Re: Gecko differences

Thanks for the well thought out post. Really clears it up for me.

A question I have now is about the Z.

With a plasma table, what is typicaly used on the Z axis ?

I can see the X and Y would benifit from the high speed of servo's
but does the Z need the same level of speed ?

I'm thinking that the slighly higher cost of the step multiplier and
encoder and servo make each axis more expensive than a stepper
(espically since I already have the steppers)

Dave

> > > >
> > > >What makes a 340 better/different than a 320 ?
> > > >
> > > >dittto for the 201/210
> > > >
> > > >
> > > The 340 and 210 have the G901 pulse multiplier on their

inputs.

> > This
> > > allows them to multiply incoming pulses by 2, 5, or 10 (or 1),

so

> > the
> > > motor moves faster. It also gives the ability to work with a
> > common
> > > ground or a common supply for the optoisolators.
> >
> > I'm looking at a wood router and a plasma table.
> >
> > Both will probably be either rack or open belt, more likely the
> > plasma will be rack and the router will be open belt.
> >
> > Any idea on what type of machine the multiplier is needed ? Or

is

> it
> > more for higher resolution or faster rapids....
> >
> > I'm trying to see the added benifit for my applications.
> >
> >
> > Dave
> >
> >
> >
> > Any ide
>
> Dave you need to do a little math to decide is a multiplier is
> needed. There are a couple of givens: The max pulse rate you can
> get (depending on the PC used) is between 25,000 and 45,000 pulses
> per second. The next number you need is the maximum speed in IPS

you

> expect to need. Now you have to calculate the gear/belt ratios.

The

> final number is the line count of the encoders in the case of

servos

> or the microstepping ratio in the case of steppers.
>
> Lets' work through a scenero.
>
> 1. The max pulse rate of our PC and software is 35,000 pps (pulses
> per second) If we are spinning a stepper we know most take 200

steps

> to turn one revolution. Now if we are using a Gecko 201 it
> microsteps at a 1/10 increment meaning that if we send it a pulse

it

> moves the motor 1/10 the normal distance so to get the same single
> revolution we have to send ten times as many pulses or 2000 to turn
> it one time.
>
> Lets assume that your belt/gear/pinion ratio is 2:1. You have to
> calculate how far the load will move given one revolution of the
> motor shaft. If one rev will move the load .5 in then it will take
> 4000 pulses to move the load 1 inch. We now have a normalized

number

> to use. Now lets say our target speed is 200IPM (3.33 ips) then

all

> we have to supply to get there with the given ratios is 13,333 pps;
> well within our 35,000 max rate.
>
> If your numbers came out that you need to supply 65,000 pulses per
> second to get your target speed then you would need to use a pulse
> multiplier so that one pulse from the computer puts 2 or 5 or 10
> pulses into the circuit.
>
>
> The problem starts to be amplified when there are servo's

involved.

> First it's a good idea to operate servo's at higher RPM rather
> because of the torque characteristics. Then usually means your
> gearing ratio needs to be higher. A 3 or 4:1 ratio is often used.
> The good news is that the speed is now converted to increased

torque

> (like dropping the old pickup into a lower gear). You have to know
> the line count of the encoder used for feedback.
>
> Lets use the same basic setup as before but add in a 3:1 belt
> reduction before the rack reduction of 2:1 and the fact we have an
> encoder of 500 lines.
>
> We first need to multiply the 500 line encoder by four since we get
> four distinct transitions (counts) per line. Called Quadrature it
> increases the resolution but forces us to deal with the higher
> number. Since a pulse from our PC will move the servo on count we
> now need 2000 pulses to move one complete revolution on the motor.
> Apply the reduction ration of 2:1 (rack) and 3:1 (toothed belt) for

a

> total of 6:1 speed reduction and now we need 12,000 pulses to move
> the load 1 inch. to move the same 3.33 ips as before we need

39,960

> pulses per second.....woops our PC will only do 35,000 maxed out.

A

> multiplier is the best way to get hit our top speed.
>
> What you find in real life is that you have to play with the
> numbers. Every change effects something else. The torque vs speed
> issue is one. That's why you see a lot of motors that are

oversized

> since you would have to change the entire design if you were trying
> for a given force at the tool. Often in our world of HSCNC you
> design around "found objects" so that might dictate the other

numbers.

>
> I found some nice ballscrew assemblies with belt driven servos
> attached. The encoders were 500 line. With a 5 TPI screw ratio, a
> 3:1 belt ratio and 500 line encoders and my target speed of 150

IPM,

> I needed 2000 * 3 * 5 to get an inch or 30,000 pulses to move ips.
> The target speed is 2.5 times that so I would have needed 75,000
> pps. The Gecko 340 was called for. I ran the multiplier at 1/5 so

I

> only need to supply 15,000 from MACH2.
>
> As for the gain or loss of resolution that is in the numbers as

well

> but as we have noted true accuracy is not the resolution of your
> machine but its ability to be in the exact positon commanded. MAny
> things effect that number.
>
> My advice is to first decide on your drive method and motors then
> start running numbers to see how fast you will have to spin the

motor

> shaft to hit your target speed. Once you have that and if the
> numbers are reasonable then you can determine which model of Gecko
> you would need.