Ballscrews or ACME for Z-axis?

[digits]

Hi guys, I am trying to design a heavy duty Z-axis to support ~20kg of mill head, and I was planning to use a 16mm screw and a gas-strut to take some of the load, and to run it along a pair of 35mm HiWin rails with one block per rail as the blocks are 110mm long and good for >1kN of force!

I plan to use a stepper to drive it, but I can't decide whether ballscrews or trapazoidal/ACME would be better, as I am worried that the non-locking nature of ballscrews would cause the head to hit the table when the power is turned off. Is backlash a problem with ACME z-screws anyway - doesn't the weight of the Z-axis keep the screw pressed against the nut?

Sorry if these are silly questions - I did a quick search but didn't see anything relevant.

Cheers.

[NC Cams]

The amount of weight that the screw/motor has to support in concert with the coefficient of friction in concert with the gearing and the screw geometry determines if the screw (regardlass of type) can "hold up" the weight of the Z axis.

IMO, you'd actually want to couterbalance the Z so that the screw isn't keeping Z from dropping/lifting ALL of the weight up when/if you have to move it = the motor would then only be moving the weight suspended, not completely hold it up 24-7.

Pretty sure the following thread gives insight into the reason why math (in lieu of SWAG estimates) comes in so handy with regard to sizing steppers and servos:

http://www.cnczone.com/forums/showthread.php?t=17707

The weight/braking issue is a reason why supplemental actual clamping brakes are used on some commercial mills - gravity sucks, no matter what your budget is and you have to deal with it someway or some how when you design/integrate the components into your CNC'd whatever.

[digits]

Originally Posted by NC Cams The amount of weight that the screw/motor has to support in concert with the coefficient of friction in concert with the gearing and the screw geometry determines if the screw (regardlass of type) can "hold up" the weight of the Z axis. IMO, you'd actually want to couterbalance the Z so that the screw isn't keeping Z from dropping/lifting ALL of the weight up when/if you have to move it = the motor would then only be moving the weight suspended, not completely hold it up 24-7. Pretty sure the following thread gives insight into the reason why math (in lieu of SWAG estimates) comes in so handy with regard to sizing steppers and servos: http://www.cnczone.com/forums/showthread.php?t=17707 The weight/braking issue is a reason why supplemental actual clamping brakes are used on some commercial mills - gravity sucks, no matter what your budget is and you have to deal with it someway or some how when you design/integrate the components into your CNC'd whatever.

Thanks, I do intend to do some maths before sizing steppers and screws etc, but is the self-locking and non-backdriveability of an ACME nut actually an advantage in the Z-axis?

[NC Cams]

The friction helps and hurts you.

It helps in preventing droping BUT it reacts against you when/if you want to move fast. Any time you ADD friction ALL the time (acme vs ball screw), you have to overcome it one way or another.

When you want to or ask for fast moves, the added friction of an acme (compared to ball screw) hinders performance.

Nothing in life is free. There is ALWAYS a compromise....

[Geof]

Originally Posted by digits ... but is the self-locking and non-backdriveability of an ACME nut actually an advantage in the Z-axis?

As soon as you get a bit of vibration, tool chatter, etc, the self-locking is no longer self locking; you need a brake or other method of holding the axis in one place. If you counterbalance so there is minimal net load a stepper will probably hold.

[ViperTX]

Well....I would tend to disagree with Geof....because your head is only 20kg most any milling is going to try and pull the head down or backdrive it....that is where an ACME screw is advantageous...I would certainly look at having a servo drive the ACME screw.

[Al_The_Man]

Originally Posted by ViperTX that is where an ACME screw is advantageous...I would certainly look at having a servo drive the ACME screw.

I am of the other opinion, personally I do not think Acme screws and CNC go together, with backlash and added friction (BS 90+% - Acme 65%eff.) , I believe you are defeating the object.
You will have to increase motor size because of it, especially if you require rapid accel/decel.
I am on the side of some sort of counter-balance system, you can always add a brake for when the m/c is off.
Al.

[NC Cams]

The interrupted nature and/or varying loads of a cut will deflect the tool - GOD SAID SO. Thus, ANY clearance/slop/backlash/hysterisis in ANY axis will result in deviations from the intended cutter path/finish. Again GOD SAID SO.

If you want the ultimate in performance, you choose the method/equipment that will NOT subject your work to cutter path deviation. THis is common sense but that comes from GOD too.

Now, this servo/balls screw/brake sort of Z axis control is typically done in high end machines. If you don't want to spend that kind of money or this is a typical DIY project and don't need that sort of accuracy/finish/etc, you could surely use "something other" than a preloaded ball screw and a super tuned servo and get it to work acceptably. How acceptably depends on how well you integrate the motor and axis drive.

Depending on how the "something other" is tuned/built/spec'd, you could end up with equivalent or less than equivalent performance. It all depends on how well you engineer it, spec it, assemble it and tune it AND HOW MUCH YOU WANT TO SPEND.

Ultimately, the forces that you plan to generate will have to be calculated or considered or SWAG planned for. Otherwise, it is anyone's guess if a SWAG selected anything will work well in a haphazardly engineered and/or assembled guestimated servo or stepper drive of ANY axis.

The "how good do you want it to work?" question has not be spec'd yet as far as I can see in this thread. Until that happens, opinions will continue to be expressed/shared and probably be quite varied.

The net and ultimate decision will still depend on how accurate do you want the machine to be and how much do you want to spend????????..

[digits]

Thanks guys - it is very interesting to hear all your opinions.

My current machine is all ACME screws and steppers - I am looking for more speed an accuracy in my next machine, but not at an infinite cost! If ballscrews are the way to go for all axes, I guess I'll have to find the extra £'s - I just don't want to over engineer it!

[digits]

Originally Posted by NC Cams The interrupted nature and/or varying loads of a cut will deflect the tool - GOD SAID SO. Thus, ANY clearance/slop/backlash/hysterisis in ANY axis will result in deviations from the intended cutter path/finish. Again GOD SAID SO. If you want the ultimate in performance, you choose the method/equipment that will NOT subject your work to cutter path deviation. THis is common sense but that comes from GOD too. Now, this servo/balls screw/brake sort of Z axis control is typically done in high end machines. If you don't want to spend that kind of money or this is a typical DIY project and don't need that sort of accuracy/finish/etc, you could surely use "something other" than a preloaded ball screw and a super tuned servo and get it to work acceptably. How acceptably depends on how well you integrate the motor and axis drive. Depending on how the "something other" is tuned/built/spec'd, you could end up with equivalent or less than equivalent performance. It all depends on how well you engineer it, spec it, assemble it and tune it AND HOW MUCH YOU WANT TO SPEND. Ultimately, the forces that you plan to generate will have to be calculated or considered or SWAG planned for. Otherwise, it is anyone's guess if a SWAG selected anything will work well in a haphazardly engineered and/or assembled guestimated servo or stepper drive of ANY axis. The "how good do you want it to work?" question has not be spec'd yet as far as I can see in this thread. Until that happens, opinions will continue to be expressed/shared and probably be quite varied. The net and ultimate decision will still depend on how accurate do you want the machine to be and how much do you want to spend????????..

Sorry - I read your reply after my last post!

How much do I want to spend? TBH I am still coming to terms with how much all linear motion components cost! I am in the UK and everything seems to be priced as if £1=$1, even though it's closer to £1=$1.9. eBay UK is also pretty useless, so the cost per axis for even 16mm ACME screws,nuts and supported round rails is about £300/$570. Ballscrews add another £50-£80 per axis - which while painful I would consider if they'd make a positive difference.

I don't really fancy upgrading to servos as well right now - but if I can design and build a machine that's accurate and sturdy enough to benefit from them, I'll probably start saving up!

My aim is to build a machine capable of cutting aluminium with a 1/4" endmill at a depth of about 0.1" and full-width cut at a rate of 20-40 IPM with about 0.005" error across 12", and a maximum travel of 20" in any axis. This may prove impossible or at least impossibly expensive for me!

[Al_The_Man]

It sounds to me that you are trying to build a machine from scratch.
1/ It is not always the most economical way to go.
2/ You end up spending alot of money and time and the results are unsatisfactory, especiall on a tight budget.
If this is so, you generally are better off looking at a retro-fit candidate that has already the CNC features on it. I am no longer in touch with the UK market, but in N.Am. Machines with dead controls can be had for a song, (usually the shipping exceeds the price).
Al.

[digits]

Originally Posted by Al_The_Man It sounds to me that you are trying to build a machine from scratch. 1/ It is not always the most economical way to go. 2/ You end up spending alot of money and time and the results are unsatisfactory, especiall on a tight budget. If this is so, you generally are better off looking at a retro-fit candidate that has already the CNC features on it. I am no longer in touch with the UK market, but in N.Am. Machines with dead controls can be had for a song, (usually the shipping exceeds the price). Al.

Thanks Al - I haven't ruled out retrofitting an existing machine - I just haven't seen a machine with the sort of travel and spindle speeds I want that isn't 600lb, £2500 or both! I had almost talked myself into buying a very nice looking manual mill and CNC'ing it when I noticed that despite the £2200/$4200 price tag, ballscrews were an optional extra!