Accuracy of Flexible Shafts in Motion Transmittion 2

[farzadinjast]

Hi fellows,

I am working on design of a servo mechanism that needs shifting the rotation axis from one axis to another parallel one. I wonder If flexible shaft can be used to transmit the motion. Are flexible couplings precise in transmitting the motion or there is significant backlash or dead-zone between input and output angles? I need accuracy of less than 1 degree.

The motion is bidirectional. The mechanism should work in high temperature (300C). The speed and torque is relatively low and mechanism is used to positioning a vane.

P.S: What is the main difference between the flexible shaft and the universal shaft?

 

[drawoh]

farzadinjast,

As far as I know, flexible shaft are multi-strand wire cables. These are flexible from side to side, and I would expect significant amounts of torque wind-up.

Look into electroplated bellows. I would expect these to have way more torsional stiffness, although they may do interesting things when they buckle. They make couplings out of them.

A universal shaft has a couple of U-joints, which usually have have backlash.

--
JHG

 

[bdeuell]

how far are you trying to shift this axis and is ?

I have seen a couple types of flex shafts:
- sheathed braided steel cables, tend to have some backlash due to clearances between the sheath and cable and torsional rigidity is not great
- helical spring, lower rigidity the more flexible the shaft as the spring can wind up under load
- shape memory alloy, these use the super-elastic properties of SMAs, used in medical applications for ease of sterilization (not cheap)

there are also lots of flexible shaft couplings but these are usually only intended for small misalignment in the shafts, a couple degrees and a few thousandths concentricity, generally speaking the more misalignment the shorter the coupling life. there are also oldham couplings which can handle higher concentricities

Universal joints can handle fairly high angles 25-30 degrees but they do not not have a linear output angle i.e. as the input is rotates the output rotation will vary sinusoidally. This effect will get worse with greater angles of shaft misalignment, however i believe you can use two U-joints (as you would need for a parallel shaft) to cancel each other out if positioned correctly.

There are also Constant velocity (CV) joints that eliminate this nonlinear velocity issue with U-joints

but why not just use a timing belt if the shafts are parallel?

 

[3DDave]

How much shift?

 

[tbuelna]

For reliable operation at 300degC and no backlash your best option would be some type of metal flex coupling, such as a diaphragm or disc coupling. The only issue with these types of couplings is angular misalignment capability.

 

[farzadinjast]

Many Thanks for the replys.

The shift in the axis is about 10 Cm in 30 Cm length (The longitudinal distance between the ends of the flexible shaft is 30 Cm).

The two ends of the shafts can not meet on a common plane so I can not use any timing belt or gear.

In general does any significent nonlinearity exists in the flexible shafts?

 

[BrianE22]

I believe that they are torsionally stiffer in one direction. This would affect how much gain you could use if this is used in a closed loop servo system.

 

[farzadinjast]

Unfortunately, I have not found this difference in torsional stiffness in any catalog.

 

[MikeHalloran]

A Schmidt coupling might work in the envelope you have specified:

http://www.zero-max.com/flexible-shaft-couplings-schmidt-couplings-c-1_8-l-en.html

Mike Halloran
Pembroke Pines, FL, USA

 

[bdeuell]

"In general does any significent nonlinearity exists in the flexible shafts?"

backlash and stiffness will contribute nonlinear components to the shafts rotational output velocity. other than that i don't think there would be any significant nonlinearities provided the scale of the flex shafts structure is relatively small compared to the overall size of the shaft. for example if you are using a flex shaft with a coiled spring design and the pitch of the coils is relatively large the center line of the spring when bent would not follow exactly the same path for every rotational angle, i.e. the location where the end of the spring wire terminates could begin to play a role in the dynamics. Basicly if the flex shafts structure is small in scale compared to the shaft size it can be viewed as a homogeneous structure and should not exhibit the significant nonlinear velocity output that U-joints do.

I would not rule out using two u-joints for your design, as long as they are phased properly you can cancel out the non linear velocity. at low speeds you wouldn't need to worry about vibration from the intermediate shaft inertia. backlash in a u-joint comes primarily from clearance in the bearings, if roller element bearings are used there should be little or no backlash.

if your loads are light enough (in particular the amplitude of the variation in the loads) a helical spring type flex shaft may be the simplest design (fewest moving parts ... other than a SMA shaft).

 

[BrianE22]

The following addresses the torsional stiffness in each direction for flexible shafts:

http://machinedesign.com/mechanical-drives/flexible-shafts-eliminate-obstacles

 

[farzadinjast]

Offset coupling seems wonderfull for my application. Have you ever used it? Is it reliable enough?

 

[mechedoc]

My experience with the schmidt style coupling was in a lab setting with high torque, high rpm, short duration, impact loading. In that application it survived through our entire project - including the "calibration" phase where we crashed the platen in the table - but it really was only used for 1000 cycles at most.

My concern would be with the effect of the high temperature on the bearings. While our setup was in an oven - the coupling was outside of the insulated zone. Though I'm sure you could contact one of the speciality suppliers (such as

http://www.schmidt-kupplung.com/en/engineering/schmidt-kupplung.html),

or work with a local machine shop to utilize bearing of yoru own specification.

 

[farzadinjast]

Thank you very much