasting long and the tiller not moving anywhere fast.
 
Then you hook onto a rock or root and you get a long hard pull that gets the tiller moving. That long hard pull is likely to be fewer Gs than the bouncing around, so I don't think looking for peak Gs will work.

Build a circuit that will ignore the bouncing around but is sensitive to the relatively long lurch.  That means a low pass filter on the output of an accelerometer, connected to an amplifier that drives a relay.

Here is how I would do it
  • Use a battery - at least until the rest of it is working. You can screw around with the magneto if you want later.
    Use a single axis accelerometer with analog output - you only care about the forward and back motion
    Feed analog signal to a buffer (amplifier) - unless the sensor has one built in
    Send that signal to a active multi-pole low pass filter with a 3db cutoff around 4 hz. Use filterpro free from TI.com to design the filter. (http://www.ti.com/tool/filterpro)
    Send that signal to a op-amp circuit, make the gain adjustable using variable resistor
    Drive a relay with the op-amp output (buffered through a N-MOSFET).

The gain setting on the opamp will set your threshold.
If you swap in a variable resistor on filter, you can also dial in your frequency, but 4Hz sounds OK to me if it is a steep filter.
The relay will have the mother of all hysteresis's so once it shuts the motor down, it should stay down for several hundred milliseconds. This will allow you to set the threshold high enough that it isn't cutting out every two seconds.

When you're done, patent and sell this feature to the tiller makers.  When you're rich, I want a few bucks in return.

Last word - turning the ignition on and off will put fuel into exhaust system, and then detonate it. Wear appropriate eye and ear protection.

Offline wstr75Topic starter

  • Beginner
  • *
  • Posts: 2
  • Helpful? 0
Re: Want To Build Accelerometer to Control My Dangerous Rototiller
« Reply #2 on: October 22, 2012, 07:37:11 PM »
Awesome and thorough answer, Billy!  Thank you! 
Bill

Offline Soeren

  • Supreme Robot
  • *****
  • Posts: 4,672
  • Helpful? 227
  • Mind Reading: 0.0
Re: Want To Build Accelerometer to Control My Dangerous Rototiller
« Reply #3 on: October 29, 2012, 11:00:35 PM »
Hi,

Questions:
2.  Is the supercapacitor arrangement described above logical?
Supercaps are made for low current drains and they're usually 2..5V max.

Apart from that, if you are new to electronics, a digital/microcontroller solution will be just as easy as the suggested analog one and can be made a lot smaller, with really steep filters if needed.

However, with only a faint knowledge of tillers (we have people for that kind of work ;)), and not absolutely sure of the error scenario, I get the idea that measuring the angle would take care of the really bad cases and things like speed of acceleration as well as max. RPM can be handled by the ignition interruption, without any extra sensors at all (just the ignition pulses, which can be detected in several ways.
It's better to interrupt every n'th ignition pulse, than letting it rip and then interrupt the ignition for a longer time (for both the motor and your sense of control).

Perhaps I've got it all wrong, but a tiller-killer should be small potatoes.
Regards,
Søren

A rather fast and fairly heavy robot with quite large wheels needs what? A lot of power?
Please remember...
Engineering is based on numbers - not adjectives

 


Get Your Ad Here

data_list