|
Dance Officer posted:2.5D and 3-axis or 3D milling are in practice very similar to each other. You're working in 3 axes in both cases and the programming is generally the same, but with 2.5D you're more limited in the moves you can make. The big thing you can't do with 2.5D milling is helicoidal moves that are often used for getting into pockets with optimized roughing strategies. There may be some chamfering or contouring that needs 3D movement. And helicoidal moves are often handy for tightly toleranced holes. This is my experience. I learned on V-Carve and Fusion 360, and the fundamentals are very much the same. The UI is different, but once you know how to set up an operation in one, select your tool and feeds and speeds that’s the game on most CAM programs.
|
#
?
May 24, 2024 14:11
|
|
|
|
| # ? May 26, 2024 05:30 |
|
|
I have a weird request that could help me a whole bunch. If you use an industrial CNC machine at work with a fairly common industrial control, (Fanuc, Siemens, Hurco, any of em) would you take a photo of the video cable for the screen? I am building a project for remotely monitoring my Fadal using screen caps and I am wondering if it could be used elsewhere.
|
|
#
?
May 24, 2024 14:18
|
|
|
Ethics_Gradient posted:I see that they cover 2.5 axis milling; how much is involved in making the jump to 3 or 4 axis, and/or metal? There is actually a second course offered in the Spring that covers 3 and 4 axis machines as well as aluminium and plastic (definitely interested in learning to work with metal) that I’d love to take, but this year it was only offered in the daytime so I don’t like my chances that there’ll be a night one next year. OTOH there’s a good chance I’ll have access to their shop as a student so may be able to learn on my own, or worst case can join my local hackerspace which has a couple CNC machines available. Vcarve is a great tool for 2.5 axis milling and it will get you familiar with the basic tool flow (design->CAM->gcode->mill). This is a tool meant for sign shops and unsophisticated users so it is pretty easy to use. I used it more when I had a router and mostly for engraving work now. The jump to 3 axis programming/machining in something like Fusion360 is significant. In my experience, Fusion CAM is pretty easy to use as these tools go but would have a steep learning curve so someone starting out due to all of the background machining knowledge needed. If you don't have machining knowledge, I highly recommend taking a few manual machining courses to get a feel for how metal behaves. Going to 4 and 5 axis simultaneous machining is not a linear progression as more like an exponential increase in difficulty and cost. The consequences for failure (crashing) go up steeply and you really have to know what you are doing to get good results. CAM software costs also increase massively. 3+1 and 3+2 axis machining is more of a linear progression but you still have the very expensive consequences. Ethics_Gradient posted:Was also wondering what sorts of jobs might this open up, and what else might I want to learn on my own if I wanted to increase my employability? I’m currently on a contract in an unrelated field through the end of the year and am likely to try and jump back into regular teaching if I can nab a midyear gig (it probably pays the best of my options right now, and I'd really like the summer off for an extended trip to Japan), but I’d also be willing to take a pay cut if I could find a CNC gig where I was learning a lot on the job. To be honest at your level of experience, the best job you are likely to get is "machine operator". I.E. you hit the green button and only hit the big red button if the machine makes a funny noise. Programming jobs would typically require more training/experience. Machine operator is as close to unskilled labour as the shop can make it. The investment in training someone is significant and unless the shop thinks you are committed, I would think operator would be the best you would get. I have heard that https://academy.titansofcnc.com/ has some decent resources but I haven't checked them out myself. Ethics_Gradient posted:Tangental to all this, but a lot of my interest in CNC comes from wanting to eventually get into making dies and molds: molds for casting a sort of resin stuff (not super high temp, but probably too spicy for a resin 3D print), and the other for stamping thin sheets of brass. They'd both be relatively small and have fine details. Curious if anyone in here makes either or had some resources to share. I have made molds, stamping dies, and many other things like that. There are lots of resources out there but you are asking a pretty open ended question and you need to get the basics down. These aren't conceptually difficult but the devil is in the details.
|
|
#
?
May 24, 2024 14:30
|
|
|
CarForumPoster posted:I have a weird request that could help me a whole bunch. If you use an industrial CNC machine at work with a fairly common industrial control, (Fanuc, Siemens, Hurco, any of em) would you take a photo of the video cable for the screen? Most modern controls (LCD based) do not have separate monitors and there are no video cables. The monitor is integrated into the control panel and connected to the PLC through proprietary bus cable. The old SECIOS control on my machine had what was effectively a MDA cable between the processor and the monitor. It used a funky obsolete connector that now costs $100s each.
|
|
#
?
May 24, 2024 14:35
|
|
|
HolHorsejob posted:Anyone know how much laser you need to cut 1/2" plywood? A friend is considering a laser cutter Possible on a 100w if you don't mind taking all goddamn day, but if that's your main use case you really want a cnc router, or a little practice with a band/scrollsaw if the design isn't super complex Buncha my shop tools are down so I've been spending time at the local makerspace and istg I'm the only person there who uses these things to cut anything but, like, one big rectangle A Wizard of Goatse fucked around with this message at 15:02 on May 24, 2024 |
|
#
?
May 24, 2024 14:55
|
|
|
ZincBoy posted:Most modern controls (LCD based) do not have separate monitors and there are no video cables. The monitor is integrated into the control panel and connected to the PLC through proprietary bus cable. The inside of the control panel is what I’m hoping to get a picture of. For example my fadal has a funky proprietary monitor but there’s a BNC with component video from the control card. If you could take a picture of where you might expect to find a video signal, it’d be very helpful. At some point, the control is sending a signal to a screen that the OEM purchased. I want to see what that point of video transmission looks like on various machines. Maybe it’s HDMI/DVI that’s passed through a ribbon cable, maybe it’s a BNC, etc. I want to essentially man in the middle the signal.
|
|
#
?
May 24, 2024 15:02
|
|
|
CarForumPoster posted:I have a weird request that could help me a whole bunch. If you use an industrial CNC machine at work with a fairly common industrial control, (Fanuc, Siemens, Hurco, any of em) would you take a photo of the video cable for the screen? How old is too old? I have a vmc equipped with a fanuc 0-m control, circa 1987
|
|
#
?
May 24, 2024 15:05
|
|
|
Ethics_Gradient posted:Tangental to all this, but a lot of my interest in CNC comes from wanting to eventually get into making dies and molds: molds for casting a sort of resin stuff (not super high temp, but probably too spicy for a resin 3D print), and the other for stamping thin sheets of brass. They'd both be relatively small and have fine details. Curious if anyone in here makes either or had some resources to share. I've done this at the hobby scale and the hard part is designing a good mold/stamp, not telling the robot what to do. Metal stamping at any level of detail or complexity is especially brutal, or I'm especially bad at it - backyard casting is pretty doable and there's a lot of resources to help you, I'd start with something like this albeit it's a bit focused on metals
|
|
#
?
May 24, 2024 15:22
|
|
|
Ethics_Gradient posted:Was also wondering what sorts of jobs might this open up, and what else might I want to learn on my own if I wanted to increase my employability? I’m currently on a contract in an unrelated field through the end of the year and am likely to try and jump back into regular teaching if I can nab a midyear gig (it probably pays the best of my options right now, and I'd really like the summer off for an extended trip to Japan), but I’d also be willing to take a pay cut if I could find a CNC gig where I was learning a lot on the job. On the less technical advice front: go talk to the shop teachers at your CC. They'll know what the local industry is looking for and what level of experience you'd need to get hired. If you're excited enough they may (depends on the people!) be more flexible about teaching and willing to just let you hang out and learn. They might also be able to point you to other resources in your area: some CCs or universities will run machining "bootcamps" or the like. And if you're really interested in education they can probably provide some tips for how to get into it. For high school education specifically, there's a lot of initiatives to expand shop class and get people into the trades. If you've got teaching experience and can get up-to-speed on CNC machining there's absolutely teaching jobs out there. However: it's also not something where you take a community college class or two and you're good to go: the best programs have teachers with industry experience, sometimes a lot of it. If you're looking at a school without a program then the bar might be lower, but then you need to be able to bootstrap (which probably means finding grants to buy equipment!) There's a lot of money and resources out there, but that also means that there's a lot of schools applying for it, so again you need to be competitive and driven and have a plan.
|
|
#
?
May 24, 2024 15:32
|
|
|
CarForumPoster posted:The inside of the control panel is what I’m hoping to get a picture of. For example my fadal has a funky proprietary monitor but there’s a BNC with component video from the control card. If you could take a picture of where you might expect to find a video signal, it’d be very helpful. The old CRT based systems were like this to an extent. The SECIOS cable (from 1988) had a "video" twisted pair and vertical and horizontal deflection signals The CRT did not generate its own deflection from sync and it was a dumb amp. It is possible to convert this to a format you could feed to a capture system but it would require some pretty specific circuitry. Older LCD based systems may or may not have standard video interfaces. I would expect many of the integrated monitors to be connected by LVDS or other proprietary interfaces. Once they switched to full colour interfaces the chances of a standard interface increase greatly. Most older lcd panels did not have standard video interfaces and required a funky parallel bus with specific timing. Modern (windows) based controls will almost certainly have standard HDMI or similar though it may still be a LVDS interface like a laptop. The connection used is not likely to be standard however. The easiest way to do this would be a web cam with a super wide angle lens so it doesn't need to be in front of the monitor. You could set it up so it was just a few inches above and in front of the screen. Then do distortion correction in software. This would give a universal solution that doesn't require reverse engineering each control.
|
|
#
?
May 24, 2024 16:58
|
|
|
A Wizard of Goatse posted:Possible on a 100w if you don't mind taking all goddamn day, but if that's your main use case you really want a cnc router, or a little practice with a band/scrollsaw if the design isn't super complex Yeah, the level of scorching here will be pretty intense, plus kerf issues with focal lengths, and smoke/VOCs from the glues and other poo poo in the ply.
|
|
#
?
May 24, 2024 17:31
|
|
|
ZincBoy posted:The old CRT based systems were like this to an extent. The SECIOS cable (from 1988) had a "video" twisted pair and vertical and horizontal deflection signals The CRT did not generate its own deflection from sync and it was a dumb amp. It is possible to convert this to a format you could feed to a capture system but it would require some pretty specific circuitry. Having pulled the back off my machine's control, it's built like a laptop and good luck getting anything in there without trashing it somehow. Seconding the fisheye webcam and bracket idea.
|
|
#
?
May 24, 2024 17:40
|
|
|
What happens when the lighting changes? Lens gets coolant on it? Operator needs to use it? A webcam introduces a huge number of failure modes, probably just wont work, and is WAY more complicated to support in the long run. EDIT: If you guys are able to "fix it in software" to make a sub $250 webcam reliably read text in a huge number of dirty, misty, humans involved environments you're some god tier engineers and should be at boston dynamics hastening skynet IMO. fins posted:How old is too old? I have a vmc equipped with a fanuc 0-m control, circa 1987 I'll take whatever! CarForumPoster fucked around with this message at 22:47 on May 24, 2024 |
|
#
?
May 24, 2024 22:40
|
|
|
that would presumably be why webcams run $10-$20, while there is no general aftermarket solution doing what you want and the proprietary single-device ones run thousands of dollars
|
|
#
?
May 24, 2024 23:29
|
|
|
CarForumPoster posted:What happens when the lighting changes? Lens gets coolant on it? Operator needs to use it? A camera is the easy generic solution. I though you just wanted to replicate the display for remote viewing but adding OCR to scrape the screen is not that much harder given that the screens are low resolution and pretty high contrast. Modern OCR in things like opencv is amazingly good. It would be pretty easy to do even with a raspberry pi and the integrated cam. If the screen is readable to a human than OCR will work (and with the catchpa war, OCR can be even better than most humans). I am not sure you understand just how many video standards there are. There is no way you would make a generic solution that could work on anything from MDA to VGA to HDMI for any reasonable amount of money. This is ignoring the fact that most controls will have proprietary interfaces between the control board and LCD panel that would need to be reverse engineered for each version of the control. CarForumPoster posted:I'll take whatever! Here is the manual page with all of the info from the 1988 SEICOS control that was on my machine. I tossed all of the internals when I retrofitted it so I don't have pictures of the connector but it was an obsolete rectangular thing that only seemed to be used in Japan. I am 80% sure that it is some form of MDA signalling given the processor was an 8088. No guarantees the timing is the same though.  Note the two different connectors that were used for this one control depending on the application.
|
|
#
?
May 25, 2024 04:47
|
|
|
ZincBoy posted:... ??? they are just male/female honda connectors. specifically "Right-angle DIP type male/female connector with locking block for PCBs"
|
|
#
?
May 25, 2024 11:03
|
|
|
ZincBoy posted:A camera is the easy generic solution. I though you just wanted to replicate the display for remote viewing but adding OCR to scrape the screen is not that much harder given that the screens are low resolution and pretty high contrast. Modern OCR in things like opencv is amazingly good. It would be pretty easy to do even with a raspberry pi and the integrated cam. If the screen is readable to a human than OCR will work (and with the catchpa war, OCR can be even better than most humans). You’re guessing at a lot of things I might want to do. What I want is for people to post pictures of their video connectors. I’ve spent >1000 hours making OCR and image recognition apps and my opinion on opencv via video feeds is very different than yours. Afaik it has no ocr capabilities and would need something like tesseract or another model.
|
|
#
?
May 25, 2024 13:18
|
|
|
E:Quote isn’t edit
|
|
#
?
May 25, 2024 13:32
|
|
|
CarForumPoster posted:You’re guessing at a lot of things I might want to do. What I want is for people to post pictures of their video connectors. I’ve spent >1000 hours making OCR and image recognition apps and my opinion on opencv via video feeds is very different than yours. Afaik it has no ocr capabilities and would need something like tesseract or another model. Yeah, OpenCV would be more for detecting if your part monkey is dicking around on his phone instead of deburring parts for OP2 (why isn't the machine doing it for him?).
|
|
#
?
May 25, 2024 22:18
|
|
|
|
| # ? May 26, 2024 05:30 |
|
|
ebay: cnc hmi sellers tend to post pics of ports and connectors (though usually only the external ones). Two other possibilities I haven't seen mentioned: 1) If any of these HMIs have a DMA-capable interface like firewire or a pci slot, you could DMA the bits straight out of the framebuffer. That'd be pretty cool. 2) Find one running WinCE or similar, crack it and install vnc. Like those people who install DOOM on $100k HP oscilloscopes.
|
|
#
?
May 26, 2024 03:48
|
|
