|
They're going to want to check that before they send you the part and it's a whole lot easier to hand someone a caliper and a print than monkey around trying to measure off a digital model
|
#
?
Jul 13, 2017 20:48
|
|
|
|
| # ? May 23, 2024 13:36 |
|
|
rawrr posted:If something requires particularly strict tolerances (or, say, tapped holes) then I'll include a drawing to call it out, but otherwise it's a waste of time to dimension everything as if something is being manually machined. It's not like when they're programming the part based on the step file that they need to know that the holes are 5mm and 20mm away from eachother. Like I said, its not the dimensions they care about, its the tolerances on them. This is why you need drawings unless you say "every feature is +/- .005-inch in location and size" Unless your shop is checking with a CMM in which case a solid model is nice and you can get a nice report on what all the features are.
|
|
#
?
Jul 13, 2017 22:17
|
|
|
acrylic, day 3: you piece a poo poo calling something so chip/shatter-prone "plastic" is a cruel joke e: poo poo when does my Metalworking Thread membership get revoked, do i have to run an aluminum part once every two weeks so it doesnt lapse or
|
|
#
?
Jul 13, 2017 22:31
|
|
|
Ambrose Burnside posted:acrylic, day 3: you piece a poo poo Protip: don't store your acrylic in the freezer 
|
|
#
?
Jul 13, 2017 22:57
|
|
|
mekilljoydammit posted:I'm at the point where I'm tempted to just not specify tolerances at all since nobody follows them anyway. I try, and most the time I can pull it off, but these loving forgings are worthless. You're locating off of an unqualified surface. First op can't get them to run consistent because the forgings are inconsistent, and every other op is hosed because they are dependent on the first op. Imagine having a 6 inch long part with dimensions in the middle that are +/- .002 in height, but that height is referenced off of one feature on the far left, and one feature off the far right. If the part isn't drat near perfectly square when we load it, then at least one side is off by a mile. The part gets screwed down to a plate, and the plate gets loaded into jaws. So starting from the beginning, any variance in the height of the jaws will affect squareness, any variance on the surface of the plate that is loaded into the jaws will affect squareness, any variance on the pins on the plate that the part rests on will affect squareness, any variance in the previous op that machines the surfaces that rest on the pins will affect squareness, and any variance in the forgings that get run on the previous op will affect squareness. Yet we're supposed to run them at about .001 of variance in square over 6 inches. gently caress you. The set up has been going on for probably 30 hours now, and probably still won't be first pieced when I come in tonight. 
|
|
#
?
Jul 13, 2017 23:00
|
|
|
Volkerball posted:I try, and most the time I can pull it off, but these loving forgings are worthless. You're locating off of an unqualified surface. First op can't get them to run consistent because the forgings are inconsistent, and every other op is hosed because they are dependent on the first op. Imagine having a 6 inch long part with dimensions in the middle that are +/- .002 in height, but that height is referenced off of one feature on the far left, and one feature off the far right. If the part isn't drat near perfectly square when we load it, then at least one side is off by a mile. The part gets screwed down to a plate, and the plate gets loaded into jaws. So starting from the beginning, any variance in the height of the jaws will affect squareness, any variance on the surface of the plate that is loaded into the jaws will affect squareness, any variance on the pins on the plate that the part rests on will affect squareness, any variance in the previous op that machines the surfaces that rest on the pins will affect squareness, and any variance in the forgings that get run on the previous op will affect squareness. Yet we're supposed to run them at about .001 of variance in square over 6 inches. gently caress you. The set up has been going on for probably 30 hours now, and probably still won't be first pieced when I come in tonight. If youre milling, this is what probing routines on a 4th or 5th axis with established tolerances are for. Find some feature that should roughly be something (large dia holes are nice) get it where it should be by probing the mounted part, erroring out if the casting/forging is out of parameter with regard to this feature. Adjust until all points are within params. Machine that feature with a helix bore/face/whatever makes it a decent reference. Now remount and adjust your WCS to use this true plane, inspecting the forging with the probe before continuing. This seems like a lot of work but compare to what youre describing, its maybe a 4-5 hour probing and figuring out the first 1-2 job and then you run all the parts consistently only scrapping the ones that were already scrap because gently caress the forging houses. Fast edit: also don't load into jaws unless its a one off or theyre machined soft jaws with expected reference points like casting or forging tabs. 2nd Edit: Large means 1" DIA and >1" deep in my mind here. CarForumPoster fucked around with this message at 23:12 on Jul 13, 2017 |
|
#
?
Jul 13, 2017 23:09
|
|
|
CarForumPoster posted:gently caress all forging houses and foundries forever
|
|
#
?
Jul 13, 2017 23:12
|
|
|
Can I use a scroll saw for brass? Like, up to 1/2". Because band saws are expensive
|
|
#
?
Jul 13, 2017 23:18
|
|
|
CarForumPoster posted:If youre milling, this is what probing routines on a 4th or 5th axis with established tolerances are for. Find some feature that should roughly be something (large dia holes are nice) get it where it should be by probing the mounted part, erroring out if the casting/forging is out of parameter with regard to this feature. Adjust until all points are within params. Machine that feature with a helix bore/face/whatever makes it a decent reference. Now remount and adjust your WCS to use this true plane, inspecting the forging with the probe before continuing. It's so much easier to just run junk tho, so that's probably what they will have us do. These parts never come back so whatever I guess. It's one thing to come into a job like that that's already up and running, but trying to get a part out that can be first pieced is a nightmare.
|
|
#
?
Jul 13, 2017 23:21
|
|
|
rawrr posted:My pet peeve is shops that still require fully dimensioned engineering drawings even though the step file already has all the dimensional data :< N+1 to I need tolerances. Also the shop I'm at now has 3 people (myself included) who use solid models in cam software, 2 mill one lathe. Everyone else, 40 people give or take, is either conversational on the hurcos or using a prototrak on knee mills. We actually just bought more computers and software so people who need dimensions can get it themselves. I'm teaching 3 people how to use inventor and autocad from scratch right now. Its working pretty well and I'll probably wind up teaching everyone that wants to learn in time.
|
|
#
?
Jul 13, 2017 23:33
|
|
|
Lol, just lol if you don't use GD&T for every part, even silicone masking patches.
|
|
#
?
Jul 14, 2017 00:01
|
|
|
Metal Geir Skogul posted:Can I use a scroll saw for brass? Like, up to 1/2". Because band saws are expensive Yes, go for it. I remember youtube superstar click spring using a coping saw to cut his brass and a scroll saw is like a motorized coping saw. Half inch is thicker that what he does but it will work.
|
|
#
?
Jul 14, 2017 00:07
|
|
|
That's what got me on it. I don't really need a band saw, and the cheapest "acceptable" versions of each are $76 for the wen scroll saw and $225 for the wen 10" band saw. The 9" wen is $125 but for some reason it scares me.
|
|
#
?
Jul 14, 2017 00:19
|
|
|
Maybe it's just because the machinery we have and the work we so is a naturally very accurate process, but we generally won't touch a part if we don't have a print with tolerances. Solid model is great for making a program, inspection not so much.
|
|
#
?
Jul 14, 2017 00:58
|
|
|
We completely make our own print that only has the dimensions and tolerances we touch. Even if the customer provides a print, we always, always, make our own. It insulates us from the "but the print says 16 ra there" on a feature we don't touch. That and sometimes we get horrible prints and it forces folks to focus on just the critical stuff. I've been commissioning an upgraded centerless grinder and man, it's awesome to see a machine come together. It's cool to be all like "comp .00002" and then have to use the laser to verify the part diameter change. The only bummer is the chiller is undersized so our capability is now constrained by thermal stability.
|
|
#
?
Jul 14, 2017 01:29
|
|
|
Yooper posted:We completely make our own print that only has the dimensions and tolerances we touch. Even if the customer provides a print, we always, always, make our own. It insulates us from the "but the print says 16 ra there" on a feature we don't touch. That and sometimes we get horrible prints and it forces folks to focus on just the critical stuff. We do the same thing occasionally, but you still need some kind of print from the customer to get the tolerances in the first place, no?
|
|
#
?
Jul 14, 2017 01:34
|
|
|
Metal Geir Skogul posted:Can I use a scroll saw for brass? Like, up to 1/2". Because band saws are expensive Sure. If you can cut it with a hacksaw, you can cut it with a scroll saw -- and you can cut p much everything with a hacksaw. Get a blade with relatively fine teeth and go slowly. Also, time to mention my favorite tool in the entire world again: a jeweler's saw. Like a coping saw but the blades are far thinner, like wires. So useful in so many places. Ambrose Burnside posted:acrylic, day 3: you piece a poo poo Yeah acrylic blows. Even laser-cutting it sucks, because while a laser will cut it nicely, it stinks to hell, and if you don't cut it all the way through and decide to just crack the piece out, wear some gloves bc it will fracture just like glass and make a razor sharp edge. If you need to machine a clear plastic, polycarbonate is way way nicer. Anyway, have you tried using a mixture of dish soap and water as a cutting fluid? That's what we generally do when working acrylic, based on the recommendations of an insane old plastics engineer, and it seems to help a fair bit. There's still occasional chipping but it's reduced. Oh and keep the stuff far far away from alcohol. A really cool thing to do is cut a piece of acrylic -- looks fine -- and then wipe the dust off with isopropanol and watch over the next few minutes as it spontaneously fogs over and cracks along the cut edge. 
|
|
#
?
Jul 14, 2017 03:59
|
|
|
A Proper Uppercut posted:We do the same thing occasionally, but you still need some kind of print from the customer to get the tolerances in the first place, no? Sometimes, but our process is so specific to the part that we have the same tolerances for 99% of our product. It's nice to get a print, but in those occasions where there's DDC issues we work with whatever. Anyone have any experience with production scheduling software? We've got a custom package right now, but I'm curious what's out there.
|
|
#
?
Jul 14, 2017 13:13
|
|
|
Metal Geir Skogul posted:That's what got me on it. I don't really need a band saw, and the cheapest "acceptable" versions of each are $76 for the wen scroll saw and $225 for the wen 10" band saw. The 9" wen is $125 but for some reason it scares me. Check Craigslist. You should be able to get a scroll saw for way less or a 12 inch craftsman bands saw for around a 100 bucks.
|
|
#
?
Jul 14, 2017 14:08
|
|
|
Volkerball posted:I try, and most the time I can pull it off, but these loving forgings are worthless. I'm at the opposite side of the thing - in an R&D/development lab for a small engine company where the production is in China. You design stuff with datum features, you do testing and research to determine the correct surface roughness of things, etc etc etc... and then the factory just randomly switches to some other supplier without telling anyone downstream and you wonder why all of a sudden the stuff coming over is garbage. "Gee, which spec did the purchasing guy for the factory get paid off to ignore from the new supplier?" I half suspect but can't prove.
|
|
#
?
Jul 14, 2017 14:23
|
|
|
So the parts i'm running are milled from 1/2" acrylic plate. I clamp the blanks with toe clamps, the program drills some countersunk holes that line up with the tooling plate, it stops and i remove the toe clamps and add screws so that the whole surface can be fly-cut. It adds something like 8-10 minutes to a 40-minute program. I'm tryna think of ways to expedite it, and working out a single workholding solution would be the most obvious way. Side-clamps were my first thought, but the plate gets cut through almost all the way through in several places so i suspect it'd start buckling/lifting off once those cuts are made with side-clamps. There are also some significant slots and holes drilled so vacuum workholding might not be viable (never used one tho). Any other ideas for workholding approaches that don't clear the top of the workpiece and exerts primarily downward force? It's a run of maybe 400 parts. Ambrose Burnside fucked around with this message at 21:09 on Jul 14, 2017 |
|
#
?
Jul 14, 2017 17:35
|
|
|
mekilljoydammit posted:I'm at the opposite side of the thing - in an R&D/development lab for a small engine company where the production is in China. You design stuff with datum features, you do testing and research to determine the correct surface roughness of things, etc etc etc... and then the factory just randomly switches to some other supplier without telling anyone downstream and you wonder why all of a sudden the stuff coming over is garbage. "Gee, which spec did the purchasing guy for the factory get paid off to ignore from the new supplier?" I half suspect but can't prove. I think you're probably dead nuts. I worked in a similar situation for a huge company that produces using Chinese OEMs and even though there was a very large team of us both in country and in the US to try to monitor and help improve their processes, you couldn't get them to configuration control anything, including g code. You reminded me, subbing to a subs subcontracted sub is enough of a problem in the US...I can't imagine that over 10 years the MTBF of Chinese aircraft systems isn't going to go to poo poo for exactly this reason.
|
|
#
?
Jul 14, 2017 17:42
|
|
|
I'm not sure if this is a metal work question or more of a pure material science question: I'm trying to find ways to mark weld seems in a large water tank. You'd think it would be simple, but it has to be visible through 80 feet of water and last as long as the stainless it's attached too. Gotta be applied on site after the welding (still dry, luckily) and can't reduce the strength or corrosion resistance of the stainless in anyway. Colorful Epoxies got put on the back burner because nobody can seem to find to find one that will last decades. I've found a ton of Cermet coatings that would be great, any sort of metal matrix coating seems like it would be a winner. I don't know of the methods to grow them in the field though. Thermal spraying? Some sort of mobile box that does vapor deposition? Someone's gotta have done this before, right?
|
|
#
?
Jul 14, 2017 19:36
|
|
|
M_Gargantua posted:I'm not sure if this is a metal work question or more of a pure material science question: I still think your best bet is colorful epoxies. Whats the function of it? Crack when the weld seams crack? Does it need to be metal based for some reason? This comes to mind and I am pretty sure theres a military version of it that likely has qualifying data: http://www.3m.com/3M/en_US/company-us/all-3m-products/~/3M-Scotch-Seal-Tamper-Proof-Sealant-1252?N=5002385+3293194232&rt=rud I'd also try contacting 3M and Henkel and see if they have something meant for what you want to do. Remember anything that has metal has to be galvanically compatible with the stainless.
|
|
#
?
Jul 14, 2017 19:43
|
|
|
Ambrose Burnside posted:So the parts i'm running are milled from 1/2" acrylic plate. I clamp the blanks with toe clamps, the program drills some countersunk holes that line up with the tooling plate, it stops and i remove the toe clamps and add screws so that the whole surface can be fly-cut. It adds something like 8-10 minutes to a 40-minute program. I'm tryna think of ways to expedite it, and working out a single workholding solution would be the most obvious way. Can you add an op, fixture up just to drill the countersunk holes on another machine, and then your machining center is just doing the finish work? We do this in a turning cell with a Bridgeport paired up to it. The Bridgeport has a 3-jaw chuck on the base and the operator center drills the casting. They can drill about 20 while the machine is running. Before we'd have to wait for the slow-rear end tailstock to move out of the way, center drill it in the lathe, then move the tailstock back in. It sucks adding an op, but the time savings might justify it.
|
|
#
?
Jul 14, 2017 19:51
|
|
|
Yooper posted:Can you add an op, fixture up just to drill the countersunk holes on another machine, and then your machining center is just doing the finish work? Not viable, unfortunately- one man, one machine for this job, everything else is tied up. (it's a prototyping-geared shop that had a relatively big run of parts fall into its lap)
|
|
#
?
Jul 14, 2017 21:08
|
|
|
Ambrose Burnside posted:So the parts i'm running are milled from 1/2" acrylic plate. I clamp the blanks with toe clamps, the program drills some countersunk holes that line up with the tooling plate, it stops and i remove the toe clamps and add screws so that the whole surface can be fly-cut. It adds something like 8-10 minutes to a 40-minute program. I'm tryna think of ways to expedite it, and working out a single workholding solution would be the most obvious way. I'm actually using a vacuum chuck for the first time today. So far so good but im no expert. When I was researching it they sell plates you can bolt on top and machine your own gasket channels so it won't lose vacuum when you break through. Edit: Pierson workholding makes the chuck we got.
|
|
#
?
Jul 14, 2017 21:25
|
|
|
CarForumPoster posted:I still think your best bet is colorful epoxies. Whats the function of it? Crack when the weld seams crack? Does it need to be metal based for some reason? Everything I've found doesn't claim to hold up under water. The epoxies in current use, while an older formula, need to be replaced every few years. It's mostly for an identification system for the areas around the weldments and what's in/under them as an operator aide.
|
|
#
?
Jul 14, 2017 23:58
|
|
|
M_Gargantua posted:Everything I've found doesn't claim to hold up under water. The epoxies in current use, while an older formula, need to be replaced every few years. It's mostly for an identification system for the areas around the weldments and what's in/under them as an operator aide. Stencil and aluminum oxide blast?
|
|
#
?
Jul 15, 2017 00:05
|
|
|
burnished gold foil?
|
|
#
?
Jul 15, 2017 00:06
|
|
|
Sagebrush posted:Oh and keep the stuff far far away from alcohol. A really cool thing to do is cut a piece of acrylic -- looks fine -- and then wipe the dust off with isopropanol and watch over the next few minutes as it spontaneously fogs over and cracks along the cut edge. Really? Sounds crazy... 
|
|
#
?
Jul 15, 2017 00:08
|
|
|
M_Gargantua posted:Everything I've found doesn't claim to hold up under water. The epoxies in current use, while an older formula, need to be replaced every few years. It's mostly for an identification system for the areas around the weldments and what's in/under them as an operator aide. You can't just use paint?
|
|
#
?
Jul 15, 2017 00:31
|
|
|
mekilljoydammit posted:I'm at the opposite side of the thing - in an R&D/development lab for a small engine company where the production is in China. You design stuff with datum features, you do testing and research to determine the correct surface roughness of things, etc etc etc... and then the factory just randomly switches to some other supplier without telling anyone downstream and you wonder why all of a sudden the stuff coming over is garbage. "Gee, which spec did the purchasing guy for the factory get paid off to ignore from the new supplier?" I half suspect but can't prove. We're one factory with maybe 80 employees and we make nothing but our own products. The R&D and process engineers share the same office, and finance is right across the hall. This shop is just a disaster. It's the worst one I've ever worked at when it comes to communication and general competence. For context, those parts had no first piece stamp when I came in, and nothing had been changed, but evidently we were running them now. One time, we ran 1,000+ parts that run at 4-5 parts an hour on one op, and about 6 on EDM at the next op. Something ended up being dicked up with how the parts were anodized or coated, and all of them cracked. All junk. So they change some stuff around, buy a new machine so we can have two operators pumping them out because now we are way behind, and run 1,000 more parts. Well they were supposed to be universal, but evidently they were only functional in certain models. All junk. Run em again. So we just finished another 1,000. And this is a part we've had for over 20 years. It's kind of impressive, really. Volkerball fucked around with this message at 00:48 on Jul 15, 2017 |
|
#
?
Jul 15, 2017 00:44
|
|
|
Yooper posted:Stencil and aluminum oxide blast? Too hard to see at a distance, plus they don't want anything to potentially weaken the main structure, so any sort of abrasion marking system was also ruled out for that purpose. shame on an IGA posted:burnished gold foil? Too easy to scratch off. CarForumPoster posted:You can't just use paint? Doesn't last long enough, same as the rest of the epoxies tried. Hence the idea to see if there was a metallic matrix/cermet that would last full length of the facility. But everything i've found is about applying them to small tool heads or turbine blades during the manufacturing.
|
|
#
?
Jul 15, 2017 00:45
|
|
|
A line of colored glass frit bonded in place with a torch?
|
|
#
?
Jul 15, 2017 00:59
|
|
|
shame on an IGA posted:A line of colored glass frit bonded in place with a torch? I haven't heard of that before. A quick wiki doesn't mention anything about it being used on stainless. And wouldn't the glass crack over time?
|
|
#
?
Jul 15, 2017 01:11
|
|
|
I'm not sure I understand the goal but maybe string a cable along the weld and as a path from the manhole like the stop bell on a bus.
|
|
#
?
Jul 15, 2017 01:35
|
|
|
A stainless flag welded in next to the main weld.
|
|
#
?
Jul 15, 2017 03:19
|
|
|
Metal Geir Skogul posted:A stainless flag welded in next to the main weld. The first kind of solution I tried to get them to bite on. Stainless and welder hours are cheep compared to other line items in the budget. Will report back with what they go with. And maybe report back in 10 year increments about how it held up.
|
|
#
?
Jul 15, 2017 03:27
|
|
|
|
| # ? May 23, 2024 13:36 |
|
|
Been thinking of making a propane forge (from an old compressor tank I got) but the refractory material I was thinking of (ceramic wool) I am reading has asbestos like properties wrt cancer and health, so I don't think it's something I want to drag home.... Any tips on alternative forms of insulation, I've heard all kinds of stories of people using stuff like cat litter which is some kinda clay.
|
|
#
?
Jul 15, 2017 07:10
|
|
