[angelw]

hello again, in both cases, the result may be affected by an error :
... r syntax : center may be shifted
... ijk syntax : r may be shifted
* even so, effects happens at a scale that is too low, to matter

considering that code is generated ok, and the controller can run both ijk and r, did someone actually hit into a case where 'ijk' was better then 'r', or viceversa ? and if i may, what was exactly the issue ? if you wish to share the code, i will test it for you / kindly

A 360deg circular move can't be programmed in one Block using R Format. In every case where I've seen a full circle programmed using R Format, the user has programmed two Command Blocks, each dealing with 180degs of the full circle; logical on the surface, but not so. Being able to programming the circle in one Block is the first plus for IJK Format in my book.

If there is an error, caused by rounding, minuscule as it may be, when the Control calculates the centre of the arc and shifts the arc centre, there will be a doubling of the error as the shift of the centres will be in opposite directions in each of the 180deg arcs, but it seems that theoretical accuracy matters little to you.

[deadlykitten]

A 360deg circular move can't be programmed in one Block using R Format.

hi / yes, that is true, but a full circle is a particular case, and also R<0 may deliver arches > 180*

my book

aaa, do you have a programing book ? i don't know, maybe you do ... if you would have one, i would read it

theoretical accuracy matters little to you

i use code as accurate as it could be, folowing error is never 0, there is backlash, and most dimensions have a tolerance >1um

even when folowing error is 0, cnc motion is not a "real" circle, and also encoders motion <> turret ( or spindle ) motion, since there is a precision lost around the mechanical chain

It is not a very accurate statement. The arc starts with the start radius. If it cannot reach the target point due to some inaccuracy, it makes a small linear move in the end to reach the end point, provided the radius error is within the specified limit.

hi sinha, i know my "accuracies" ...

          if swd[wo] then t:='G02'
                     else t:='G03';
          if full_circle 
           then 

about that "small linear move" :
... if next block contains an axis move instruction, then the machine won't reach the "end point", but will get close to it until distance < in-position; and there is a big chance to have the in-position > g02 code precision
... if next block does not contain an axis move, or it contains whatever instruction that is telling the controller to reach the end point, then, if there is some inacuracy, the machine will move to the end point, not among a linear segment, but among a "non-interpolated segment", that is executed on such a short distance, that the servos won't be able to reach the programmed feed, so servos won't be in sync, but each one will reach the final point on it's own; is possible that only the servo will rotate, but the turret will be allready stationary; kind of a succumbed motion, when the cnc can't reach the programmed feed / kindly

[angelw]

hi / yes, that is true, but a full circle is a particular case, and also R<0 may deliver arches > 180*

But less than 360deg. If one were to use the R Format for a full 360deg, a more accurate overall result would be achieved by programming, say, a 359.5deg arc, and 0.5deg arc. However, is there was an error caused by using rounded numbers in the calculation, a shift of the arc centre would still result.

aaa, do you have a programing book ? i don't know, maybe you do ... if you would have one, i would read it

Typical halfwit DK comment.


What Sinha stated is quite correct. You clearly stated in the Post Sinha referred to that both IJK and R Format are shifted. This is not the case with IJK Format as the arc centre is provided by the part program and is not calculated by the control for that arc specified from the Start Point to the Arc Centre. If the End Point specified results in a different Radius to the specified Arc Centre, to that of the Start Point, and the difference is less than a tolerance set in parameter, the action is as Sinha stated,

There is a clear pattern of you making erroneous statements and when pointed out, you come back with waffle that simply highlights your lack of knowledge and makes you appear a bigger buffoon.

[deadlykitten]

You clearly stated in the Post Sinha referred to that both IJK and R Format are shifted.

actually i did not said that both formats are shifed, but that they may be shifted, thus there is a probability

and the difference is less than a tolerance set in parameter, the action is as Sinha stated

again, here is a probability : it may happen, or not; and when it happens, the movement is not an interpolated segment, because programmed feed may not be reached for that short segment

There is a clear pattern of you making erroneous statements and when pointed out, you come back with waffle that simply highlights your lack of knowledge and makes you appear a bigger buffoon.

you are also so full of wisdom ...

ps: 2252-01 ALARM-B DATA WORD circle calculation A side 1