Some people build Models, others build CNC PLASMA machines

[rodw]

This thread might take a fair while to develop. Its been 12 months in the making already. My Arduino Rotary table controller project got me interested in CNC but I did not want to do the obvious and convert my mill to CNC. I eventually got interested in CNC plasma cutting and I started about 12 months ago by cramming the electronics into an old Plasma machine case complete with a Torch height control. There are 3 layers of electronics in the case.

I got very disollusioned becasue my cheap plasma cutter died and the parallel port Break Out Board refused to talk to the PC.

So it sat on my desk for the next 10 months or so until one day I remembered how much time and effort went into it so I decided to kick start the project again. So in June, I purchased a 50 amp Everlast Plasma cutter with CNC interface and a CNC torch for it.

After hundreds of hours designing a machine, it is finally coming together. After machining 20 or so parts, finally, the Z axis is complete and I am sure it will work nicely.

Originally, I was going to make a ghetto build machine but since I got the idea, I've purchased an interest in a business that could use a plasma cutter so I am sparing no expense to do the job right. As I am surrounded by fabricators, the table will also be of commercial quality.

These are just quick pics taken with my phone, I will post up some better ones one day if I find the time to pohotograph it properly.

It has a 150mm ball screw which gives total travel of about 100mm. You can see that there are 2 linear rails. One is for the Z axis itself, and the other is a short floating section that senses the material.

Plasma is a bit different to other CNC machines as there are no cutting forces. Here is what happens at the start of a cut.
1. The torch rises until it trips the home sensor which you can see below the black bearing mount on the stepper side.
2. The torch hits the material and moves up along the floating rail. This trips the upper proximity sensor.
3. The torch height is adjusted for the switch hysterisis.
4. The torch raises to cutting height (around 250% above cutting height)
5. The controller fires the torch.
6. Once the Arc is established, the Plasma cutter sends an ARC OK signal back to the CNC controller
7. The torch is lowerd to the cutting height and off it goes.
8. The CNC controller monitors the arc voltage sent back though a voltage divider from the Plasma cutter as the voltage is a function of torch height.
9. The torch height is maintained constant based on this voltage as parts can tip up or warp while being cut.

The red micro switch is a limit switch so the ballscrew does not fall apart. In theory, it should never be tripped as the material sensor should always be triggered before it.

The other thing that could happen is the torch could hit something that could damage the machine. To protect against this, I've added a breakaway magnetic Torch mount.

This has 5 x 12.5mm dia magnets that are 3.5mm thick. These have been pressed into holes cut with a 12.5mm endmill so they are just proud of the mounting surface. They are secured by 3mm countersunk screws. The locating pins have been machined using my boring head ball turner documented elsewhere on this forum which are a press fit (helped with some retaining compound. This is the second iteration of the magnetic breakaway mount. Initially I wanted to use a pair of North and South oriented magnets so that one magnet was sitting below the surface and the other located into the recess. I found the magnets were not dimensionally accurate and also turned out to be a press fit. Breaking an M4 tap on th elast hole was the final straw but I decided the part needed a total redesign and went with a steel plate located by the pins.

If the torch ever falls off, it makes sens to stop the machine immediately so I've drilled and tapped the back side for another proximity sensor which will return an e-stop if it ever triggers.

The reason why the torch mount is a lot lower than the carriage is that I hope to be able to build a cover for the axis to keep out metal dust.

This machine will be controlled by LinuxCNC and I have purchased an Ethernet Mesanet 7i67E interface card and the matching MESA THCUD torch height controller daughter board. Unfortunately, the Mesa board is significantly larger than the Breakout board in the yellow case, so everything needs to be stripped and remounted into a 500mm x 500mm enclosure I've purchased.

Today, I managed to drill and tap 44 x 4 mm holes into a piece of Aluminium RHS on my little SX3 mill. This proved quite a mission as the X axis handle sits proud of the mounting table so the RHS had to hang off the table to get about 2mm clearance from the handle.

Its a bit hard to see but there is a length of 75mm x 50mm steel mounted to the table to act as a fence so I could slide the material along. I used a short piece of scrap linear guide when I had to reposition the material (table travel limited me to 7 holes at a time and there are 22 on each rail.). I've screwed this scrap down with 2 screws and located the undrilled hole with a close fitting punch. This worked pretty well but on the first side, I started at one end and was out a bit by the time I got to the other. On the second side, I started in the middle to halve any error.

The problem with this setup is that the holes have been referenced on 2 different sides so I don't really know exactly how far apart the rails are. That won't be hard to sort out.

I clamped one rail to the back fence when mounting it. The steel fence is dimensionally accurate, but I've forgotten the accuracy spec. The second rail is still loose and will be tightened up once I get a carriage mounted up.

I'll try and post up a few more photos tomorrow as I hope to get the Y axis close to finished. The plan is to have 2 linear rails on the front face of the RHS and a gear rack for the rack and pinion drive mounted underneath to shield it from dust as much as it can be. This is going to be a bit of a pain as its designed to be mounted from the rear into M6 holes in the rear of the rack. I'm going to have to drill access holes on the top surface to fix the rack. I'm glad thes holes are much further apart than the 60mm on the linear rails!

 

[Cogsy]

Good luck with your build - I know how tricky they can turn out to be. Here's a pic of one I built a few years ago.

 

[rodw]

Good luck with your build - I know how tricky they can turn out to be. Here's a pic of one I built a few years ago.

Cogsy, thanks for the encouragement. Hopefully all the bad luck got left behind last year in the yellow box. My table will be similar to yours.

Did you have to add gearing to your steppers for the rack and pinion or did you get away with direct drive? My pinion will give 30mm per revolution and I was hoping to get away with a direct drive.

I thought after I bought my linear rails that it probably only needed one based on their torque holding specs.

Anyway, I did not have as much time as I had hoped today but I wanted to get the rails in position and confirm they were parallel.



Here's what I ended up with.

The part connecting the linear carriages is just some scrap as I did not want to disturb my mill setup until I have mounted the gear rack.

I used transfer screws and a tap with a hammer to locate the holes. There is only 6 in the set but more than enough for this exercise.

Once this plate was nice and secure, I went along and tightened up the second rail one bolt hole at a time with the carriage positioned right beside it to set the width.

It turned out really well. So happy, absolutely no binding along the full length of travel (1240mm Y axis width).

So now all I have to do is to mount the gear rack, make the mounting plate to replace the temporary one, finalise the stepper motor pinion drive setup and mount the Z axis to have the Z and Y Axis done and dusted over the next week or 2.

 

[ShopShoe]

Rod,

That's an impressive project. For all-around work it looks like it's going to be a very useful tool. I'm looking forward to posts with some of the things you will make with it when it's done.

Thanks for posting.

--ShopShoe

 

[Cogsy]

Did you have to add gearing to your steppers for the rack and pinion or did you get away with direct drive? My pinion will give 30mm per revolution and I was hoping to get away with a direct drive.

Both X & Y were direct drive via timing belts, no rack and pinion at all (although I'd have preferred it). Z was just direct drive ball screw. You probably know it already but you want pretty much the fastest travel you can get out of it in X & Y so try for direct drive if you can.

A nice blast shield around the torch nozzle really helps keep the grit out of the drive systems and allows you to watch the cut without going blind as well.

One of these days, when I have the time, money and somewhere to put it, I'll make a half-sheet one for home.

 

[rodw]

Rod,

That's an impressive project. For all-around work it looks like it's going to be a very useful tool. I'm looking forward to posts with some of the things you will make with it when it's done.

Thanks for posting.

--ShopShoe

Thanks for the encouragement. I'm trying hard to make it commercial quality all the way.

Both X & Y were direct drive via timing belts, no rack and pinion at all (although I'd have preferred it). Z was just direct drive ball screw. You probably know it already but you want pretty much the fastest travel you can get out of it in X & Y so try for direct drive if you can.

A nice blast shield around the torch nozzle really helps keep the grit out of the drive systems and allows you to watch the cut without going blind as well.

One of these days, when I have the time, money and somewhere to put it, I'll make a half-sheet one for home.

Cogsy, Thanks, you see so many with reduction gearing but the maths say to do it pretty direct. I did see it said the ideal speed was around 1" per rev so 30mm per rev is pretty close to that.

Z axis is running a 5mm pitch ball screw.

Been busy the last couple of days so no progress, I need to think out the Y axis wiring and stuff before I build too much more. Bought some more material today.

 

[mjonkman]

I'll be following this thread. I want to build a combination cnc plasma cutter and cnc router next spring. I don't have room for one that could take a full sheet of metal so I'm planning on building mine with a cutting area of 4'x4' which is probably still too big for my small shop but I'd hate to go any smaller. Definitely be trying to understand your design to see if I can pickup any tips for when I start designing mine.

 

[Wizard69]

I'll be following this thread. I want to build a combination cnc plasma cutter and cnc router next spring. I don't have room for one that could take a full sheet of metal so I'm planning on building mine with a cutting area of 4'x4' which is probably still too big for my small shop but I'd hate to go any smaller. Definitely be trying to understand your design to see if I can pickup any tips for when I start designing mine.

Not to throw water on the fire but trying to build a combo machine leaves you with a lot of trade offs that are less than ideal. Plasma machines really benefit from light weight designs that allow for high accelerations and high speeds at low costs. For a router rigidity is more important as you have to counter real reaction forces from the cutter plus surface finish quality increases with rigidity.

Im not saying it can't be done, just that you need to realize that if you design for one use you will loose for the other use. Of course your expectations are a factor here.

By the way I understand the frustration of a small shop! I'm stuck in a basement at the moment and need to work miracles just to make use of a bunch of rather compact metal and wood working tools. If you are already tight on space I'd suggest laying out the machine in a cad view of your work shop. Either that or build a CAD (cardboard aid design) template of the foot print and see how it will fit your shop. Actually a full so all card board template will highlight usability issues right away as you can walk around the "machine" to get a feel as to how convenient a real unit will be.

I only bring this up because siting a machine creates huge problems in small shops. For example I bought a small jointer as I don't expect to be doing massive wood working projects. Even so I ran into usability issues because even with a small jointer you need three side access.

By the way there is the option of placing a plasma unit outside. It would have the advantage of fire safety and you can build any size you want. Build the majority of the machines frame from aluminum and rust isn't an issue.

 

[rodw]

I'll be following this thread. I want to build a combination cnc plasma cutter and cnc router next spring. I don't have room for one that could take a full sheet of metal so I'm planning on building mine with a cutting area of 4'x4' which is probably still too big for my small shop but I'd hate to go any smaller. Definitely be trying to understand your design to see if I can pickup any tips for when I start designing mine.

Further to Wizards comments, the difference between the two is that plasma is built for speed (up to 10 metres per minute) and the router is built for torque so it travels much slower. By comparison, a mate's CNC mill travels at 1.5 metres per minute while I'm aiming for 10 metres. He's still scratching his head about why I'm using rack and pinions when he'd use ballscrews. The difference is 30mm per rev on my rack v's 5mm per rev for a ballscrew.

Whilst a multi-purpose machine can be done, the difference is likely to become a compromise due to totally different design considerations. (light and fast v's slow and heavy).

Regarding full size sheets, I am hoping to have a simple pull out trestle on wheels at one end that will support a full sized sheet when required.

I did not get much done over last weekend other than make a shaft adapter to mount a pinion to a stepper. I needed to do that to work out some key dimensions.

After conferring with my CNC mate and mentor, I tried 2 different Y axis drive designs (top mount and bottom mounted racks) and was still not happy. I think I've worked out how to make a front mounted rack that sits centrally between the linear rails as the best option. There is so much to consider to ensure that limit switches and cables are mounted and routed nicely. If I'm lucky, I'll have it drawn up during the week ready for the weekend.

 

[mjonkman]

If I do build it, it will be built primarily as a CNC router and secondarily to work with a plasma torch. Hadn't really considered the need for speed on the plasma side. Obviously a lot more research to do, just toying with ideas in my mind at the moment.

Mark

 

[rodw]

If I do build it, it will be built primarily as a CNC router and secondarily to work with a plasma torch. Hadn't really considered the need for speed on the plasma side. Obviously a lot more research to do, just toying with ideas in my mind at the moment.

Mark

Depends what you need to cut. Thin stuff needs the speed. Find the Hypertherm 45 manual and look at reccommended speed tables in it before you get too excited. If you are only cutting say 3mm plate as a minimum, you may be ok.

There is a world leader in plasma and routers here in Brisbane and their commercial tables are built on a router gantry and you can add a drilling head with tool changer. But the one they showed me cost $135k and the drill was another $15k....! They started with routers and migrated to plasma and are found in shipyards around the world. The router table design was rebirthed as a plasma design.

 

[rodw]

Sorry this thread has been inactive for a while, but with the Z axis complete, I've really been struggling to work out how to build the Y axis. The main consideration was where to mount the gear rack. I built near complete designs with the rack mounted on both the top and bottom of the Y axis RHS beam but found that when you took into consideration all the wiring and the like, they were not going to work. There is sooo much to consider.

So then, I came up with a design that mounted the rack on the front face between the linear rails.

Through this process, my research indicated that to control backlash, you needed to run a spring loaded pinion drive like this one available from CNC Router Parts.

See: http://www.cncrouterparts.com/images/CRP301-00-01_1920.jpg

But the problem is that this design also introduces a lot of wear in the rack so I really wondered if this was the right way to go.

So eventually I asked some questions on the LINUXCNC forum and as I suspected, people were getting away with direct mount pinion drive even on commercial products. You can follow the discussion on this thread.

https://forum.linuxcnc.org/forum/30...-you-really-need-a-spring-loaded-pinion-drive

On this thread, I proposed this design

And people said I should support the Pinion drive to protect the bearings in the stepper motor. I remembered, I had a surplus bearing block so I modified it

Andy said, just build it and so I made a start,

but stuffed up reading my own drawings so will have to remake the part. So you will have to wait for another weekend.

 

[rodw]

I have not had much time the last few weeks, but I finally got in some quality shed time over the weekend and got this done

This is the drive for the Y axis on the gantry. From the earlier renders, you should be able to work out where this goes.

I am very happy with how this turned out. The shaft, stepper and bearing alignment is absolutely perfect. The pinion turns nice and freely. I'm still not sure if have clearance between the bearing block and the guiderails but the CAD model said I just made it. If I don't I'll mill some of the bearing block away.

It will be a couple of weeks before I can fit in another shed session but there are only a few more parts to make (and one to remake :fan

 

[rodw]

Its been weeks since I've had time to tackle anything further on this.

I made a bit of a mistake with my design. My calculation was out by a factor of pi (3.1417) so when I thought I was turning 30mm per revolution, I was actually turning 94.25 mm! This was going to be a major issue in terms of machine accuracy so I had to redesign the Y axis drive to put in a 3:1 reduction drive system to get it back within the ballpark at 31mm per rev.

You can see I've used a WBK ball screw bearing assembly. You can also se the slotted belt tensioning adjustment.

This reduction drive assembly is a bolt in replacement for the original stepper motor and bolts on where the stepper used to.

I still need to turn up the shaft. Its going to be tricky and I'm not looking forward to it as it will involve cutting an M12 x 1 thread in one pass.

I could not mount the stepper square on as the belt would foul the front legs. I decided to try mounitng the stepper clocked 45 degrees and hope the leg did not foul the large pulley.

I wish I had found the belt feature in my CAD software beforehand. I knew the stepper posts were going to be close to the timing belt cog but it actually fouled the cog.

Anyway, in the revised design, I've clocked the stepper leg mounting plate around a bit further and have clearance now in a revised design that rotates the legs about 20 degrees.

I only need to re-make one simple part that the legs attach to. It has 8 holes drilled and tapped so it is easy enough to make.

One of these days, I'll actually have this gantry complete!

 

[rodw]

I finally got the Y axis pinion drive built, the rack and pinion installed and the Z axis installed. Everything moves quite smoothly by hand. I still need to fit up the Y axis limit switches but know how I'm going to do that now. I will get onto the electronics in the next couple of weeks.

Clearance for the belt

Pinion Drive Shaft detail.

This also shows the full length 20mm x 6mm packer under the rack so the bearing block cleared the guide rail. You could probably avoid the need for this if you made the shaft longer. In practice, making the packer strip and drilling all the holes in it before hand made it easy to drill the holes into the gantry. There was exactly 50mm between the two guide rails. My end stops I had machined to be 15mm high which was the exactly the required distance each side of the 20mm wide packer. I used them to set the position of the packer. Initially, I drilled and tapped holes for each end of the packer and bolted it down securely, then I used my improvised 15mm spacer to align the packer perfectly while centre punching with a transfer punch to get the punch centred on the hole. With all the holes drilled, I the drilled the threaded holes out and bolted it all together.

Because this gear rack has threaded mounting bolts on the reverse side and I was mounting it to a 80mm x 40mm RHS, I had to drill clearance holes right through so I could secure the rack using socket head cap screws using an allen key. A blob of grease helped hold the bolt on the end of an allen key as I inserted it.

To tidy it all up, I added some rubber grommets on the reverse side.

Its going to be a while before this will be moving. There is still a fair bit to do to the gantry to add limit switches and all of the wiring. I've got most of the stuff I need here.

I'm very pleased to get the gear rack mounted up. The supplier said it was 150mm hole pitch but it actually worked out to be exactly 5.75" (which I promptly converted to metric to build the darn thing! I will say drilling 9 holes 146.05mm apart was a bit crazy.

 

[rodw]

Well, its been ages since I posted any updates here. I've had a few trials and tribulations along the way.

I found that I needed to use the preemptive kernel for Linux to get the Mesa Ethernet card working properly and this proved a major tasks. I installed the kernel on my 32 bit PC and the latency went out the window so I had to buy a new PC

Then a week of frustration followed and I finally installed Linux Mint 17.3, upgraded to the later 4.13 kernel, applied the preemptive patch to the kernel source code and compiled the LInux Kernel from scratch. Then I had to resolve all of the dependencies and compile LinuxCNC for the different flavour of Linux. As you can see, this is a massive PC. With the single Network interface reserved to talk to the Mesa card, the internal Wifi card introduced huge latency so I am running a USB wifi card to talk to the world from my garage.

While this was happening, I put a fair bit of work into the enclosure

I added green, amber and red indicator lights. While the machine is running it is green, if it is idle, it is amber and if in estop, its red.

A small blue LED on the far left indicates when the Plasma ArcOk signal comes on.

The red and green switches act as a run/stop and pause/resume switches.

I would like to make the green light flash when its in a pause state but I decided I had too much to do to get it functional.

I must take some photos of the electronics now its almost all wired in.

I got the gantry complete and running at about 21 metres per minute (well over 800 inches per minute for those that are metrically challenged)

The movement is spot on and there is no backlash. I let it loose for 20 metres up and down the gantry and it finished exactly where it started off on.

I decided I better tidy these bolts up

and ordered 20 M5 dome nuts from my bolt shop.

THe bill turned up and to my amazement, what you are looking at is AUD $158.60 worth of dome nuts! Well at least they were solid brass! After a terse email, the price dropped back to a more reasonable $7.60 or so. I was glad about that becasue I had ordered the wrong size and already had the correct M4 dome nuts to hand.

There is an insane number of sensors on the gantry (7 actually) so I decided to cheat and buy a 5 metre DB9 cable and build a breakout board. Using 2 wires for 24 volt power, I had just the right number of wires for the sensors I needed.

I made a bit of a mistake as I bought NPN proximity sensors only to find out that I should have used PNP sensors with the Mesa hardware so I had to add some pull up resistors which are also soldered onto my breakout board as you can see above. I also added a 24 volt LED indicator that pokes up throuh the top of the gantry to help with trouble shooting.

I have 3 proximity sensors in a row

The outer ones are limit switches and the inner one is the home switch.

These see the limit/home switches

The home sensor is a lot lower than the limit switch so it sees the edge of the mounting bracket. Once LCNC sees this, it backs off, approaches slowly to find the edge precisely and then there is a HOME_OFFSET setting that then lets me park the sensor wherever I want. I use it to park it as close to the end of physical travel as I can without th elimits being triggered.

The Torch breakaway sensor has been OR'ed with the external e-stop switch.

So if the torch falls off, the machine stops immediately.

Last week, the 25mm Hiwin linear rails and carriages arrived for the X axis and a day later the these NEMA 34 timing belt gear boxes arrived

They are about 4:1 ratio and they drive a 1.5 module helical pinion that engage with a 1400mm long rack you can see in the left of the photo above.

With all of this arriving, I spent the last few days working flat out on finalising the table design

And this morning I prepared 15 files to hand off to my laser cutters and I've emailed that off for quotes today. I hope it is cheaper than the dome nuts!

I've also ordered the NEMA34 stepper motors to suit the gear boxes so I hope to have the table together by the end of January. Its been a piece of cake sorting out the table and X axis in comparison with the gantry and Z axis. I always knew they would be the hardest bits.

 

[rodw]

Well, things are moving along pretty quickly now. I've ordered all of the laser cut parts to finish this off but with the Christmas break, I don't expect to see them until February which is a bit annoying. Hopefully they will surprise me. The NEMA 34 stepper drives for the X axis have been shipped and should be here Monday. There is about 80 holes to tap. I was going to get the Laser guys to do this but they were pretty expensive so I bought a magnetic hand tapping guide on Friday.

The money I saved paid for the tapping guide and a motorguard air filter that are meant to be the ducks nuts for plasma cutters which don't like dirt or water in their air supply.

I also managed to get one of 2 hand encoder wheels running. The other will just be a copy of what I've done. This is where the Mesa card starts to come into its own as it has 2 MPG (manual pulse generator) encoder inputs on the board.

I've just mounted the MPG wheel on a piece of scrap for now. This is a 100 pulse per revolution encoder. I will use one on the X axis and 1 on the Y axis.

The other part of the encoder is to have a switch to select the scale it uses. To do this you can use a 4 input binary switch and the Linuxcnc mux4 component. This proved quite tricky but eventually I worked it out. Using the mux4 component allows you to support 4 options with just two inputs by using binary logic like this.

0: 0 0
1: 0 1
2: 1 0
3: 1 1

So on the switch, you need to make sure wherever there is a 1, you have a voltage signal so I wired them to 24 volts.

So with this done, I told LinuxCNC to use the following scale settings using the mux4 parameters as follows:

0: Off/Disabled
1: 1 mm per pulse
2: 0.1 mm per pulse
3: 0.01 mm per pulse

And it worked!

So when I do this on the Y axis, I am going to add another switch so I can select between jogging the Y and Z axis. I Wan't going to bother with jogging the Z, but it will be so trivial to support, I might as well as I have not run out of inputs yet!

The other thing I wanted to achieve was in addition to a Pause/Resume switch and a Run/Step switch was a series of warning red, green and amber lights. Ultimately, I want to have a set of stack lights on top of the machine that will act like a set of traffic lights so you can see the machine status from across the room as follows:
Red = E-stop error
Amber = Machine turned on at the LinuxCNC console
Green = solid on - running Gcode
Green = flashing on - in pause or single step mode.

This wasn't easy to achieve after first having the idea but the guys on the LCNC forum helped me sort it out

Anyway, I got it done this morning and here is a video of it and if you are into this sort of stuff, it is very cool.

https://youtu.be/5NYhmYO4mCs

So I'm getting close to completing the LinuxCNC config now. The next step is to configure the Torch height control module. I have not properly set up all of the axes yet, but I don't think that will be hard. Thats about the only 2 parts of the config left to do.

So that brings this thread up to date.

 

[ShopShoe]

Slight OT

Rod,

I don't think I've ever seen a magnetic tapping guide from any of my sources here in the U.S.A. Tell me more about it: It seems like a useful thing to have.

--ShopShoe

 

[rodw]

Slight OT

Rod,

I don't think I've ever seen a magnetic tapping guide from any of my sources here in the U.S.A. Tell me more about it: It seems like a useful thing to have.

--ShopShoe

ShopShoe, I had a chuckle when I'm posting about the 100's of hours of effort and all this cool technology and it is a hand tool that captures someone's interest

Anyway, there is not much to tell as I have not used it yet! I got it from here
http://www.machineryhouse.com.au/T0119
It grabs the flats of the tap between two flat surfaces which slide from side to side.

And the tap is held central by the chuck tightened by a knurled knob.

The sliding holder is held captive by 2 bolts that engage a slot in each half of the vice like tap holder.

The tap is held very securely!

Anyway, here are a few more pics of progress.

I've been putting off mounting any connectors to the base of the enclosure as it will mean I won't be able to stand it up while I play with it.

So all the wires are coming out of the air vent (which is just a plumbing fitting with a grille cover)

There is a heap of diagnostic wires still attached that makes the internals quite messy. Shown here are a large blue LED warning light attached to the torch on relay that replaces the switch on a hand plasma torch. It is enabled by a Gcode M3 (spindle-on) command. That initial relay click was one of the most satisfying sounds I heard in the build! Above it is the jog wheel scale selector switch discussed earlier. Its a bit hard to see sitting on some stepper motor control cable, but there is also a 100k potentiometer I've attached to test the voltage sensing circuit of the THCAD board that reads the torch voltaqe.

The electronics showing from the right, the Mesa 7i76e ethernet card, the matching THCAD accessory board and a switch to mimic the Arc OK signal sent from the plasma machine

Here is a closer look of the Mesa hardware. BLue ethernet cable enters at the top right. Its just a patch lead that leads to an industrial RJ45 connector on the side of the cabinet.

The first section includes the 5 volt section and there are 2 x 26 pin connectors which are unused. These are pin compatible with normal breakout boards and there is a heap of cool mesa cards you can attach to them as well. eg. Adding 2 non ethernet 7i76 boards would let me control a total of 15 steppers.

The next heavilly used conector are 4 of the step generators that go to the stepper controllers. There is a 5th unused step gen on the next connector.

The orange power connector is for 24 volt power. The small green connector is for spindle control and the THCAD frequency output is attached to the encoder input. (The THCAD is a voltage to frequency converter.)

The next 2 24 pin connectors are for field inputs and outputs for sensors and warning lights etc.There are 32 inputs and 16 outputs (maybe equivalent to 3 x parallel port breakout boards).

So the next small board is the torch height sensor. It takes torch voltage from the plasma torch at the top (in my case via 50:1 voltage divider in the plasma itself but it could take raw arc voltage). It draws 5v power at the bottom of the card and where the frequency output also resides. It can output at up to 1 Mhz (1,000,000 times a second) but I've enabled the 1/32 divider on the board as I'm only going to sample it at 1000 times a second.

There is a LCNC software module that takes this input and converts it to a voltage. The idea is that voltage is proportional to arc length (eg. torch height). So if LCNC can read the actual torch voltage, it can control the Z axis to maintain a constant voltage (or height). This lets it handle any warpage from the heat on thin stock or a thicker sheet or table that is not perfectly level. I have about 85 mm of Z axis travel which should be heaps.

So when the gcode says M3 start your spindle, the final setup will lift the torch to pierce height, start the arc and wait for a pierce delay. The plasma machine establishes an arc and sends an arc Ok signal back to the CNC machine and the torch is lowered to cutting height and off it goes.

 

[rodw]

The other thing that I have got working is my warning lights. I needed a little bit of help with these but I think they are very cool.
Watch this video and see what you think.
[ame]https://www.youtube.com/watch?v=5NYhmYO4mCs[/ame]

I have ordered a stack light similar to this one which I want to mount on top of the machine so you can see it all going on at a distance when you are doing other stuff in the workshop.

So at this point, I've got all of the config set up with the exception of the Axes which need some work and the stuff to interface this to the LCNC GUI display.

Some of the setup has been quite hard, and I'm really happy I've finally got it almost all done.