HBM Profi 45 conversion

I bought an HBM Profi 45 a couple of days ago. The main reason for selecting this machine from many options was its ISO-30 taper.

My plan is following:

• Closed loop motors
• Ball screws with double ball nuts
• Full enclosure
• Flood, mist and air cooling
• Original spindle, gearbox and motor but VFD drive
• One shot oiler
• Counterweight for Z axis
• Touchscreen for control
• Power drawbar if easily possible
• At least a provision for tool changer if power drawbar is built

So far I have pondered a few design choices:

• Mach 4, LinuxCNC or some other control software?
• Closed loop steppers or actual servos? I want high performance but is it really worth putting servos on a non production machine?
• Direct drive or belt drive motors? Normal way seems to be direct drive, but with belt drive the motors could be moved to a more protected position.
• Fieldbus based control system or traditional? The fieldbus based control systems are really interesting. They seem to have much fewer parts than traditional systems which also compensates the higher cost of the components.

I'll begin by disassembly and CAD modeling of the machine as it gets delivered. My goal is to build full CAD model for easier design of the drive system, enclosure, base etc.

It sounds like you have a good plan for the most part.

I'd say belt drive is the way to go. Especially if you want a power draw bar.

Check out UCCNC for a control, I think it's better than Mach 4 by a long shot. There are distinct advantages to Windows controls as there are for Linux controls. It just depends on what you want and what your comfort zone is.

I'd definitely recommend modeling your machine. Down the road it will be invaluable. I found a model of PM25 that is pretty close for my G0704. I use that model for all sort of stuff, that includes my enclosure.

Late to answer but this may help some others with the same question.
If you are talking about belt or direct drive for the axis motors, its all going to depend on your motor selection and screw pitch. Usually a servo will need to be large to direct drive, and would provide a very high (un-useable) speed on a machine this size when coupled with typically available ball screws (5mm pitch). Small servos can be belt drive and use the belt ratio to reduce the max speed but increase the torque. Steppers are usually direct drive as they are high torque, but their torque falls off with speed, so are not usually belt connected as any belt reduction also magnifies this problem and adding belts at 1:1 adds cost and complexity for little gain.


I built my 12Z (big RF45) years ago with relatively small servos and belt reduction. Would I do it again now with 10+ years experience on it? Probably not. KISS is a good principle on a hobby machine and adding unnecessary complexity using servos with encoder feed back to the control cabinet adds more things to go wrong, plus the added expense of belts and pulleys. Steppers are beautifully simple, and the main downfall is they don't fault when they miss a step. Sizing them properly and not exceeding their capabilities are the keys to steppers. Adding an encoder feedback to a stepper just brings it back up to the electrical complexity of a servo and closer to the cost, without the belts. I've had some electrical failures over the years, all due to me not putting the proper care into protecting the encoder signal cables. The belts have been fine.
I do want to say that although I wouldn't do belt drive servos again, I'm not disappointed I did. Its been pretty good and any issues have been down to my laziness and lack of protection of the motors and connections.

I've been a bit slow on the project. Actually I've only started to model the mill so far.

I have however new ideas on the machine. LinuxCNC supports position feedback. That would be great because lower end ball screws have variations in thread pitch. Maybe it could even replace double ball nuts? Tuning might be probelmatic though, to avoid oscillations due to backlash. It would make mechanics easier, but would naturally require linear scales.
Nonetheless it is an interesting possibility.

Regardless of the control system you need to reduce or eliminate as much backlash as possible. Backlash will kill a tool in a blink of an eye especially a small diameter tool.. If you're climb milling and it pulls in the backlash then you're cutter will be gone cause you can easily double or triple the chip load on it. Another way to deal with it is to tighten your gibs up to increase friction so its doesn't get pulled in but then you're creating new issues (increased wear, higher load can equal lower top speed, etc.)
So I'd recommend doing as much mechanical solutions to reduce backlash including double ball nuts, and leave linear scales and position feed back for bigger machines. The machine isn't rigid enough to truly benefit from position feedback, unless you're machining $10000 items, which then begs the question why are using a hobby machine.

As for variations in screw pitch, you don't need live feedback to compensate for that. Linuxcnc supports screw mapping which compensates based on a lookup file. So you measure the deviations at many points, create the file and then Linuxcnc compensates for you automatically. I believe other control systems support this as well.

I'm about to start sizing the ball screws and motors. What should I use as an estimate for cutting force?

Uou sound like your doing what I did 10 years ago.


I tried calculating everything, and designing to optimize the mill cnc conversion based on math.
What I was missing was experience, and knowledge of some of the finer points regarding milling in general and cnc milling specifically. I'm still missing alot but here is what I have learned:


If your asking what are the cutting forces, you can find them by using almost any feed and speed wizard. here is one: https://zero-divide.net/?page=fswizard
example 1/2" endmill ~ 82lbs cutting force in hot rolled steel at 3.74in/min, 0.875 DOC" and WOC is 0.147" and power required is only 0.28hp


But that only tells you a small piece of the story.


The reality is that its far more about deflection than absolute force, and unless you do a really great job modelling the mill and are a wiz at analysis your going to miss the point.
The calculations will say that the mill can move (accelerate, decelerate and push) the cutter into the work (or vice versa), but reality will bite you and your cut will look like crap or your cutter will break or both.
This is due to the deflection within the mill between the cutting tool and the workpiece.


The real limitations of this mill (at least to me) is the spindle top speed and the z axis dove tails. For cnc work, ideally this mill would have a much higher speed spindle to enable faster, lighter cuts, reducing the deflection and improving the overall performance.
After that, the heads excess weight deflecting the dovetails that creates slop or excess friction is another weak point.


So to condense down what is a long story - don't worry about cutting forces, and don't even bother with calculating the forces. In general following another persons build will get you there, the nominal ball screws used on virtually all these benchtop conversions (i.e. 1605 or 2005) have typically 100% margin compared to what is needed under worst case conditions.

Unless math is fun :)


Also, don't waste any money on cutters bigger than 1/2" (exception being a face mill or fly cutter) - and honestly I've taken to going down to 3/8" (10mm) and smaller for the stuff I do - as a hobby its not time sensitive so using a reliable smaller cut is better than pushing the envelope with a large cutter.

Should I use the ready made double ball nuts (DFU series) or make my own from single nuts? Two single nuts seems to be to more popular choice. Why?

Availability, the single nuts were far more readily available than the factory doubles, get the factory made ones if possible to save yourself the hassle, plus it may actually be cheaper to get one double than 2 single.