[deadlykitten]

Hello, I have run my 1st circle arches on the mill ... i managed to control the short arc and the long arc between 2 points

While running, i stopped (reduced potentiometers to 0) the machine at the blue points, and i wrote the coordinates.

It seems that all points, relative to the theoretical circle, are aproximated with ≤0.002mm

how can i decrease precision, so aproximation to be ≤0.03mm ?

i don't need the machine to move all the time at it's heighest precision, for example, for parts with ±0.1 tolerance.


on the other side of the mirror, does this matter ? kindly!

[Algirdas]

i stopped (reduced potentiometers to 0) the machine at the blue points, and i wrote the coordinates.
Can you explain, why you do that?

[Christian_S]

Why would you want to decrease the precision, unless it is as a result of increasing speed or using more economical tooling?

[deadlykitten]

Can you explain, why you do that?

hy, i started learning some IJ codes for millig, so to better handle arches, and i tested the code by writing coordinates, to see how code with/without IJ would behave ... and i observed that ≤0.002mm precision

[deadlykitten]

Why would you want to decrease the precision, unless it is as a result of increasing speed or using more economical tooling?

hy, i don't know ... i asked ( on the other side of the mirror, does this matter ?) so is there any reason for that ?

from your answer, i guess i can:
...increase speed, and i wonder how much do you know how much ?
...use tools more economical... i don't understand this ...

there are these 2 reasons from you, but i just wonder if there is something else ... machine is too precise for my actual parts, and i just wonder if is it ok to go on like that, or if there is a way to decrease it's precision, and what comes with it ? ... for example, just saying, HD mode comes with wear, while LOW_D hase some benefits ...

so to put HD mode into stand-by, and just activate it when necessary ... or HD all the time, because it's way too normal ?

by normal i don't reffer to the way a machine is used, but at some behind_the_scene stuff

[Algirdas]

I think I got the idea. I offer you to perform a test on an unneeded work piece. You write one part program with bunch of short G01, making arc with desired inaccuracy. The next part program the same dimension arc using G02 or G03 - unneeded accuracy. You compare the machining time of each and let us know the result.

[deadlykitten]

discret interpolation of complex shapes runs well for 0.1 .. 0.3 segment .... when segments get smaller, time increases because of acc/decc times, machine stopping at each position

this can be solved only on mills, with nurbs

[Algirdas]

you can choose the biger radius in order to increase length of segments.

[deadlykitten]

hello, if there is a bigger hole, not so tolerated, you can go with bigger segments, but this aproximation method does not run as smooth as a G2/3; by smooth i reffer to motors diagram accuracy, not to toolpath

@ for segments, machine stops at each end point and starts at each start point
result = [ rough surface ] + [ a bit stress on the machine ]

@ G2/3 is built in, and machine goes smooth, at 0.001precision
result = [ nice surface ] + [ no stress on the machine ] + [ motors at default accuracy=DA ]

... i was just wondering if i can setup G2/3, at 0.03precision
result = [ rough surface ] + [ no stress on the machine ] + [ motors at 30% DA ]

... if this can be solved, than same question for G01

[Algirdas]

I recall back my comment. I don't get the point.

[deadlykitten]

I recall back my comment. I don't get the point.

you are a race driver, you know best race line, and you can stay on it like nobody and is easy for you

but you are 1st, and is boring, so let's wait for the guy from 2nd place ... you can not just brake, and wait, so you start getting away from that race line ... more like cruising instead of racing

when you see the opponent in your mirror, you push the pedal, and go not as best as you can, because this will put distance again, and you will be lonely ... you go a little in between, you are still1st, you have some fun, and you don't choke

stuff should not be perfect, is enough if it works

if a machine can deliver A class quality, than you can product on it also lower class products B, C, D , by using the machine on a decreased quality setup

......humans do great things when they are not tired

......machines perfom better with new parts, comparing to weared ones ... they actually performs better after .... google can't translate it ... how do you call when you have a new car and must drive in low rpm for about 5000 km ? so parts to get a functional wear ?

......so why not decreasing encoder precision ? and still get the work done ? something like increasing the " following error " ? hmm ... perhaps DROOP ? i must deep into it

[Superman]

Hello, I have run my 1st circle arches on the mill ... i managed to control the short arc and the long arc between 2 points

While running, i stopped (reduced potentiometers to 0) the machine at the blue points, and i wrote the coordinates.

It seems that all points, relative to the theoretical circle, are aproximated with ≤0.002mm

how can i decrease precision, so aproximation to be ≤0.03mm ?

i don't need the machine to move all the time at it's heighest precision, for example, for parts with ±0.1 tolerance.


on the other side of the mirror, does this matter ? kindly!

Decrease precision ??
It seems you want to un-invent the wheel, turn it back to being a square
- I believe the original CNC's moved using point to point... & you want to go back in time ?

If I tell the machine to goto a defined point....I would want it to be on it ....not within 0.03mm
- it's a machine, it has to go where it is told. If you don't need it that accurate, then you have to program that inaccuracy


- when programming an arc, you have a start point, ending point and an arc centre ( IJK or an R component )
The machine software controls the motion of the motors to maintain the arc sweep
- parameters also control the "In position target" signal to allow motion for the following block to begin.... this helps to generate continuous motion


What is it you are really trying to achieve ?? ...... saving time ??
Do you have "Tolerance control" on your machine ???
You could turn it OFF ( or use a coarser setting ) for roughing, & ON for finishing situations

[deadlykitten]

hello

- while movement occurs, machine always pings / verifies this difference " calculation value - actual value "
- this difference, being calculated at specific time intervals, is called the " following error "
- and this " following error " should be ≤ " allowable limit "
- where is located this " allowable limit " ?
- all this info is from attached image

- about this " allowable limit " i believe that is not what i am looking for, but is somehow related to it
- this "parameter P" what i am looking for, should target movement generating accuracy ( the "drive mechanism" ), while " allowable limit " targets the "control mechanism"
- when position is wrong, is it because of the "drive mechanism", and is sensed by the "control mechanism"

- if possible, you can set :
..........."drive mechanism" accuracy at "parameter P"=T [ mm ], and
..........."control mechanism" accuracy at "allowable limit"=T±5% , or 0.85T < "allowable limit" < 0.90T

Do you have "Tolerance control" on your machine ? You could turn it OFF ( or use a coarser setting ) for roughing, & ON for finishing situations

- where is this " Tolerance control " ?

[Superman]

You put up a picture of a DIFF OVER alarm.....usually this appears when your machine has problems
- not sure of the parameter location

Have you looked up the "In-position" tolerance, & what it means ?

"Tolerance Control" I had on a 7000 control, It was a lower option than "Super-nurbs"
- where each endpoint ( line of code ) was under a little control using Accel / Decel factors

Info for the 7000 control was in the "Special functions" manuals:

[deadlykitten]

hy superman ! god, i hate manuals

[Superman]

hy superman ! god, i hate manuals

Ahh, that explains ALL the posts you are putting up.........you can't read :stickpoke


I noticed that you ended up contributing ( Permission Group - Super Member )..... only way you could change the "User Title" ....you cheated ??? .....

[RCaffin]

Hi Kitten

i don't need the machine to move all the time at it's heighest precision, for example, for parts with ±0.1 tolerance.

I don't think you understand what a CNC machine does. It makes almost zero difference to the machine whether it goes to within 0.01 mm of its final position or within 0.1 mm. It goes where it is told, end of story.
The only place where 'tolerance' comes into play is when setting the parameters for taking corners under Constant Velocity settings. Even then, the differences are like the difference between a really sharp edge on a knife and a slightly dulled edge.

Cheers
Roger

[deadlykitten]

It makes almost zero difference to the machine whether it goes to within 0.01 mm of its final position or within 0.1 mm

hello :) somewhere inside i do believe that ... i am still curious about what means "almost zero difference"

The only place where 'tolerance' comes into play is when setting the parameters for taking corners under Constant Velocity settings

far as i know, this one ain't droop ?

Ahh, that explains ALL the posts you are putting up.........you can't read :stickpoke

hello :) please, may i joke ? if you read the manuals, than that explains all your posts :)

What's with this fanatical desire to change a title ???

just one word : design ;)

[RCaffin]

i am still curious about what means "almost zero difference"

The ONLY difference is when CV is enabled at corners. That gets technical, and you would need to read the manuals about that. If you hate manuals ...

I am assuming of course that the positioning resolution of the CNC HW and electronics is better than 0.01 mm. That is not difficult.

Cheers
Roger

[OkumaWiz]

I have had success in decreasing cycle time by "changing accuracy" of the machine. for example if you are on a horizontal mill and you rapid from left side to right side, you will see significant delays at the end of every rapid move "back, over, in" while the machine confirms that the DIFF is small enough to proceed with the next block. This can be helped by changing the "in position window".
What I did to make this easy is to create a sub routine that changes the values and then assign it to a G-code macro. I used G119 and G120 and assigned them to the subs OSLOW and OFAST. Switch to G119, cut, switch to G120, rapid. I knocked over 3 minutes out of a long horizontal program that was an hour long.

Here is the code:

$RAPID.LIB%
(SET G119 TO OSLOW AND G120 TO OFAST)
(THIS WILL CHANGE IN POSITION WINDOW AND WILL HELP SPEED UP RAPID MOVES)
OFAST
VINPX=0.0 VINPY=0.0 VINPZ=0.0 (VINPA=0.0)
RTS
OSLOW
VINPX=.0001 VINPY=.0001 VINPZ=.0001 (VINPA=.003)
RTS
%

This was installed as a library file and my post was modified to add in the G codes when switching back and forth between feed and rapid.

I have revealed another "ancient Japanese secret" to you, so please send root beer! ;-)

Best regards,