Anything welding

Ugh

 

So, I'm just adding as a filler here for anyone bored - not looking for help, suggestions, etc.

Simply typing out the issue, what was done, and measurable differences (if applicable).



Crossfire Pro
Hypertherm 45XP
Duramax machine torch
45 amp mechanized consumables (genuine HT brand)

Issue -

*Pronounced bevel on the top edge of an item, X axis, but not the opposing side

Troubleshooting -

*All gantry tubes ensured for level and square, placing a level even on a slat in the waterbed or the 3/16" plate sitting on the slats show it is flat and level with the table itself

*Panning Y axis and measuring both lead screws ensuring even, parallel movement between both sides

*Ensuring correct stepper movement, parts are accurate to within .005" of CAD

*Machine torch was completely dismounted and remounted over 20+ different times including; completely resetting all set screws, undoing all carriage bearings, and tramming Z axis to square the torch mount to the gantry tube itself, and ensuring the torch mount holder and torch itself is perfectly 90 degrees perpendicular to the slats and/or metal being cut

*Added shims to introduce artificial canting made little difference, as the pronounced bevel on the top edge remained

*Rotating the torch itself within the torch mount, ensuring the torch itself isn't cutting at an angle

*Swapping consumables again

*Rotating the CAD design by 180 degrees, bevel still occurred on top edge

*Running different air pressures

*Running different torch speeds as low as 60ipm to as high as 90ipm

*Running smartvoltage for a programmed height of .025" which equates to a physical cut height of .060"

*Running nominal voltage at 123.5v, 124.5v, 125.5v, 126v, 129v (book), and 130v, which range from a physical cut height of .020" all the way up to 0.085" with no change to the excess top bevel, only affecting the sides slightly

Running with torch height control completely off

*Bypassing refrigerated air

*Measuring the resistance from torch clamp tip to the dinse connector, ensuring less than 4ohms resistance

*Providing Langmuir customer support a DXF file, having them post process through Sheetcam, and providing said cut files for me to test with various items changed at their discretion



I know I mentioned bevel, and even a novice will say "oh bevel, either wrong torch height or canted torch mount" because really those are the two things to diagnose right away. Well the edge opposite of the bevel is completely square. In the event of a canted torch, the beveling on one side would be that of an / with the opposing side being that of a \.

The fact that all of the above still produces one side beveled (outside of normal tolerances) while the opposite side is mostly square is what's so bizarre.

I even cut one of those bed stiffener parts, and compared it with that of a set I have here already; the bevel is absolutely noticeable on just the top edge.

The next troubleshooting steps for either tomorrow or the following days are, wait to see what Langmuir says when they call me after speaking to others more knowledgeable, and then having a go at running it with my Hypertherm hand torch to eliminate the machine torch from being faulty. Technically, we already ruled this out having rotated the torch within the mount itself but we'll see.

And lastly, reinstalling the CUT60 and seeing how it performs.

 

I'm intrigued. Please have good luck figuring it out.

 

It's been bothering me since Sunday.

Like, I literally help people every day regarding CNC so I know how to diagnose and what to look for upon looking at parts being cut.

The fact that I exhausted all of my own knowledge and experience, and turned to -

My good friend that runs an STV hours on end each and every day

Langmuir Support

Two people from FB group I talk to often that also own the same setup

Every single one of them is stumped and couldn't even add suggestions to my troubleshooting at all with the exception of my friend saying "run without THC" and another person saying "try the hand torch".

Everything else mentioned above was shit I did before reaching for help

 

I don't know shit about your process, but I did think of a couple of questions:

Does this happen everywhere on the table?
Do internal cuts come out clean?
Should I just STFU?

 

Torch and mount









X gantry





Y gantry (both)





Water bed/slat



Many cutouts, which showcases the top edge bevel occurs the same on different areas of the table



Today's consumables, all genuine like yesterday's set to showcase correct part numbers as well as the use of genuine parts

 

Yes to the first question

Yes to the second, you can pick up any random cut out on the floor and see a nice rounded circle which also is a result of X and Y both panning correctly













Ignore any crud you see poking out from the backside; dross becomes more or less apparent as variables change in the cutting process

 

At this point all that is left is to try rotating your table 90 degrees…

 

Here's the gotcha - issue began when I replaced the CUT60 with the Hypertherm

Everything else, physically, has been left alone

 

I was kinda wondering if the full moon could effect this.

 

When you're leveling it, are you using round earth or flat earth levels?





Sorry though, sounds frustrating as hell, I hate when stuff like that happens for no apparent reasons. Just spitballing, have you checked for any chaffed cables with some crosstalk or interference somewhere?

 

Again, the gotcha here is -

CUT60 and UPM-105 cut nice and square as usual, then a day later 45XP and Duramax torch installed

Hypertherm combo got dialed in pretty quickly and produced a really nice cutout at 126v, 80ipm outside, 42ipm inside, .041 kerf, 100psi to the machine

Yesterday, I decide I want to test that cut again so I open the same exact file, running the same exact air compressor, filtering, etc on the same piece of steel, cutting a test piece within 2" of the last test piece before that, only to suddenly get top beveling.

The top beveling occurred all day yesterday and all day today.

No one lives here except the wife and I (and dog and kid).

No one touches the table.

Nothing was even moved a mere millimeter.

It's like all of a sudden it just isn't doing what it's supposed to.

 

I also wondered if there was a possibility of electromagnetic interference, but that wouldn't follow if the cut were turned 90 or 180 degrees.

 

Thing about interference is that, it just happens.

The issue here is consistent.

Also, THC is normally the most sensitive thing to EMI as it will raise the torch as it cuts until it's all the way up.

I've sat my phone in front of my computer to record a cut and voltage readings could be as consistent as 1v through the run - 1v means as little as .01" and we've ruled that out having run as little as 123.5 to as high as 130v (with smartvoltage) as well as disabling THC altogether.

 

Basically not trying to sound like a smartass, but address the troubleshooting steps performed that would answer any theories proposed.

I've ran these same theories in my head hence why I conducted as many test cuts as I did before even ranting on here like I did.

 

It sounds like your methods are sound. It doesn't make sense to me how your table can make a cut, say, left to right just fine but have a bevel when the cut goes right to left, on an outside cut and be fine either way on an inside cut...if I'm understanding you correctly. I'm watching to see what this turns out to be.

 

@Zebinator

I'm sure you're on the forums but in case you're not - https://forum.langmuirsystems.com/t/6-in-1-easy-backgauge-calibration-procedure/38623

 

Mechanically, it seems like everything is square and free from wiggles. Is there anything software-wise that could cause something like this to happen? Something possibly incompatible with your new Hypertherm unit? I've seen software do crazy-assed stuff that shouldn't have happened. Like I said, I really don't know shit about the intricacies of this, just have a basic working knowledge of what you're doing. I'm kinda pulling wild ideas out of my ass to maybe give you another way to look at what's happening.

Is there a possibility that the new unit has a glitch?

 

I sent the exact DXF file to Langmuir as well as a person I talk to on FB.

The item cut out fine, using Sheetcam and Fusion on their end on their table (using the exact same machine and torch).

The file, to eliminate the issue being on my end, was processed into G code by Langmuir and emailed back to me. They did this with settings they themselves dialed in on a 45XP.

Same result.

So we eliminated the software part of things.

As for it affecting the machine itself, no.

The Hypertherm only does two things as far as communication with the control box - it is told to turn on or off based on a circuit that closes, and two leads send out a signal that is stepped down from the output voltage coming directly from the plasma cutter's torch.

It would be the equivalent of pulling the trigger on a hand torch, and measuring voltage with a multi meter.

If the new machine was glitched out, it would be odd that it would only ever produce an issue on the top edge of any cut. The torch only turns on or off and would not know at all its position on the table or which direction it's going.

 

I usually just read posts here and don't think I have much to add since I bought the og crossfire I only ran it a hand full of times, don't have THC, using mach3, converted a gouge torch for machine fire and it's never been dial in per say with my lack of time. I need to baseline the machine for sure since small holes isn't circular. Most likely a leadscrew play issue.

But I do have a hypertherm 45xp and there is a bevel on one side that you are mentioning. I just accepted the bevel cuts vs doing it by hand and thought it maybe because I don't have thc or didn't level my table. I been meaning to tinker with it more but life is in the way. Hope you can resolve your issue soon. Is there a reason you went with the xp vs the sync now that it's available? I hope it isn't an xp thing now that it seems like you the expert has similar issues to me.