hi folks let's walk through how we made this connecting rod a link bar piece for project egress looks like a relatively simple part but some really good lessons here on workholding on optimizing your tool paths and we're using a Harvey lollipop cutter to machine the undercuts for the universal joint buckle up we make a few mistakes but hopefully everybody learned something welcome to another Wednesday widget holding the part in our mod vise giving a quick deck off with a super fly and then we're trying to do the majority of the work with a few tools for this part fewer tools is always a good thing if you were to changes and better consistency and height and accuracy we also ran this part a number of times I wanted to find out what's the hardest that we could push this recipe in these tooling but stay within a good process reliability framework I'm machining the left edge of the stock only as a backup our plan when we flip the part is to use the center of the bore for a work coordinate system but that edge gives me another machined datum that I can reference if needed next up 0d adaptive to open up those bores running not at the normal ,0 rpms on the 0 but rather at we have a little bit more torque here so we're able to maintain that inch to feed per tooth and a lot of times I would have pre-drilled this but again I wanted to minimize the use of additional tools here and there's nothing wrong with doing a helix or ramp in style especially on a free cutting material like aluminum a 0d contour to clean up that inside geometry especially important on this bore because we're going to use that as the datum when we flip it so we duplicated that 0d contour and we're doing the first one really just to clean up the geometry for our lollipop cutter which has to machine a bit past the part because of the nature of that cutter but then I do a second 0d contour that's all the way at the stock bottom to give us a good clean geometry when we flip the part and then surfacing in this Philip right here so this kicked my butt and it shouldn't have to but it was a really good time to just pause and figure out how to get the machine motion the way we wanted it and the key is a good tool path with a good point distribution that then matches the way your controller wants to see that tool path if you watch our channel you've probably heard me mentioned smoothing before I'll duplicate this scallop right now I'll call it smoothing example right click Edit passes smoothing if we read that hover up menu smoothes the tool path by removing excessive points in filtering arcs where possible smoothing is used to reduce the code size without sacrificing accuracy subject to your tolerance smoothie works by replacing collinear lines with one line so if we can collapse two separate lines into one line and never deviate to our tolerance similarly tangent arcs to replace multiple lines in curved areas so that's nice and said of a bunch of lines we can do with one nice smooth arc motion this all has to do with what the motion profile your machine once and how it's going to send the signals to those motors to turn it and make that machine go if you've got older CNC machines the code size the file size can be a huge limitation on modern machines you've got to pay attention to the kind of code your controller once so that's why I wanted to play around with path pilot here turn smoothing on will make it say five times greater than our tolerance and the tolerance here is how far the tool path can deviate from the solid model the smoothing happens there after it takes that tolerance tool path that says hey can I collapse some of this down subject to your smoothing accuracy so we take a look with no smoothing on we have more points here we have less we take it to the extreme we'll drive our tolerance down to one ten thousand so this tool path really accurately has to follow our solid model but we'll make our smoothing tolerance something quite large in fact fusion will give us this warning that it's significantly greater it's just a warning no problem running that if you wanted to and now you can see that dense cloud of black points has gone to almost nothing but here's the thing you don't even have to do this on path pilot because path pilot does a really good job of handling this on its own with its g command if you want to nerd out we'll put a link to the linuxcnc detail on the trajectory planning when tormach replaced Mach 0 with path pilot they did a lot of work on the trajectory planning and then shared that back with Linux C&C; which is what Pat pilots based on but I think this sentence really sums it up a g-code program can never be fully obeyed if you program a g0 x0 F that means go to position of x0 add a feed rate of well it can't go at a feed rate of because it has to accelerate it has to decelerate so it's funny to me that that G code that we spend so much time making is actually just another step in the process to how your machine actually moves in the controller itself and it's hardware actually does more work and that's what this G command is so what does G do and what does the P and the Q do G alone means keep the best speed possible no matter how far away from the program point you end up which is actually kind of a bad thing if you think about it because you're sacrificing accuracy to maintain speed but a lot of times we're more concerned with accuracy than speed so that's where we can add P's and Q's the P tolerance means the actual path that the machine actually moves will be no more than PLA so if it can slightly round a corner or an arc to maintain speed and that rounding that off doesn't deviate more than P will preserve our speed and by the way speed is more than just how fast the part gets done has to do with the fluidity and the motion in the quote unquote smoothness of the machine itself Q has to do with collapsing lines I think that tormach explanation is actually better here Q if present is the maximum deviation that will collapse a series of lines into a single line to exaggerate if we have two lines like this and it can collapse those into one line like this and the deviation is less than Q so that here that would really be the distance between this and this kind of show like there it'll go ahead and collapse it that's one less line of code for the machine to parse and for the motors to have to react to you can see here the default tormach settings on the tormach post infusion include a g that's a good thing because i think there are benefits to it too bad thing because what I just mentioned it's going to sacrifice accuracy at any cost to maintain speed so what we did on this file was I did a bunch of horrendous scalloped tests a bunch of times trying to experiment with that mix of P's and Q's for the path pilot control as well as the various different smoothing and tolerance settings within fusion and what I found is that the fusion code doesn't really matter actually kind of reminds me of a Heidenheim control which Hyden Heinz really liked a lot of points because they say hey we'll handle you know we the control will handle those points on our own just give me a bunch of points that seems to be helped at pilots working here which is a good thing what I found was the p-value wasn't that sensitive but I do like that to stay relatively low because that does relate back to the tolerance and accuracy of the part cue when I had it at three or four thousand seemed to do much but when I hit five foul really made a performance difference so you can see here if we strip out the g it's not something you would normally happen you can see that motion profile now will run it at the regular G which is probably how your machine has been run if you haven't tweaked this and then finally we'll go back to that footage of the scallop operation with the modified G P's and Q's much nicer motion profile next up of the lollipop which reminds me to mention that I believe the only toolpath in fusion 0 that currently supports undercut or lollipop style end mills is 0d contour good news is it's incredibly easy to use we have our two chains selected the left and the right shout-out to Rob Lockwood for the tip on your additional offset should be slightly more than whatever you have in your tolerance that can help you avoid errant toolpath behavior as an example here's a fusion 0 sample file which has unacceptable tool paths where there water falling over the side of the part on this sample file the tolerance under the passes tab is set to four ten thousandth of an inch and we have no additional offset so what we've told Fusion is it can deviate from the solid model by up to four ten thousand seven inch which is why in somewhat random locations that toolpath falls over the side of the part the fix for it is quite easy under the geometry tab we set a negative additional offset that's slightly bigger than that tolerance that pulls the toolpath back in to make sure it's inside any of that tolerance deviation and there we have a perfect tool path flip the part over forget before we put our part in that we first need to measure the top of the parallel because that's the location of our Z on our coordinate system probe in our hole and we're doing a both ways adaptive this time we're doing a reduced optimal load on the other way which is conventional machining ran fine it does save some time do notice though we have a little bit of gouging as a byproduct of the fact that the way we're making this part is by tabbing it or window machining it so when that adaptive comes in we've controlled the stock zone with a sketch so it thinks what you see is the purple area is the stock so it thinks it can plunge slightly outside of that face off our part to get that final surface finish ramping down just enough to expose the surface we need to come back and do our scallop again and then finally tabbing it on any 0d contour under geometry tab we have built-in tabs you can set them by distance or you can switch that to points and you can actually control the exact points you want your tabs at really slick feature we made a number of these parts and we kept just goofing a little bit on our tab Heights because I'm getting greedy trying to get that Teleca balance on the thin tab but still having it work hold enough I think the next time I do a part like this what I'll do is ensure we have sufficient tabs left 0 is fine here I don't have a problem with the initial tabs but then as you clean those tabs up I think I'd leave those initial tabs thick enough so that you could then really finish machine at least one if not two of the other remaining tabs then you can come in and whittle away at the very last tab staying off the actual workpiece with a radial stock to leave that way your cutting tool in the tool pressure is only on the tab and the raw material it's not actually contacting your workpiece tabs can be a tricky you've got to balance a thick enough tab so that the part remains secure and it can handle whatever cutting pressure you're subjecting the part to as you're doing that finishing operations but too thin and the part can break away or chatter so stick around and click that subscribe button we've got some other videos coming on three axis and five axis tabbing but here's an easy trick when you're just doing a one-off type of part use hot glue it's back we've got our four tabs left we can now use hot glue to secure the part in place you can usually use one end mill to come through and clean those up or a pro tip keep a dedicated hot glue end mill it tends to load up you can clean them out but use that tool is a sort of a roughing tool that removes most of the hot glue from the tabbed area then you can come back with a good finishing in the middle and clean that part up because you've already removed that hot glue you'll have much better chip evacuation and you'll get much better finishes and then when you're done you can use force heat or acetone to break away and clean up the hot glue and your part is done as always folks I hope you learned something hope you enjoyed take care see you soon
تعليقات: 0
إرسال تعليق