CVT - Continuously variable drive for RF30 type Drill/Mill

[BillTodd]

Since I acquired a Hardinge HLV with a continuously variable belt drive, I started thinking about making one for my RF30 type Drill/Mill.

A friend gave me the parts for Piaggio 125 scooter CVT. They looked like they might fit so I thought I'd give it a go...


I looked at various ways to fit the two pulleys in place of the originals before settling on this design. See attached pictures.

I had to make a new shaft to hold the Piaggio variator (front pulley) to the motor. I used the spring from the Piaggio's clutch pulley to close the variator (the scooter uses centrifugal weights to control the speed)

The Piaggio's spring loaded clutch pulley now becomes my variator and is mounted on top of the RF30's intermediate pulley.

ATM for testing, I am varying the drive with a simple threaded wheel screwing down from the top. I will have to make another shaft for the intermediate pulley than will support the bearings in the CVT pulley (ATM the CV pulley is simply screwed to the IP)

The new IP shaft will be drilled through, so a pull rod operated from a hand-wheel under the casing will alter the speed.

I was pleasantly surprised that it all fitted neatly in to the existing belt case

[BillTodd]

A few more pictures:

Although it is running ok the back of fenner final drive belt just catches the post causing a horrible buzzing. I'll need to source a slightly longer CVT belt, so that the intermediate pulley arm can swivel out slightly to allow the belt to clear the post

[BillTodd]

YouTube - RF30 CVT1

YouTube - RF30 CVT1

[BillTodd]

I had to bore a hole down the length of the 200mm (8") bar to make the intermediate shaft. This was accomplished by drilling as far as possible (~70mm) into each end with 5 and 6.5mm drills. I then drilled a hole in to the end of a 6mm steel rod with a 5mm drill, then ground a notch into the end of the shaft of the drill and inserted it into the hole, centre punching into the notch was all that was required to hold it in place while drilling the remaining ~60mm in the middle of the shaft.

All I had to do then was turn the shaft and fit the four bearings.

[BillTodd]

The longer scooter belt arrived today, so I set about finishing the project...

I knocked up a handle from a cassette machine flywheel with a 6mm SS bar for a handle. I threaded one end of the 4mm pull rod and screwed that into a short section of SS threaded rod ( The threaded rod screws into the intermediate shaft nut, pulling the top sheeve down [reducing speed] as it is unscrewed)

The new longer 925mm belt allows the intermediate pulley to swing around making room for the front Fenner belt to clear the post.

The result is a nice easy to use vary speed machine; with a quick spin of the handle I can change the speed over a 3.6:1 range without loss of torque

Changing the front belt I can get speeds from 175 to 1860 in four over lapping ranges.

Pulley Vmin - Vmax
1 == 548 - 1860
2 === 393 - 1409
3 ==== 290 - 977
4 ===== 175 - 633


[BTW the second picture has, by chance, frozen the belt size 925mm x 22mm by 30 degrees ]

[posix]

bill I was thinking about the CVT mechanism recently but couldn't quite figure out how to automate the process somehow. is there a way to integrate the weights into the system so it "auto-shifts gears" so to speak. but then I'm still struggling with how it would work. one way of using it that I could think of is to reverse the whole thing so it behaves like so: the faster the motor spins the spindle keeps constant rpms but then you get a whole load more torque at the spindle. just an idea.

[BillTodd]

bill I was thinking about the CVT mechanism recently but couldn't quite figure out how to automate the process somehow. is there a way to integrate the weights into the system so it "auto-shifts gears" so to speak.

If you wanted to automate it, the simplest way would be to drive the screw with a stepper motor or similar. To integrate it with a machine controller (mach or emc) it would need something (i.e. a simple microprocessor) to convert the controller speed signal to an absolute position of the sheeves.

I'm not sure auto-shifting would be such a good idea; Most cutting processes require a (ideally) fixed linear speed (i.e. feet/min) not fixed cutting force (or torque). That said, I tried adding extra weights to the scooter variator (the bit on the motor above) instead of a spring, but the belt tensioning force generated was nowhere near high enough (the scooter revs to 10-15,000 rpm, my motor does 1480!)

[posix]

I believe that you will find that ALL cutting processes require a fixed chip load, which can be gotten by either "plunging" your (powerful) cutting bit deep into material and moving slowly or "skimming" the material with (not such a powerful) cutting bit and moving fast. High chip load requires lots of power. Power is torque at rpms and to vary power you either vary torque (change motors) or vary rpms (change pulley ratios). I don't know, somehow CVT makes perfect sense to me: for light loads slow the motor down for high loads (hard materials) crank it up, in both cases your cutting bit is running at the same rpms (but with different torque) and in both cases running at the same feed rate (ipm). I need to think.

[BillTodd]

for light loads slow the motor down for high loads (hard materials) crank it up, in both cases your cutting bit is running at the same rpms (but with different torque) and in both cases running at the same feed rate (ipm). I need to think.

Hmm, I'm not sure that isn't upside down ...


Typical AC electric motors run at fixed speeds (with some slippage under an increasing load, until they stall). The available torque will depend on the maximum power of the motor (or the power supply). The applied torque, and hence power used, will depend on the load (i.e. the cutting force).

By using gears or belts, either the available torque can be increased at the expense of speed, or the speed increased with a reduction of available torque. The available power remains unchanged (albeit minus some losses in the transmission).

By using a CVT, it would be possible to sense the load and adjust the gear ratio to ensure the maximum output speed was obtained for a given load (i.e. the speed would reduce as the load increases) . It would probably only require an extra spring in the variator to make my CVT load sensitive.

If you were to arrange a mechanism to increase speed as the load increased then this would require a squaring of the input power and, since the load is itself speed dependant (i.e. load increases with speed) there is a great danger of 'run away' i.e. load increases so speed increases, therefore load increases therefore speed increases - power used goes up until something goes bang.

[BillTodd]

Updated the video:

RF30 CVT and other modifications - YouTube

RF30 CVT and other modifications - YouTube

[wizard]

Reminds me of a snowmobile.

A technical point, AC induction motors don't develop torque unless they slip. The magnetic field generated by the stator needs to "cut" the windings in the rotor. This induces a current in the rotor and thus a magnetic field.

In any event it should be interesting to see how this holds up over time. Even on conventional shop drill presses variable speed is a huge convenience. On a mill it should be even more useful. I'm somewhat surprised though that you didn't try to increase your top end speed.

[BillTodd]

I'm somewhat surprised though that you didn't try to increase your top end speed.

The only easy way I could see to do that would be to increase the size of the motor sheaves.

If I come across a cheap Piaggio clutch/variator (the same as the one fitted to the intermediate pulley) than I might be tempted to make another motor shaft for it. A larger motor pulley would also increase the variable range.

The one trouble with scooter, and vehicle transmissions in general, is that the speed range requires for road use is covered by a 3 or 4:1 variation in the gear ratios, so not really that useful for a tool. My Hardinge lathe, for instance, gives a 8:1 adjustment.

Bill