Second CNC table build, here we go again....

Hi all,
I am now starting my second table build, this time a bit bigger at 1250x1250mm cut area, I have also upgraded to a Hypertherm Powermax 45 cutter. My first table is now too small and cannot handle the speeds needed by the PM45.

I have just about decided to go with ballscrews all round, 16 or 20mm lead with anti-backlash nuts, motion is on HiWin 15mm rails and carriages with scrapers and wipers fitted.

Here's a couple of sketches of the gantry ends, just ideas at present but would seem to fit well..
http://www.mycncuk.com/attachment.ph...id=15629&stc=1http://www.mycncuk.com/attachment.ph...id=15630&stc=1

The main table will be 50x50x5mm box section steel, fully welded, the motion supports will be bolted on so I can shim them up and avoid welding warp etc.

The top rails or Y-axis will be above the bed so I can slide a bigger sheet in if wanted, this will also keep the tracks and screws above the pierce splatter.

Drive is with my existing Nema23 3.1Nm steppers, if that does not play nicely then I can upgrade easily to Nema23 servos.

I am aiming to get a speed range of 100mm to 9000mm/min so I can use the full range of the PM45 without messing around.

Feel free to offer any advice :)

A plasma table has reappeared on my potential project list, and I do have a couple thoughts about mitigating dust.
.
I wonder if mounting the rails on the underside of the side frame would help dust from settling onto the rails, then have the screw mounted below the rails/gantry/carriage? On the inside edge you can then have a bit plate hanging straight down to stop dust coming of the cutting area, and arrange a couple brush strips on the outside to minimise dust blowing around and in.

Yes, that could be a good idea, i'll sketch it up.

it would mean a small loss of Y travel or making the table a little bigger as the carriages will hit the support legs, no real issue there though.

A bit like this?
http://www.mycncuk.com/attachment.ph...id=15632&stc=1

Might be better, Not totally convinced though, the rail is now facing the pierce splatter zone.

I have people advising I go with rack and pinion and I have people saying go with ballscrew - both with good reasons for and against.

How do i decide what method to go with???

All the ones I've seen have been RP, share their thoughts with us, I would have thought ballscrews would be abit overkill given it's contact free cutting? I don't know but RP just seems like it would be a better fit, I guess it could get pricy if you got really good stuff but dose a plasma based machine need that kind of acuracy ?

.Me

Just done a rough sketch of what I was thinking.
.
If you wanted to add support to the side rail, you could use a wider box section, and either move the rail nearer the edge and have the legs/supports mounted on the opposite edge, or use a double box section, with the rail on one and legs/support on the other.
.
Dust could be mostly eliminated by adding a couple brush strips, or suitable strips of metal with minimal gap either side of the rail to further reduce the ability of dust to settle onto rails.
.
And having just looked at my sketch and thought about how to protect the screw from dust, mounting it at the side of the rail so you only have one 'area' you need to minimise dust ingress might be an option.

Regarding the R&P v Ballscrew, both have their pros and cons.
As Lee mentions, a plasma doesn't really need the accuracy of ballscrews, but the biggest drawback is coming up with a design to keep dust out, as I'm sure you're well aware, plasma dust is pretty abrasive and will quickly kill ballscrews.
However R&P on the other hand, provided it's designed well enough, will be relatively unaffected by dust, but you lose accuracy, and have the issue of dealing with wear, especially uneven wear which will be challenging to adjust out to minimise backlash while avoiding binding. Off course the solution to this could be simply changing the racks when worn, building in some form of spring loading of the R&P, or using anti-backlash pinions.
.
Both options have pros and cons, both will work, and both can be designed to have long service lifes. As with most things, it comes down to implementation.

Hmm, thanks, I have doubts now, not good;)

I wanted accuracy for small part cutting in thinner materials, so high speed without wiggle on the torch.

Putting the screw and rail under the beam is ok but is filling the gap I wanted left clear for large sheet entry, likely only a few inches clearance left and the covers will be at risk of taking a battering from sheets.

I see machines like PlasmaCam (which I think have a good name?) running simple plate/skate runners but they do use servo drive, they seem to work ok. Other production builds like Torchmate and PrecisionPlasma use various other simple systems ranging from belt to R&P. All mostly with not much regard paid to keeping the muck off of transport parts.

But there again, none of them using ballscrew either!

I'm not sure where this project is going now, back to stage 1.

Compromise.
Or just make a water table so you don't have to worry about dust :-)
.
If you do go for ballscrews, then you are going to have to ensure they're well protected, whereas R&P doesn't have to be. Yes you'll get backlash using R&P, but you need to remember that Plasma is not that accurate a method, so even 0.5mm of play isn't going to have much effect on most items.
.
Personally, I'd be designing the table so it sits above the side rails, that way the only thing in the way during loading would be the gantry. You then only have to handle dust from above, instead of spray back/dust coming up the way as well. You will still get some dust coming up the way, but hopefully a lot less.
.
Have you had a look at the swift-cut machines for some ideas?

Thanks,
If i had the cash i would buy that 1250 table tomorrow! It is exactly where i want to be, has the full cut speed range I want, pity the pictures are not clearer ;)

I see that have R&P drive but use V-Rails all round, Can't seem to find that stuff?

I like the below-table idea too for the transport, places more stress on the gantry support but obviously works.

The motors are not much above what i have - they use 4Nm and I have 3.1Nm, both hybrid so thats a good sign.

Now, can i build that table with Linear rail or does anyone know where I can source the V-rails????

Google has the answer - http://www.linearonline.co.uk/acatal..._Bearings.html

(although it did take a few attempts to find it!)

Oh, and their youtube videos give a little bit more detail if you pause at the correct moment ;)

Thats the stuff, thanks

Now i have even more options;)

Trouble is, what way to proceed???

Vee rails with R&P drive
Linear rails with R&P
Linear rails with ballscrews
Flat plate & skates with R&P
Other...

Quote:

---

Originally Posted by Davek0974

Thats the stuff, thanks

Now i have even more options;)

Trouble is, what way to proceed???

Vee rails with R&P drive
Linear rails with R&P
Linear rails with ballscrews
Flat plate & skates with R&P
Other...

---

sorry about this but what are the advantages and disadvantages of these different types??

Vee rails with R&P drive - seems to be pretty common, reasonably easy to build, precision depends on rail mounting, backlash in rack, adjustment of rollers
Linear rails with R&P - very accurate transport, precision depends on backlash again
Linear rails with ballscrews - highest precision, highest cost, hardest build
Flat plate & skates with R&P - a very common method, precision depends on quality of plate used and how its mounted, easy to replace, lowest cost, needs to be kept clean from grit etc

Quote:

---

Originally Posted by Davek0974

Vee rails with R&P drive - seems to be pretty common, reasonably easy to build, precision depends on rail mounting, backlash in rack, adjustment of rollers
Linear rails with R&P - very accurate transport, precision depends on backlash again
Linear rails with ballscrews - highest precision, highest cost, hardest build
Flat plate & skates with R&P - a very common method, precision depends on quality of plate used and how its mounted, easy to replace, lowest cost, needs to be kept clean from grit etc

---

What is backlash?
And when you say highest cost re ballscrews how much are we talking??

Backlash is the slop or lost motion you can get with rack and pinion drive, normally gears are run with a small clearance between the teeth, when used in motion control you take all the clearance out by spring-loading the pinion against the rack so there is no clearance at all, this works well until the pinion or rack wears and backlash comes back. This will cause steps in the cut finish where the direction changes or circles to come out not quite round etc, basically its a bad thing.

Ballscrews can have anti-backlash nuts fitted which are two nuts with tension between them.

Cost for a 1.4m 20mm screw with end bearings and nut in the UK is about £350-£400, length is also limited.

You do realise you can get anti-backlash gears?
Essentailly two gears mounted together and spring loaded out of alignment. You could do something similar using two gears mounted apart with a spring loaded belt between them, however you have to remember high acceleration forces could overcome the anti-backlash spring and allow backlash to develop briefly until the forces reduce again.
Then you don't get prefect motion with a straight cut rack, as you get a slight notchyness as each tooth engages/disengages with a resultant increase/ (there is a proper term for this, but it's essentially what makes straight cut gearboxes so noisy), but for a plasma cutter it's not really an issue. If you're really concerned, you could use a helical rack, but I wouldn't worry about it.

Thanks, i have read of some innovative ideas re backlash removal.

Without sounding rude to anyone here, I have just had a reply from a very well respected heavy-duty expert in plasma cutting process and machinery via email, his examples of work are what i would like to see and are well known.

Due to my budget constraints, it is felt that I should be looking at V-Rails and rack and pinion, or, ground, flat plate and skates with rack and pinion for my build.

This makes a lot of sense and is in line with 90% of the commercial production built machines available - no point in re-inventing the wheel.

I will price up and sketch out a v-rail build.

There is another option for drive - open-ended toothed belt with an omega loop drive - this is simply a length of timing belt stretched tightly from front to back with the motor on the gantry and the belt wrapped in a loop over the sprocket. It is supposed to be a very good system as long as steel corded belt is used and it's pulled tight.

My existing build uses a belt but its looped which doubles the stretch possibility and its too small at only 10mm wide, also due to the poor design at the shafts, i cannot stretch it very tightly.

Its too easy to forget things like a budget at times :(

Need to get back on track now I think ;)

In summary -

1 - Go for the belt option again this time, but get it right! I think belts, designed properly would outlast a rack, easy to replace too.
2 - Using Linear rails is a possibility, pretty much same cost as V-rail but easier.
3 - This time I will go for two motors on the Y-axis, this will remove having a long cross-shaft in the gantry and the resulting torque twist it suffers, at a risk of racking the gantry, but plenty of builds have two slaved motors and work ok.
4 - Keep the rails above the table, with a gap under for sliding in bigger sheets. Having the rails up high means a low gantry - good for stability etc
5 - Stick with the motors/drives I have, they are Nema23 size so easy to upgrade to Nema23 servos later on if needed.

Out of your list, personally I'd say go for two servos now for the twinned axis, rather than buying an extra stepper/drive.
Servos are what is going to make the biggest difference in achieving higher speeds while still maintaining accuracy.

Hmm, would be an idea but i already have the spare motor and drive :)

I bought an extra last time round.

Torque and RPM...


Is it best to use a small drive pinion on the rack with higher motor rpm or larger pinion with lower motor rpm??


A 20t pinion will move 62.83mm/rev and at 3:1 reduction I will need 429.73 motor rpm for 9000mm/min travel.


A 32t pinion will move 100.53mm/rev and at 3:1 reduction I will need 301.58 motor rpm for the same speed.


I am fairly certain a larger pinion meshes better than a smaller one, but what size would be best considering my 3.1Nm motors.


My existing table is belt drive but works out at 100mm/rev travel and 300 motor rpm for 9000mm/min speed. When i did my high speed test it was running at 450 motor rpm so I know they can reach at least that.


I have a heavier gantry on the bigger machine, but also twice the motor power as i'm using two slaved motors.


Any views guys?

Smaller will give you better acceleration, and bigger higher speed.
If you know what torque you'll have, and the weight of your gantry, then you can work out your (theoretical) max acceleration. As I'm sure you'll be aware, there's no use having a high top speed if you don't have the acceleration to make use of it.

I'm now looking at direct drive, a 17t pinion will give me 53.39mm of travel so I will need 168rpm to reach my top speed of 9000mm/min.

The gantry will weigh at a guesstimate 30kg and will have two 3.1Nm motors, direct driving it.

This might work better as it keeps the motor rpm way down in the high torque area.

Any chance you could illustrate the maths needed for the acceleration estimates???

Quote:

---

Originally Posted by Davek0974

I'm now looking at direct drive, a 17t pinion will give me 53.39mm of travel so I will need 168rpm to reach my top speed of 9000mm/min.

The gantry will weigh at a guesstimate 30kg and will have two 3.1Nm motors, direct driving it.

This might work better as it keeps the motor rpm way down in the high torque area.

Any chance you could illustrate the maths needed for the acceleration estimates???

---

2pi/.05339 gives you a 117.68 advantage over a 1m pulley

So 2 x 3.1 x 117.68 = 730N

So a 30kg gantry will accelerate at 730/30 = 24 m/s/s about 2.4G on a good day with a following wind.

Assuming I haven't totally cocked up, someone should check this :subdued:

Thanks, if that is indeed correct then I have slightly more acceleration than needed, even allowing lavishly for other losses:)

I have read that plasma cutting only needs around 0.2 to 0.5G for quality so it does sound encouraging.

So what is the bending force on your gantry from accelerating the Z axis components?
Suppose that is 3kg of motor, slide, torch, screws and stuff. Wild over estimate.
We know that 1kg force will accelerate 1kg mass at 1G because that is what happens when you drop it.
So to accelerate 3Kg mass at 0.5G requires a force of 1.5kgf
Why do you need 30kg of gantry to withstand 1.5kg in X and Y?
Perhaps it is the 3kg in Z you are worrying about?
Perhaps you should take a few gantry length bars, support them at either end, put a weight in the middle and see how far they actually bend.
Maybe 30kg is not necessary after all?
Maybe 30kg is only there to accelerate the 27kg it contributes all by itself?

Ideally, it needs a structure that resists torsion moments - this is where my current machine fails - the gantry is a length of 44x44 structural aluminium which is ok in the beam but poor in tension as the centre box is only 18mmx18mm.

The torch and axis are not balanced and when accelerating it twists one way and the opposite on decelerating - causing ripples in the cuts.

What is ideal is a simple, light structure that ca withstand torsion or twist as well as support 3-4kg in the centre.

What that is, I have no idea, most machines just use box section because it's there I guess.

It needs a thick enough wall section to take 5-6mm screws firmly so I guess 5-6mm thick.

Without getting a structural engineer on the job, what stock item would be best suited here?

When I started making one I got a piece of 4"x4"x10g aluminium box and made the gantry as the long axis. Across the front I bolted a linear rail, using the bolt from behind kind so the dust could not get inside the block via the bolt holes. I used 4 1/2" bolts and spacers so I could bolt everything from behind without crushing the box. I cut the box ends dead square. I inserted 1/2" square plates with 8mm steel nut inserts dead centre from either end. The gantry end plates had 4" x 4" recesses and an 8mm bolt hole. When I tightened that sucker up it became incredibly rigid. The gantry end plates got double round linear blocks running on 30mm round linear rails which held the assembly in place simply by being bloody heavy. I liked my design but it all went horribly wrong because the 5mm pulleys were such a crap fit on the 5mm belt. I lost heart.

I tried to add a picture but failed miserably...

The message you have entered is too short. Please lengthen your message to at least 10 characters.

http://www.mycncuk.com/attachment.ph...id=15642&stc=1

That looks ok, the 5mm belt is very light, more suited to 3d printers i think.

What happened in the end??

Dave I'd stick with belt drive on Linear bearings and just protect it good. V-pulleys are a pain in the arse on routers and with Plasma dust etc I'd think they would be a just has big a mare.

Belt with give you best of all worlds. Higher resolution, Higher efficiency, less maintenance, Smoother action. R&P is ok and works but more to building it to do properly and it becomes expensive. Belts will be cheaper and easier to replace if damaged.!

If was using a Ballscrew (which I probably wouldn't) I'd put it down the middle and run it thru a tube tunnel with slit in bottom. I'd use profiled linear bearings on underside like you have shown.!

Thanks, I have opted to try a direct drive version now, linear guides, rack and pinion with 15T pinion.

Part of the issue I have now is caused by the belts so I don't want to add that back into the mix.

I have also got some larger motors coming - 4Nm with DSP drivers and a higher voltage PSU, that all means i can build more of the new table before removing bits from the old one.

Ballscrews are out for the present.

If the direct drive fails then an easy upgrade to belt reduction boxes is made. Rack and pinion seems to be the industry standard, a few are direct drive, we will see what happens I think.

The gantry is looking like 50x75x3mm steel, with a strengthening strip inside where the rail and rack screws on.

Aluminium was looked at, and is lighter, but stiffness suffers as a result and I want it rigid this time ;)

Frame bracing???

First frame idea, overall size is around 1500mm x 1500mm x 1000mm to bed level.

Should i be looking at bracing or gussets anywhere???

http://www.mycncuk.com/attachment.ph...id=15647&stc=1

Quote:

---

Originally Posted by Davek0974

Aluminium was looked at, and is lighter, but stiffness suffers as a result and I want it rigid this time ;)

---

Aluminium has the huge advantage that it is usually quite straight. Steel box section can be quite horrendously bent, as far as CNC building goes, but still well within tolerance. Even slightly bent can make it hard to align a rack. I agree with Jazz, belt for a plasma cutter.

I find the thinner wall box to be pretty straight usually, especially down in the 2-3mm section.

I have belts on my existing table and they cause issues, not solely down to the belts but as 90+% of commercial builds use rack, I am going that way first this time.

Quote:

---

Originally Posted by Davek0974

Part of the issue I have now is caused by the belts so I don't want to add that back into the mix.

---

Very much doubt the Belts caused the issue.!! Belts are super accurate and very close to ballscrews for efficeincy so if you had any issues they probably came from how it was designed or put together rather than the belts.

One thing I do know is that an old mate of mine builds Plasma Machines and his first test machine used Direct drive R&P and didn't work very well. Low resolution and low efficeincy needing large motors, think it used 8Nm or 12Nm Nema34.?
Anyway it was crap.! So then went to sprung loaded into rack with 3:1 ratio using timing belts with Nema 34 motors. This worked but he couldn't reach high feeds as the ratio and slower spinning motors worked against him. He was also having issues with R&P binding etc at higher feeds.

Last machine I seen was using gearbox's and what looked like Nema23 motors.!

Reality check time...


How do you decide what table is the right one?


I know the 2'x2' i have now is too small, it works ok for most of my jobs but i can only cut with the plate square on the bed so no quick twisting it to get another cut out etc.


I can physically fit the 4'x4' (1250x1250) into my shop, thats ok.


I cannot lift a sheet of steel that size, have little possibility of lifting gear etc. as its only a double garage. A sheet of 3mm which is my most popular is about 36kg so only just about ok but the other one i use a fair bit of is 4mm and would be 50kg, no way i can lift that.


Would it be better to go 3'x3' instead?


Cost difference in build is minimal, but what makes sense - having the real estate and not using it or saving floor space etc??