Hello Guest it is March 28, 2024, 03:32:56 PM

Author Topic: fastest way to drill.  (Read 5948 times)

0 Members and 1 Guest are viewing this topic.

fastest way to drill.
« on: March 18, 2013, 03:23:28 PM »
hi all.
I have around 200 points to drill. I need them to drill as fast as possible. but no meter what feed I give it always takes same time. let me explain.
I have 200 points with 0.5 mm gap between them. my z axis retract is set to 0.5 mm. feed set to 2000 mm/m. all motors are set to 2000 mm/m velocity with 1800 mm/s acceleration. but it took around 3 minutes.
then putting mach3 into offline mode, I changed motor velocity to 10000 mm/m with acceleration to 10000 mm/s. and feed also kept 10000 in my drilling gcode. still mach3 completed program in same old 3 minutes. why it didn't completed program much earlier even after increasing feed, velocity, acceleration 5 times.
plz help.

Offline BR549

*
  •  6,965 6,965
    • View Profile
Re: fastest way to drill.
« Reply #1 on: March 18, 2013, 03:36:01 PM »
What does your Gcode look like ???  If the moves are so short that MAch3 cannot get up to speed in that distance then you are limited to the time base of the moves.

IF you can drill at wide open speed then a straight gcode program(no canned cycles) drilling with G00 will be the fastest possible means.

Post a copy of your gcode and I will take a look at it.

One way to look at it is take all the distances of all the moves and divide by the time you will then get the average DISTANCE/time that the machine is doing.

Just a thought, (;-) TP
« Last Edit: March 18, 2013, 03:42:35 PM by BR549 »
Re: fastest way to drill.
« Reply #2 on: March 19, 2013, 05:57:40 AM »
Quote
What does your Gcode look like ???  

attached gcode.

Quote
If the moves are so short that MAch3 cannot get up to speed in that distance then you are limited to the time base of the moves.

you can say moves are short. all point are in straight linear way with gap of 0.5 mm between them. also z axis just retracts 0.5 mm. but i wonder how does it affects mach3. AFAIK speed in such small move can be affected by acceleration rate of motors. as mach can immediately deliver pulse train to highest frequency. or like motors, mach too is bound by rate of acceleration in delivering pulses. if this is the case then what is the rate of pulse mach can deliver.

Quote
IF you can drill at wide open speed then a straight gcode program(no canned cycles) drilling with G00 will be the fastest possible means.
thats not possible. but one thing i noticed. when i keep motion mode to exact stop, it performs almost double the fast compare to constant velocity. now thats strange, as i read many places that CV is faster then ES. below giving you results of different settings i checked,

Velocity    Acceleration   Time in CV mode     Time in ES mode
1000        800               3.43 Min                2.15 Min
3000        2400             3.00 Min                1.31 Min
6000        4800             3.00 Min                1.31 Min
9000        7200             3.00 Min                1.31 Min

notice that after increasing feed above 3000, there is no increase in time for even single second.

Quote
Post a copy of your gcode and I will take a look at it.
Thanks. i really appreciate your help.

Quote
One way to look at it is take all the distances of all the moves and divide by the time you will then get the average DISTANCE/time that the machine is doing.
for ease of calculation, i am counting from after drill has been done to point 1 and z is at 0. so movement will be as follows.
1) retract z                         0.5 mm +
2)move to next point           0.5 mm +
3)drill to that point               0.5 mm =
                                         ---------------
    total movement               1.5 mm *
     number of points             400
                                         ---------------
    total linear movement        600 mm
Considering if it had to move 600 mm linearly, entire file should have finished less then 1 minute, even if we take slowest feed that is 1000mm. but its not linear move, but then also is it possible to do that much fast???

« Last Edit: March 19, 2013, 06:00:26 AM by janak »
Re: fastest way to drill.
« Reply #3 on: March 19, 2013, 12:12:29 PM »
Hi all,
nice challenge  ;D

Everything you do in life takes some time also in the g code. Also the speed of light is not infinite.

To keep the processing time as short as possible leave everything away what is not necessary.

In your case it may look something like this


G64
G90 G0 X0 Y0 Z0.5 F2000
g91
m98p1l200
G90
M30
%
O1
g1z-0.5
z0.5
x0.5
m99
%

The system needs at least 13 seconds for the 200 points ( someone less ? ) ;D

The settings for this are at least:
Steps per MM = 400
Velocity = 1800
Acceleration = 1500

Regards
Alex


Offline BR549

*
  •  6,965 6,965
    • View Profile
Re: fastest way to drill.
« Reply #4 on: March 19, 2013, 05:47:23 PM »
You have to remember that every move has to accel to speed then deaccel to a stop. IF the move is so short that mach 3 cannot get to full speed then you are limited to the move's time base due to accell/deaccel.

Yes in this point Exact stop may be faster. YOUR moves are ONLY .5mm long but as you see you still run up against the time base of the short move. the machine will NEVER get any faster than it can start/stop on a short move.

Looking at the drilling feed speeds you are trying to "achieve" why can you NOT use a straight GO feed to do the drilling?? WHAT are you drilling?

At some point you HAVE to except the physical limitations of the machine and control software.

(;-) TP
Re: fastest way to drill.
« Reply #5 on: March 21, 2013, 04:54:49 AM »

The system needs at least 13 seconds for the 200 points ( someone less ? ) ;D

The settings for this are at least:
Steps per MM = 400
Velocity = 1800
Acceleration = 1500

Regards
Alex

I wish 200 points can be done in 13 sec. but that's not happening. I tried removing unnecessary lines from my gcode. you can also try, i've uploaded gcode file in 3rd post.
Re: fastest way to drill.
« Reply #6 on: March 21, 2013, 05:20:46 AM »
You have to remember that every move has to accel to speed then deaccel to a stop. IF the move is so short that mach 3 cannot get to full speed then you are limited to the move's time base due to accell/deaccel.

Yes in this point Exact stop may be faster. YOUR moves are ONLY .5mm long but as you see you still run up against the time base of the short move. the machine will NEVER get any faster than it can start/stop on a short move.

Yes, i can understand in short moves accell/deccel plays important part. but look at tests i had conducted. anything above 3000mm feed and 2400 accel, does not finish job early even for 1 sec. I had tried this offline, that means it is limitation in mach3 or my pc, that is unable to accel/de-accel pulses at faster rate. but i've not read anywhere about what is the rate of accel, mach3 can deliver. do you have any idea?

Quote
Looking at the drilling feed speeds you are trying to "achieve" why can you NOT use a straight GO feed to do the drilling?? WHAT are you drilling?


basically i am not drilling, i am trying to decorate bangles using diamond faceting tools. toolpath action is same as drilling. this kind of drilling gives shiny circles on surface of bangles. if i use straight feed then it will leave arcs, rather then circles.

Offline BR549

*
  •  6,965 6,965
    • View Profile
Re: fastest way to drill.
« Reply #7 on: March 21, 2013, 12:10:19 PM »
Here is an example of straight feed code.  How long it takes will depend on how FAST your machine can move. This is the fastest possible code example for your process. Here in SIMULATION it took 30 sec.

G21
G0 G17 G40 G49 G80 G90
T20 M6
G0 G90 G54 X0. Y0. A0.
S1145 M3
G43 H20 Z.5
G91
M98 P1 L796
G90
M30
%
o001
G0 Z.5
Z-.5
X.5
M99
%
Re: fastest way to drill.
« Reply #8 on: March 21, 2013, 12:41:49 PM »
The same example without G91 ( some do not like it )

G21 #1=0
G0 G17 G40 G49 G80 G90
T20 M6
G0 G90 G54 X0. Y0. A0.
S1145 M3
G43 H20 Z.5
G64
G0 X0 Y0 Z0.5 F2000
m98p1l200
M30
%
O1
#1=[#1+0.5]
g1z0
z0.5
x#1
m99
%


 :)
Re: fastest way to drill.
« Reply #9 on: March 22, 2013, 09:35:50 AM »
Thanks BR549, Thanks cncalex,
You guys solved my half problem. Tool-paths did finished in less then 30 seconds. (Though still trying to understand how each line works. ;-) )

But for my other half of the problem, what if drill points are non linear. Just to give you an idea, attaching a picture. You can see 2 lines of circles making shape of rotated "S". I've also attached gcode of single line "S".
So this drilling pattern is not limited to "S", we can have any kind of shape for designing, and not necessarily it repeats.
Any clue how to do that as fast as we did linear one?