Advice needed on actual cycle time against Est cycle time

[rmtucker]

Been doing some testing on UCCNC today concerning cycle times.
And i am having a problem finding any correlation between my estimated times coming from Aspire and the real time taken on the machine.
Take this part as an example:-

Est Cycle time from Cad/Cam = 13.55 mins
Actual cycle time on machine = 38.5 mins
I have tried loosening the cv parameters by as much as 3 times the original amounts but it does not help much.

MCE 0.1
MLE 0.1
MLAL 1
MLUL 2

to

MCE 0.3
MLE 0.3
MLAL 1
MLUL 2

The feed rate in the program is set to 3000mm/min but it seems to never get above 1200 to 1300 (Bit difficult to see the spikes in feed rate).
I have my accel set to 250mm min.
I have tried it at 750mm min (even though my machine would not actually run at this accel)but this still only reduces the cycle time to 27 mins
I came across one of batwells video's on youtube which shows the same problem.
He has his feed set to 6000mm/min but most of the time it is cutting below 2000mm/min
https://www.youtube.com/watch?v=9OhaldtZbMQ

Tried to attach the file but it is over 2mb.

[rmtucker]

Just a little further info.
If i slide the feed rate override down, it does not have a constant feed rate until it reaches around 600 mm /min.
How on earth can you work the chip load out for the cutter if it is wandering around so much?

[ger21]

What you should be doing is adjusting the scale factor in V Carve/Aspire, so that the estimated time matches the actual time.
Not trying to get the actual time to match the estimate.

[ger21]

How on earth can you work the chip load out for the cutter if it is wandering around so much?

The short answer, is bigger motors, or more power. Our router at work can get to 20,000mm/min in about 1 second.

Most hobby machines are vastly under-powered, due to cost restraints.

[rmtucker]

What you should be doing is adjusting the scale factor in V Carve/Aspire, so that the estimated time matches the actual time.
Not trying to get the actual time to match the estimate.

Yes i understand that but then the next job is more or less complicated or maybe 2d 'v' carving and that does not correlate any more.
I thought that the cv would give you a constant chip load but it is doing far from that wandering from slow feed to fast feeds all the time.

If i try to work on a .1mm per tooth chip load to stop the cutter burning then the machine is reducing the chip load down to .01 at times how can i cater for this?

[cncdrive]

Well, target velocity is reached via acceleration and decceleration.
With complex 3D toolpaths the movement vector direction is often changing and then the axes of the movement have to accelerate and deccelerate.
The axes cannot do the acceleration and decceleration instantly, because first of you setup a max.acceleration parameter for them in the axis setup and second with infinate acceleration setting your motors would loose steps as they cannot accelerate and deccelerate the axis mass in zero time.
And so the acceleration up and decceleration down phase requires time to reach the target speed and if it is changing often then it is ofcourse possible that the machine cannot reach the target velocity at all.

You can make things better with setting the acceleration parameter of your axes higher if your motors and machine can handle it.
You can also adjust the constant velocity tolerance parameters on the general settings page to higher values to let the CV optimiser do a more optimised path to reach the target feedrate, however the higher the tolerance is the more details you will loose from your 3D toolpath, because with those settings you telling the software that it can use higher path deviation in order to reach the target speed.

In summary, the constraint is your settings and the constraint of those is the physics. (mass and acceleration)

[Battwell]

a customer of mine did a huge lithoplane in corian around 8 years ago. estimated time was 28 hours- it took 96!
was done as 1 huge panel 10ft long 4 ft wide
the machine was run continuously for the whole time with only 1 stop to replace the cutter half way through.
the job was put on the machine just after i had converted it to mach 3/galil combination as it was originally analog.
customer said i hadnt done the settings right so i had to explain what cnc drive said above .
anyway- take estimates as wont be correct estimates!

[cncdrive]

I don't think that the Vectric software even knowing the acceleration parameter of the machine, so it's estimation cannot even be correct.
It just knows the feedrate and counting as if the axes could immadiately reach that target speed which is ofcourse not possible in real life.

[ger21]

Yes i understand that but then the next job is more or less complicated or maybe 2d 'v' carving and that does not correlate any more.

That's just a limitation of Vectric's formula that they use for estimating time.

I thought that the cv would give you a constant chip load but it is doing far from that wandering from slow feed to fast feeds all the time.

To get close to "actual" Constant Velocity, you need to have very high acceleration. Otherwise, the machine has to slow down to change direction, or make a more gradual change, causing a lot of lost detail/
As CNC Drive said, you can't break the laws of physics.

If i try to work on a .1mm per tooth chip load to stop the cutter burning then the machine is reducing the chip load down to .01 at times how can i cater for this?

If your bits are burning, you need to lower your rpm, probably by quite a bit. Most people run way too high rpm's than they need to.

Most hobby level machines are not rigid or powerful enough to get anywhere near recommended chip loads, so chip load is a somewhat meaningless target on hobby machines. And with tapered or V bits, the chip load varies along the length of the bit, with the tip of a V bit having zero chip load.

On our work machine, we cut at very high speeds and very high chip loads to maximize tool life.
On my machine at home, I never look at chip load at all. I just try to keep the rpm as low as possible, and feed rate as high as possible, to get the best possible finish. Finish quality is far more important than chip load.

[rmtucker]

So is it not possible to add a cycle time estimation which uses the Accel and max velocity in the configuration?.
I have tried Dan911 plugin but it also returns 13mins instead of 39mins.
This make costing a component impossible unless you run it through Uccnc in real time.
Which is a bit of a problem on jobs with 4hour cycle times.