"Stepping" in it

ndzied1

Lifetime Supporting Member
Join Date
Aug 2002
Location
Chicago, Illinois
Posts
2,857
Here's one I can't seem to wrap my head totally around. I think what the customer is asking for is impossible but maybe someone has an answer.



I have a customer with the system shown below.

step.JPG




He already has purchased ballscrews with 5mm lead (i.e. One nut revolution moves the screw 5mm). Now he says he wants one step (or an integer number of steps) on the stepper motor controller to be equivalent to a move of 0.0001".



I can set up the Step Controller for any number of steps/revolution that is a multiple of 200 (up to 100000 steps / rev). I can play around a little with the pulleys given the standard number of pulley teeth for L series pulleys:

http://www.tbwoods.com/pdf/tim_sprk_blt.pdf

10 12 14 16 17 18 19 20 21 22 24 26 28 30 32 36 40 44 48 60



Then he tells me he's using an IDEC Micro-Smart for step generation which has a max frequency of 20kHz. This comes into play because he requires a speed of 1.5 in/sec. So, if he could get his exact 1 Step = .0001” he would need 15kHz for 1.5in/sec. This rules out using anything higher than 1 step / 0.0001” as even 2 steps / 0.0001” would require 30kHz.



They started w/ Pulley1=36T and Pulley2=12T for a 3:1 reduction. Running through the math gives:

1 nut turn = 5mm = 0.1969”

1 motor turn = 5mm/3 = 0.06562”

Setting the control for 600 steps/rev gives 1 step = 0.00010936” travel.



Again, I think the use of the 5mm lead screw makes what he wants impossible but maybe someone else here has some ideas….
 
With both feet

You can't meet all the criteria. It isn't clear if having exactly .0001 inch resolution is desired or required.

You are the control guy. Make the mechanical guy earn his pay by designing a system with some even numbers. Mixing metric and US units will not work. We are cursed with the .4 in the 25.4 mm/inch
conversion.

I also think you will need to consider another controller. 20khz is crude. Most of these steppers have very coarse ramping and use a simple divider. This reduces your speed choices.
 
Unfortunately, I'm only getting to supply widgets for this job.

Right now the .0001" is required, but I think it will soon be only desired when they realize it is impossible if they stick w/ the 5mm screw. They bought the screws (I'm sure they saved 5 cents a piece over an inch pitch) when they thought they were using a different control.

This is a weird situation with an OEM, integrator, distributor (me) and the OEM's parent all with their hands in different parts of the design.

They have a smaller machine using the same 20kHz PLC step/direction control. For what it is doing, it should be fine except they blew the balscrew sizing!
 
THE CLIENT IS NUTS !

First of all, the lead screw will never give him .0001" of accuracy or repeatability due to local variations in the screw, the nut and the belt.

Second, a micro-stepping drive will NOT position down to the micro-step, more like to 1/4 a full step.

You can get 0.010 accuracy / repeatability if you can map (compensation table) the screw. Maybe a bit better if it's a precision ground short one. However, if he's asking for .0001 resolution, then my guess is he wants .001 accuracy 'cause it's nice to have a 10:1 ratio between raw resolution and needed resolution.

Tell him to furnish a gear set that has a 2.54:1 ratio. You don't want an even ratio as that'll cause wear patterns.
 
bob53 said:
THE CLIENT IS NUTS !
I thought the customer is always right (even when they're wrong)

bob53 said:
First of all, the lead screw will never give him .0001" of accuracy or repeatability due to local variations in the screw, the nut and the belt.
This is for a surface grinder so over a localized area it is plenty fine. Again, this system is working (with an inch screw) on a similar machine right now with acceptable results.

bob53 said:
Second, a micro-stepping drive will NOT position down to the micro-step, more like to 1/4 a full step.
We've done much better than that, about 3% of a full step or 0.054 deg. (full step = 1.8 deg or 200 steps / rev).

bob53 said:
Tell him to furnish a gear set that has a 2.54:1 ratio. You don't want an even ratio as that'll cause wear patterns.
Not a bad idea, the first possibile combination is 127/50.
 
ndzied1 said:
Not a bad idea, the first possibile combination is 127/50.
That was my first thought... Mechanically convert mm to inches directly. But remember, you also have that 'multiple of 200 steps / rev.' which must also wind up in the ratio. This might help (or hurt) you... :confused:

I would think there's some combination that would work, but I doubt it's possible with just one set of pulleys. I think you'd need a jackshaft (or two) to wind up with the exact ratio.

Thankfully, 25.4mm is considered an 'exact' inch. A 'US Survey Inch' is actually 25.4000508001016002032004...mm. Try and achieve THAT ratio with a handful of timing belts!... o_O

As Peter said, don't mix metric and inch and the problems disappear. IOW, tell them they MUST buy a ballscrew with an inch pitch, period... :nodi:
🍻

-Eric
 
I think backards

y'all know this is over my head but I was wondering something. The customer wants .0001 precision which equates to approximately .00254mm, why not just do everything in metric? Obtain the pulleys in metric (if possible) to give you the the ratio per step that is needed in metric. Just because he states it in English reference doesnt mean you have to use that reference in your algorithm(s).

I dont have a clue if what needs to be done can be done but personally I would ignore the English reference and work with the metric aspect since a metric component is being used.
 
Not backwards, good idea

That is why I asked about a step being exactly .0001 inch. I was thinking about 2.5 microns per step and doing it all in metric units. 5 mm / (2.5 um/step) = 2000 steps. The 3 to 1 gear reduction means Norm would have to get 2000/3 steps per revolution at the stepper motor which isn't going to happen.
Can you get by with a 2/1 reduction.

You must get rid of all numbers that cause fractions that can be represented with just a couple decimal places.
 
rsdoran said:
Obtain the pulleys in metric (if possible) to give you the the ratio per step that is needed in metric.

I don't see how this would help. The only thing metric about the pullies would be the pitch and belt width. You're still dealing with whole number ratios based on the available tooth-number pullies.

Dunno.gif


It's strange that they need that kind of resolution on a surface grinder, unless it's controlling the downfeed rather than the longitudinal feed. (Which may indeed be the case.)

Paula
 

Similar Topics

G'day Gents, Been given an opportunity to play with Logix designer and a Compactlogix controller to learn a bit more about Ethernet/IP CIP...
Replies
2
Views
1,567
Hello.. I am having touble trying to single step a program so I can monitor the values. I can remember I had the same trouble last time, but for...
Replies
2
Views
1,370
Hello boy's and girls, I was wondering if someone could tell me if this is normal or not. This is the first time that I have managed to single...
Replies
1
Views
1,274
Using a Sequencer like the example from this link http://forums.mrplc.com/index.php?showtopic=6455 It is working great for controlling a Pick...
Replies
7
Views
2,781
Hi to all. Is there a way to single step through a program using RsLogix 500 to run a program step by sep? Thank you for your help.
Replies
3
Views
2,419
Back
Top Bottom