|
Brekelefuw posted:A 1/4" piece of aluminium plate is so cheap. Why bother making some wax? What if you want to make something thicker than 1/4"?
|
# ¿ Aug 31, 2016 01:57 |
|
|
# ¿ May 22, 2024 09:20 |
|
Got a lathe? 10 minute job.
|
# ¿ Sep 2, 2016 18:43 |
|
Well, that sounds perfect for those of us who choose to use metric in their personal projects, like 95% of the human population.
|
# ¿ Oct 29, 2016 22:35 |
|
Used to be a man'd get banned for that, yes sir. Times sure have changed
|
# ¿ Nov 30, 2016 09:13 |
|
polyfractal posted:Just starting to research hobby CNC machines. Before I get too deep in the weeds, any recommendations for a machine/setup that emphasizes precision over speed/size?
|
# ¿ Jan 17, 2017 00:29 |
|
Well it's one bigger.
|
# ¿ Mar 14, 2017 07:31 |
|
the first thing you make should be a better control box
|
# ¿ May 12, 2017 02:20 |
|
There's a way to figure out the wiring by measuring the resistance of various pairs of leads. I can't remember the sequence but it's easy to look up online.
|
# ¿ May 30, 2017 17:43 |
|
Volkerball posted:so what happens if it over travels? DRDRDRDRDRDRDRDRDRDRDRDRDRDRBRBRBDBRDBRBDBRBDRDBBBRBBZRBZBRBZBBRZBBRZBRBZZBRBBZBRBZBRBZBBBBBBDRBDBRDBRDZBDBRDBRBZR
|
# ¿ Jul 19, 2017 05:18 |
|
I forget, was this posted here already? https://www.youtube.com/watch?v=BAtb7O_yk8c
|
# ¿ Jul 19, 2017 07:10 |
|
I think my favorite is the one at 2:21 where the initial plunge drives the tool clear through the part, then it dwells for a second like "hmmm...well, whatever" and turns on the coolant.
|
# ¿ Jul 19, 2017 17:55 |
|
Mister Sinewave posted:Yeah, for those of us who don't live and breathe this stuff it's a reminder of the fearsome power behind these things. biracial bear for uncut posted:I like the one where the spindle bends as one of the horizontal axes moved once the bit was in the material. I was casually flipping through the manual for one of our Haas mills the other day and saw on the spec sheet that the X and Y axes are capable of producing four tons of lateral force. Yep
|
# ¿ Jul 19, 2017 22:56 |
|
Depends on the design of your machine, but yes, sometimes that can happen in LinuxCNC or Mach3. 1) 3D model geometry is analyzed to generate 2) toolpaths, which are then post-processed into 3) g-code, which is interpreted by 4) the machine host or driver (LinuxCNC, etc), which generates a 5) timed sequence of motor steps, which are sent to the 6) motor driver ICs, which convert the step pulses into 7) alternating pulses of the proper voltage on the motor coils, which 8) rotate the motor shafts by the specified amount. I haven't used Mach3, but with LinuxCNC in step (5) the software bit-bangs the motor pulses out over the parallel port, and the driver interface box takes those pulses and commands the driver ICs to actually spin the motors. This is why those homebuilt CNC machines still use parallel ports and why they're so picky about them -- the software needs extremely low-level access to the port to be able to time the pulses correctly. USB-to-parallel adapters frequently don't support that kind of operation. In a more fully integrated machine, like an industrial CNC or a 3D printer, the g-code is interpreted onboard by the machine's firmware and the motor drivers are run from the processor's GPIOs.
|
# ¿ Oct 2, 2017 06:09 |
|
Many (most?) CNC machines are open-loop and count steps of the motors to keep track of their location. If the motors slip, well, then you're hosed. That's why you keep an eye on your spindle load and axis loads, which high-end machines will continuously measure and report. You can of course buy encoders for each axis, just like what you'd use for a DRO, and I'm sure there's some way to feed that data back into linuxcnc or whatever. For the record, our Haas Mini Mill can produce 8 tons of force in each axis, so missed steps mean you did something really, really wrong Sagebrush fucked around with this message at 06:53 on Oct 2, 2017 |
# ¿ Oct 2, 2017 06:49 |
|
Also, a lot of those online calculators are meant for maximum-performance machining of aluminum using flood coolant and carbide tooling, where every second saved is money in the bank. As a hobbyist, you don't need to go that fast. The general rule is that you should be optimizing for a specific chip load. In a perfect world, larger chips are better because they carry away more heat and it's faster. In reality, the largest chip you can make depends on your machine rigidity, spindle horsepower, and of course just the mechanical strength of your tool. If your chips are too small, though, you'll just end up making a bunch of dust, and that gives a poor surface finish and overheats and dulls the tool. So as previous guy noted, something in the .001-.002 feed per tooth (i.e., every time a new tooth is exposed to the material, you feed forwards that distance) is reasonable for starting out. For small mills, I'd say to start at like ~5000 RPM and work out an appropriate feedrate, then move up from there in both feed and speed until you find the limits of your machine's rigidity. Take shallow cuts (.020 or less per pass) and take advantage of the CAM software -- don't try and plunge down and cut the whole part in a single go. Listen to the tool and you'll eventually be able to tell by the sound whether it's cutting properly or not.
|
# ¿ Jan 13, 2018 20:21 |
|
If the two lines are connected, the automatic relations and guides should automatically do what you're looking for, or something very close. If they're not connected then yeah you might have to manually insert the relation but it's very fast when you get the hang of it. Draw two lines, drag a box from the lower-right to the upper-left that crosses over both, click "parallel" on the featuremanager, done.
|
# ¿ Jan 28, 2018 23:56 |
|
Karia posted:For perpendicularity, auto-relations can take care of it, but from a quick test only if the start point of the line you're drawing is not on the line you want to be perpendicular to, which I think is an odd design decision, but not hard to get used to. You can still do it in that case 1) draw the base line at whatever angle 2) start the perpendicular line on the base line 3) bring the cursor down so that the new line is collinear with the base line (they overlap) and hold it for a fraction of a second to set the focus 4) the auto-relations referencing that base line should pop up, including the dashed yellow perpendicular reference 5) snap to that reference to insert the relation Sounds way more complicated than it is -- try it out.
|
# ¿ Jan 29, 2018 00:55 |
|
? there are a lot of different variations of the sword-and-thunderbolt insignia that are used by different military groups. His appears to be the insignia of this division https://en.wikipedia.org/wiki/United_States_Army_Civil_Affairs_and_Psychological_Operations_Command
|
# ¿ Feb 24, 2018 09:41 |
|
Might work with a razor-sharp tool and high speed/feed. Soft materials in general need sharper tools than hard materials simply to cut at all.
|
# ¿ Mar 1, 2018 22:51 |
|
You can't engrave (anything but lines) with a drag knife, though. A super sharp (HSS), one- or two-flute end mill spinning like 20,000+ RPM could probably engrave into the surface of a piece of cardboard pretty well. It could cut too but the material is so flexible that I bet you'd have trouble keeping it in place. Vacuum table? Really a laser cutter is the way to go, though, yeah.
|
# ¿ Mar 2, 2018 04:15 |
|
His Divine Shadow posted:There's another, cooler design I also am trying to model, it's easier to model, but the joint configuration of the "spherical" hinge part is beyond me at the moment. This is super cool and now I want to model one just to 3D print it for coolness' sake.
|
# ¿ Mar 7, 2018 06:48 |
|
It's the portable Voigt-Kampff scanner from the new Blade Runner movie. e: or just an iris scanner, I guess. Still, it's for detecting replicants.
|
# ¿ Jan 9, 2019 18:50 |
|
You could make something very similar out of aluminum bar stock and one of the inexpensive desktop CNC mills from eBay or AliExpress. They're slow, and aluminum is about the hardest material they can handle, but that is a very tiny part so the cutting forces will be low, and it's only like two operations. You could make a hundred of the parts over a week or two without much trouble once the machine is set up. I have no experience personally with the cheapy chinese mills, but I've bought other products from SainSmart before so their 3018 model might be worth a try: https://www.amazon.com/SainSmart-Genmitsu-Control-Engraving-300x180x45mm/dp/B07DBY44KJ/ With a mill like that, you might also consider using an engineering plastic like delrin, which is strong and tough and machines very nicely on low-powered machines with sharp tools. Delrin is also self-lubricating which is great for a sliding mechanism. It's going to be more flexible than metal but that might not be an issue for the design, and it will absolutely be tougher than a cast resin part.
|
# ¿ Jan 9, 2019 20:27 |
|
There are definitely a million ways of making it work as it's supposed to on the cheap, but given that the thing is a replica movie prop (and he's been developing this in various threads for like a year or more now), I'm guessing that having the exact appearance is pretty critical.
|
# ¿ Jan 10, 2019 16:56 |
|
I would probably screw a dial indicator into a 1-2-3 block and slide it along one rail while measuring the other, yeah.
|
# ¿ Mar 6, 2019 00:40 |
|
He's measuring the separation between two rails, though. Doing that with handheld calipers isn't going to be super accurate unless you're inhumanly good at compensating for cosine error. And it doesn't matter if the table is flat. Stick the dial indicator on a 1-2-3 block, push the block up against the inside of one rail, align the indicator on the other rail, slide back and forth. Tiny bits of vertical movement won't noticeably impact the reading unless the rail surfaces are non-perpendicular.
|
# ¿ Mar 7, 2019 01:44 |
|
Volkerball posted:It doesn't take super human skills to tell that the gap is .003" wider here than it is over there with decent calipers. I can't really picture his part since he didn't give any details, but 90% of the time, checking it in a few places with calipers will be enough to determine if you have an issue and what that issue is. I would only really do what you're describing if the rails are so flimsy that they actually flex under the force of opening the calipers, and/or there was parallelism callout of .003 MAX or tighter, which is uncommon. Yeah but what I'm getting at is that handheld calipers aren't suited to measuring this particular dimension to the required level of precision. What you're suggesting is mathematically valid, but it doesn't work as well in real life. Here is why: I don't know exactly what sort of rails we're talking about, but I'm imagining two rails or precision surfaces 6 inches or so apart that need to be perfectly parallel along their length, like you'd have in a CNC motion platform or something. I'll take your example of the parallelism needing to be better than .003. If you measure the separation between these surfaces with a pair of calipers and you don't hold them perfectly square to the rails, you'll measure a longer dimension than intended because you're now reading the hypotenuse of a triangle. That is called cosine error. A bit of math says that, in order to not skew the measurement by more than our .003 tolerance, the calipers need to be held at less than 0.18 degrees off of perpendicular. Can you see a misalignment on that scale? I don't think that I can. There are tricks you could use to get around it, like slowly rocking the calipers back and forth several times and taking the minimum value, or holding it against a machinist's square that is held against the rail. But if you're going that far I would just do the 1-2-3 block thing because it's a more robust and repeatable setup.
|
# ¿ Mar 7, 2019 21:52 |
|
Why does it have to be geared? That's a range where a nema 17 stepper motor is totally appropriate, assuming you can set up the proper drivers. You could also look at hobby brushless motors designed for radio-controlled airplanes and stuff. E: oh, you said inch pounds, not ounce inches. Nevermind
|
# ¿ May 30, 2019 17:02 |
|
What material? Aluminum? I assume by "mil" you mean millimeters and not the SAE mil (0.001") A part that size in aluminum I would probably stick to a spoil board with double-sided tape or super glue and machine without tabs. It'll get floppy otherwise. Alternately, if your machine is squared up nicely, you can leave an onion-skin layer on the bottom by telling your tool to leave the bottom .010" uncut. That's thick enough to hold the part in place but thin enough to quickly zip out with a coping saw and file off the leftovers.
|
# ¿ Jun 10, 2019 02:15 |
|
In Sweden a mil is 10 kilometers
|
# ¿ Jun 10, 2019 02:31 |
|
What feed and speed are you going to attempt when making that full thickness, full width cut? Looks like a half inch tool? On a part and mill that small, I would probably go for five passes at .050 depth leaving about .020 radial stock, and then one finishing pass at full depth to bring it to size. It's a CNC machine so there's no extra operator effort in doing that, and the lighter cuts will be less likely to clog up the tool or cause deflection in the thin stock.
|
# ¿ Jun 10, 2019 15:52 |
|
Do you have a vise? You should get a vise. Also, for thin tabs like that don't use a plasma cutter; a coping saw will do fine.
|
# ¿ Jun 16, 2019 16:35 |
|
It's fine for the shaft to get jammed into the bearing as long as it's not such a tight fit that it's deforming things. You want the shaft to rotate on the balls , after all, not on the inside bore. Try putting the metal rod in the freezer for a while first.
|
# ¿ Jun 28, 2019 18:07 |
|
Well, either 1) the shaft is larger than it's supposed to be 2) the bearing bore is smaller than it's supposed to be 3) they're both correct, and you aren't used to what a press-fit means in this context Gotta measure the first two, and if they're both correct, get yourself an arbor press or whatever to do the third. Sagebrush fucked around with this message at 06:02 on Jun 29, 2019 |
# ¿ Jun 29, 2019 05:58 |
|
0.06mm~0.002" is plenty of clearance for a precision ground shaft going into a bearing. In fact for a true press-fit you'd usually have it the other way around -- the bore would be a couple of thou smaller than the shaft, and you'd either press the shaft into place or use heat and cold to get a shrink fit. Are you sure there isn't a ridge or nick somewhere on the shaft or bore that are preventing them from fitting together? Feel them carefully with your fingertips and see if you detect anything that isn't perfectly smooth. Try using some crocus cloth on the ends of the shaft and the rim of the bore to clean off any little burrs. If there's nothing sticking out of either surface, then I suppose one of the two components must be non-circular. It seems unlikely that McMaster-Carr bearings would be defective, so I'd start by measuring the shaft in a couple of different axes (measure it, turn it 90 degrees, measure it again) to see if it's somehow ellipsoidal.
|
# ¿ Jun 29, 2019 17:19 |
|
shovelbum posted:Those are incredibly cool but I don't really understand the point of them unless you're doing really detailed field modifications to something, having to trace out your cuts on a CNC by hand seems to defeat the purpose. Well the use case there is that if you're not doing large runs, you can (theoretically) make large pieces just as accurately as a proper CNC machine without the dedicated machine footprint. I don't know if it's genuinely as precise as that, but it seems to work pretty well.
|
# ¿ Jul 30, 2019 03:08 |
|
I'm not sure I completely understand what you're trying to do, but it sounds like on the physical level you are talking about a data bus and on the communication level you're maybe talking about packetization or multiplexing. Maybe Google those words?
|
# ¿ Nov 24, 2019 19:01 |
|
Steel up to a quarter inch thick you can bend in a vise with a hammer.
|
# ¿ May 16, 2020 00:07 |
|
I would disassemble the box. Even if you had a router with enough vertical clearance to handle it, it's going to be hard to fixture a hollow box in a way that it won't deflect under the tool forces. What is the box made of and how thick are the walls? Is it just holes, or are there slots that would require side cutting?
|
# ¿ Jul 6, 2020 05:43 |
|
|
# ¿ May 22, 2024 09:20 |
|
If it's just one box and it doesn't have to be laser-precise I would just do it by hand. Lay it out with a ruler, a scriber and a center-punch. The slots can be done with a drill and files. If it's a lot of boxes and/or it has to be laser-precise, get a mini CNC that has the correct amount of Z clearance and make some sort of fixture/spoilboard that goes inside the box, locates it precisely in place, and supports the wall to be drilled from underneath. It's a bit more setup but honestly 90% of good CNC work is having a good job setup before you turn anything on.
|
# ¿ Jul 6, 2020 06:06 |