Some people build Models, others build CNC PLASMA machines

[joco-nz]

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

Don't worry mate I'm captivated by all the techie stuff. I am intrigued where you are going to fit this monster. Clearly not in your shed.

I am looking at what you are doing and using it to inform ideas for when I convert my BF20 to CNC.

Looking forward to videos of first automated cuts. :thumbup:

 

[rodw]

Don't worry mate I'm captivated by all the techie stuff. I am intrigued where you are going to fit this monster. Clearly not in your shed.

I am looking at what you are doing and using it to inform ideas for when I convert my BF20 to CNC.

Looking forward to videos of first automated cuts. :thumbup:

Yes, currently living in the garage beside my wife's car. I'm sure you can relate to that!

Can you imagine how momentous the first automated cut will be if I got excited by hearing the torch on relay click!

D day won't be far away. Definitely before Easter.

Personally, I would recommend that you bypass Mach3 as its well and truly obsolete and go with LinuxCNC. Not knowing Linux is not an impediment as they have a complete .ISO install image that installs Linux and LinuxCNC at the one time.

There is also the stepconf program included which will build your config in a graphical environment and for a simple machine no hand editing of files will be required.

For a mill your size, I would not bother with the Mesa hardware. I'd grab something like this
http://www.omc-stepperonline.com/3-...r-m542t-driverr-kit-for-cnc-router-p-183.html
(My 3rd order from these guys is due to arrive on Monday. ) For you, just order from their global site and skip any GST. It just takes a few more days to arrive.

Grab any Parallel port breakout board. Unlike Mach, there are not any restrictions on parallel ports and one on the motherboard will work just fine. Look for a Celeron J1900 board or similar that runs on 12 volt power and treat it just like another component to wire into your enclosure. Any more grunt and you are wasting money. 2 mb is fine but hard to buy now so I'm using 4 Mb.

Breakout boards start at this one which I have here (Came free with my stepper drives)
http://www.ebay.com.au/itm/breakout...762289?hash=item3d204a8f31:g:xMgAAOSwqfNXjxQq

This is pretty crappy, a better board is a C10 which is not much more and has bidirectional ports so you can add additional inputs.

http://www.ebay.com.au/itm/C10-BI-D...015749?hash=item237af95445:g:w6AAAOSwJ7RYScNZ

Cheers mate, wear some paint off those machines you've bought and turn of that infernal plastic extruding contraption!

 

[joco-nz]

Yes, currently living in the garage beside my wife's car. I'm sure you can relate to that!

Hell yeah - that just described my whole shop. :-D

Personally, I would recommend that you bypass Mach3 as its well and truly obsolete and go with LinuxCNC. Not knowing Linux is not an impediment as they have a complete .ISO install image that installs Linux and LinuxCNC at the one time.

Cool - I'm a Linux guy from back when configing and building your own kernel was the norm. Very comfortable with Linux and associated techs.

Cheers mate, wear some paint off those machines you've bought and turn of that infernal plastic extruding contraption!

I will. I have made a few supporting bits and pieces with the Lathe but in the middle of the mill table build at the moment. Lots of cutting steel and MIG sparks. Once mill is operational then have a nice little beginners engine to make. Plans all printed and ready to go.

 

[ShopShoe]

Rod,

Thanks for the answer. I chuckled at your chuckle.

I like what you are doing with the high-tech stuff as well and I have a background which allows me to follow your descriptions of what's going on (just). I would love to start building something using Linux CNC myself, but the toy budget is in slow growth mode right now and life is in the way of much shop time. I am working on a stepper-driven indexer and rotary table and have made use of your posts on that topic.

Specifically, I like the panel layout and the way the lights work to keep the operator aware of the operating state of the system. The "Traffic Light" should work well in a multitasking shop (shed?).

Keep Building: You're keeping my interest.

--ShopShoe

 

[joco-nz]

Rod - looking at the h/w you have suggested for a "light weight" setup versus the more industry path I believe you have taken, if you were to summarise, what do you see as the key advantages that the Mesa controller kit offers?

Cheers,
J.

 

[rodw]

Rod - looking at the h/w you have suggested for a "light weight" setup versus the more industry path I believe you have taken, if you were to summarise, what do you see as the key advantages that the Mesa controller kit offers?

Cheers,
J.

Just a few.

1. 24 volt system not 5 v to combat noise often prevalent in industrial environments particularly on plasma system. No one uses 5v in the real world
2. Unlimited inputs and outputs by addding additional daughter cards for different purposes.
3. Built in jog wheel encoders
4. USD $69 torch height hardware supported by LCNC vs up to USD $1000 systems
5. Ability to use ethernet to separate PC from plasma noise (optional)
6. Mach3 is a dead and obsolete environment.
7. Parallel port is dead and obsolete
8. Windows cannot offer a real time environment, LCNC can.

So never having used Mach3, even with a mate in my street who knows Mach3 backwards after building his own CNC lathe, I could not see any compelling reason to go with it.

I chose LCNC over Mach in the first instance. I had a go at using a parallel port breakout board and failed so when I kick started the project again, I upgraded to Mesa. I now know that a faulty ribbon cable I made was the reason why the parallel port did not work initially.

The choice to go to ethernet substantially increased the complexity of the Linux install but thats another story.

For a small mill or lathe, a parallel port would work fine without spending $200 on a smooth stepper card and buying. M3 license.

 

[rodw]

Rod,

Thanks for the answer. I chuckled at your chuckle.

I like what you are doing with the high-tech stuff as well and I have a background which allows me to follow your descriptions of what's going on (just). I would love to start building something using Linux CNC myself, but the toy budget is in slow growth mode right now and life is in the way of much shop time. I am working on a stepper-driven indexer and rotary table and have made use of your posts on that topic.

Specifically, I like the panel layout and the way the lights work to keep the operator aware of the operating state of the system. The "Traffic Light" should work well in a multitasking shop (shed?).

Keep Building: You're keeping my interest.

--ShopShoe

ShopShoe,

Speaking of High Tech, this is where the idea for the lights came from. My old workplace, designed by me from the ground up. You can see a stacklight mast on the machine out of focus behind Mark's head at about 2:30 into the video.

https://youtu.be/PSbTux4CY3U

The machine in the background is one of 2 black and white devices we had running at 160 pages per minute. There were capable of running at 320 copies per minute and speed limited by a software licence key that cost about $80k to remove. We also had a colour device that ran at 70 pages per minute. The paper punch is a 100% mechanical device, no electronics anywhere! A testament to German engineering. All up about $1.0m of equipment.

We achieved 99.7% on time delivery in a 48 hour window, way beyond industry best practice. I think Xerox themselves worked on about 90%. Maybe thats why there is none of their equipment in the factory.

 

[joco-nz]

Thanks for the pluses for the Mesa kit. I do rather like the look of it but then I like robust solutions. I'll focus on getting the mill working manually first then worry about this stuff. I did find a local supplier who do some of the Mesa h/w. The same lot I got a lot of my 3D Printer gear from. A good bunch.

http://www.makershop.co.nz/CNC/Mesa

 

[rodw]

Thanks for the pluses for the Mesa kit. I do rather like the look of it but then I like robust solutions. I'll focus on getting the mill working manually first then worry about this stuff. I did find a local supplier who do some of the Mesa h/w. The same lot I got a lot of my 3D Printer gear from. A good bunch.

http://www.makershop.co.nz/CNC/Mesa

Its great you have a local source. I got mine from John Thornton in the USA, one of the key Linuxcnc forum members from here. http://mesaus.com

John has contributed some of the LCNC plasma torch height control code. Freight was reasonable and service was prompt. He has a lot of LCNC resources here http://gnipsel.com

I'm still not convinced on a small machine that you need to spend the money on Mesa as you can get by with a lot fewer inputs and outputs. Eg. Shared limit switches or even no limit switches. It does make life so much easier to wire up though!

 

[joco-nz]

Its great you have a local source. I got mine from John Thornton in the USA, one of the key Linuxcnc forum members from here. http://mesaus.com

John has contributed some of the LCNC plasma torch height control code. Freight was reasonable and service was prompt. He has a lot of LCNC resources here http://gnipsel.com

I found John's site in the last 24hrs, quite useful. Even though I found a local supplier I can still land the gear from the USA cheaper than sourcing locally, so long as I keep under the GST radar.

I'm still not convinced on a small machine that you need to spend the money on Mesa as you can get by with a lot fewer inputs and outputs. Eg. Shared limit switches or even no limit switches. It does make life so much easier to wire up though!

Well there is what I could get away with then there is what I want to be able to do. I quite like the idea of a 5th axis. I definitely want limit switches up the wazzo. I now what a pain crashing a 3D printer can be due to only having one side of an axis limited. I'm not taking that risk on something spinning at high RPM and costing a hell of a lot more dollars to fix if it really goes bad. Then through in half formed thoughts of a automated tool change, even if just to experiment with and the Mesa gear starts to look like the way to go.

I am intrigued with the ethernet over "parallel" cable choice. Is that really more relevant to the plasma cutters due to the level of electrical noise they throw out? It looks a lot tidier from an architecture/design angle .i.e. no extra card in the PC meaning you should be able to go to a really small form factor mini/micro style ITX board with the associated small case.

Cheers,
J.

 

[rodw]

With Mesa, you have a choice of either an ethernet solution (my 7i76e for steppers or its cousin the 7i77e for servos) Or a PCI/PCIe solution. In that case, you plug a card PCI in your PC (probably a 5i25) and then you use a 7i76 or 7i77 daughter card which connects to the PCI card via a parallel port cable. The PCI interface has multiple DB25 connectors for the daughter cards.

The 7i76e is basically a 5i25 and 7i76 rolled into one card connected by ethernet. It supports 2 daughter cards so I could add another 2 x 7i76 cards which would let me control a total of 15 steppers. I think then I could add a RS485 interface board that would allow me to add another 4 I/O cards to increase the number of inputs and outputs if required.

Aside from loading a different driver, PCI vs ethernet is totally transparent when you use it.

Tommy my LCNC plasma mentor on the forum says that putting your PC in there with the electronics with plasma is a surefire recipe to spend countless hours troubleshooting recurring random problems due to noise.

Most Plasma system use a high frequency start that generates a high voltage, high frequency spark to kick the plasma off. Mine has a blowback torch where the arc is started in the torch and once established, the tip blows back or retracts and thus does not generate the noise or HF. That style is much preferred for CNC systems as few torch height systems can handle the HF stuff. Mesa's THCAD card can...

 

[rodw]

i.e. no extra card in the PC meaning you should be able to go to a really small form factor mini/micro style ITX board with the associated small case.

Cheers,
J.

Sorry. Did not address this. Yes, I'm using an ULTRA SMALL FORM FACTOR PC (USFF).

The only issue is that to get support for hardware, you need to run a later version of Unix than the LCNC distro. This for me meant compiling a newer kernal and then of course compiling LCNC from source.

The kernel upgrade can be reduced to a single script but LCNC compiling takes a bit of work due to the large number of dependencies that need to be installed.

The onboard wifi cards on these cards cause terrible latency so you need to use a USB wifi dongle to connect to the internet as the ethernet port is in use. IF I built another, I would find a USFF PC with a dual network interface.

 

[joco-nz]

Thanks Rod. When you get back to posting on build progress I, and perhaps others, would be just as interetsed in the software/electronics side of the build as the medhanical. E.g. Wiring diags, models/versions of MBs/CPUs used, performanc characteristics/jitter/etc, good suppliers you have discovered.

A lot of what you a working through on the LinuxCNC side and associated electronics and controls I'm very interetsted in as I can see its application to any router style CNC table and much that can apply to a LinuxCNCing of mills or lathes.

I just cant let the wife get a hint of another possible project fasination right at the minute. :hDe:

Hanging out for more build posts. :thumbup: :bow: th_wwp

 

[rodw]

I just cant let the wife get a hint of another possible project fasination right at the minute. :hDe:

Hanging out for more build posts. :thumbup: :bow: th_wwp

Sorry, but I'm writing this on a Chromebook that crashed and lost my post. Very Annoying!

My wife would die if she knew what I've spent on this project!

The detail is on the LCNC forum here
https://forum.linuxcnc.org/show-your-stuff/32029-rods-spaceship-scratch-built-plasma-cutter-build

Make sure you follow the links I post as many of them refer to links where I ask a specific question in another sub-forum to make sure the right people see it and help me!

Anyway, today was 36.2 degrees according to my weather station just above my shed which means it would have been 46.2 degrees where I was working in a tiny corrugated iron shed.

I managed to fit the steppers to the 5:1 timing belt reduction drives before I expired in the tin sauna. I emerged in a lather of sweat and was able to plug them in to my control panel which meant for the first time, I had all 4 steppers functioning.

I can't really do much more until I get the gantry ends installed and the gantry up on the table as I need limit switches and homing switches so I can home it. Until I get that done, I can't run any gcode to tune it.

But at least all of the motors run in the right directions.

Off to ebay to order 6 more proximity switches. This time I will make sure I buy the right type (PNP)

 

[rodw]

Well, I have not got any more of the machine built but I have been working on the software side of things which has been quite time consuming

All of the important buttons have been connected to the hardware now which was really hard work. This particular screen has a couple of buttons that I probably don't need (as they are designed to feed into the Gcode but I can set that in the Gcode that is generated from the nesting software. (Usually with plasma cutters, you are cutting multiple parts out of a sheet of steel that are nested and all cut at the one time.) Anyway, I'll probably clean up the interface as time goes on.

I've also installed a new larger (10.5amp) power supply and rewired the NEMA34 steppers to run in parallel mode. This means they use a lot more power but will run at faster speeds so it looks like despite the lower gearing on the reduction gear boxes I bought, they will keep up with the Y axis at 18 metres per minute. Previously, they were flat doing 10 m/min. So I am happy with that!

I've also spent a fair bit of time researching how PID controls work. Now I've got into it, the torch height control in LCNC is pretty basic but people say it works well. I've got some ideas on how to recode this but time will tell if its necessary or if it actually improves performance.

Hopefully I will have some laser cut parts next week and the fun will really begin!

 

[rodw]

Its been a long time since I've updated this thread but a lot of progress has been made.

My plasma cut parts turned up

and the

parts all mounted up as they should have

very pleased about that

And I got these huge 1.8 metre long 12mm thick rails that had abojut 50 holes I needed to drill and tap to moiunt the gear rack and the 25 mm linear rails. The small part will become the home and limit switches based on my already proven design.

The gantry ends are folded from 6mm aluminium plate

and whilst its hard to see, the pinion clearance was perfect.

Then on Saturday, I borrowed the work ute and grabbed all of the steel for the table

Once it was cut up, it did not look to menacing. The table frame is 100mm x 4mm wall SHS.

and I managed to get one of the gantry ends mounted up and moving.

I had to drill and tap a couple of holes in the stepper mounting plates that are carefully aligned with features in the casting. This let me mount a piece of ally angle to mount the stepper wiring plug to.

I had a lot of problems calibrating the movement but manged to sort it out by trial and error against a 1 metre steel rule. On the previous rack I installed, I just worked out the maths and it worked out perfect. It is clear this rack and pinion was non standard. Anyway, once I got it sorted, I sent the gantry up and down for 20 metres as seen in this video

https://youtu.be/r8YuW3JrFNE

To find I was only out 0.02mm so 0.001 error per metre will do me!

So I think I will call that a wrap!

Today, I went and bought some 330 kg caster wheels and need to work out a couple more things before I start welding up the table. ITs probably going to weight about 400 kg when done so I had to have some casters other wise I would never be able to move it.

Also, on Friday evening a chap from the CNCzone forum tracked me down to see my progress. I think he got a bit of a shock how much work there is in a project like this. (I seem to remember I did too a year or so ago!)

I also found a bug in Linuxcnc and I spent a lot of time sorting that out in the code and on the weekend, the responsible developer released a fix based on the code I had written. I am pleased about that as I did not want to be maintaining code.

I have to say I've become quite a Linux nerd through this and have even checked out different development branches of the code that contained some cool new features that I am sure will work its way into an improved torch height control in time.

Anyway, that brings you up to date.

 

[rodw]

Well, I still have not tackled the table yet as life has been a bit busy. I also have been designing the last few parts for the table because this thing will be so heavy, I need to be able to get it on the 330 kg casters I've bought for it to move it round.

So I made a mount at the stepper end to hold a wiring socket. This required drilling and tapping the stepper mounting plate

I don't think I showed you guys the air regulator with the dessicant to dry the air.

After the air goes though these filters, it will also go through a Motorguard "toilet paper" filter which are often used on plasma cutters.

Which reminds me, I have not worked out how these will get mounted to the table

Somewhere along the line, I found a super cheap welding fume extractor with a 5 stage filter but I don't think it will have enough airflow for the plasma table.

Last weekend, we went to see Adele here in Brisbane. 60,000 people the largest concert ever staged in Brisbane and the first hosted at the Gabba cricket ground since about 1978. It was really enjoyable.

Sunday, I wired in the last stepper driver and soldered up the cables to it.

I did a few CAD drawings in the evenings this week and I checked them last night and thankfully I found the error before I got it laser cut!

One of the things I had to check was the voltage divider on my Everlast plasma cutter. It has the option of raw volts or divided volts at either 16:1 or 50:1. The divider is selected by a jumper on the CNC board inside the plasma cutter.

Plasma voltage maxes out at around 300 volts. My Mesa THCAD card has a 0-10V range so 16:1 was not going to work and 16:1 would not have great resolution as it would only go from 0-6 volts. I had this great idea to use a voltage divider so I could use the 16:1 divider. This would give me a full scale reading of about 9.375 volts so it will be so much better.

Peter from Mesa set me straight on the LCNC forum about how to do this. You don't need a traditional voltage divider circuit, you just need to add a resistor in series with the voltage signal. The value of the resistor is based on a formula in the manual. My Maths was way out and Peter kindly stopped me from buying a 0.1 ohm resistor when the correct value is 100,000 ohms!

It is rare to see a manufacturer to be so involved in a forum and the support Peter offers is simply amazing even building custom firmware for people using his hardware.

Last night, I pulled my plasma cutter apart and checked out what voltage was selected

.

As it turns out, I did not need to becasue it appears to be set to the correct value based on the position of the jumper on the bottom right. I still have to put it together again. This morning, I had to drive past Jaycar electronics to pick up some material for my web site (10 metres of 60mm dia plastic that needs to be cut and faced into about 180 parts). So I stopped and lashed out all of 55 cents to buy a packet of 100k resistors.

So there you go guys, you are now right up to date and I am ready to build the plasma control cable now. I will have no trouble tucking the tiny resistor inside the cable backshell.

 

[rodw]

Progress has been pretty slow. While I can kinda weld, Lee at work does it for a living so there is no comparison between what I can do and what he can do. Finally, the table is coming together.

The jig we are using is one we use to make truck bodies for utes and pickups. To get the linear rails square, we welded up one side as square as we could and then mounted the linear rails. Then I fitted some angle iron brackets to the linear carriages and welded up a crossmember that spanned the two linear rails. YOu can see it in the pics. Once we got the second side clamped in position, I could push the rails along with one finger. So at that point, we tack welded the second side in position.

Unfortunately Lee has been pretty busy so has not had the time to finish this off.

I've also been plagued with some problems with spurious limit switch errors on the gantry sensors (there are 7). I think today I found the problem.

Circuit boards are meant to be flat right? I have to rethink how to do this...

 

[joco-nz]

Yeah they are meant to be flat. That really doesn't look healthy. Was it on a mount? Or was there some serious tension on that plug?

 

[joco-nz]

Rob - can't see if this was covered in your prior posts. If so just say and I'll re read in detail. For the plasma torch, is it using a pilot arc or HF start? Asking as my little 40amp job is HF and wanted to know if cnc for it on the future was viable or not. I would be looking at a SMALL table, 1m x 1m max size. But not sure if viable with HF start. Hence the question to draw on your experiences.

Cheers,
J.