Rods "Spaceship" Scratch built Plasma Cutter build

Tommy, thanks, It was Gremlins as I had it working earlier. Unlike you old hands, I have to read the manual for every setting and had no time. But I appreciated your help as it will save me some time when I get to review it all. I looked at the Lam Technologies stuff but they were hard to find down under. It did look very impressive.

The OMC steppers I'm using are much easier to source as we are so close to China and they hold some stock in my country. On my next build, I will probably use servo motors and step up one more level. I'd probably source them from China. I think there is an opportunity in Australia for a reasonably priced plasma table as there seems to be nothing under $20-$30k. I've got CAD files for every part so the next one would be all cut for me and after tapping a few holes it would go together in no time. But then again, the spaceship might not take off into that orbit.

So far, I am so happy with the quality of the build. Everything is really precise and accurate except for a little bit of sideways slop in the magnetic mount (which I think I have a simple solution for).

The only thing I've had second thoughts about is that as it currently stands, the rails mount to the table legs and therefore it might be very hard to get them square as the legs will need to be perfectly aligned. I might revisit that so they mount to a lateral beam but I can't change the mounting rails.now I've ordered them as they are hundreds of $ each.

Its been annoying that I've had to spend the last 3 nights sorting out network issues but I finally found 5 minutes to look at the problems Tommy diagnosed earlier.

So after slowing the Z axis down top a more leisurely pace, both up and down arrows were moving in identical direction. I realised it had to be a wiring fault and quickly found the loose wire.

Now I've got a homing issue on the Z axis which has a shared limit/home switch at the upper limit of travel
I get a "Hit limit in home state 7" error. This stumped me for a while until I realised that i had forgotten to add.
Now it looks like I've shared the home switch with the wrong limit switch as I had to hit estop as the Z axis hit its upper hard stop. I thought I just have to swap the Z axis limit switches around between min and max but that did not work. Its too late now so I'll revisit it over the weekend.

Here's what I've got in HAL.

Now I see that Z-home has no prior value so I'm guessing joint.3.home-sw-in needs to be set to be the same as min-z

rodw wrote: Now I see that Z-home has no prior value so I'm guessing joint.3.home-sw-in needs to be set to be the same as min-z

I'm glad I went to bed when I did last night. Just a stupid error.

How on earth can an axis home itself if HAL does not have a home switch?
I had not sent the home switch input anywhere, so this did the trick.
I think that means I have both Y & Z axes working so i need to wire up my NEMA 34 steppers and get on with the dual joint axis.

I'm a bit annoyed at my laser cutter. I gave him some files nearly 2 weeks ago and earlier this week he said he could not fold the gantry ends in 10mm aluminium as it would crack (I think that is risk limiting jargon for might crack).

Anyway, I reworked the gantry ends in 6mm and that took hours as I had to rework another 2 parts. in the end, I got it done, but cursed my inexperience with 3D design software! I'll probably weld in some gussets to stop the ends from twisting but I have not bothered about them until I know where all the wiring done!

Racedirector said it was 46 degrees in his shed today and I'd believe it! My weather station only registered 36 degrees here today but in a courrugated iron combination shed/sauna, I'm sure it would have been 46 degrees where I was working...

Anyway, got the X axis stepper motors mounted up
image missing http://i210.photobucket.com/albums/bb270/rodweb/Mobile%20Uploads/2017-01/20170114_174517_zpstqmvgzee.jpg

image missing http://i210.photobucket.com/albums/bb270/rodweb/Mobile%20Uploads/2017-01/20170114_174526_zpsiwt50vpz.jpg

This time, I had the config sorted OK and they worked straight away. Its going to be hard to wait out the time until my gantry ends to arrive.

I've been doing a bit of research on torch height control and I found a journal article that said torch voltage varied during a cut as the "anode spot" (being the opposite and of the arc to the torch tip) varies up and down through the thickness of the material during the cut. They found with some signal processing there was a linear relationship between minimum voltage (eg when the anode spot was closest to the torch - eg. at the upper surface). I'd like to experiment with this but I think with the PC hardware I have has too much latency to sample fast enough to implement their algorithm.

Anyway, they found that there was a ripple in the voltage signal becasue the voltage was sourced from a 60 Hz 3 phase AC power source at around 360 Hz so they needed to ensure that the sample period was large enough to capture this voltage spike and they settled on 40 milliseconds. But to find the lowest voltage, they sampled at 40 microseconds. eg they take 1000 samples per 40 milliseconds and save the lowest voltage they find! So if you are using a 50 Hz single phase plasma cutter like me, I think the rectification ripple will be at 100 Hz (eg. 2x the AC frequency per phase)

So this got me thinking about redirecting the THCAD frequency output to an Arduino to do the signal processing before sending the processed minimum voltage to LCNC. It certainly can handle the 40 usec sampling period and in fact you can prescale the Arduino clock to run an interrupt at at 4 usec so sampling can be done every 10 clock cycles. Nice and simple!

Anyway, when I get this working the next step is to get the processed signal to LCNC. I think the best way would be to just send the processed signal onto LCNC by outputting a modified/corrected frequency that substitutes the THCAD output. The alternative would be to send a numeric value back to LCNC but I'm not sure if that that can be achieved in the timings we need to meet (but the servo thread is quite a bit slower than the sample timings! Any ideas on how best to tell LCNC the arc voltage would be appreciated.

Not happy here today. I managed to fit an upgraded power supply and it was so hot, I had to wipe out the sweat in the enclosure before I was game to turn the power on.

image missing http://i210.photobucket.com/albums/bb270/rodweb/Mobile%20Uploads/2017-01/20170115_202642_zpsyambzghc.jpg

Once I did, I found the new 10.5 amp supply (on the bottom) was dead on arrival. I bought it with my NEMA 34 steppers because I need more power. This one only cost USD $45 so hopefully, they will send me a new one.

I also ordered 6 proximity sensors. I could not find a well priced eBay source so I ended up spending $140 for normally closed PNP sensors..

ON one bright note, I managed to move from the axis GUI to gmoccapy and I found it surprisingly easy until I tried to hook up the THC to the gmoccapy plasma module and could not work it out. I've asked for some help on another thread so hopefully I will sort it out.
forum.linuxcnc.org/gmoccapy/32164-plasma-component-in-gmoccapy

The other problem since I tidied up all my axes is that while the home switches all work, on a G28, I get an error that the Z axis can't go beyond its limits and it errors This was there in axis. I don't understand this as it happens even if the Z axis is already homed.