Ball Leadscrew sizes & Motor sizes

[richardandtracyParticipant @richardandtracy]

I am to convert a Clarke CL430 to cnc this year (I hope!). I would appreciate it if you could comment on some of my current thoughts.

The machine is to be used for cnc machining of pens. I had hoped the machine I got in this thread: **LINK** would work, but unfortunately flexibility of the gantry has meant it's not as successful as I had hoped, with many more operations needing to take place on the lathe than I had hoped. It has, however been very useful to reduce cnc-fear and give me the confidence to convert a machine myself.

I plan on creating a 3 axis machine, X, Z and A, where the A axis can be released with a clutch so the normal lathe motor can be used for 2 axis cnc lathe work.

Now the CL430 has an axial leadscrew that's 20mm diameter and is moderately fine in pitch, think it may be 2mm, though I've not really investigated yet (my Warco WMT300 at 2.5mm pitch is visibly coarser). There is a fair bit of friction on it though; not bad for hand operation but a bit of a load for a stepper motor. The obvious replacement is a 20mm diameter, 5mm pitch ballscrew (2005). However, the smallest diameter on the current leadscrew is 14mm where it goes into the power feed dog clutch, which implies to me that actually a 1605 ballscrew would probably be suitable. The 8mm diameter reduction on the ballnut (36 to 28mm on the main diameter) would give more room to play with under the saddle. So I think a 1605 ballscrew is the one I'm going to use (the fact it's half the price is nice, but not really relevant). Can anyone see any problems with this?

On the cross feed I was thinking of 1204 (12 dia, 4mm pitch) ballscrew feed. I think the ballnut will have to attach to the far side of the saddle as opposed to the current position of under the saddle, and be attached using only half the attachment screw positions (as per the method JS used on a Seig mill conversion).

As for motor sizes. Was going to use the Nema 23 frame size, and go for a 3Nm motor for the axial feed and use a 2:1 reduction. Having said this, I have a slight concern that the reduction should actually be 4:1 given the increase in leadscrew pitch the 1605 ballscrew would demand, getting 1.25mm advance per motor revolution instead of 2.5mm. Any comments on the gearing ratio?

On the cross slide I was planning on a Nema 23 frame, 1.26N.m/180 oz.in stepper and a 2:1 ratio, though for simplicity it might actually be better to go for a 2.4N.m motor and do away with the gearing. The cross slide is not going to be used as heavily as the axial feed, so a slower actual feedrate than programmed isn't going to be a problem.

At the moment I am unsure what motor I will need for the A axis when it is used as a X/Z/A machine. I was thinking along the lines of a Nema 23 frame, but the torque is debatable. The rotation velocity is not going to be great in this configuration, in the region of 10 degrees per second maximum, but it's likely there are going to be lots of changes in direction. The spindle will have an MT4 collet chuck, so the inertia isn't going to be high and I don't think a big motor to accelerate the spindle will be needed. I almost feel a 4:1 ratio and a 1.9N.m motor would be adequate. Any thoughts?

Looking forward to any comments.

Regards,

Richard.

[richardandtracyParticipant @richardandtracy]

[MuzzerParticipant @muzzer]

When it comes to reduction ratios and stepper motors, it's instructive to look at the torque – speed curves for stepper motors. As JS pointed out in the linked thread, the torque falls off rapidly with speed and you soon also run out of volts to make it go any faster.

If you plot the motor torque – speed curve for a few typical stepper motors and then look at the effect on the final torque of the same motors running quicker (through a higher reduction ratio at a correspondingly lower torque), you'll often find that you are no better off. You may as well just drive the screw directly in the first place, if the layout permits.

There's a lot of willy waving and downright BS on the home CNC forums. Some basic sizing calculations might be helpful to get into the right ballpark. A 5mm pitch leadscrew with a 2Nm stall torque applied through a 4:1 reduction would generate a force of 10kN (aka 1 tonne). That should certainly suffice for moving a table back and forth with no actual cutting forces to contend with. Not convinced a 1605 ballscrew would be comfortable with that load, mind.

I suspect speed and acceleration are more important than maximum force. You might be better off spending your money on a more sophisticated solution like a closed loop stepper or a lower powered servo drive so that you get better security against loss of position and faster response and rapids.

Some of this comes down to your approach to budget. Is this to be done on a shoestring or do you want to spend a bit more to get a better solution?

Murray

[John HaineParticipant @johnhaine32865]

Just as a benchmark my Super 7 conversion uses a pair of NEMA 23 steppers. I'll have to check the torque spec, but they were the larger size that ARC supplied a few years back, with 1/4" shafts. It uses the standard 8 TPI leadscrew driven through a 2:1 reduction by toothed belt. I started using a standard screw on the cross-slide with 2:1 reduction but replaced that with a 12mm dia / 5mm pitch ballscrew with direct drive as I was getting small cyclical feed errors from slight eccentricity in the pulleys. I've never had any problems of not having enough torque on either axis.

I assume that you want to motorise the headstock to use it as an A-axis with a rotary tool in the saddle? You may want to have a higher reduction to get a force on the workpiece equivalent to that exerted by the Z and Z axis drives – it needs some calcs.

If you fit a ballscrew for the Z axis you'll want to keep the swarf off it. Dyson vac hoses are reputed to be good for this – as used by cncyourmyford.com.

I'd shop around for the lathe – similar machines are sold by various suppliers but in different levels of finish. I suggest visiting an exhibition and looking at least at Warco and Chester. When I've looked at Clarke machines in Machine Mart, frankly I've not been very impressed.

A final thought, given it sounds like you have a gantry router, and presumably an A axis already, have you considered getting a proper CNC mill rather than a lathe? I picked up a Denford Novamill chassis, without electronics, on eBay for ~£600 and fitted modern stepper drivers, and I'm delighted with the machine. If you look at the Denford forum, there is an A axis someone has made for it – one day I'm going to use up the spare stepper drive I built into my electroncs and fit my stepper-controlled dividing head.

[richardandtracyParticipant @richardandtracy]

Muzzer,

I'm aiming for a solution that will work. If there are definite advantages to going to a more expensive solution, I'm for it. I am looking at the Planet CNC controller as it has processing capability on board and doesn't rely on the PC so much. Costs a bit more, but has a definite advantage over a board that cannot interpret its own G Code – I admit I'm inferring this latter capability, but it's based on the ability of the controller to run a program without a PC attached using only a program on an SD card.

Don't want to throw money away – my general approach is to go the cheapest way that will work, but if something will (work better)/(give definite advantages) for a bit more, I'll go for it.

The root diameter of a 1605 ballscrew is around 14mm – about the same as the minimum diameter of the current CL430 drive train that I'd be using (want to put the drive in the belt box if I can so as to keep the axial feed handle etc untouched at the right hand end of the lathe – need to check dog clutch backlash though). The CL430 does have a shear pin, about 5mm dia, in its drive train which will be the weak spot – I intend to keep that, and it should prevent the ballscrew being overloaded.

Regards,

Richard.

[richardandtracyParticipant @richardandtracy]

John – you posted as I was thinking about the reply to Muzzer.

I have the lathe – I found a nearly new CL430 at a good price on the Bay (and by 'nearly new' I mean it may possibly have been used once or twice). I went for that as it's basically the same machine as my Warco WMT300 and all the tooling will be usable on both machines. The Clarke is cheaper than the Sealy version too.

Thanks for the info about your drives and the leadscrew shielding. I will be using it mostly for pen making, and the coloured plastic pen blank swarf gets everywhere. It is light enough to float in the air on occasions, and I can see it getting in the ball nut.

When using the A axis I expect to be engraving using a 300W router with up to 3mm ball nose router or something similar, and it'll be into silver, bronze or plastic. The diameter of the work will be 20mm max, and I'm used to feeds of 50mm a minute for the tiny routers as it stops them overheating the plastic.

Regards,

Richard

[MuzzerParticipant @muzzer]

Have you been following JS's recent thoughts and trials with Chinese CNC controllers? They incorporate an FPGA for the high speed processing – with the additional benefit that you don't need a PC. Everything is built-in, including the display.

I assume the PlantCNC controller still requires a PC / monitor etc to load, configure and control the machine, even though much of the processing power is offloaded to the controller?

[richardandtracyParticipant @richardandtracy]

I have been following JS's thread and I am in two minds. I like the idea of a full controller, but finding enough (or any!) information about them is difficult. I need to be sure that 1) it will control a lathe and 2) I can achieve a 3 or 4 start thread before I commit to one

The Planet CNC controller does need a PC to configure, set up and proof a program. Only when that's done can it then be used independently.

At the moment I am not 100% committed to the PlanetCNC solution. I was, as a result of Blowlamp's video cutting a thread, but since have been slightly dissuaded by the lackadaisical approach taken towards support when someone has a problem. Blowlamp has on the PlanetCNC forum, highlighted a problem (as have other people) and the response time is dreadful. If there is one.

Regards,

Richard

 

Edited By richardandtracy on 09/01/2017 13:07:35

[Martin 100Participant @martin100]

For motor sizing the articles by Dick Stephen on his conversion of a Sieg X3 mill to CNC in MEW a few years ago might be useful as it goes back to fundamentals rather than just making a wild stab in the dark.

Part 1 (2nd link on this page)

Sieg X3 Mill – CNC Conversion

Edited By Martin 100 on 09/01/2017 14:17:58

[richardandtracyParticipant @richardandtracy]

Martin, thanks for that. It's a series I was unaware of, thanks for sharing.

Regards,

Richard

[Martin ConnellyParticipant @martinconnelly55370]

I use Leadshine hybrid stepper servo motors. They have an error output if they get more than 200 pulses (18 degrees) out of position (factory setting but configurable). The attached encoder ensures no lost steps under normal conditions and the error signal can be used to stop the cnc program if a crash or other unexpected overload occurs.

Martin

[MuzzerParticipant @muzzer]

Yes, these were what I was referring to as closed loop steppers. I've actually got one of the integrated "Easy Servo Stepper" motors. The drive is integrated with the motor. You simply provide it with step/dir signals and 36V (for the 2Nm version) and it does the rest. Because it has closed loop operation and can flag up loss of position, you don't need to ensure as large a design margin as you would with a conventional open loop stepper drive. I'm pretty sure I read that they actually use vector control, so they are a lot more sophisticated than many stepper drives. And as it has a high resolution encoder, it can genuinely resolve microsteps without loss of torque etc.

Another interesting option would be a proper digital (AC) servo drive. Many of them are quite costly but there are a few affordable products such as DMM Tech or CNCdrive. Much better torque characteristic (flat up to perhaps 3000rpm), closed loop, high resolution, fast response etc. Lower headline ("stall&quot torque so you may need to introduce a reduction ratio of perhaps 2-4 or so but the torque is real and usable.

Murray

[richardandtracyParticipant @richardandtracy]

Thanks for these suggestions. I will take a good look at them, but at the moment they seem to me to be quite a bit extra for the limited benefit they provide.

In the next day or two I'll put my maths head back on and work out the hard way what torque motors I'm going to need, and make a wild stab at the A axis torque needed.

Thanks for your suggestions,

Regards,

Richard.

Edited to correct tripe-writing error.

Edited By richardandtracy on 10/01/2017 14:13:48

[Author]

Posts