What have I gotten myself into!: "Tooling up"

Instrument panel #4

What I’ve been able to cut with this table so far is more than good enough for the way in which plasma cutters are normally employed, making parts for welding jobs, or components for other types of loose fitting assemblies. Because these parts do not usually require the high precision of a CNC mill, it makes them suitable for the plasma cutting process.

The million dollar question still remains though… is this CNC/plasma cutter combination accurate enough to produce component to size right off the table?

As usual, the answer is not as clear cut (pun intended) as one would like, and while it might eventually be that accurate, there are so many variables at play in this type of machine, that time and practice will be needed in order to home-in to the ideal settings.

First there is the question of the table accuracy. Nothing else will matter if the table won’t reliably get the cutting tool to the exact location time after time, and the cut will never be right regardless of whether one is using a plasma cutter, a laser beam, or… whatever!

Then there is the material. I have already found great differences in the way aluminum reacts to being cut compared to steel.

Dross on mild steel

Dross on 2024 Aluminum (right). Backside up.

Then there is the thickness of the material, the dimension of the part being cut and how heat is dissipated (or not), use of a water table (or not), the kerf (width of the plasma jet), the shape of the kerf (like a cat’s iris), the height of the nozzle from the material (engaging different parts of the kerf), the amperage, the type of nozzle, the state of the consumables, the air pressure, the dryness of compressed air, the travel speed of the cutter, the acceleration/deceleration of the table, the entry/exit point of the cut, lead in/out distances and angles, the direction of the cut (hard to believe, I know), the pierce delay time, inside corner vs outside corner, to say a few.

How amperage and travel speed can impact dross formation on the same material (parts flipped up)

I decided to begin with the basics, and start testing the CNC table to try to discover, and then minimize any errors it might have. The two big ones to be on the lookout for with any new CNC machine are backlash, and size distortion (I made up this last name).

Although the Langmuir table uses ACME screws, the preloaded nuts cut the backlash to 0.001” or less, so we are all good there, and since I have already explained backlash here let's move on to size distortion.

Size distortion happens when you ask your CNC table to move, say… one inch, and though the DROs (digital read out) say the tool did move one inch, the actual measurable amount is different. This is usually caused by having the wrong stepper motor’s steps per inch selected, AKA how many impulses the computer transmits to the stepper motor to get one inch of machine travel.

As far as Steps per Inch, the Mach3 (controller software) Langmuir’s configuration defaults to 12800 spi (steps per inch) which is mathematically correct, but does not take into account the physical reality of the components. The only way to account for all the intangibles, is to conduct many a test on the table, and modify settings accordingly until the error is minimized, or eliminated.

The way to do this is to gather the biggest object (to minimize compound errors) that can fit on the CNC table and that can be measured very precisely with a caliper, and try to measure it once more using a dial indicator in place of the cutting tool. Any difference between the two measurements is an error due to the steps per inch selected, and needs to be iteratively adjusted and retested, until the error approaches zero.

Measuring the size of my turn table using a dial test indicator and Mach3's DROs

Indicator is set to zero at point #1, then moved in Y+ to #2 until it reads zero again (Yes I know it's 0.0005" off in this photo)

Testing in X direction

New Y axis SPI value 12818 found after a series of 9 test measurement runs

Y axis SPI value set to12818 in Mach3

New X axis SPI value 12872 found after a series of 7 test measurement runs

X axis SPI value set to12872 in Mach3

With the more accurate Steps per Inch figured out, the table is now as precise as it’s ever going to be, and although it’s still a couple of thousands off (we are really asking the cantilevered gantry setup for a lot), this is quite good enough for any of the work I am planning on doing with it.

I might have mentioned that the main thing I wanted to be able to do with this machine was to cut my own instrument panels (in seconds instead of days), so let’s try doing just that on an expensive sheet of 0.090” 2024 Aluminum. Bear in mind this cut was done before the above adjustments were completed, and before figuring out more reasonable speeds and feeds.

Plasma cutting an instrument panel

Pretty exciting stuff, right?! Well, I thought so, at least after I remembered about the plastic on the aluminum sheet.

Doh!

With this first panel I did have to do a bit more filing than I was hoping for. Hopefully, with the newer settings I was offered online and a little more experimentation, I will be able to shorten the post-processing phase.

Anyway, here are a few pictures of how the panel turned out, I think you will approve…