TL;DR

Torque to turn a large Rotisserie

Hi - I'm trying to design a Rotisserie for fabrication of large UC beams, and could do with some physics help if possible please.
The idea is... I'll create collars to allow a UC beam to turn on a set of rollers during fabrication process.

The largest item I have weighs in at 4225kg including the collars I'm fitting around to make it circular for turning.
The Collar has a radius of 450mm.
The C of G is pretty much axially aligned, but is 3.5mm out due to some features on the beam.
The collar with the beam in sits on 2 off bearing shafts with R45 wheels upon it

What torque would I need to turn the beam in its rotisserie from rest?

I'm looking for force required to turn it in small increments from rest, as I reorient it to welding requirements. Its not spinning at a constant speed etc.

Could anyone assist?
Thanks in advance if you could.

Baluncore said:

Welcome to PF.

Torque is needed for two separate reasons. One is to overcome the 3.25 mm eccentricity of the mounted beam, the other is the acceleration or deceleration needed to start or stop the beam rotating.

When the column is rotated, what angle will it rotate through, and how much time is available to achieve that?

Hi,
It will probably be turned in 90deg increments each time, then welded for an hour or so and turned again.
Time to turn through 90deg is not critical.
If it took 30secs or even a minute.

I was really trying to figure out if I can power this manually, as I've already got a 10:1 reduction via the collar to bearing wheel ratio, or if I'll need further ratio reduction and possibly an electric drive and gearbox.
I'm hoping to keep it manual for cost reasons.

Nick09 said:

I'll create collars to allow a UC beam to turn on a set of rollers during fabrication process.

Nick09 said:

The largest item I have weighs in at 4225kg including the collars

4225kg or 4225N?

Nick09 said:

The C of G is pretty much axially aligned, but is 3.5mm out due to some features on the beam.
The collar with the beam in sits on 2 off bearing shafts with R45 wheels upon it

What kind of bearings? What friction numbers are we talking about from the bearings? What will you use to hold/clamp the rotational motion in the worst case where the CoG is offset horizonally from the main axis?

berkeman said:

4225kg or 4225N?

What kind of bearings? What friction numbers are we talking about from the bearings? What will you use to hold/clamp the rotational motion in the worst case where the CoG is offset horizonally from the main axis?

Hi Berkeman,
its 4225kg for the mass of the longest variant of these we make, the 10m.
Re bearings - its still a WIP and could change but at the moment, I've put in these agricultural bearings
https://simplybearings.co.uk/shop/p...wcB&gbraid=0AAAAAD-_g0HWf303iiEQY35z9YsWJgVNi
There are 2 off per shaft, so 4 off bearings per Trestle. 8 off total in the 2 off Trestles.
They don't give a friction rating in the tech info - but I have seen other calcs with this value inserted...
Roll co eff, steel on steel (ball bearings) - 0.004m

re turning - holding in position,
I haven't yet fully developed it, but but my plan was to add 20tooth chain wheel to each driving roller and link or connect via a 3/4" drive chain (green in image below) - then have a second drive chain (purple in image below) with either a second 20 tooth chain wheel, or I can optionally drop to 10 tooth. This would be driven by either a hand cranked handle, or motor and gearbox, depending on how much torque this needs to turn it.
if I stick with 20 tooth wheel - the 10:1 reduction ratio would remain.
if dropping to a 10 tooth - this would further reduce reduction ration to 20:1.

The rotation can currently be locked off in 15deg increments via an insertable pin. You can just see the handle of it (coloured silver in the imaghe below)

section through a drive shaft also included below - bearings are blue, main collar = purple

Baluncore said:

I see no reason why it cannot be manual, just like a steam railway engine out of a roundhouse, on a turntable.
https://en.wikipedia.org/wiki/Railway_roundhouse

You will need another factor of ten gear reduction, with a right angle gearbox. You may need to link the two supporting rollers at one end with a chain, to prevent slip, but that will depend on friction.

If an overhead crane was used to load the column, then it might also be available to roll the column, by wrapping the hoist chain around the column.

Hi Baluncore,
Thanks for the advice - yeah i'm still working on this, currently adding chain drive to link the 2 rollers.
I'm thinking 20 tooth wheel on each drive shaft, with 3/4" chain linking (see green line) - then a second 20 tooth, taking a second chain upto either a hand crank lever and handle, or motor and gearbox (see purple line) - 2nd cog on the purple chain could remain 20 tooth to keep the existing 10:1 reduction ratio, or even change to a 10 tooth cog to further drop the ratio to 20:1.

I'm confident I can get it working - but where I'm at, you get one shot or they loose interest, hence adding a few options and reducdancies.

Re cranes - we currently fabricate these UC beams using saw horse trestle and o head cranes, but we have to move to a facility with no o head cranes. hence planning these rotisserie.
Plan is to load them via a fork lift, then fab and weld in the rotisserie.
Current method of rolling 4t Beams using offset crane chain, on saw horse trestles is not ideal, as it really shocks the crane, when these roll and fall, so a rotisserie should improve safety too.

Nick09 said:

where I'm at, you get one shot or they loose interest, hence adding a few options and reducdancies.

BTDT, so I'm speaking from experience. You need a sufficient safety factor in the torque calculation. I suggest calculating torque based on 35 mm eccentricity between CG and center of rotation. With a safety factor that large, and your selection of bearings, you can ignore bearing friction because it will be small compared to the eccentricity torque.

Then the torque = 4225 kg X 9.8 X 0.035 m = 1450 N-m. Your concept of bearings, wheels, and chains looks good. I suggest driving the system with a worm gear reducer with a hand crank handle on the input shaft. Worm gears are self locking, so you turn to the desired position, and it will stay there. Specify the reducer for the 1450 N-m output torque. Check the chain pull on the reducer output shaft to make sure you don't overload that shaft. Or you can directly couple the gear reducer output shaft to one of the drive shafts.

If the users complain about the time or effort needed to rotate the beam, remove the crank handle and connect a variable speed reversible electric drill.

jrmichler said:

BTDT, so I'm speaking from experience. You need a sufficient safety factor in the torque calculation. I suggest calculating torque based on 35 mm eccentricity between CG and center of rotation. With a safety factor that large, and your selection of bearings, you can ignore bearing friction because it will be small compared to the eccentricity torque.

Then the torque = 4225 kg X 9.8 X 0.035 m = 1450 N-m. Your concept of bearings, wheels, and chains looks good. I suggest driving the system with a worm gear reducer with a hand crank handle on the input shaft. Worm gears are self locking, so you turn to the desired position, and it will stay there. Specify the reducer for the 1450 N-m output torque. Check the chain pull on the reducer output shaft to make sure you don't overload that shaft. Or you can directly couple the gear reducer output shaft to one of the drive shafts.

If the users complain about the time or effort needed to rotate the beam, remove the crank handle and connect a variable speed reversible electric drill.

Thanks jrmichler. That's fantastic advice.

Baluncore said:

@Nick09
The bearings you have referenced are not specified to be self-aligning. Self-aligning ball bearings have a spherical outer race, but are more difficult to seal, and more expensive than deep groove ball bearings. The blocks shown have no external grease nipple, just a screw to clamp the insert and prevent rotation.

Self-aligning bearings can have a spherical external surface on the outer race, that seats in the spherical inside surface of the pillow, or plummer block. Identify them by the external grease nipple.
https://www.ebay.com.au/itm/133934882190
https://www.ebay.com.au/itm/235479806155

Deep groove ball bearings would be sufficient, but cylindrical roller bearings could stand more abuse during the forklift delivery of the UB. To self-align they would need to be mounted in a lubricated spherical block.

Thanks Baluncore. Great advice.

Just wanted to say thanks again to everyone who helped - its very much appreciated. Got the 1st off built up last friday and used it for a week now with ever progressive lengths of Universal Column 'UC', and it works well (not perfect yet). I ended up fitting a 20:1 chain drive, and its cranked by the handle on the side. One operator can quite easily turn even a 10m long UC with not a great deal of effort by hand.
The lessons learned form first build....
When spacing the trestles - perfect alignment is critical to how much effort required to turn. You could fit them on rails, but I have a very tight budget on this. if 10mm out of align it doubles the effort to crank, if 20mm out of align it won't move, and they're a pain to fine adjust - short lengths a 5lb sledge works, but 10m and you need a fork lift.
So.... my next plan - if I add a swivel under where the collar sits into the cradle, no matter where the trestle legs are placed, once the UC is dropped into the collar, the swivel will turn the collar under its own load weight to perfectly axially align with its neighbor trestle, so no pre loading alignment needed.
One more prototype to build, then onto full production of these.

Thanks again all.