Hello Guest it is September 23, 2023, 04:29:52 PM

### Author Topic: Servo ratios and drives..  (Read 9731 times)

0 Members and 1 Guest are viewing this topic.

#### ART

• 1,702
• Tough as soggy paper.
##### Servo ratios and drives..
« on: October 12, 2006, 11:13:36 PM »
Hi all.. continuing on with this very rambling topic..

Today we discussed motors. We're going servo as I said earlier, but the question comes up , how to connect the..ballscrews.. belts..ect..

The idea is to have as low a mass as possible to move, but with the rigidity required to get a good cut. Ive seen many designs, and we're not
close yet to deciding. Brian mentioned an interesting design where a timing belt is flipped inside out on a pulley and the pulley runs on a track with
a series of tapered slots on it. This woudl kill backlash problems as the rubber of the belt compresses into the track slots.. interesting.. might have to try something
like that. Ball screws are great and tight, but they do have a rotational inertia and that adds some acceleration time to the movement. Hi acceleration is the
ultimate target. Like all things in life, tradeoff's abound in the design of a router table. The stiffer the heavier, but the lighter the higher the acceleration,
the trick woudl appear to be getting just the right balance. Since we'll be using Servos that at fast speed are about 3000RPM, the perfect design is about 3000
rpm ratioed to give the highest speed we would want at the drive.  The required acceleration then would simply be a function of the mass of the gantry. Y on the router tabel is
the long axis, so it will push the whoel gantry, its the worst one, the X is usually a cakewalk as it has less than 1/2 of the mass of the Y, and the Z is in a class of its own
as it has to lift or drop the weight of a carriage. Even with counterweighting, there is the mass to consider on that one. (Just because somethign weighs nothing, doesnt mean
you dont have to take its mass into consideration.. )

Id like about 12,700mm's per minute. Or about 500 inches per minute as my rapid. It may be overreaching, but we'll crunch some numbers and see. That means on a 4" pulley, Id need about
2*3.14*2 or about 1 foot per rotation on a 4" pulley with  reversed belt, so an end speed of about 500/12 = 41rpm , so a reduction of about 75:1 woudl seem great under that scheme.
Gonna have to look around to see what I can find and make a compromise I suspect. More thought required. Just thought Id mention all this so you get an idea of what to consider in
some of the planning of driving the gantry about to get a better result.  Ill keep you informed on the progress...

Art

• 361
• When the going gets weird, the weird turn pro.
##### Re: Servo ratios and drives..
« Reply #1 on: October 13, 2006, 12:49:54 AM »
HI Art,

Is permitted to discuss things here or is it only a log for you? I have a few comments based on my own design that I would like to share. And some problems that I have run into in this area.

If not please feel free to nuke this post.

#### ART

• 1,702
• Tough as soggy paper.
##### Re: Servo ratios and drives..
« Reply #2 on: October 13, 2006, 08:16:28 AM »
No, feel free to post. In the end we will probably reorganise it all to some logical oreder, but kick in at will..

Art

• 361
• When the going gets weird, the weird turn pro.
##### Re: Servo ratios and drives..
« Reply #3 on: October 15, 2006, 02:57:15 AM »
"Id like about 12,700mm's per minute. Or about 500 inches per minute as my rapid. It may be overreaching, but we'll crunch some numbers and see. That means on a 4" pulley, Id need about
2*3.14*2 or about 1 foot per rotation on a 4" pulley with  reversed belt, so an end speed of about 500/12 = 41rpm , so a reduction of about 75:1 woudl seem great under that scheme.
Gonna have to look around to see what I can find and make a compromise I suspect. "

Hi Art, This is pretty close to what I have set up on my mill, and in retrospect I would have done it differently.
My motors are a little slower than yours and I am using a 50:1 planetary gear reduction to a belt drive. It is all very heavy duty but i still have a little backlash ~.025 in the x (my long axis) .01 in the y and .002 in the z (ball screw). My gantry with the z weighs around 1300 lbs. There is lots of mass there, and as a result to avoid oscillation (i am suspecting belt spring) i have to have the accel to around 20. full out the machine is capable of 550 ipm on x and y and 160 on the z.

You are indicating that you would like to have around 72:1 for a reduction. To accomplish this you basically have two options. First a \$1600 low backlash gear head or a big mess of belts and pulleys. for 72 to 1 it is going to end up probably being a 4 stage belt pulley assembly. that turns out to 8 or nine not cheap pulleys 4 belts lots of bearings and shafts and a bunch of machining for the transmission case. Belts do stretch a little and do have a little backlash and the more you have the worse it gets. Especially if you have a bunch of mass to deal with at high speeds. Don't forget you have to drive the long axis from both sides otherwise you will get racking in the gantry, that adds to the complexity and cumulative backlash.
I opted for option one with the gear drives. My plans this week are to definitively find the backlash in my x. I have a feeling it is a combination of belt stretch 6 arc min of backlash in the gear head and a twisting of the drive 2" shaft.

Any way I said if i had it all over to do again it would be different. What i think would be a better route is to go bigger on the servo and do a one stage 4 or 5 to one with a 2" belt. This gets rid of the backlash in the already expensive gear head, gives more speed and makes for a simple transmission. We are planning on converting the entire thing over to ac servo and the plan is to put around 3hp on the x,y,z and use as little reduction as we can get away with.  We recently got a Haas tl1 and after carefully digging around it and getting ideas i realized that they aren't doing any reduction with the exception of the ball screw pitch. The servos are 750 watt and are directly coupled to the screws. This seems to provide plenty of power and i have a feeling they have limited the rapid electronically to 200 ipm as it should be capable of way faster rapids ( not that that would get you anything other than shaking your machine apart, at least on this one).

To summarize i would suggest saving the money on fancy planetary gear heads or tricky pulley belt creations and put it into a bigger motor and a simple reduction..

Just my \$.02

#### ART

• 1,702
• Tough as soggy paper.
##### Re: Servo ratios and drives..
« Reply #4 on: October 17, 2006, 10:09:15 PM »

Well, I managed to find some 20:1's here, so Ill probably go with those. Ill design around them anyway I think. They happened to be on my shelf and look like I can adapt them to my AC servos', so Ill probably give that a whirl.

Art