Building a plasma table...servo's or stepper's? (Torch head?)

[carbidecraters]

I am wondering if I need the accuracy of the servo's. I am looking at CandCnC.coms plasma pack 4 with servos and a voltage height controller. My question is my table will only be 5 x 8 with nice thick linear ways. Can I get really nice accuracy with the 34 step steppers or is it always better to go with the servo's? I am trying to get some decent accuracy on some parts we make and the older stepper machine doesnt seems to do it all that well (Maybe the THC control...not sure).

BTW I really like tha fact that Tom puts it all together for you and all you need to do is run your wires and hook up your motors right (Way cool)

[Torchhead]

The choice of servo or stepper is more based on the overrall use and size of the table. Accuracy is a function of proper grearing, low backlash and rigid construction.

I have a Stepper driven router table that has good resolution and accuracy due to 5TPI leadscrews. antibacklash nuts and smooth performance from the Gecko drives. It will hold about .0015 in repeatability and absolute accuracy. I traded speed for torque and resolution.

Steppers exhibit the characteristic of having good torque (close to their specs) at low RPM but that drops off with speed. There is a temptation to just use the low RPM and direct couple the motors to a Rack & Pinion. Speeds can be impressive but torque to take sharp corners at higher cutting rates (acceleration) suffers and resolution is dismal. A pinion ratio is PI (3.1416) times the DP of the pinon gear so you get an increase in speed and comparable reduction in torque and resolution.

The "perfect" combination for a stepper running with a pinion or 1" or smaller is a 5:1 step down. Steppers properly driven will turn about 800 RPM. Take the 3:1 STEP UP of a 1" pinion and a direct coupled stepper will haul ass at close to 2400 IPM ;it will actually stall sooner than that but your accuracy won't get much better (no matter how tight the mechanics) than .020. Adequate for plasma if you like ragged arcs. Add the 5:1 and you gain back the lost accuracy and increase the resolution down to about .002. The torque at higher speeds is better and you still have 480 IPM top speeds.

If you factor in a servo system to that same machine and increase the belt ratio to 10:1 (requires a two step belt transmittion) then you start to see even more accuracy, torque (at speeds). Given that the RPM of the servo can hit 3000+, you can see that direct coupling is out of the question (9000 IPM...insane). The 10:1 moves it back close to 900 IPM and with gobs of torque across the full range of speeds.


Soooo. If things are designed right you can have a moderately fast and accurate stepper. If you want higher end performance, better acceleration, and high speed Rapids the servo has the edge.

Just like a Diesel VS a gasoline engine, each has a place where it's features can best be exploited.

I run tables with both types of motors (why shouldn't I, we make both types of controllers). My steppers serve me well and run non-stop for hours.

A table is only as accurate as all of it's errors combined. Servos start out with more raw resolution and their faster RPM's allow you to "trade" speed for more resolution by proper belt reduction.

Servos need to be considered when:

1. The gantry mass becomes great enough to demand the added torque needed to do the fast acceleration and direction changes. You can buy yourself some added torque with steppers or servos by using a dual drive (motor on each side) approach. Even them steppers start to fall down if you want both speed AND torque.
   
   2. The size of the table starts to make the upper speeds of rapids a factor in production.
   

############################################

All the hooey about not losing steps and being more accurate (or my favorite, that Steppers are "old technology" and servos's aren't) is generally marketing hype. Steppers are dependable when the system (power, voltage, gearing, and tuning) are done right. Problem is that "right" can be hard to quantify on paper.

Servos will help save a poor design. They just have enough power to punch through most design mistakes. They still need adequate power (watts) and require a more complex interface and setup procedure.

There are pros and cons of each type of system including cost. Servo motors (DC brushed types) are not much more than larger steppers, the motor drives are not much more, but the support electronics are more expensive and the needed power is a lot more.

For a source for prebuilt belt drives (5:1 and 10:1) see http://www.CNCPartsKit.com.

TOM CAUDLE
www.CandCNC.com

[carbidecraters]

Your the man Torch head! I am 99% sure I am going with Servo's. Do you offer a 6 axis setup? I would like to

X Axis (Long) 2 servos
Y Axis 1 servo
Z Axis plasma 1 servo
Z Axis router 1 servo
A Axis rotational 1 servo

The table I am building is using 2 2" linear ways on the Y and 2 2" linear ways on both sides of the X so it will be a very stout table and I just need to be able to use the plasma and the router in different variables without having to switch heads. Engraving and then plasma cutting the parts out would be very usefull. I wonder if I would have to have 2 different X0Y0 offsets or if I could calculate the spacing between the two heads....got any thoughts on that?

[carbidecraters]

I probly could get away with 1 servo on X but I really would like the torque of two servo's since I am really building a stout table. Am I overbuilding the servo system here?

Table will be 8' x 13'

[Torchhead]

Originally Posted by carbidecraters Your the man Torch head! I am 99% sure I am going with Servo's. Do you offer a 6 axis setup? I would like to X Axis (Long) 2 servos Y Axis 1 servo Z Axis plasma 1 servo Z Axis router 1 servo A Axis rotational 1 servo The table I am building is using 2 2" linear ways on the Y and 2 2" linear ways on both sides of the X so it will be a very stout table and I just need to be able to use the plasma and the router in different variables without having to switch heads. Engraving and then plasma cutting the parts out would be very usefull. I wonder if I would have to have 2 different X0Y0 offsets or if I could calculate the spacing between the two heads....got any thoughts on that?

We have 5 axis and 5.5 axis (6 drivers & motors with one hardware slaved) units.

Look at the Plazpak 5 and 6 http://www.CandCNC.com/PlazmaPakPriceChart.htm

The bigger the table the more mass you have to move. Big gantries need more torque.

TOM CAUDLE
www.CandCNC.com

[David Bord]

Originally Posted by Torchhead We have 5 axis and 5.5 axis (6 drivers & motors with one hardware slaved) units. Look at the Plazpak 5 and 6 http://www.cnczone.com/PlazmaPakPriceChart.htm The bigger the table the more mass you have to move. Big gantries need more torque. TOM CAUDLE www.cnczone.com

Tom, for some reason, I cant connect to the above URL. I see your sig now shows cnczone.com, but I cant seem to find the link to your online store with the motion control parts.

David

[ger21]

Try this.
http://www.candcnc.com/PlasmaPakPriceChart.htm

[Torchhead]

Wow three typos in row! My bad...musta been late....

the correct link is http://www.CandCNC.com/PlasmaPakPriceChart.htm

[Torchhead]

Gawwwwww! THe Zone is changing my url link after it posts!!!!!

replace cnczone.com with CandCNC.com

[Torchhead]

Okay, which brilliant programmer came up with that idea?......spell checking and correcting a URL?

It's my website "C and CNC dot com" (take out the spaces and put in a period for the "dot". The rest of the URL is correct.

TOM CAUDLE
www.CandCNC.com

[DRL]

Originally Posted by Torchhead Gawwwwww! THe Zone is changing my url link after it posts!!!!! replace cnczone.com with cnczone.com

Wow, thats weird. I have visited your site many times using the links(bookmarked now) and they worked, they must have changed something lately

[carbidecraters]

We are going for a 7' wide by 13' long table using 30/30 extrusions for the ways and gantry and building an extra table for material support so we dont tweak the gantry. I will definately be picking up a plasma pack 4 at the least in the next 30 days or whenever we are ready for the electronics. Tom can you install longer leads for us if we decide to go bigger?