[Speed57]

Hi All,
I’m looking for some advice on parts to use for a PM-45(RF-45) cnc conversion. I have been reading and reading all the RF-45 buildthreads. Seem like the more I read the more confused I get.

I would like to start converting my PM-45 Mill. Here is the component parts I am considering using.

Ballscrews

Ebay guy -Linearmotionbearings2008

RM2005 – 20mm x 5mm lead X and Z axis
RM1605 – 16mm x 5mm lead Y axis

Motors
DMM Technology
3 – DYN3-H drive
3 – 750W AC servo motor w/ encoder 3000RPM Rated torque 2.5Nm Peak 7.2Nm

Other electronics
PMDX-126 Breakout board
Ethernet smoothstepper motion controller.



Questions:
1. Do I need double ball nuts on the ballscrews.
2. Who makes a high quality but reasonably price ACbearing.
3. It there a benefit to using a double row AC bearingor is it overkill?
4. Should I use to a belt drive when using these ACservo’s or can I direct connect?
5. Are these servo’s the correct size for this size of mill?
6. Are these drives and motors the best choice.

Comments or advise on the parts I have selected would be greatly appreciated.

Thanks

[SCzEngrgGroup]

1. If you want zero backlash, you need double ballnuts. The inexpensive ballscrews will typically have 0.001-0.003" backlash with a single nut.
2. Look at VXB.com. I've used their A/C bearings on a number of projects, and all had zero backlash, despite the low cost (<$10 each).
3. There is no advantage to double-row in this application, and they will take up more room. Plus, they're not adjustable.
4. Absolutely. Servos are capable of very high RPM - more than you can use in direct-drive. But servos also generally have lower peak torque than a similar-size/cost stepper. Using a belt reducer lets you take advantage of the full dynamic range of the servo, while spending less on the motors and drivers than you would for a comparable direct-drive setup. I would suggest looking at the peak RPM, decide what max rapid speed you want (anything over 200IPM will be very scary for your first few years, unless you're an experienced CNC machinist), and pick the belt ratio accordingly. 3 or 4 to 1 is typical.
5. Those would impress me as serious over-kill. As a reference, my knee 9x49 mill uses DC servos rated at 810 oz-in peak, 140 oz-in continuous. With 2.5:1 reducers, I can do 350 IPM with ease. And that's with a considerably larger, heavier table. I use the same motor with a higher reduction to move the knee, which weighs about 700#, and it'll do 75 IPM.
6. AC servos are definitely the way to go, if you can afford them. I just yesterday took delivery on a brand new machine with AC servos, and am really looking forward to seeing it in action. DMMs have a good reputation. I'd just consider going down a notch in motor size, and going with maybe a 4:1 reducer to keep the speeds reasonable. Extremely fast rapids on machines this size are not of any great practical value, and GREATLY increase the chance of accident, and serious damage, if you're not paying 100% attention every second. Running into the stops at 350 IPM *will* break something, and you can't always count on limits to save you.

Regards,
Ray L.

[zamazz]

I have a G0704, not a RF45, but the concepts are the same.

1. Depends what accuracy you are looking for. The Chinese ballscrews seem to have around .001 of backlash, which added with machine flex and other sources of backlash will get you tolerances around .003, if you're careful. Preloaded ballnuts can reduce this, but require quite a bit more planning. I decided to use them, its a fairly good return for the investment.

2. Most people (myself included) use VXB bearings on ebay. My 12mm AC's were $8, and I needed 6. They seem fairly high quality, although I have yet to test them thoroughly.

3. You either need double row or two AC's per axis. I choose two per axis so I could adjust the preload.

4. You will need a belt drive. The 750's may be able to direct drive, but they will spin at an RPM greater than you need. For instance, at 3000 RPM and 5mm per rev your rapids will be 590 ipm, which is pretty useless on a machine like this. I would definitely recommend a belt drive, probably 2:1 ratio. I'm using 2.5:1 on my G0704 with 300W DMM servos.

5. Probably overkill. Especially once you add the belt drive, the 400W should be plenty. Once again, I don't have a 45 sized mill, so I'd get a second opinion on this one.

6. DMM is about the best value there is. You won't be able to find many comparable, new servos for the same price. They get a bunch of use here and I have yet to see someone who was disappointed.

Also, I'd go with the SS over the breakout board. Just personal preference, but it opens you up to many more expansions down the road. Also, it eliminates the need for a parallel port, which had given me trouble a couple times on our engraver. Good luck!

[SCzEngrgGroup]

Also, I'd go with the SS over the breakout board.

He'd need the SmoothStepper *and* a breakout board. Ethernet SmoothStepper is preferred over USB - seems more stable and reliable for most people.

Regards,
Ray L.

[Speed57]

Hi Guys,
Thanks for the replies.
Just so you know. I’m not a trained machinist or an engineer. I’m learning as I go. I’m an IT guy who works with computers and computer networks all day. But I love making chips! So please excuse some of the uninformed questions. I really appreciate everyone’s help.

The thing that confuses me is some of the guys are talking about using 900W AC servos and at the time waiting until the higher voltage drive as available to turn them at 3000RPM. The DYN2-H drive from DMM will turn the same motors at 1000 rpm with the same torque specs. I don’t know how to calculate the IPM based on RPM and the ballscrews. It’s my understand that servos develop more torque at the higher RPM vs. the stepper is just the opposite.
I’m thinking 200 IPM is plenty fast for me. I was just looking at the list price of the DYN2 vs DYN3 driver and the DYN2 is about 70% less then the DYN3. Changing to the smaller 400w motor only saves about $50.00 each.
I’m guessing that if I went with the DYN2 drive I would have more torque in the range I needed it.
So my thought would be to stay with the 750W motor and use the DYN2-H driver. It should allow the motor to develop peak torque at 1000 RPM. If I went with this combination could I use a direct drive and get the 200IPM and have the torque needed?
I’d like to avoid using with pulleys and belts if I can. But I do not want to compromise performance either.

The PMDX-126 BOB is setup so that you can added either model of the smooth stepper to it as a daughter board. The PMDX ever supplies the power to the SS board. It works very well. I used it on a lathe conversion. Steve the owner of the company is very friendly and helpful.

[TiagoSantos]

To calculate IPM based on RPM and ballscrews -

Your ballscrews will very likely be 5mm pitch, which means your table/saddle/head will move 5mm per revolution of the ballscrew. At 3000rpms, you will move 3000 x 5mm = 15000mm/min. In inches, around 590ipm.

I also think the Dyn3-H drives are overkill. I would either use the 750W motors with the Dyn2-H drive (which limits the motor RPM to 1000rpm = 196ipm) and direct drive, or the Dyn2-H with 400w motors and belts. That's what I have.. I have a 3:1 ratio on all three axis, which gives me more than enough torque and around 196ipm maximum, theoretically. I moved the Z axis, with the 50lb 2hp motor in there, at 180ipm easily with no air springs or counterweights. No tuning either, just typed in the settings they gave me.

Anyway - the 400w motors and 3:1 ratio belts end up providing more torque than the 750w driving the screws directly. As far as cost, it probably evens out, the 750w motors are $50 more, but that's about the difference in price of pulleys/belts vs good oldham couplers. I do prefer belts for the packaging options, it allowed me to mount the motors "inward" and keep the enclosure a bit more compact - the wall that my enclosure sits against is JUST big enough, I couldn't make it even 1/2" bigger.

[ninefinger]

Check out Kelly's build - he uses the DMM servo's on a MD-001 (big RF-45)
http://www.cnczone.com/forums/bencht...e_-_md001.html

Don't spend money where its not needed. Save it for tools!

Mike

[Speed57]

I sent DMM as email asking about using the 750W motors with the DYN2 drives. Below is the reply. I think I want to use direct drive for now. I can change to belts at a later date if I need to save space or multiply torque. I don't understand all the formulas he has in his email. Does this look correct for a PM-45 sized mill? Thanks for all the advise!


For sure, with the DYN2 servo drives and motors at ~1100rpm, a 5mm lead ball screw will give you a speed output of ~220IPM of rapids. The load estimate is as follows:

Je (Equivalent Inertia) = M (Load Mass) * (L (Travel) / 2pi )^2

Where L = 0.5cm / rev

Our 750W Motors have a Jr (Rotor Inertia) of 2.05 Kg-cm^2 and under normal CNC loads, Je/Jr<5. So Je<2.5

2.05 = M*(0.5/ (2*3.14) )^2

M = ~400kg = 880lbs

So this setup will be able to support a steady state load of 880lbs including load mass, friction and cutting forces. Servo motors are able to instantaneously (momentarily) overload their inertia output to ~10 times their rated output, so this should give you a very smooth performance even during rapid acceleration/deceleration while carrying a lot of load.

The torque characteristics / S-curve profile of our servo motors are very smooth and consistent throughout the entire speed range. This means the motor's peak output torque are 100% available whether moving at 5rpm, or at 1000rpm.

[TiagoSantos]

The Jr (rotor inertia) is a characteristic of each motor. They've worked out that the 750W motor has a rotor inertia of 2.05Kg/cm2. He's also assuming (and I have no reason to disagree) that under normal CNC conditions, the ratio between the "equivalent inertia" and the actual rotor inertia should be below 5. If Je / Jr < 5 ----- that is, if the ratio is under five --- then solving for Je means diving 5 by Jr (which is 2.05)---> 5 / 2.05 < 2.5..

Then he solves the formula of equivalent inertia, which has variables for the load that is being moved, the pitch of the screw and the equivalent inertia he just calculated above. He does this to work out the maximum amount of load that those motors should be able to carry under normal circumstances

I can't say if the calculations are accurate or not (I have no way of testing the rotor inertia, for instance) but I have no reason to not believe them. The 750W servos will definitely work fine for your machine, though!