[LaBelleNote]

Hi,
Sorry for a maybe dumb question, but I'd like to ask it anyway:
I'm wondering what limits currently the capabilities of my machine, especially in terms of feed rate.

I have a CNC6040Z desktop CNC, with a 2.2kw water cooled spindle, an ER20 colet, ball screws.
It's driven by an undocumented USB board "RTMUSB-2.2V1", as far as I understand it is from Wixhc Technology and must be similar to the MKX-V XHC CNC controller.
The stepper motors are 57BYGH76-3A

The control software is Mach3.

Currently, the (as sold) software configuration in Mach3 says :

- Kernel Speed : 25000Hz : what is it, what kernel ? The Pc's kernel ? what does it limit ?

- The motor tuning for X and Y says Velocity 2000 mm/s and "steps per" 320 , acceleration 200mm/sec2 , G's 0.02
How can I know if it's optimal ?
Does it mean in my gcode any feerate can not exceed 2000 ?

- There is also a configuration for backlash : X : +5 , Y: -5, Z: 1 , Backlash speed 20 : is it necessary with ball screws ? Does it mean it changes the feedrate to 20% each time there is a change in direction on these axes ?

I would like to understand as well what limitation do I have with the USB interface.

What would be the benefit of changing the control card to an Ethernet one ? Which of these parameters would I be able to improve with Ethernet ?

I actually have a laptop with no native Ethernet interface. I could add an external usb to Eth dongle, but would it be any better than what I have currently ?

Sorry for so much questions, but that's something I would really like to clarify.

Thanks for reading me !
C.M

[jaguar36]

Your rapid (non-cutting) feedrate is going to be limited by your stepper motor/ballscrew combination. If you find the torque curve for your stepper motors, you'll see that the max RPM is probably in the 1200-1500 RPM range. That plus the pitch of your ballscrews will determine how fast you can go. That is likely where the 2000mm/s comes from. If you send the controller G-Code telling it to go faster than that, it will just cut it to whatever the max is set too.

Your cutting feedrate (and possibly your rapid too) are going to be limited by how rigid your frame is. The generic 6040Z frames are not anywhere near rigid enough to take advantage of a real 2.2kW spindle.

The controller/interface doesn't matter for a straight feedrate, as its just a single line of gcode. It may be an issue if you are doing complicated 3D surfacing toolpaths though.

[LaBelleNote]

Your rapid (non-cutting) feedrate is going to be limited by your stepper motor/ballscrew combination. I......

Thanks for your answer !
Yes, the stepper motor/screw pitch factor makes sens, I have read more since your answer on limiting factors for stepper motors in relation to inductance, power supply, etc ...
So, even if in the worth case I can get 580rpm from my motors x 6mm, which I think is the pitch for my screws (need to verify), it would be max 3480 mm/s.
Would it be safe if I try to increase the velocity parameter in Mach3 for my motors ?


Any thoughts about this backlash tuning as well in Mach3 , is it useful for ball screws ?

One last question about the communication between PC and control board, you explain that a simple movement takes only a single line of gcode to send, I can understand, but what about the DRO feedback ?
(sorry again, this may seem a bit naïve, by I have a lot of interrogation on what happens really between Mach3 and the board.
I anyone could have some good readings about that, I would be happy.

Thanks a lot !

[routalot]

I would be absolutely astonished if a hobby machine could ever achieve 2000mm/second!Just think about it for a moment,where would you derive the power to move the mass at that rate?Most,although perhaps not all of us,use mm/minute for feedrate. I would guess that the limiting factor with your machine will be the structure flexing under load,the numbers in boxes within the Mach 3 configuration can't be understood by a collection of parts and it will do it's best to comply with the instructions that arrive.Which may cause the magic smoke to escape or deformation of the structure.

[LaBelleNote]

Sorry, my mistake, yes, obviously, it 's 2000mm/mn, not per second, it's just a typo.
Yes, I understand and fully agree that the main drawback of this machine is the lack of rigidity, especially in rotation around the X axis.
I also understand now that I should have chosen a less heavy spindle, and 2.2kw is overkill, but anyway, I'm still happy with the machine and despite all these drawbacks, I have managed to do interesting things so far.

I'm quite sure the next machine will be one I will build by myself, maybe from the components from this one, so it's a reason why I want to get a deeper understanding of all the aspects.

I was primarily focusing on the controller board, the fact it's a no name, usb one.
Would it help to upgrade with respect to increasing the feed rate : I understand: not so much given all the other aspects.
Would it help as well securing the reliability of the thing ?
It's not been often, but a couple of times, I experimented with loosing steps, and I'm not sure what phenomenon caused this and what could I do to secure my operations.
I was supposing it could be because of this usb interface, but what else could cause loosing steps and what could I do to improve my system and secure this ?

Thanks

[hanermo]

OP LaBelle.
It´s a rabbit hole, and far deeper than You probably think.

One of the best controllers is the cslabs csmio-ips, that I use.
For about 6 years.
After trying 5 different ones.
Lathe and vmc.
It is absolutely perfectly reliable, and endless io and good sw support and very good sw features in firmware, like microcontroller probing, limits, mpg.

Because the critical things (probing, limits, estop) run in firmware, they are very very fast and reliable.
They run at 4 MHz, not 25 kHz, and in firmware inside the controller processor, not in sw on the PC.
So the controller will fault instantly, if anything goes wrong.

It has saved my very expensive lathe hardware at least 6 times.
Because the controller faults, estops, before the big very very strong hw breaks (32 mm ballscrews, 750/2.5 kW servos at 500 kHz, 25 mm linear guides, stuff).

But it´s expensive.
Endless cheap controllers exist, some midrange options exist, like centroid acorn.

If You get a top-end controller, likely soon You will find out limitations in your stepper system.
And in your mechanicals.
It´s very floppy, bendy, and not too good at repetition.
So probing -- tool sizes, and limits/homing are variable.

Anything can be fixed, not too difficult, for medium amounts of money per issue.
The end result may or may not be what You wanted, and will be surprisingly expensive if done well.

I recomment not doing "medium" upgrades.
This ends up paying for it 3 times, poor, medium, good, and endless hours of fiddling.

The machine needs to be about 100x more rigid than You think and expect.
Rigid is not strong and strong is not necessarily rigid, but mostly is.
A small 12x24" lathe with 100 kgf loading is 100.000 kg "strong".

Nothing is flat or straight or rigid or accurate -- you are just not measuring accurately enough.

[joeavaerage]

Hi,
your steppers are 5mH inductance which is medium-high. Its probably realistic to expect no more than 500rpm from them.

With 5mm pitch screws that results in 2500mm/min, and that in turn comports with what the manufacturer has suggested.

Inductance severely limits the speed a stepper can run before it loses steps or stalls. If you are buying steppers then:

23/24 size 1mH -2mH, 1 mH preferred and reject anything over 2mH
34 size 2mH-4mH, 2mH preferred, reject anything over 4mH.

The other classic method to counter inductance is to use the highest voltage drivers and power supply you can afford. The current norm amongst serious drivers
is 80VDC....so get 80VDC drivers and an 80VDC supply.

With say a 1.5mH 24 size steppers and an 80VDC drive you could reasonable expect 1000rpm to 1500rpm and with 5mm pitch screws 5000mm/min to 7500mm/min,
which is quite good for a hobby machine.

Craig

[joeavaerage]

Hi,
fast rapids or traverses do not often end up resulting in reduced cycle time. 3D toolpaths have a extremely large number of very small moves.
For such toolpaths high acceleration is the key, and acceleration is highly dependent on the rigidity, or lack thereof, of the machine.

I would not concern myself about speed limitations......find out how it works in practice ,and the effective limitation in practice, is often rigidity(or lack of).

Craig

[LaBelleNote]

Hi,
your steppers are 5mH inductance which is medium-high. I.......
Craig

Thanks a lot Craig for the information, that's exactly what I was looking for. I'll keep that in mind, maybe for the next machine, because as you say, given the lack of rigidity of this one, no need to accelerate or shake it too much.
But I like to understand how things work, that's really useful !

If I try now to focus on securing my work, in order to be 100% sure my that my gcode does not induce loosing steps, what are the key things to guaranty that ?
Is there actually a single explanation for this phenomenon of loosing steps ?
(digging yet a little deeper into the rabbit hole)

Thanks,
Christophe

[joeavaerage]

Hi,

f I try now to focus on securing my work, in order to be 100% sure my that my gcode does not induce loosing steps, what are the key things to guaranty that ?
Is there actually a single explanation for this phenomenon of loosing steps ?

There is no magic solution.....you just have to use the steppers within their capabilities. A stepper either misses steps or stalls when it does not have the torque to take that next step.
While it might be good at slow speed its torque diminishes as it goes faster, so there is always a point somewhere when the required torque of the load is equal, or worse greater, than the
torque of the stepper.....and that's when it misses steps or stalls.

The only solution is to tune the machine max velocity and max acceleration to be less than that threshold.

Many manufacturers will try to sell you 'closed loop steppers' being faster, more powerful, never miss steps....BS. They are still steppers and lose torque the faster they go. Sure the closed loop driver
will insert an extra step to catch up the step that got missed, but guess what, the extra step is just as likely to be missed as an ordinary step.

What a closed loop stepper can do is if it detects that its missed too many steps it will fault out and Estop the machine rather than go on and make a misshapen part as open loop steppers might,
so they are useful, even if not nearly as good as the manufacturers claim.

If you want genuine closed loop performance get AC servos....they'll eat any stepper ever made.

Craig