Servos redux

Hi all, I'm still somewhat floundering on my XY servo table. If you EVER think this group is OT, go to alt.machines.cnc, so...I ask here. At your
advice I have specified servo motors for the job and have a few conundrums. In order to design a system I have to start somewhere but I have been reading. I know the 50 lb. table has to move 1" in X or Y in 3 to 4 tenths of a second. The table is on ball bushings and I plan on a 5/8" x 5 tpi ballscrews. Therefore the motors must accelerate from 0 to 1,300 rpm and decelerate back down to 0 in that time. I have run Parker Hannifin's "Motion Sizer" software to get a feel of what's needed. I do want it very simple and cost effective, but I want to reproduce the system on 4 more machines as budget allows. I've looked at and talked to the Rutex rep. about their drives and was recommended to use brushed motors as he wasn't comfortable with their brushless drives yet. I like the Rutex setup as it's modular and has IO and runs from a parallel port on a PC. He recommended #23 double stack or # 34 NEMA motors with about 15 in/lb torque with 250 line encoders. Parker's software shows somewhat less torque needed.
It seems that every vendor I've talked to didn't seem to grasp one part of this project or another and I felt uncomfortable. I know I'm not the best communicator in the world but...I don't want to blow my whole budget on a system that won't do it or is too complex or un-reproducible.
So, the motors need to be a good fit with the motion and mechanicals, the drives need to fit the motors, the PC needs to fit the drives the software needs to fit everything.
* Does it seem I'm on the right track? * What comes first, the chicken or the egg...What do I specify first then build around? * Is there any reason not to use brushed motors? * Encoders, resolvers? Rutex only takes encoders; OK? * Any thoughts about Rutex? * Any thoughts about software to drill 9 rows of 12 holes getting a signal from the drill crank for the next move? * Any sources for motors and power supplies and such? * I do have a great source for ballscrews http://www.nookindustries.com/ and they are 3 blocks away. * Anybody have any of this stuff for sale?
( sorry abou the OT but I promice to cut a LOT of metal)
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Sat, 30 Oct 2004 15:36:33 GMT, "Tom Gardner"

Greetings Tom, Have you called Gecko? There is an engineer there, I think the only one, because he's the one I always get. He will give you TONS of information about motor size, torque required, etc.. He has helped me a few times in the past. I have bought the Gecko servo drives but have not been able to install them yet. Eric
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
I've read about Gecko and they get good reviews. The only advantage to Rutex is the modularity and IO and higher voltages = more options on motors.

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
In article <5yOgd.18060$5b1.4069

Load > drive (ball screw or timing belt) > motors > amps > controller. But there's bound to be some back and forth as you optimize.

Besides the maintenance issues of the brushes themselves, brushless motors are generally capable of higher peak torques than brush motors. This can be significant in an app like this which requires high accelerations.

Encoders are fine as long as you pay attention to noise issues. Differential outputs are far superior in rejecting noise.

For moderately priced systems, I've had good luck driving Bodine brushless motors with Copley amps. You do have to do your own cabling and motor commutation if you go mixing amps and brushless motors. Brush motors are simpler in this regard. Until you've done a few, commutating a motor can be very frustrating.
Build your own power supply--transformer, bridge rectifier and filter cap. No regulation required. Hammond and Signal Transformer are good sources for this sort of transformer. Hammond is cheaper, Signal has more available.

Unless you need high accuracy, you'll get higher speeds with a timing belt positioning system than with the ball screw. For fast, short moves you'll find that most of the motor's torque is being used to accelerate the motor's rotor and the ball screw (unless the screw is very short), rather than the load.
You should be able to get within .010 - .020" of the commanded position with a reasonable length timing belt, and repeatability will be somewhat better than that.
Urethane timing belts with steel cord are best for positioning apps. Breco is one mfr. I've used a lot for this sort of thing.
http://www.brecoflex.com /
Ned Simmons
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Tom Gardner wrote:

I think the first thing is the motors. The reason is everything else is pretty much a commodity item, but motors are all different, and you do have a somewhat demanding application. But, real motion system manufacturers have done stuff WAY more demanding than this. I have some Yaskawa motors that are designed to accelerate from zero to 2500 RPM, move 1/4 of a turn and then stop again, all in about 10 ms! The drivers for these motors are several times larger than the motor itself!

What is the duty cycle of this machine? If it is for heavy production use, then you certainly want motors with replaceable brushes, at the least.

Encoders are a lot more common, and easier to interface to.

Drill crank? Why not just use a 3-axis CNC system, and have everything under unified control? If you want to change anything, it is all programmed in one file, you just change a number, and it is taken care of.

Well, having just done some retrofitting, I think I have a product that would do it just fine. I'm not necessarily trying to sell something, but just pointing out that there are a lot of common CNC system that should do exactly what you want. You can see some photos of this last retrofit at : <http://jelinux.pico-systems.com/pwmbox.html You will need somewhat larger motors for your required performance. Surplus Center (Lincoln NE), C&H (Pasadena, CA) and Herbach&Rademan (NJ) are some of the places to look for servo motors. And, of course, eBay.
Jon
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

The machine needs to do @ 350 pieces per 7.5 hr shift. 350 x 96 = 33,600 moves. (target speed, I might settle for less though who wants a slow printing press printing dollars.

Well, the rest of the machine runs from the overhead shaft and a bunch of other stuff is going on; staple feed, staple cutting and forming, wire bending and insertion...etc... It's just easier to ignore those actions because they all take place in the time to drill the hole.

The problem with eBay is "hit and miss" and it might be hard to reproduce on the next 4 machines.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
Tom Gardner writes:

That's 200 inches/minutes, so you need a peak speed of 400 inches per minute, reached in 0.15 seconds, or 160K inches/minute^2 acceleration.
Dunno what this table weighs, but that is gonna take some serious horsepower, and serious controller and power supply to deliver it.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
That was my reaction too, but I was surprised to discover that it's only 5.756 lb force to accelerate (and decelerate) his 50 lb table as described.
I estimate the torque necessary to spin up an iron rotor 4" long x 2" dia to 2500 RPM in 0.15 sec to be about 8.1 lbf-in. "Estimate" means I just used rotor volume and specific gravity of iron to calculate estimated rotor mass and then moment of inertia. Power at this torque at 2500 RPM is 238.7 watts or 0.31 HP. The rotor dimensions used are about those of a 50-volt servomotor I have sitting here 4"" dia x 7" long, don't know its power rating. The leads coming out of it look like they would easily handle 8 amps.
No idea what other losses and loads might be present (ball screw efficiency, sliding , friction, etc) but I'd say about 1/2 HP of system power would handle the acceleration loads.
On Sun, 31 Oct 2004 01:26:02 -0500, Richard J Kinch

Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Site Timeline

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.