R vs. I and J

[ScottyB]

This may be a simple question, but...

I've been spoiled by writing programs for my Fanuc controller which uses x,y,R for 7 years. Now I have to deal with a Multicam that uses I,J,K. After digging into my notes from class in 1996 I re-edimicated myself as to how to use this. Is I,J,K just old technology or is there an advantage to it that I don't see.

I ask because the Multicam is only 4 years old. I didn't realize how spoiled I was with R.

 

In other words, I,J,K sucks...why is it used.

 

ScottyB programmin' old school

[Rick46]

In my opinion i,j,k is better output that xy r output.. we have some machines that go crazy using the xy r out put if its to tight of a arc in the path it needs to follow.. when we changed to ijk output in our posts no more problems... just my opinion though that the ijk vectors are better....

[mayday]

my guess,

in a G18/G19 plane all I have seen is J/K or I/K

dont know what setting the parm to read an R will do in 3d. our fanuc's will do it either way in 2d. I leave em at I/J/K

[Rekd™]

The only time I use R is in Drilling Cycles. IJK are more flexable, more complicated and I've had several headaches as a result of using R's.

 

[dig]

Sounds to me like you might bet better off with a simpleton software like FreeMill if that's your approach to usage; (It's easy, it must work fine!)

[/dig]

 

j/k

 

'Rekd

[Surface]

Scotty,

The I, J method is prefered because of numeric rounding. To prove this, create a half circle of a known radius. Then try to create the matching half-arc with a radius .0001 larger using the create/arc/endpoints method.

Notice the center is way different than the origional. Now try with an arc .0001 smaller than the origional. Can't be done. Machine error.

Thats why.

[Rekd™]

quote:

---

Machine error.

---

No, Programmer error..

 

'Rekd teh why program un-machinable geometry?

[Philcott]

ScottyB,

 

I,J,K usually negates radius errors because they designate the distance from the start of an arc to the arc's center point. They are signed (+/-) which tells the control exactly which way to go. With "R" the next line of code is what tells the machine which way to go (CW or CCW).

 

Phil

[Harryman]

I and K drove me back to the classroom. Hand writing Gcode for a Fanuc 5T with no CAD/CAM support and no tool radius comp on the control. I can tell you that mathing out a tool path for an angled line tangent to two arcs including tool radius comp is a lot tougher than our speed and feed problem of a couple weeks ago.

 

The vectors locate the center of the arc, and the machine does not have to calculate the center. The older ones flat could not make the calculation.

[Guest CNC Apps Guy 1]

IMHO - I,J,K is better where I wirk because I like to do a lot of Helical Interpolation and with I,J,K I can sweep a 360 whereas with R you're limited to less than that.

 

JM2C

[crazy^millman]

No commnet just say Rekd you cover all I got to say.

 

Crazy Millman

[BK]

When using I,J,K values on a machine tool, the control is checking the math on the fly. You're basically telling it that from where I am right now, do an arc and oh yeah, by the way, here's where the arc center is. Whatever your circ interp tolerance is set at in the machine if your math is off (usually .0004") the machine will alarm out.

 

There is no problem with using R's as long as your numbers are correct (i.e., from MasterCam).

 

Here's the good part. If you use the R word and your math is off, the machine will just "fake" the move. Bascally it knows you want to do an arc and you've given it the start and end points but your math is well, not so good. On most machines the math can be off by up to two times the programmed radius value and it will still do the move!

 

Makes for some really crappy looking parts.

 

HTH

 

Bob

[LNDCAD]

I,J, K is often more complicated to program manually, since the center of the arc often be calculated from the print. The radius is usually given on the print. However, when programming with R values, keep in mind that the machine has two solutions to a given radius between a fixed start and end point. One solution is less than 180 degrees, and one is greater than 180 degrees. To prove it, draw two points at 3 o'clock, and 6o'clock. You can draw a CCW arc with the center at the center of the clock-face which is 90 degrees. You can also draw a CCW arc with the same end points, and radius, where the center is directly below the 3 o'clock point, and to the right of the six o'clock point. This arc will be 270 deg. The controller will usually pick the smaller arc, unless you program a negative R value. Then most controllers will pick the larger arc. Also, a single command full (360 deg) circle can not be made with the R value format. By programming the actual center of the arc (I,J,K) all of these problems go away. You are correct that the R value is more forgiving. If you miss the center location by a few thousandths of an inch, the controller will pick it up, and give an error. However, if you miss the radius by a few thousandths, if possible, the machine will just move the radius until it touches both points, and machine it (if the radius was originally tangent, it is not tangent anymore). Do you really want the machine to go ahead if you make a mistake? Finally, I, J, and K define the location of the center of the arc in the X, y, and Z axis respectively. Usually they are measured relative to the start of the arc, although some machines will allow them to be located from part zero if the program is using absolute coordinate (G90). Sorry this got so long winded. Hope it helps.

[Rekd™]

quote:

---

Here's the good part. If you use the R word and your math is off, the machine will just "fake" the move. Bascally it knows you want to do an arc and you've given it the start and end points but your math is well, not so good.

---

This is one reason why I quit using Rs. I've seen machines that 'think' they know what I want, so they 'fake' it..

 

Also, very nicely stated, LNDCAD. It's hard some times to explain to some folks how these things work, (arcs, crc etc) Good job.

 

'Rekd

[CNCme]

Good Day,

 

 

I am very surprised no One gave the theory

of circular Int. ( "IJK" VS. "R" word )

 

I taught MDI or hand written nc prog.

when Mcam was an infant, for over 10 Yrs.

 

( Keep in mind when it comes to Mcam, the posted

Progs., Whether in "IJK" word or "R" word, both

are identical.

 

IJK word programming:

 

In this type of programming,the arc center

must be defined almost exactly ( within .0001 -.0002 ) or the machine will alarm out.

 

R word programming:

 

In this type of programming, the arc center

is not defined, ( it is what I call the lazy

mans circle int.) The arc center is a result

of a combination of the start, end point and

The radius. You can be off on your start and

end point and radius and still connect.

 

________________________________________________

Here's the good part. If you use the R word and your math is off, the machine will just "fake" the move.

________________________________________________

 

the only problem is it will connect, but it may

not be pretty (possible out of tolerance)

Most likely non-tangent arcs to lines.

 

( Keep in mind when it comes to Mcam, the posted

Progs., Whether in "IJK" word or "R" word, both

are identical.

 

 

HTH

Tony G

Almost Employed Senior Programmer

N.E Massachusetts - Southern New Hampshire

_________________________________________

End mills and tooling are like The "AMMO"

And coolant and chips are like the enemy

Under your boots as you advance in the

Manufacturing Battle

--------------------------------------------------

[crazy^millman]

So Tony G still no Job huh sorry to hear that. Everytime I see you post I keep hoping to see that sitution change and hope it does.

 

Crazy Millman

[Craig Madsen from CM3D]

I have to add my 2 cents here.

 

1. The Error indicated above with the arc created by endpoints does illustrate the error.

 

2. The error will most likely not be caused by Mastercam, the programmer, or other sources with any kind of good reputation.

 

3. It WILL, generally, be caused when the radius you are cutting using "R" designation is within a 2 or 3 degrees of 180 degrees. Your math may be perfect, but the control may be operating with limited precision, or may be converting between inch values and metric output. There are many instances where the conversion is between "Even" numbers and the control will introduce a error that is multiplied during interpolation.

 

4. I've seen the results of this many times, with many different brands of controls.

 

5. It may not always happen. Just like I can drive 80mph and not get ticketed. But, it will happen, and I will deserve the result.

 

Craig Madsen

[Guest CNC Apps Guy 1]

quote:

---

5. It may not always happen. Just like I can drive 80mph and not get ticketed. But, it will happen, and I will deserve the result.

---

Craig, that was an AWESOME illustration, especially the

quote:

---

...I will deserve the result...

---

part.

[CNCme]

Crazy...Thanks for the reply...Im keepin'

BUSY

 

Craig,

________________________________________________

1. The Error indicated above with the arc created by endpoints does illustrate the error.

________________________________________________

 

Reply:

1. Im not to sure which error you are reffering

 

 

________________________________________________

 

3. It WILL, generally, be caused when the radius you are cutting using "R" designation is within a 2 or 3 degrees of 180 degrees.

_________________________________________________

 

Reply:

 

3. When will this condition occur...177 deg or

183 Deg.? + what type of part would use this

type of arc (If the machine had a problem with

180 +/- 2 or 3 deg., I would use

"breakarcs :1 quadrants "

 

The condition I reffer to in my last post:

The " R " is mostly for corner blending

not circle cutting ....

NOT( full circle or 180 deg x 2).

 

AND

 

(ALSO only on SLOPPY hand written progs...

 

NOT Mastercam Code!!!!!!!!!!!!!!!!!!!!!!!!!! )

 

Tony G

Almost Employed Senior Programmer

N.E Massachusetts - Southern New Hampshire

_________________________________________

End mills and tooling are like The "AMMO"

And coolant and chips are like the enemy

Under your boots as you advance in the

Manufacturing Battle

--------------------------------------------------

[ScottyB]

WOW,

I feel like I got sent back to grade school. I've been using R's for years without any problems. But I don't disagree with your points. I will start working with IJK for the more critical parts.

This subject came to mind when I was trying to trouble shoot a program that someone else wrote. The program looked right, but the machine was way off on the radius corners. I knew the part needed a 4" radius, so when looking at the program I was hoping to find the start point, end point and R4. That would have made it easier to see a red flag. I still don't know what the machine saw, but I tossed that program anyway.

Thanks for the input.

ScottyB.

[heeler]

With our testing of machines and doing Ball Bar test, the only place that we noticed a problem was if we programed an arc with a diameter value with an odd number for the fourth decimal digit. You can also verify this in MCAM. Create a 4.1255 Dia. arc, then create a 2 pt. arc from endpoint to midpoint, you cannot create one with a 2.0627 (1/2 the Dia. rounded down). But you can with a 2.0628(1/2 the Dia. rounded up). Look at the results. The Fanuc control by default will run the smaller arc. It all about math. If you will either break arcs into quadrants, or make sure that your Dia. always have the fourth, or fifth decimal be an even number you should have no problems.

 

HTH

Glenn

[mmonica]

ScottyB, keep in mind, I squared + J squared = square root of R

[mmonica]

opps, I squared + J squared = R squared

[Guest]

I never realized IJK vs. R was such a hot issue. I have always used IJK. I don't know why other then that was the way I learned it. I have always known about the R way but chose the IJK way. Sounds kind of biblical, Don’t use the R road take the IJK road.