Grizzly G3617 Conversion - Timing pulley sizing and servo/stepper questions.

[turbostang]

I finally got my mill so I can get underway with my addiction/new hobby.

I've been reading and reading and reading and can't find what I'd call a definitive answer on timing pulleys, specifically what ratio I should be using and what kind of motor I should get. (The machine is pretty big by most cnc hobbyist standards and it will take pretty big motors I'd assume)

I have a few emails out right now to various suppliers like Kelling, but no replies as of yet.

I was just curious as to which motor style I should get - a stepper or a servo with encoder. I constantly hear of missed steps etc are the servos that much better? What size of motor should I be considering? I want a stronger/bigger motor than you'd normally use because I plan on using it a LOT.

Am I to understand it correctly that a 4200 RPM servo motor, 4 pitch ballscrew and timing pullies with about 4.28:1 (14T/60T) - in theory should produce 245 IPM in a perfect world? (under no load, wide open)

Just in case anyone was wondering how I came up with those numbers...

4200 RPM X .250/rev = 1050 IPM with no gear reduction.

1050 IPM / 4.28:1 gear reduction = 245 IPM rapid?

Or am I not even in the ballparK? ideas? Suggestions?

Edit - for the record, this machine is just like Chich2's HAFCO shown HERE.

[Bubba]

Turbostang,
Here is some information that I got from Mariss (Mr Gecko) about sizing of servos, max speeds etc. It helped me when I was trying to figure the ratios etc.

"1) Start by finding out what your motor's no-load speed is at its rated voltage. Use a strobe tach if you don't have the data.
2) Design for a speed equal to 80% of your motor's no-load RPM when your mechanism is moving at its maximum speed.
3) Calculate the RPM of the driven element (leadscrew, etc, etc) when the mechanism is moving at its max speed.
4) Divide that RPM into 80% motor no-load speed. That will be the optimal reduction ratio from motor to load. Use a toothed-belt for anti-backlash reduction.
5) Determine what is your software's maximum step pulse rate. It'll be published somewhere. Convert your motor no-load RPM into revs per second (divide RPM by 60)
and divide that by 4. Divide that into your software's max step frequency. The result will be the maximum encoder line count.
Pick the next lowest standard mfg encoder line count. In other words, if your result is 566.66 lines, pick a 500-line encoder."

[davo727]

Hi, Im doing a RF45 clone and was advised by many people to just go with servos and encoders from the start, that steppers were ok for a mini mill but not for what I wanted to do. I have been advised that the 1100 oz servos from http://www.camtronics-cnc.com/ would work great for my rf45 mill. It has a 32 by 9.5 in table. I would think that for your X and Y axis they would work fine also . You might check with the guys in the bridgeport forum here what size servo they use on their Z axis knee. The Z might need more gear reduction than the other 2 axis. Maybe some other members will advise what pulley ratios they are using. Read this also : Lathemaster Square Column Mill ? I will post more later. Dave

[turbostang]

Bubba - thanks for the great info, now I think I've confused myself even more!

Dave - That is the same thing I consistently hear about my mill as well...1100 oz. servos seem to be the ticket, especially when coupled with ballscrews.

My confusion now stems from how do you arrive at the proper stepdown gearing? It seems I've seen a bunch of threads wtih similar machines ranging from 3:1 to 4:1.

Now, how do I know which Gecko to use? What is needed besides the Gecko's? I.e. a power supply, geckos, PC case to house it in...and what else.

[davo727]

Hey Turbo, Going by info from MCG who make the 1100 servo for camtronics http://www.mcg-net.com/electric_moto...rs/ID33004.pdf The servo driven with 80 volts max from a gecko g320 will turn 3360 rpm. I got this using mcgs figure of 23.8 volts req per K rpm. So using Bubbas calculations 80% of this max is 2688 rpm. Edit - I goofed the other calcs , hang on a minute ... Dave

[Bubba]

Turbostang,
Ok, now you have opened the whole can of worms!

First off, DON'T rely on "peak torque" as this is the point that is close to the self destruct point of the motor (servo). My preference is servos and know less than nothing of steppers.

Again, I will defer to Mr. Gecko and have put together a spread sheet that will allow you to "play" with the various values. Here you can play with the servo voltage, the parameters of the actual servo motor, and the "load". everything is a trade off.

OK, you have decided on your motor, now you will need both a driver (gecko?) and a power supply. Oh, by the way, you will need a method to connect your computer to the driver(s), the power supply, and the machine to be controlled (servos).

POWER SUPPLY:
This can be an unregulated power supply of proper voltage and many have figured the amperage to be ruffly equal to 2/3 the sum of the amperage of the axis supplied (eg. X Y Z axis each requiring 5 amp= 15 amp and therefore a supply that will do at least 10 amp).

BOB (Break out board)
I personally think it is advisable to have a bob to make all the necessary connections between the computer, the drives and peripherals. (you will find many that differ on this one)

CASE
Finally, you will need something to put all this stuff in to minimize the potential for swarf and coolant from getting to the electronics!

In the spreadsheet, I have used a load of 500 Lbs as a suggestion, YMMV:})
This file has been uploaded previously and i "assume" Mariss as reviewed before, so this is my interruption of what he has said. If I am wrong, well so be it:})

If I can help further, let me know.

[davo727]

Going by Bubbas formulas here: 4 TPI ballscrew times 200in per min rapid = 800 divided into 2688 rpm = 3.36 pulley reduction required. ANOTHER EDIT:3360 max servo rpm max div by 60 = 56 revolution per second multiplied by 600 pulse per revolution from a 150 cpr- count per revolution encoder will give you 33,600 pulse per second required. Mach3 software will pulse 45khz so this will work. If you go with finer resolution like a 500 cpr (2000) ppr encoder it would put you at 112,000 pulse per second required which is over twice what mach3 will output so your rapids would be 2.5 times slower. Dave

[davo727]

Turbostang: As far as the pulley ratios I think with mine and yours possibly there might end up being some trial and error. If something doesnt work take the offending pulley and sell it on ebay and try another ratio. Not the end of the world. But I do think that your Z axis will need a greater amount of gearing reduction than the x and Y due to the weight involved.

[turbostang]

Hey guys, This is some good info and hopefully it will turn up in a search for someone else - now it will turn up in a search for me, because I'll definately have to come back to it.

Bubba, I took your advice and called Marris. That guy is too smart for his own good! We talked for nearly 1 hour, he rattled off so many numbers it was unbelievable.
He told me which Gecko drive I needed, all the info I needed for my power supply, the pullies I needed to start with in regards to desired rapid moves... everything!
Long story short, the 1125 oz motors are welcomed overkill, and can be geared down to 3-4:1 on Z and live a long happy life. I've ordered the motors and hopefully they'll be here mid next week. Once the arrive and I confirm they are what I paid for, then I am calling Marris back to order the drives. He was so helpful, it was unbelievable.

Edit - Marris also mentioned gearing down the Z some, to help out with the weight and to help slow down the rediculously fast rapid moves.

[davo727]

So did he reccomend the G 320? or ? Thanks, Dave EDIT: Also did you discuss encoder resolution with the servoguy?

[turbostang]

So did he reccomend the G 320? or ? Thanks, Dave

Over the phone he quoted his own post I read in a different thread, LOL!

The G340 is marginally more useful than a screen door on a submarine.

Its only function is to turn your fine 1,000 line encoder into a 100, 200, or 500 line encoder. Why would you want want something like that? Only if you are already stuck with a high resolution encoder, a wimpy step pulse source and you need to go fast.

Break any part of that chain and you are far better off with a G320. It costs a lot less, in fact, just about the cost of an encoder from www.usdigital.com .

Its only advantage is sevro stiffness. A 1,000 line encoder will have a +/- 0.045 degree dither while all moves will be in 0.9 degree increments.

It's a specialized drive meant to meet specialized needs. Don't waste your money on it unless you meet the requirements listed above.

How do I know? I designed and manufacture the drives.

Mariss

[davo727]

Hi All, Somebody correct me if Im wrong but it looks like for this particular servo at its max of 3360 rpm at 80 volts that using a 200cpr (800ppr) encoder is the max resolution you can use before running into the Mach3 limit of 45 khz. EDIT: Also here is the link for encoders : http://www.usdigital.com/products/in...encoders.shtml I think that we need to use E5S or E6S depending on shaft diameter. So basically you are looking at about $45 to $60. each if you have to replace them due to wrong resolution. Dave