New steel mill/router design. Feedback appreciated.

Hi Everybody, I've been on this thread a few times, this might actually be my first post, I can't really remember. Anyway, I've been working on a CNC mill/router design over the last year or so, and I've really pulled the design up over the last few months. This is my first build, so be kind.

The design started off as an aluminium T-Slot build, but I was convinced to build it from steel, which in the end looks like it will save me about $400, and still have the same cut space.

Basically I want to build a machine that is capable of cutting aluminium, but will mainly be used to cut through woods/plastics/fibers etc. I need to be able to build this machine with the tools i already have (welder, grinder, drill press tap/die screw drivers etc), so a very efficient design is required. I have found a guy that can help with plasma cutting the plates and some small milling of the parts that need to be straight, at a cost of course.

My cutting space is X: 720mm, Y: 950mm, Z: 200mm
The main drive and guides im using are:
Y Rails: Hiwin HGR20, with HGW20 carriages
Y Drive: 1610 leadscrew, with SFU1610 ballscrew
X Rails: Hiwin HGR15, with HGH15 carriages
X Drive: 1610 leadscrew, with SFU1610 ballscrew
Z Axis: THK33, what ever I can find on EBay really.
1.5kW water cooled spindle with VFD
All steppers (4 off) are 387oz-in, driven by a gecko G540.
PSU 48V 10A
A full list of materials can be found here.
You can see on this list which parts I've already purchased, like the motors, G540 etc..
My aim is to keep my entire build under 3K. I've already spent about $600, and I've got about $2100 remaining, so the budget looks good so far.

My design is a cartesian style mill with 2 drives on the Y and a single drive on the X.
here is a snap shot of the CAD model.
Attachment 239570

You can also download a 3D PDF of the model at the bottom of this post.

OK, so I believe I have done all my calcs right, and my running gear should be strong enough to cut aluminium.

The full weight of the gantry is 55kg, but I have rounded it to 60kg is my calcs, the following are the results of my calcs for a 1605 screw, 1610 screw, 1620 screw and 2020 screw.
My power supply is 48V 10A. Im not really sure how I should distribute this on my machine.
The following calcs are based on 3A each for Y, 3A for X, and 1A for Z.
At 48V, 3A, my steppers output about 1.616Nm with a corner speed of 936rpm.
The critical speed for my configuration (1610 screw 1200 long in a fixed/supported configuration) is 1249rpm.
Im basing this on a required cutting force of 10N.

• With a 1605 screw, to get the the feed rates of 4500mm/min I need 900rpm and 1.692Nm. This is actually above what I have available, so lets try the 2020 screw in my configuration
• With a 2020 screw, to get the the feed rates of 4500mm/min I need 225rpm and 2.298Nm. This is even further outside my motor limits, so lets try a 1610 screw
• With a 1610 screw, to get the the feed rates of 4500mm/min I need 450rpm and 1.431Nm. Just inside my limits.
• With a 1620 screw to get the the feed rates of 4500mm/min I need 225rpm and 1.887Nm. Which is outside the limits.

references:
What size stepper motor do I-need.
http://www.geckodrive.com/step-motor-basics
FSWizard - Free Advanced CNC Speed and Feed Calculator

The 1610 screw is the only screw that suits. It fits my requirements, and I can still lower the 3A down as low as 2.7, allowing more amps to go to other steppers.

I'd like to know if my calcs sound right, and if anybody has had good success with a similar build. Feel free to make comments or suggestions as you wish. This is the 5th iteration of this design, so I really don't want to re-design this that much.

Thanks in advance,
Nick

No way are you gonna run 900 rpm at 48V.
No way to make a 10 mm rise screw work.
The cutting force is 50-500 N.

You have vastly too little resolution.

Quote:

---

No way are you gonna run 900 rpm at 48V.
No way to make a 10 mm rise screw work.
The cutting force is 50-500 N.

You have vastly too little resolution.

---

Thanks for the reply hanermo, is there any chance you would be able to explain?
Are you suggesting I should be going for a 1605 screw?
The cutting force is a product of the cutting depth? So if I cut shallower, shouldn't the force work out OK?

Maybe my calcs are incorrect for my motor corner speed.
191xV/(current x inductance)
=
191x48/(3.5/2.8)=935.51

That's not the speed my motor will turn, I don't fully understand it, but I believe it's the speed at which the motor starts to destruct, or something.
read this: https://www.geckodrive.com/gecko/ima...cs%20Guide.pdf

Some more details.
Good routers with high-rise screws use high-resolution servos.
Even then, accuracy suffers by use of the higher rise.

The gantry mass does not matter. BUT.. your acceleration suffers a lot, because the gearing is wrong.
There is poor mass matching.
In steppers, you must use your real world torque for calcs. That will be about 1/2 to 1/10 of your stall torque.

You also need about 10x more steps than your planned cutting resolution.
Otherwise, you will get low rigidity ie the cutter will shake and jitter.

Also, your screw is too thin for 900 mm.
I had a 0.750", or 19.05 mm screw (for 600 and 1200 mm).
Now replacing with 40 mm.
The thinner screw is not stiff enough.

4M/min is ambitious, unless you sacrifice resolution.
Poor resolution is a bad idea.

Quote:

---

Also, your screw is too thin for 900 mm.

---

My screws are fixed on one end and supported on the other. Does that change things?

The critical speed for the 16mm screws 1200mm long, fixed at one end and supported at the other is 1249RPM.

My resolution is calculated as 200 steps per turn. 10mm for turn.
10/200=.05mm/step

If you want me to increase the steps per rev by 10, the calc would look like this
1/200=.005mm/step.

having a 1mm pitch screw is a bit excessive isn't it? Or are my motors under-powered to get the speeds I need.

I think you'll be fine with that setup. It similar to what many people have been using for several years now.

387oz motors give the best balance of speed and torque with a G540. And hundreds of machines built has shown that a 10mm pitch is about perfect for steppers. With steppers, the faster you want to go, the lower the resolution will be. The higher the resolution you want, the slower you have to go. You can't have both, due to the steppers fixed resolution.

It's nearly impossible, though, to calculate motor rpm with a formula. It's really not that simple, as there are too many variables involved. You could use three different drives, and all three will give you different performance.

I suggest going with a 2010 ballscrew or 2510. Long screws tend to bend and whip at fast speeds. By increasing the diameter of the screw you reduce the whip that occurs.

Quote:

---

Originally Posted by nekidfrog

I suggest going with a 2010 ballscrew or 2510. Long screws tend to bend and whip at fast speeds. By increasing the diameter of the screw you reduce the whip that occurs.

---

I would disagree. 950mm is not that long, and increasing the diameter will greatly affect your acceleration, due to the higher inertia of the bigger screws.

Quote:

---

Originally Posted by hanermo

No way are you gonna run 900 rpm at 48V.
No way to make a 10 mm rise screw work.
The cutting force is 50-500 N.

You have vastly too little resolution.

---

I was able to rapid my small machine at 400ipm with 180in-oz steppers at 48v with 1/2"-8, 2 start leadscrews (4tpi.) That's 1600rpm, and I even tested to 500ipm, though with binding at the extremes of the table.

That said there's always a tradeoff between speed and resolution when it comes to steppers, because the decreased torque output at high speed may not be able to move your axes sufficiently. They do make high-voltage steppers that retain good torque even up to 3000rpm, but they require special drives.