Titivating A CNC Plasma Table

[awemawson]

Since I first got this machine I've not been overly happy with the Z Axis. It's main mounting plate had been bent at one time, and both the normal Z movement and also the 'floating Z Axis' used for material sensing only ran on single  linear rails, resulting in a bit of side to side wobble and tilt. Don't get me wrong - it works - but it's not overly robust in it's design and there were features that I wanted to incorporate - namely:

A/ Dual parallel rails for the main Z and the floating Z movements

B/ A Magnetic Breakaway Torch Mount

C/ Emergency Stop to be asserted if the Magnetic Breakaway is triggered

Now this is all very well, but I didn't want to strip down the existing Z axis and hence stop me using the machine, while building the new one. So I decided to start from scratch and make a modular Z axis that would be on plugs and sockets and fit existing mounts on the machine.

I procrastinated for a long time, looking at economical ways of getting the linear ways, ball screw and nut, bearings etc until eventually a 'Prototype Z Axis' turned up on eBay. It had been put together but obviously never used, and would need re-making, but it had the linear rails, ball screw and nut, bearings, stepper etc just as I wanted, but for far less than the total of buying the separate items 

[awemawson]

The end mounting plates will need re-making to spread the rails further to suite the existing machine mounts, the coupling between the stepper motor and ball screw was rigid, and I'll have to mount a new floating Z axis on it with suitable magnetic breakaway torch holder and limit switches.

So - buy an Oldham coupling and bore it to the correct sizes - simple (and done  )

Draw up a CAD drawing to show how all the components potentially interfere with each other, to make sure that the new upper and lower plates will work - done 

Buy a handful of limit switches - done 

[awemawson]

Buy a big slab of 10 mm thick aluminium alloy ground tooling plate to make the bits - done 

Source a (VASTLY EXPENSIVE  ) Magnetic Breakaway Torch Holder - done and customs duties paid 

(I had intended to make my own, and even drew it out in CAD and bought the magnets, then I realised that the 'real ones' had been carefully developed to hold on in normal situations, and let got if the torch catches the plate and the amount of magnetic force needed was a great imponderable, and with torches about £600 each I didn't want to carry out tests !)

[awemawson]

Source some smaller THK-10RM rails and sliders for the floating Z - done but will need cutting in half 

So now it is a case of making the end plates - (simple but need doing accurately) so I extracted them as individual parts from the CAD montages and will probably CNC machine them.

Still quite a few details to work out re the floating Z and limit switch mounting, but I'll get the end plates made so I have something to prod and poke !

[awemawson]

Incidentally, the pre-existing mounting slots that I previously referred to are these:

[philf]

So - buy an Oldham coupling and bore it to the correct sizes - simple (and done).

Andrew,

I'm not usually a pedant but that coupling is a Lovejoy type coupling and not an Oldham!

An Oldham coupling wouldn't work in that application because they have no lateral stiffness and would need a bearing to support the screw.

http://www.oepcouplings.com/assets/ocanimationsmall.gif

The Lovejoy type couplings allow angular misalignment only and no (or very little) axial misalignment.

Cheers.

Phil.

[awemawson]

I bow to your knowledge Phil 


The vendor referred to it as a Plum coupling 

[awemawson]

Well the upper and lower plates are 'VIRTUALLY' done but no time yet for the real things, also need to resolve the usual fixturing issues 

(I marginally altered these from the CAD drawing above, reducing them to 145 x 68 to take a bit of weight out of the assembly and also give a bigger  'waste' for the milling, as the source plate is only rough sawn at a nominal 150 mm wide.)

[awemawson]

I got time today to finish the upper and lower plates for the Z axis.

Each was a two stage process - boring and drilling the holes, then making a profiling fixture and profiling their periphery giving the corners a 3 mm radius.

[awemawson]

This has allowed me to get to 'first trial assembly' so that I can better decide what layout to use for the travelling stage that mounts the 'floating Z axis' and the actual torch holder

[awemawson]

A bit tied up this morning collecting and distributing 'self assembly lamb kits' (ready for the freezer) but I did manage to do a bit of drawing up the 'front plate' that will hold the 'floating Z axis slider'.

This is a bit of reserved design, in that I am undecided on exactly how I will locate both the HSR-15 sliders and the ball nut location, so each have three alternative positions - being CAD then CAM it's as easy to let the machine do all three alternatives and I can pick and match as it evolves - then if I'm REALLY keen I'll make another with just the selected holes in it.

Meanwhile the plate will look like a bit of Emmental cheese 

[AdeV]

What you need is a lump of plastic-like material which can be machined very quickly and easily in order to prototype the finished article. You know, something that doesn't cost a fortune, and doesn't break cutters if you get a tool-crash, and who's chips can be melted down and reformed into new blocks.... 

[awemawson]

 

Got a cupboard full of machinable wax Ade 

Not really applicable here though - it would break 

[AdeV]

Got a cupboard full of machinable wax Ade 

You might very well think that, I couldn't possibly comment! 

Actually, you'd be surprised how thin you can go and have it hold together; but yes, one overtightened bolt and before you know it you've got a pile of bits all over the floor....

[awemawson]

... No I really DO have a cupboard full of the stuff Ade, but now we know that you do as well   

[RodW]

I am undecided on exactly how I will locate both the HSR-15 sliders and the ball nut location,

I don't know if this will help you, but this is how I made my floating head Z axis using HGR15 rails.



This is the floating head

I've added an adjustment screw on the other side so I can keep the switch movement before triggering the sensor to an absolute minimum. Its just a countersunk screw with a lock nut to keep the plates apart. One day, I'll replace it with a nice knurled knob...

I know how you feel working out the design. What I finally built was probably about version 87!

I've got everything built except the table but its all cut up and ready to weld.

[awemawson]

Thanks Rod.

At long last I've found a bit of time to go from a rendered model of the front plate to making a physical prototype. This will allow me to develop the Floating Z bit, and finally decide on which alternative mounting positions to use for the rail slides and the ball nut mount.

Fairly obviously the final version needs to be a tad wider to accommodate the counter bores for the M4 cap screws fixing the rail sliders, but with this mounted I can develop my ideas a bit further

[PekkaNF]

Thanks Rod.

At long last I've found a bit of time to go from a rendered model of the front plate to making a physical prototype. This will allow me to develop the Floating Z bit, and finally decide on which alternative mounting positions to use for the rail slides and the ball nut mount.

Fairly obviously the final version needs to be a tad wider to accommodate the counter bores for the M4 cap screws fixing the rail sliders, but with this mounted I can develop my ideas a bit further

Nah, just open then couterbore out with a slot drill, mechanically it will fine and look the part.

Out of curiosity....the "floating" feature is arranged with stepper motor control? Like Z-is sent from program and then there is a whizbox that superimposes torch control over that setpoint and drives the stepper?

Pekka

[RodW]

Out of curiosity....the "floating" feature is arranged with stepper motor control? Like Z-is sent from program and then there is a whizbox that superimposes torch control over that setpoint and drives the stepper?

Pekka

Plasma is a bit of a funny animal when it comes to CNC. The floating head is to find the surface of the material when initially piercing. The torch basically hangs on a section that can slide up when the torch head hits the material. This triggers a float switch (proximity sensor in my example above).  So the torch then moves up to adjust for switch hysteresis and the material height is recorded. The torch then retracts to pierce height (say around 3-5mm) and turns on. The plasma machine sends a signal back to say the Arc has been established (ArcOK) and then the controller counts down for a pierce delay. After that, the Torch moves down to cutting height (say around 1.5mm) and then off it goes on its merry path to cut your part.

The problem then is that material is not dead flat and thin material can warp from the heat so most modern CNC Plasma controls measure the torch arc voltage as there is a linear relationship between cut height and arc voltage (voltage increases with Torch height). The controller then controls the Z axis height to keep the voltage at the desired level.

There are several other methods to sense the material. Ohmic sensing simply uses the material and the torch tip as an electrical switch (without the 300 volt arc frying your electronics).  Others monitor Z axis torque (or current) so that when it starts to stall as its driven into the material, the material height is recorded.

So in summary, the plasma table is simply an XY machine where the Z is let to do its own thing to keep a stable arc voltage.

[PekkaNF]

Thank you, I was wondering the rollercoaster drive on steel plate. Long time ago I was watching often even then old "optical" tracker of semiautomatic oxyburner to cut plates. They were not that simple to use and adjust. Sometimes the "seeker went wild" and run ofcourse torch blasting thick steel plate to art forms

Pekka

[RodW]

Thank you, I was wondering the rollercoaster drive on steel plate. Long time ago I was watching often even then old "optical" tracker of semiautomatic oxyburner to cut plates. They were not that simple to use and adjust. Sometimes the "seeker went wild" and run ofcourse torch blasting thick steel plate to art forms

Pekka

I think now most Oxy cutters use capacitive sensing between the material and a ring around the torch that never touches the plate.

[awemawson]

Hypertherm (makers of arguably the best Plasma power sources) advocate the use of 'Ohmic Sensing' as being most accurate, but my concern was I will often be cutting plate that is slightly rusty so it might not work reliably. There are systems out there that use both 'Floating Z' and 'Ohmic Sensing' - the floating Z effectively being a back stop.

Certainly Ohmic Sensing is simpler to implement and it avoids the use of a secondary slider.


... talking of secondary sliders - I'd got the hole spacing wrong on the THK-10RM rails and drilled and tapped at 20 mm centres rather than the 25 mm they should be 

However as it was all made on an alignment fixture I've been able to return the prototype to the CNC Mill in exactly the same place as before, and run a little diddy program to spot face, drill and tap the relevant four holes (and also modify the Autocad drawing accordingly !)

So my Faux Pas is recovered from without raising a sweat 

I do love making things on fixtures - this particular one just clamps in the Kurt vice, and I've bored an accurate 20 mm hole dead on centre which becomes my datum point. If I need to remove the fixture and replace it all I have  to do is lower my Heidenhain touch probe into the hole and tell it to find the centre - fast and accurate. The various parts bolt to this fixture using holes that are part of the workings.

It also means in this case that the various parts are held about 10 mm higher than the vice jaws, allowing profiling of the edges without risking cutters and jaws.

[RodW]

I know the feeling. I have one hole to fill in with weld and retap. I got all mine laser cut but until I could actually check for fit I was nervous as hell with $1000 worth of parts....

Hypertherm (makers of arguably the best Plasma power sources) advocate the use of 'Ohmic Sensing' as being most accurate, but my concern was I will often be cutting plate that is slightly rusty so it might not work reliably. There are systems out there that use both 'Floating Z' and 'Ohmic Sensing' - the floating Z effectively being a back stop.

Certainly Ohmic Sensing is simpler to implement and it avoids the use of a secondary slider.

I don't think you can get of that lightly. Everything I've read says you need a backup either sensing motor torque or a float switch for those Oh! sh1t moments when you don't get an ohmic connection. I'd like to experiment with measuring motor torque by sensing the current the stepper axis draws. Even if I had to use an Arduino as an intermediary. My Plasma has a wire for the material clamp on its CNC interface (via a 100k resistor) so I might try it later. I have drawn up a circuit that should work and lock everything out while there is an arc.

Anyway, here's a sketch of the circuit I came up with. Check it carefully.

It was based on something I found on the net and I added the loop through the torch on relay as an added safety feature.

[awemawson]

So, next thing to do is cut one of the THK-10RM rails in half, as I only need a very short movement. So - first transfer the slider off one rail onto the other one so that it can be cut without contaminating the bearing.

Now received wisdom is that this size of slider will fall apart shedding ball bearings unless the rails are pressed well together end to end, and the bearing pushed over the join carefully

Well I've news for you - they shed bearings anyway  Hundreds of them, and they're tiny  And in my case they are magnetic and stick all over the place 

Now in the last picture that minute bit of black plastic is in fact one of the track guides that reverse the balls so that they circulate. Even with tiny tweezers it was very reluctant to go back into it's recess in the end cap.

[awemawson]

So now the slider is in bits I decided that I might as well go ahead and cut the rail, so in the unlikely event that I can re-assemble the bearings at least it can be done on the rail it will eventually occupy.

Cut with a 1 mm abrasive disk on a 115 mm angle grinder held in a 'chop saw stand' (Thanks Spurry for the suggestion ) I then tidied the ends up on the Clarkson T&C grinder