PeterSteele

Well, I didn't get very far. My mill kept throwing up phantom limit switch errors, which ... well, I don't know quite what causes it. I'd guess EMI, except that I've got perfect grounds, shielded cables and ferrite cores on everything. It might be coming off the VFD - I seem to get the errors more when I'm really hogging out a lot of material. This still seems a little odd though, as yesterday when I was cutting the fixture, I was working it just as hard - same feeds & speeds, same stepovers, and there weren't any problems. I don't have time today to redo the toolpaths for smaller stepover / stepdowns, unfortunately, so I decided to CANX while I could still salvage the billet.

Finally had time to make some progress on the AR lower project. Finished designing the left side fixture to hold the lower while I machine the right-hand side. Now the plan is to cut at least the left and right sides of the lower tomorrow. I still don't have the right side fixture designed yet, so I can't cut the FC pocket, buffer tube hole, etc.

It is a simple 2D slot, yes, but (a) it's the width of the cutter, so if you do 2D contour and pick the open loop I always find that it gives the kernel fits and I wind up with no toolpath generated, and (b ) if I pick just one side of the loop then it will cut one radius too far. Either way I'd need to create more geometry ending one cutter radius shy of the end of the slot, so I usually wind up using TRACE vice 2D CONTOUR in a circumstance like this.

Make sure you have the cut going the right direction. On the geometry tab, when you're selecting the contour, click on the red arrow to force the cut to enter from the open end of the slot. If you're setting the path to be traced to start outside of the stock, then you could conceivably turn off lead in as well as lead out - you'll be on the path already when you enter the stock, lead in will be irrelevant.

Use TRACE with axial offset passes. Leave Axial Offset set to zero, but check the box for axial offset passes. Set your maximum stepdown for however deep you want each pass, and set the number of stepdowns where the closest whole number where Number = (slot depth) / (max stepdown). I use that for cutting slots for keyways.

I cant seem to make the 1/16th slot over travel on the one end. I'll take another crack at it this weekend You're using F360, right? How big is the cutter? Is it also 1/16"? If this is the case and you are using F360 or Inventor (or even Solidworks with HSM), create a sketch at the bottom of the pocket. Draw a line along the centerline of the pocket. Have it end 1/32 from the back end of the pocket at one end, and at the other end, have it end wherever you want to have the cutter start it's run. Use TRACE on the 2D panel of the CAM tab. When you select geometry, make sure that the arrow is pointing in the direction you want the cutter to be going. If it's not, click the arrow, and it'll flip around. On the LINKING tab of the trace operation, leave "Lead-in (entry)" alone, unless you want to change the radius. Un-check "Lead-out (exit)". This will make the cutter come straight up when you get to the end of the slot. Done.

I've been too busy to get anywhere yet with my AR lower. I have had a chance to cut some aluminum recently though - made a couple of things out of it now. First up was a nifty little tool to set the angle on my lathe toolpost, and then today I made a set of jaws to hold a shaft that I need to cut keyways in. Biggest thing I've been doing with power tools was mind-numbingly boring, though. I had to take a bunch of pieces of 3" bar stock and face off each end, then bore them out to .531. The drum in the picture below is, with the exception of the few things you can see sticking out of it, completely full of chips that I made in that process. Went through 3 indexable cutters, all 6 points of each one. Time consuming and boring ... but I did learn a lot about what I can and can't get away with on my lathe. Got a couple more things I want to make out of aluminum, and then it'll be time to make the fixtures for the AR ... then it'll be time to make some guns.

I may or may not be sending a copy of that picture as a BOLO to my brother in law on the job in Mesa. What's it worth to ya?

Now that you've got Inventor up and running, you'll want to look into Mesh Enabler, if you really absolutely must play with STL's. I guarantee you'll absolutely hate it, though ... It's normally for subscribers only, but there's a thing on the Inventor part of the Autodesk forum where you can request it, since you've got a student license. The guy will give you a license for it. Good info - I never actually knew the difference between 203 and 214. I just use whatever is the default when I export stuff from Inventor, TBPFH.

Another one to look at is the SmoothStepper from Warp9. The USB version more or less emulates 2 parallel ports, giving you double the I/O, and the Ethernet version gives you a third one. I'm going to be upgrading my garage computer in the semi-near future, and going probably with the ethernet version. The circa 2003 Gateway Pentium 4 is starting to misbehave, but I didn't realize just how bad it truly was until I had the brand new Pentium Dual Core in the mill at work.

Fusion 360 does 3+2, as was noted elsewhere. If you want true 4 or 5 axis machining, you need to look at Inventor HSM Pro (~10k), or Solidworks + HSM Pro (~10k + Solidworks license). Okay, now let's talk file formats. Fusion 360 - and Solidworks, Inventor, most engineering software - works with solid models. They have mass, volume, etc. They are defined by mathematical equations - a shape and an extrusion vector, for instance. This means that a rounded object is truly round, in a mathematically perfect sense, and it's very easy to calculate. STL files are more like the sort of thing created by 3D animation software. They are made up of individual polygons, each made up by connecting a set of points. This means that a rounded object is only round from a distance, and the more you zoom in, the less round it is. If you create an STL file with lots of details, there will be lots and lots and lots of individual polygons. A solid modeling software package will HATE this. Solid modeling software likes solid models - they are watertight, so to speak, and it's easy to regenerate them just by changing the math used to create them. An STL, being comprised of individual polygons - which, by the way, also have to worry about which direction they're facing - is very rarely going to be watertight. Most STL files are only solid from the front, and are open on the back. These are generally more art related. Even if an STL file is actually watertight, if the normal vector for even one face is missing or incorrect, then -boom- no longer watertight. STL is just a flat-out pain in the ass, and I would seriously like to see the person responsible for STL covered in honey and buried in an anthill. If you want the best results working with F360 or any other solid modeling program, stick with either native file formats or neutral things. F360 really likes Inventor files, and does quite well with STP / IGS, and I think it can do okay with x_t as well as a few other formats. (Incidentally, IGS is probably a good one to stay away from - STP is more modern and still currently being developed for. IGS was more or less abandoned almost 20 years ago, if I remember correctly.)

Yeah, it's the bridgeport. Turns out the issue was in the backlash compensation. Turned that off, everything was fine again. Re-did the facing operation and even though you can tell a little bit of difference visually in the toolmarks, you can't feel a difference with your fingers. I'll get some pictures a little later. It's a pretty nifty rig. Not sure if I'll have time to play with it today ... I'm trying to figure out the dynamics of some air cylinders to get all the operations I need out of an automatic packaging system I'm designing for work.

Well, this has been a confusing - and somewhat annoying - morning. This was supposed to be a straight forward facing operation. The block is aligned parallel to the X-axis, and this was supposed to be a facing operation right-to-left with .125 stepovers in the Y direction. .04" off on the first pass, then a finishing pass at .010". Welllllllllllllllllllllllllllllll ... I'm losing steps somewhere. About .098" per pass, and it's quite consistent. I'm a little confused by this. (Yes, I know the finish sucks - that cutter has just about had it, and all I needed was for the piece to be more or less flat.)

One worthwhile note - it's this way in Inventor, and I'm pretty sure it's this way in Fusion as well - create some sketch geometry on areas that you don't want the cutter to go, and you can use the sketches to define avoid areas. (Well, let me rephrase - I know you can define avoid areas in Fusion, but I'm not 100% positive that the sketches in Fusion work the same way that they do in Inventor.)

Thanks for the shout out! Looks like you're getting a good handle on it. I got lucky - the implementation of Mach3 on my mill at work requires no changes from the default Mach post. My router at home on the other hand ... didn't get so lucky there.