Bronze bushing for a low rpm high load application?

[sploo]

Not sure which would be the best forum, but I suspect a metalworking audience may have the most experience in this area...

I'm building a friction drive - basically two touching metal disks for transfer of power (think cogs with no teeth). Friction drives rely on the two disks being pushed together with sufficient force to prevent slipping at the contact point - so the axles are under a lot of radial load.

I've got a prototype with a ~15mm driven disk (well, a rod) driving a 150mm disk. Speeds will be exceptionally low - the output rotating once every 24 hours; with the input obviously taking 2.4 hours.

The intended use is for a telescope tracking mount, so given UK weather it'll likely get run for only a few hours per month.

Currently I'm running the prototype on decent quality ball bearings, but I understand that bronze bushings are generally preferred for low rpm / high load applications. Anyone have experience in this area and might be able to comment?

 

[CHJ]

You need to consider lubrication, you need to maintain a lubricant film between your shaft and the bearing.

You may need the bearing to shaft contact to be tapered (slight cone) and some means of end thrust for adjusting the clearance to allow for initial setting and future wear, although at your intimated limited running hours wear may be insignificant.

If you can fit a collar to your shaft to increase its diameter in the bronze bearing this will help on loads and wear due to increased bearing support area.

 

[sploo]

You need to consider lubrication, you need to maintain a lubricant film between your shaft and the bearing.

That is one of my concerns - as the bushings would (in order to give support) be pretty close to the friction disks; and obviously you don't want any lubricants near the disks.

If you can fit a collar to your shaft to increase its diameter in the bronze bearing this will help on loads and wear due to increased bearing support area.

The output shaft is currently 20mm diameter and the input 15mm. I could probably sleeve them to increase the diameter though. What size would ideally be required? 40mm and 30mm respectively might be possible, but not much larger than that.

 

[Droogs]

consider using phosphor bronze alloys as they can in your proposed usage be regarded as self-lubricating. Although I would be on the look out for old crown bowling balls in order to use the Lignum vitae wood they are made from as my bearing instead. this stuff is still used as the bearing for propshafts on some royal navy ships and cruise liners today

 

[sunnybob]

You can get oil impregnated bronze bushings. maybe they would help?

 

[Rorschach]

Oilite bushings/bearings would be a great choice for this.

 

[hawkeyefxr]

You could fit a small oiler pot that would drip feed. Myford have them on their lathes. Here's one on fleabay.
https://www.ebay.co.uk/i/313089598781?c ... gJvEfD_BwE

Just had another thought, you could look for needle rollers, they come in a hardend steel cage and would last forever with a light oil.

 

[MusicMan]

Bronze bushings would work, but best use the sintered bronze ones with the holes filled with lubricant or preferably PTFE (Teflon). A plain bronze bearing needs a certain speed to maintain a hydrodynamic film, which it would not get in your slow-moving system.

Friction drives are excellent for smooth, backlash-free motion. The only problem is some slippage might occur, but not if the materials and forces are well chosen. In any case, they are commonly used in servo drives, in which backlash is awful but some slippage is acceptable as they will compensate by the encoder systems.

You don't need a super heavy bearing. Roller bearings carry a huge load, needle rollers would also be fine as already suggested.

And note that you do not even need to transmit the load to the bearing. You can use a sprung pressure wheel the other side of the drive wheel, or (better) on the other side of an L-shaped rim. This takes all the load of the friction drive and the main bearing just has to keep it on axis. You use a sprung mounting, to let the wheels mesh against each other tightly without force on the main bearing.

Do google 'friction drives' to save you re-inventing them.

Mind you, all this is overkill for a telescope drive. A synchronous motor and one or two worm-and-wheel pairs will do just fine and be very smooth. The main use for friction drives is when you want extreme precision and no backlash (for servo motors). I was project monitor once for an ultra precision lathe which had precision of a few nanometers, and this made great use of friction drives. In your application the output is always moving in one direction, set by the earth's rotation, so you do not care about backlash.

Keith

 

[sploo]

Oilite bushings/bearings would be a great choice for this.

On the assumption that the Oilite bushings don't "drip" any oil then they do look like a good solution. My main worry is damaging a ball bearing (even though the pressure I need is within the specs of the bearings I have).

 

[sploo]

Bronze bushings would work, but best use the sintered bronze ones with the holes filled with lubricant or preferably PTFE (Teflon). A plain bronze bearing needs a certain speed to maintain a hydrodynamic film, which it would not get in your slow-moving system.

Friction drives are excellent for smooth, backlash-free motion. The only problem is some slippage might occur, but not if the materials and forces are well chosen. In any case, they are commonly used in servo drives, in which backlash is awful but some slippage is acceptable as they will compensate by the encoder systems.

You don't need a super heavy bearing. Roller bearings carry a huge load, needle rollers would also be fine as already suggested.

And note that you do not even need to transmit the load to the bearing. You can use a sprung pressure wheel the other side of the drive wheel, or (better) on the other side of an L-shaped rim. This takes all the load of the friction drive and the main bearing just has to keep it on axis. You use a sprung mounting, to let the wheels mesh against each other tightly without force on the main bearing.

Do google 'friction drives' to save you re-inventing them.

Mind you, all this is overkill for a telescope drive. A synchronous motor and one or two worm-and-wheel pairs will do just fine and be very smooth. The main use for friction drives is when you want extreme precision and no backlash (for servo motors). I was project monitor once for an ultra precision lathe which had precision of a few nanometers, and this made great use of friction drives. In your application the output is always moving in one direction, set by the earth's rotation, so you do not care about backlash.

Keith

I have a (used) 100:1 harmonic drive that's giving me the initial reduction stage, but I need another 10:1; for which I'm using the friction drive. Machining wheels/rods on the lathe is obviously easier than making a worm drive (plus there is periodic error with a worm gear stage).

Currently the design has an extra bearing that gets pushed against the drive rod. The drive rod is on a moving mount, so it applies pressure to the larger driven wheel. However, that does mean the load will ultimately be taken by the bearings on the axle for the larger driven wheel. I.e. if I understand correctly - I do have a spring pressure wheel on the drive wheel (though I can't immediately picture what you mean by an L-shaped rim). The drive and driven wheel axes are parallel; I have seen some friction drive designs with the drive at 90 degrees to the driven wheel.

 

[sploo]

For clarity:

• The light blue circle is the driven wheel (a 10mm thick, 150mm diameter steel disk)
• The dark blue circle is the driven axle (20mm diameter steel rod - connected to the driven wheel). This axle is fixed.
• The red circle is the drive rod (15mm diameter steel rod). This axle can move a few mm up and down
• The light red circle is a pressure roller that's forced into the red drive rod

It works; in that I've had about 15Nm of torque on the output without slipping.

 

[CHJ]

I have seen some friction drive designs with the drive at 90 degrees to the driven wheel.

A driving wheel on the face of a driven disc is usually to achieve a gear ratio calibration between the two by moving the driving wheel on a radial line across the disc. (variable gearbox)

Or as in the case of mechanical airborne navigation computers (GPU's) of yesteryear, to calculate (compute) the Ground position from multiple inputs (Doppler/Airspeed/magnetic etc.) before the days of Decca Navigator beacons, INS, and then GPS.

 

[sploo]

I have seen some friction drive designs with the drive at 90 degrees to the driven wheel.

A driving wheel on the face of a driven disc is usually to achieve a gear ratio calibration between the two by moving the driving wheel on a radial line across the disc. (variable gearbox)

Or as in the case of mechanical airborne navigation computers (GPU's) of yesteryear, to calculate (compute) the Ground position from multiple inputs (Doppler/Airspeed/magnetic etc.) before the days of Decca Navigator beacons, INS, and then GPS.

That was partly my reason for not going down that route, as whatever thickness you have in the driving wheel would result in a slightly different circumference (and therefore speed) across the driven wheel. That said, I've seen a few successful DIY telescope mounts using that configuration; though the only commercial ones I know of have the axles parallel.

 

[Cheshirechappie]

If you want to keep things simple, use bearings with a large bearing area (larger shaft diameter than normal and long bearing length), a leaded bronze bush paired with a low carbon or medium carbon steel shaft, and a high viscosity lubricant such as grease or even a dry lubricant such as graphite or a PTFE based compound.

For a lower friction alternative, rolling element bearings - but check the allowable static loads. Overspecify if in doubt.

 

[sploo]

If you want to keep things simple, use bearings with a large bearing area (larger shaft diameter than normal and long bearing length), a leaded bronze bush paired with a low carbon or medium carbon steel shaft, and a high viscosity lubricant such as grease or even a dry lubricant such as graphite or a PTFE based compound.

For a lower friction alternative, rolling element bearings - but check the allowable static loads. Overspecify if in doubt.

Thanks.

I revisited the calculations for the force required to achieve the necessary torque, and it turns out it's much lower than I originally estimated. The large gearing down is for the required low speed, but obviously results in a large potential output torque. Of course that level of torque doesn't really matter - it's the amount of output torque I need to move the telescope without slipping, and the force (pressing the friction drive wheels together) to achieve that is much smaller. As such, I'm an order of magnitude within the static load rating of the ball bearings, so I'll stick with those for the moment.