Moment of Shaft Direction Change

[ext123]

I have a machine that has a motor on one end. There is a magnetic coupler between it and an inertia wheel. The inertia wheel is connected to a brake using a ling flexable shaft. The machine spins the inertia wheel to a fixed speed, then the brake is applied. The motor decouples and only the inertia from wheel flexes the shaft. I need to know the exact point that the shaft will change directions and start to unwind. We currently use an encoder for this, but they do not last long because we are constantly slamming the shaft to a stop. This happens in both directions. Does any one have any ideas of another way to accomplish this?

 

[VEBill]

Maybe a vision system with a sufficiently high frame rate to give you the temporal resolution required. Beware of subsampling artifacts.

 

[itsmoked]

Not sure I follow your description.

I recognize that flexible shafts wind up/down but if you are braking the inertia wheel to a stop the shaft would start unwinding once the wheel had stopped? Or just before the wheel has stopped and some balance has been reached where the shaft spring constant equals the inertial wheel torque?

So you want to know when the still rotating system reaches some specific equalization? If this is the case I'm even more confused because why then would you be slamming the system to a stop, you'd rather need a slower smooth reduction so you could find this moving equilibrium point.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[ext123]

I over simpligied first discription. We are actually testing an aircraft part called an ABrake. This is an actuator that senses when a flap has been damaged on an aircraft. So in my discription the brake is actually this actuator. If the flap has been damaged it will force the actuator to slam into its end stop. The inertia wheel will spin almost a full revolution before it stops and reverses direction. We have to apply a brake to the shaft within 20 miliseconds of the direction change. We do not want the shaft to unflex.
Thanks for the interest.

 

[VEBill]

There should be some sort of a shaft position encoder where the moving part of the sensor is rugged enough to survice the abuse. Perhaps an optical encoder except where the disk is NOT made of glass (as they often are).

Or... Automotive Antilock Braking Systems (ABS) need to monitor the realtime rotation (or not) of the car's wheels and they're subject to extreme abuse and survive quite nicely. The ones I've seen are typically a tooth and magnetic sensor system (but there may be other technologies). Your application is very similar to ABS in some ways.

 

[btrueblood]

Ve1 has the right idea in his last post, with a sealed encoder or a hall-effect sensor and toothed wheel target.

 

[itsmoked]

I see. Thanks for the additional info.

I'm in with the same solution. Dirt, grit, vibration, ambient temps, all handled with modern automobile wheel sensor tech.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[sreid]

Take a look at the Renishaw RM22 Magnetic Encoder. Really rugged.

http://www.rls.si/default.asp?prod=rm22

 

[fangas]

Two carefully placed ind. prox. switches looking at the keyway.

Placement must be done such that the switches output a quadrature signal (90degs apart).

There's all varieties of COTS stuff that'll take that input and give you a control signal on a direction change.

ed

 

[fangas]

There's also zero speed switches (Retrigerable 1-shots) you could use.

just set the timing down real low.

Ed

 

[VEBill]

Not knowing the dynamic rotation specs this is a guess:

Only one set of I/Q signals per rotation (off the keyway) isn't likely going to detect rotation reversal within 20ms. I suspect the system will need more pulses per revolution to give the temporal resolution.

 

[waross]

Can you anticipate a little? Instead of waiting for a full stop, watch the decelleration of the shaft with automotive sensors and trigger the brake when the shaft is almost stopped.
I was thinking zero speed switches but I doubt that the ones that I am familiar with would be fast enough.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[fangas]

Opps, Somehow (wasn't paying enough attention) the 2nd post regarding that 20mSec response

Ed