Hi everyone,

So last year we bought a CNC router with a 3x2m table from a local manufacturer, when we ordered the machine we specified our requirements to the manufacturer in detail and he told us that he would make a top of the line machine for our application, which is signage. One of the requirements was that the machine be stable because we route a lot of thick acrylic (4cm) so that there would be no vibrations and that the cut would be as good as possible so that we don't a lot of manual labour with the finish of the acrylic. We were told to achieve that they would use top of the line components that are totally overkill for our application but just to make sure there would be no skimping on the parts. They used 25mm ballscrews (if I remember correctly) and the servos used are made by Schneider Electric with: 1.1kw of power,  IO 5.55 arms, MO - 3.4 Nm nN - 4000 rpm, lmax 17.84 Arms, nmax 8000 rpm

All in all it's a pretty nice machine, works well enough but there's a rub. For some reason the machine can only cut "well" at 2-2.5m/min. If you increase the speed any more than that you will get extremely rounded edges where they should be a straight edge (excluding the radius of the bit of course).

For example when routing the letter K at 2m/min with a 4mm bit you would get a 4mm radius but if you were to route at 5m/min with the same bit the radius is much larger.

Now here's my main question, as I don't understand as much about CNC mechanics as I probably should but why does this happen ?

I've never felt that the machine doesn't have enough power in the servos, I mean it has no problem jogging at 10-15m/min around the table when it's not cutting and these are all lightweight materials like 3mm acrylic and foamex so it's certainly not a issue with the spindle (3kw Teknomotor air cooled spindle with 24k RPM, not that it matters for this specific issue.)

I also have a theory as to why this happens, my guess is that the router can't change direction fast enough or rather that it can't stop abruptly and then start again to make the "sharp" edge and that it sort of just does a continued motion (if that makes sense) and that's why it leaves the very curved edge.

The reason I'm asking this is that it's pretty frustrating that the machine can't go faster accurately, I was expecting that a new machine with strong servos and good components would be able to do 10m/min. It's not a tragedy, we do good work with the machine but I'm trying to figure out if anything can be done to make it go faster and remain accurate.


Thanks in advance for any tips!

You need to set the acceleration values higher to account for the higher speeds . The faster you go the higher the acceleration requirements are to help prevent corner rounding.  For what you are doing I would keep raising them until I started getting position errors from the servos.

A machine that has LOW acceleration values HAS to round corners it has no choices to OTHER than to run exact stop mode and THAT slows down the cutting down  sometimes dramatically as it has to start and stop at each segment of code.

I think the real problem was your machine designers were not really machine designers just parts assemblers.

(;-) TP

You need to set the acceleration values higher to account for the higher speeds . The faster you go the higher the acceleration requirements are to help prevent corner rounding.  For what you are doing I would keep raising them until I started getting position errors from the servos.

A machine that has LOW acceleration values HAS to round corners it has no choices to OTHER than to run exact stop mode and THAT slows down the cutting down  sometimes dramatically as it has to start and stop at each segment of code.

I think the real problem was your machine designers were not really machine designers just parts assemblers.

(;-) TP

Thanks for the reply, one question though. I did have that option in mind but I had one worry, when I'm running the machine at high speeds and if it somehow does come to an abrupt stop, for example if I do a stop command instead of a pause, you can hear a fairly loud impact like sound when the machine comes to a sudden stop. Will this happen during acceleration/deceleration and if it does could it result in damage to the machine like the ball screws, rails and such ?

Thanks again!

If your machine is really solidly built, then I don't think that 1.1Kw servos are overkill at all.

As Terry said. it's all about the acceleration.

Just hope that they built it as solidly as you think, because when you increase acceleration, the forces the machine will see will get a lot larger.

If your machine is really solidly built, then I don't think that 1.1Kw servos are overkill at all.

As Terry said. it's all about the acceleration.

Just hope that they built it as solidly as you think, because when you increase acceleration, the forces the machine will see will get a lot larger.

Eh, how would someone with limited knowledge go about testing that without doing damage ?

You shouldn't be able to do any damage. What I was referring to, was that if the machine is not rigid enough, you may start to see shaking and vibrations as you increase acceleration.

I did have that option in mind but I had one worry, when I'm running the machine at high speeds and if it somehow does come to an abrupt stop, for example if I do a stop command instead of a pause, you can hear a fairly loud impact like sound when the machine comes to a sudden stop. Will this happen during acceleration/deceleration................

It makes a loud noise because you're telling your machine to try and defy the laws of physics and stop instantly. You should never use the "stop" button while the machine is moving, except for emergencies (where you should really be using the e-stop).
It shouldn't really make any noise during accelerating or decelerating, unless you have faulty components, or a flexible machine.

Alright so I just tried cranking up the acceleration 20 fold and this is the result - https://www.youtube.com/watch?v=Ox4c9hCxApU

The whole machine shakes pretty violently when it comes to a stop, you probably can't see it in the video but you can hear it very well.

As one might guess the shaking is more intense on the longer axis, Y in my case (3m long) than on the X (2m) but you can still feel the vibration with your hand on the table when you move the X as well which probably isn't a good sign.

The velocity of X and Y is 12000mm and the original acceleration for both was 250mm, in the video it's 4000. At 1000-2000mm it shakes a bit so that you can hear it to some degree and feel it but it's not bad, at 250mm obviously it accelerates and decelerates very slowly and there is no shaking.

To me the logical reasoning is that the machine is shaking because it's too light, but I'm probably missing something else as well.

Any advice on what my next step should be, I'm guessing that any vibration at all is a deal breaker, especially when the machine has to perform a lot of turns and "twists" it would probably be shaking quite a bit (still have to test this one).

If any of ya folks have experience with this issue I'd appreciate some advice, I'm gonna do a few short tests like routing small stuff at high speeds and at the normal accelerations to compare the radiuses and then return it to the original accelerations until I have more info.

Part of the issue is that the machine is not rigid enough, but also the fact that Mach3 uses a linear accel and decel, which can get rather violent at higher rates.

There's an old experimental version of Mach3 with an S curve trajectory planner, called Tempest. Unfortunately, it's very limited, with no feedhold, and it may only work with the parallel port. It will run your machine very smooth, though.

Your options are probably to set your accel as fast as you can and still get acceptable results, or try a different controller with a better trajectory planner. You don't necessarily need S-Curve acceleration, but you need a control that will allow you to specify a maximum path deviation, so that it won't round corners with a slower acceleration..

One other option might be to just try a file in Exact stop to see if the time differential is unacceptable considering the improved quality of the finished part.
It may or may not not be a dramatic difference in time .... as TP mentioned earlier.
Only a trial would tell.
Russ

Once again thanks for all the advice, appreciated!

So I just figured something, it doesn't seem to exclusively be an issue of acceleration but one of velocity as well. Even at lower accelerations but higher feed rates (5000mm/min) the corner stops are so abrupt that the bridge (gantry) shakes when it comes to the stop.

My question now is...does this really matter ? I mean does it affect the cut or wil it in the long term do mechanical damage to the machine itself (rails, bearings, ballscrews, servos even) ?


Also just as a point of curiosity so that one knows for the future, how are machines made more rigid so that things like this don't happen ?

The reason I'm asking, among other things, is that I've seen some flatbed cutters equipped with spindles that route at pretty high speeds (10m/min+) but they most certainly don't seem sturdy to the naked eye.

Case in point:

https://www.youtube.com/watch?v=4Z7gDIzx5Rk