16mm with 16mm pitch is nice and knarly. NIC 77 is the ballscrew Guru here he will see this and give some advice. As for me, I hope you post some build pics.
Sent from my SM-N970F using Tapatalk
Hi guys
Been a lurker for a long time. Made my first CNC router/mill about 11 years ago when I was in college, but have never been super thrilled about the performance.... Recently I picked up a bunch of servos, linear rails, ball screws, actuators, etc and I'm looking to use some of this to make a new machine without starting from scratch. Attached are pics of the actuators that I want to make use of. 15mm THK rails, 16mm diameter 16mm pitch ground ballscrews, 400w panasonic servos, 1um linear encoders. About 100lbs each
I'm looking to build a CNC mill that will devour aluminum, but also be able to machine steel at normal recommended feeds and speeds (instead of the just about grinding to dust I have to do on my non-stiff machine). Machine weight in the 500lb range with maybe 10" y travel and 20"x travel. I'm leaning toward making a new Y-axis base for the machine out of epoxy granite, then mounting one of these actuators on that as the X and another vertically as the Z.
My question pertains to the 16mm pitch primarily - I've been reading tons, but I can't really find great answers. I've done the calculations and it seems like the 400w servos will be fine moving and accelerating the loads at 16mm pitch, but I know pretty much everyone uses the 5mm pitch in a smaller cnc mill like this. At 3k RPM I would be moving at 1900ipm, which I couldn't really make use of, so it seems like I'm leaving a lot of torque on the table by using a 16mm lead instead of 5mm. The actuators are all assembled and extremely high precision - not sure what the specs on the ballscrew is, but based on the machine it came out of I'd guess toward the best. I could swap out the ballscrew but I'd prefer not to. I guess I'm looking for some real-world advice with regards to this setup since all calculations check out, but I could be missing something.
I also have a bunch of 750w panasonic or yaskawa servos that I could potentially use with a little bit of machining, but again I'd rather stick with the 400w if adequate.
Second question - is there any good hobby/light industrial controllers out there that would accept a second encoder input for each axis (for the linear encoders)? Side note - I installed some of the extra 1um encoders on my manual mill as a DRO and I had to limit the decimal places it was displaying to stop me from trying to hit .0000 dead nuts every move.
Thanks in advance for any and all advice/help
Similar Threads:
16mm with 16mm pitch is nice and knarly. NIC 77 is the ballscrew Guru here he will see this and give some advice. As for me, I hope you post some build pics.
Sent from my SM-N970F using Tapatalk
You would not have enough power from the 400w servos to machine steel with a 16mm pitch Ballscrew even the 750w would not be enough with a direct drive cutting aluminum you might just scrape by with light cuts most machining centers don't go above 12mm pitch and have large servo motors to cut steel
For example I have a light mill which has 5mm pitch screws and has a 400w ac servo on this at 2:1 gearing and it can cut steel at a normal speeds and feeds, can be stalled easy if the cut is too deep
Also for a comparison a Bridgeport Knee mill just gets by with a 750w servo motor and 5mm pitch Ballscrew some of these are at 2:1 gearing also which gives you some good cutting power but at direct drive they just get by
16mm Pitch screws can work well for a wood router, and light cutting aluminum
Mactec54
The information in this link is mostly Greek to me but Linuxcnc can do what you ask:Second question - is there any good hobby/light industrial controllers out there that would accept a second encoder input for each axis (for the linear encoders)?
Link ---> LinuxCNC Documentation Wiki: Combining Two Feedback Devices On One Axis
I've read Galil motion control hardware also has this ability.
There may be many other possibilities.
Anyone who says "It only goes together one way" has no imagination.
I'm not going to try to answer your question on the ballscrews other than to say that 400 W seems a bit light. This is just my gut feel, nothing to really back up that feeling.
The only controllers, in a hobby price range, that I know of that have dual encoder inputs are Galil Motion Control products. There are many units available on Ebay at reasonable prices. There are Galil plugins available for Mach3, and maybe Mach4. I have successfully used 1um linear scales with a Galil/Mach3 system. But be very careful of what you are buying from Ebay, not all of the Galil products listed are compatible.
Having said that, while there are some applications that could benefit from a dual loop system, it is my opinion that using linear scales only is the way to go on a machine tool. I don't care what the motor/ballscrew is doing, the only important position is the table position. On my mill I don't even have encoders on the motors/ballscrews, am only using the linear scales. On the other hand, I am only using the motor encoders on my lathe. Both machines are very accurate and repeatable.
Jim Dawson
Sandy, Oregon, USA
One thing I'd be concerned with in this case is the servo control. The encoder is used for real time velocity calculations by the servo drive. For example a Yaskawa servo encoder has roughly 16million pulses per revolution that the drive takes as feedback in order to provide real time adjustments to smooth out motion and damp vibrations. I don't think you could run one by plugging in a linear encoder to replace the rotary.
The motor encoder would feedback to the drive as normal, and is required by the drive. The linear encoder would be connected to the controller and all of the position control is controlled from there. In this case, the motor encoder would be roughly analogous to tach feedback from an old DC velocity system.
Jim Dawson
Sandy, Oregon, USA
@Jim Dawson
Okay yes this is exactly what I had in mind, we are on the same page. I'm actually a Galil distributor, so maybe I'll take a look. It's just hard to look at a $1000 controller when my previous machine uses a gecko g540/Mach3! Just checked and we have a 41x3 and 40x0 demo I could use for a bit. I have a bunch of these motion cards (pic) but I don't know how to use them or how to get software
Not sure what the motion card is in the picture. Never seen that one before. But here is all I could find about it EMOTIONTEK
I would go with the 40x0.
Just as an odd side note: I purchased a 4040 off of Ebay. It said 4040 all over it so I can't fault the seller, but it turned out to be a 4030. Very strange, never seen that before.
Jim Dawson
Sandy, Oregon, USA
If I were to stick with 400w servos, 16mm pitch.... what kind of cutting would I be limited to? Would I be able to use a .375" 3 flute end mill in aluminum, .375" deep, .050 stepover, 15k rpm, 200 ipm feed rate? Which contributes more to radial cutting forces... chip size or feed rate?
Doing some calculations... with the above parameters I get around 150lb radial load. Which with this set up would require 2.82Nm continuous in addition to the 1.13Nm to accel/decell at 1g. Which would still be above a 750w servo direct drive.
It seems kinda silly, but would running a belt reduction to the ballscrew be advisable? Seems kinda silly to introduce a belt instead of just getting a 5mm pitch ballscrew. But I already have the ball screws and belts/pulleys are cheap. I guess I'd be worried about responsiveness/settling time after introducing a belt. Kinda thinking out loud here, so any input is appreciated!
Another question that I've been curious about - Z axis. Do people generally have a beefier Z axis or a weaker one? I feel like my Z feeds are going to be less than my X/Y in general. And I'm going to counterweight the spindle (estimating 100lb load, 100lb counterweight). Would 16mm pitch be fine in this case if I kept Z feeds low?
Just to put this in perspective, I have a BP clone knee mill. The servos are 750W with a 2:1 belt reduction, 5mm lead ballscrews. I have not noted any accuracy problems with the belt drive. Any inaccuracy in the drive system is pretty much mitigated by the linear scales.
Jim Dawson
Sandy, Oregon, USA
If I were to do a belt drive reduction... Let's say 4:1 with now a 750w servo - am I going to be driving the 16mm lead, 16mm diameter ballscrew too hard? Accelerating 200lbs at roughly 2g? Ballscrew is 2-start
As near as I can find, the dynamic thrust load capacity on a 1616 2 start ball screw is 980 lbs. So about 2X the acceleration load that you propose.
Jim Dawson
Sandy, Oregon, USA
Correct you can't do that, there are newer model servo drives that you can feed both the servo encoder and a linear scale to the servo drive and the servo loop is closed in the drive Delta and Yaskawa can do this along with some other manufactures also, no other hardware needed but in most cases if you have quality Ballscrews and quality servos and drives you don't need the linear scale for normal machining
Mactec54
You would be fine I don't think you would need a 4:1 I would start out at 2:1 or 3:1 max and change if you had to as it is easier to do a 2:1 than a 4:1 as the timing pulleys will get big when you get up to 4:1, you have to consider the smallest diameter Timing Pulley you can use, number of teeth engaged is important for the motor pulley, this timing pulley determines the diameter and number of teeth for the #2 pulley
Mactec54