Re: Servo drives- exploding head

> Dave, a few things off the top of my head for you to consider....
>
> 1: Generally, a 1" pinion produces quite a bit of cogging and

> losses at the pitches typically used on a cnc machine, empirical

> (i.e. learned the hard way) has taught me 1.25" (30t on a 24 d.p.

> gives you the best tradeoff between fricitional loss and ratios.

> circ. is nearly 4", and a reduction ration of circa 4 - 10:1 will

> you reasonable resolution for the task at hand, and great rapids.

> ipm is 120 rpm at the pinion, so somewhere around 600 rpm for the

> motor, this is one of those tradeoffs, if you're using steppers,

> or 5 to 1 reduction, for servos, somewhere closer to 10 or 15 to 1

> ideal. (with low encoder counts on the servo, naturally)
>
> 2: On all but the smallest routers (36" or smaller), 240 ipm is
> downright creeping, while I've been running around that speed

> is only because I always have far too many non-machinery types

> the machine that have no clue what that yellow line on the shop

> means, when I had it in my shop, and normally with no one else

> 500ipm rapids were the norm (and yes, I've had a control system

> and let the gantry, all 800 or so lbs freewheel into the shock

> at 500 ipm, thought I was going to buy the farm that time, scared

> crap out of me) When cutting wood (and most plastics) even at

> you need speeds of about 200-400 ipm to get proper chipload, of

> this depends on bit, # of flutes, dia, etc. I usually run abut 200

> with a 1/4" straight single flute bit at about 30k, this gives me a
> reasonable finish, and good chip loading to remove heat with the

> instead of catching the pywood on fire. (yes, I made that mistake

>
>
> 3. Best approach I've found is simply try something (preferably

> scrounged/surplus parts), you will learn more in a "lost" weekend
> playing than all the time spent rummaging through the mailing list;

> much I can guarantee. Try some steppers, say about 1/2" to 1" per

> unless you're really a glutton for punishment, then try them at

> it's quite doable, and you'd be amazed at the acceleration

>
>
> erie
>
>
>
> >"If you apply 10 Lbs on a 5" diameter crank, you should get 400 in-

> >of torque."
> >
> >Using those as my reference,
> >
> >I can convert that from a 5 tpi at 400 oz-in to a 1 tpi at 2,000

> >and further to a rack and pinion (1 inch pitch or 3.14159 inches

> >rotation) to 6,283 oz-in on the spur gear. ( 1 inch /

> > so, 2,000 /0.31 = 6,283 oz-in when using a 1 inch spur gear.)
> >
> >That tells me the force I need to be able to handle with bearings

> >mounts for the final gear part of the drive. (ie: the final drive
> >gear much be beefy enough to handle the torque at that point)
> >
> >Working back from there, and assuming rapids of, say, 250 ipm on a
> >router or plasma, we can figure that if one spur gear rotation is
> >3.14159 inches, then a 1,000 RPM motor will be a tad too much.

> >many people want 3,141 ipm rapids, or 52 inches PER SECOND is a tad
> >fast for a 48 inch wide machine)
> >
> >A 10:1 reduction would spin the motor at 1,000 RPM, and the final

> >at 100 RPM or 314 ipm. Good enough.
> >
> >The 10:1 reduction would be required to be in two stages as you

> >get that from timing belts in one set of gears. (actually 10:1 is
> >hard to get in two stages.)
> >
> >But, although one has a theoretical 6,283 oz-in on the spur gear,

> >the 10:1, then the motor needs to be 628 oz-in.
> >
> >If, we used the 5:1 target ratio, then that would effectively

> >the 628 to 1,250 oz-in at the motor and still have the 6,283 at

> >
> >Now, the original thesis was for a person applying cutting forces

> >the Bridgeport, not transport speeds for a plasma or cutting speed

> >a router. I would assume that a metal router might use the same
> >forces, so for a metal cutting router, the 1,250 oz-in might be a
> >realistic top end. For plasma, it could certainly be less.
> >
> >Working in reverse, 3.14 inches per gear rotation and 250 IPM

> >drive shaft speed of 79 RPM and a motor shaft speed of 400 RPM or

> > That is in the range of stepper speeds.
> >
> >Now, the motor rotation is 400 RPM and the force is 1,250 OZ. So

> >do the 400 times 1,250 and divide that by 1,351 to get 370 watts.
> >Further 1hp is 746 watts so 370/746= slightly less than 1/2 hp.
> >
> >OK, I took liberties in that the original number is for cutting
> >forces. But a router table will also have cutting forces.

> >not as much as the Bridgeport, but much higher than the plasma.
> >
> >I didn't take in any factors for acceleration.
> >
> >(I should be looking for something to catch the grey matter soon)
> >
> >A servo is not great at 10% of it's full rated speed (treadmill at
> >5,500 RPM) so running it an absolute max of 400 RPM leaves

> >to be desired.
> >Even a specific duty servo motor at 1,700 RPM is missing that 80%
> >sweet spot by major amounts.
> >
> >A simple treadmill motor is rated for 1-1 /2 hp at 5,500 rpm or

> >such, but a Gecko 320 likes less than 80 volts. A 48Volt

> >as the basis of a power supply offers 66 volts.
> >
> >That's 57% of the nameplate voltage, so one gets 57% of speed and

> >So, the treadmill motor at 66 volts offers 3,157 RPM and 0.86HP
> >
> >(grey matter starting to look for exits….)
> >
> >…something happens here with torque and hp and all that….
> >
> >
> >(Quick. Look at wife, and smile…)
> >
> >change direction…
> >
> >A 370 watt stepper is also out of the sweet spot for steppers.
> >
> >OK, a 28 tooth pulley is the smallest listed for use on a 1 /2 inch
> >shaft and a two stage gear reduction with a 32 drive and 72 driven

> >each section will yield a final drive ratio of 5.06:1 so that part
> >can be achieved. OF course, there is NO simple set of pulleys that
> >move to 0.001" or 1.0000 inches. Everything is off by some amount,
> >and care needs to be taken so the cumulative error over the table

> >not too great.
> >
> >Then.. THEN,,, youse guyse toss in inertial matching and moment of
> >inertia to add some excitement to help the veins stick out just

> >little bit more.
> >
> >And you think my feeble skull can keep in the throbbing grey

> >
> >I did put this into a spreadsheet to try to make it all easier to

> >
> >Dave
> >
> >