|
Tres Burritos posted:Is it a bad idea to use linear scales for closed loop on a CNC machine? I found a post on cnczone ( My first hand knowledge is limited, but we have some custom grinders that originally had glass scales (linear encoders) for feedback. The machine operation was erratic at best. We don't think it was an issue with the scales, machine, or process, but how the software read what it saw and responded to it. In hindsight a more sophisticated machine tool control (professionally designed Fanuc or Siemens) is going to be able to handle that sort of feedback better than our circa 1998 VB6.0 PC based control. Our setup was relatively low inertia, I can only imagine the difficulties of reading position, predicting end point, making adjustments, all at some ridiculous feed rate. I chatted with a guy at the Fanuc booth at IMTS about it and there's way more than just measuring and making an adjustment.
|
#
?
Apr 8, 2019 19:23
|
|
) that indicated servos were better for closed loop stuff, but the reason given was, 'Well the software just doesn't work'. The other reason I found was that 'Setting up scales is fiddly and time consuming'. 
|
|
| # ? May 10, 2024 16:46 |
|
|
From working with other positioning stuff, in general closed loop algorithms are going to go weird if there's a lag between the output and the input changing. Weird overshoot stuff - think of it like, the control loop sees nothing's moving yet so it'll drive the motor even harder. Probably could be tuned around but it's an issue.
|
|
#
?
Apr 8, 2019 20:47
|
|
|
All of our wire edm machines have glass scales. But yes they are made by a major manufacturer with the software to go with it. Doing it on your own might be difficult.
|
|
#
?
Apr 8, 2019 21:03
|
|
|
mobby_6kl posted:I guess this would be the best place to ask this. If by hand tools you mean a drill, you can drill in to the end of the shaft, tap it and then put a screw or bolt in the end with an appropriately sized washer that will hold it on there. Use Loctite. or lock washer.
|
|
#
?
Apr 8, 2019 21:45
|
|
|
My favorite is to drill through and hammer in a spring pin.
|
|
#
?
Apr 8, 2019 22:19
|
|
|
All the siemens 840 stuff I've used was rotary encoder in the servos and it holds to 10um so it's definitely doable
|
|
#
?
Apr 8, 2019 22:41
|
|
|
bred posted:My favorite is to drill through and hammer in a spring pin. this is what i'd probably do if i had the right resources on hand, but anything involving drilling the shaft may get dicey if the shaft is hardened and tooling is limited. skillwise, grinding a flat is also far more hobbyist-friendly than drilling a single properly-aligned hole through a comparable whole lot of steel- anybody can manage a flat with a dremel and 5 minutes' practice grinding, while id expect a whole lot of broken drills under comparable conditions if you went the thru hole approach. its also easy to forget how big a pain in the rear end it can be to find "just one" roll pin if you don't have a particularly fastener-rich local hardware store. that said, if the shaft was actually reasonably fileable with standard steel hand files and the file didn't slide off like glass you should be fine to drill it with common jobber drills without undue risk of fuckuppage if you have access to a drill press with vise, and drill the hole using an appropriate RPM, cautious-but-not-TOO-cautious feed, and a lubricating oil/grease/wax on the face of the drill. Ambrose Burnside fucked around with this message at 23:05 on Apr 8, 2019 |
|
#
?
Apr 8, 2019 23:00
|
|
|
If I were to stick to the flat, as others have said, just use a flat + set screw with loosening prevented via addition of retaining nut/spring washer/loctite. and if that's basically your existing setup and the problem is with loosening, def start with what may be an extremely easy and quick improvement of your existing efforts before starting over from scratch, if all you need is a daub of loctite red then take the easy out
|
|
#
?
Apr 8, 2019 23:07
|
|
|
shame on an IGA posted:All the siemens 840 stuff I've used was rotary encoder in the servos and it holds to 10um so it's definitely doable Right, but if it's at the servos, it's not getting thrown off by backlash. Similarly, if it's a real setup with presumably ballscrews and stuff, it's not using the feedback to compensate for backlash much.
|
|
#
?
Apr 9, 2019 02:36
|
|
|
Tres Burritos posted:Is it a bad idea to use linear scales for closed loop on a CNC machine? I found a post on cnczone ( I can comment on professional-level machines from a mechanical perspective. All of this is in addition to the feedback loops and servo tuning others have mentioned, which are huge problems (though maybe not as much on a hobby machine, just due to the much lower feedrates. 10ipm is a bit easier than 500 ipm.) Theoretically, they're really helpful. In practice... it's a bit harder. Ultimately, trying to use a linear encoder to account for inaccuracy is treating a symptom, not the cause. Why is your machine inaccurate? If you've got backlash, you need to take it out: if you've got backlash, regardless of whether your software can comp for it it's going to cause vibration due to the slop. If you've got inaccurate leadscrews/ballscrews, you should replace them, or if it's extremely repeatable, make comp tables (either by laser measurement or just manually tweaking them.) If you've got other mechanical defects (axes aren't straight or perpendicular, machine distorts under cutting forces, etc), scales aren't going to help you at all. Even the case that you hear the most (thermal compensation) is over-sold by machine tool builders. If your machine's axes are generating enough heat to significantly change the length, you can bet they're distorting in other ways. And if your shop temperature is changing enough that you need glass scales to hold tolerance, then you'd better be accounting for temperature shift in the part as well. In the right applications, they can be helpful. Ultra-high precision machines (EDM, die-mold, micro-milling) use them because of the extreme degree of accuracy necessary *. But even then, scales on professional level machines are improving the repeatability of an already reliable system. They are much less/not useful to compensate for flaws somewhere else in the design or structure. The vast majority of professional users don't need glass scales. On a hobby-level machine (I assume, since you ask about steppers) this isn't going to be your limiting factor by a long shot. * Also, not all scales are made equal. The type of scales you'll find on an extremely high precision machine are not the same kind as what you get as an option for some production machining center. Different levels of accuracy and different purposes. I think Heidenhain's got some good writeups on this if you go searching.
|
|
#
?
Apr 9, 2019 05:14
|
|
|
Idk how you'd go with what's available but would a cotter pin kind of arrangement work on the sprocket?
|
|
#
?
Apr 9, 2019 06:56
|
|
|
Thanks all! I do have a drill of course but no drill press, so yeah drilling into the shaft might be a bit sketchy. It was reasonably easy to grind for the key but still since it's not easily replaceable I think that would be the last resort if filing a flat spot and Loctite won't work. It's a pretty small sprocket so something like a cotter pin might interfere with the chain if it sticks out too much, but I'll check it out.
|
|
#
?
Apr 9, 2019 09:27
|
|
|
Carpe machina my dude.  Then you can stand and admire it because you have no _actual_ need for it and none of the right size T nuts to hold anything down even if you did. mobby_6kl posted:Thanks all! I do have a drill of course but no drill press, so yeah drilling into the shaft might be a bit sketchy. It was reasonably easy to grind for the key but still since it's not easily replaceable I think that would be the last resort if filing a flat spot and Loctite won't work. Loctite blue is worth a shot. Rapulum_Dei fucked around with this message at 11:42 on Apr 9, 2019 |
|
#
?
Apr 9, 2019 11:39
|
|
|
Has anybody ever heard of... sub-commercial servo press experimentation, successful or otherwise? Google doesn't come up with anything at all, which is actually kind of surprising, because on paper it seems like low-hanging fruit for some retired engineer to take a crack at- servo presses can do almost any presswork within their operational parameters extremely effectively, even the cheapest commercial offerings are ludicrously expensive, and you could probably get a lot of preliminary gently caress-around mileage from just off-the-shelf parts tacked onto an end-of-life press frame. People have made garage-practical improvised versions of basically every other flexible-application hammer/press design I can think of and I can't think of any reason why servo presses can't follow suit. Ambrose Burnside fucked around with this message at 22:53 on Apr 10, 2019 |
|
#
?
Apr 10, 2019 22:46
|
|
|
Karia posted:I can comment on professional-level machines from a mechanical perspective. 2nding this. Glass scales give you an absolute incremental measure of the machine's motion, which eliminates ballscrew compensation, backlash, thermal growth of the screw and nut, and a bunch of other factors. In theory it provides you with an extremely accurate external source to drive the motion from, in practice, outside of specialty systems designed from the ground up to use the scales, you really don't see the benefits over just using the machine as designed. The motion systems on even an entry level Haas are miles above what you'll see on a hobbyist or converted machine. Every machine has a ball-bar test done on it, which uses a glass scale on some magnetic ball bearings to very VERY precisely check the machine squareness, backlash, ball screw comp, and servo tuning. You'll see results on fine motion to within 0.5-2 1/10,000th of an inch, and bulk motion cutting parts under load that are within 1-2 10ths. On stuff like EDM machines, Die/mold machines, where you're going to be cutting a faux leather pebble skin pattern into a set of injection mold dies, single digit microns are important, and as such the machines are designed to accommodate glass scales and the motion system is tuned for that.
|
|
#
?
Apr 11, 2019 00:24
|
|
|
Finally stumbled onto something halfway to a micro servo press: https://makezine.com/projects/one-ton-linear-servo/ tbh this seems like a much more sensible and affordable way to go about this on a shoestring budget if you don't need the super-fine stroke control of a directly-driven leadscrew, but would be perfectly adequate for "close enough" sheet metal stuff, doing a bunch of press-fitting, riveting, etc i'd also assume that this would give much faster press travel/cycling than comparable homebuilt pneumatic/hydraulic bottle jack presses, but even if it didnt you could make good use of the electronic control for doing stuff repeatably, like a low-rent version of the real deal. stick load sensors in swappable press plates and program cycles to take advantage of em like, idk, " 'rapid' to 0.10Z, continue travel at pressing speed, stop servo once load cell returns reading of X or greater, half-second of dwell, rapid to top of cycle" (can you tell that someone with purchasing power at my shop found out about servo presses but cant quite justify the MSRP and i know he's a sucker for stupid sugggestions like "but what if we built a really lovely small version, for fun, and research, and possibly even occasional revenue-generating purposes" ) e: on that note, every cheap hydraulic jack is gonna be hand-pumped, right? like with an up-down hand pump as opposed to a rotary input? 20-30 tons is just right for this application and that's easy from a consumer-grade hydraulic bottle jack but not from a scissor jack, but a servo motor can't crank a hand pump. mcmaster carr's got something compatible intended for electrically-pumped power paks but ofc it's way too expensive to justify Ambrose Burnside fucked around with this message at 01:22 on Apr 11, 2019 |
|
#
?
Apr 11, 2019 00:58
|
|
|
Karia posted:Stuff Methylethylaldehyde posted:2nding this. Thanks for the replies! It's hard to get good non-bullshit answers out of the internet some times about this stuff.
|
|
#
?
Apr 11, 2019 01:02
|
|
|
Ambrose Burnside posted:Finally stumbled onto something halfway to a micro servo press: https://makezine.com/projects/one-ton-linear-servo/ Out of curiosity, why a servo press if you arn't getting the fine stroke control? I would think that is the big advantage of a servo press over a more traditional one. If you are looking for a relatively low cost press solution, one option is re-purposing a electric/hydraulic log splitter. Pretty much everything you need for a few hundred dollars. Many of them have adjustable stroke lengths but obviously not to the level of a servo system. You could also achieve a similar effect by using a solenoid to trip the return.
|
|
#
?
Apr 11, 2019 04:12
|
|
|
ZincBoy posted:Out of curiosity, why a servo press if you arn't getting the fine stroke control? I would think that is the big advantage of a servo press over a more traditional one. "Fine" is relative, and the repeatability of a mechanical jack/press mechanism driven by a CNCed servo/stepper would still beat out any other press you could put together yourself by some exponential factor. commercial servo presses brag about being repeatable to figures in the ten-thous or better- a DIY version being repeatable to just 0.010 or 0.050" or something like that, combined with rudimentary force-feedback, opens a ton of doors to how you can use the thing, given how the alternative is "no positional repeatability at all, and a bottle jack pressure gauge for very approximate force repeatability, user-maintained, at best". I guess a concise thesis here on the specific appeal of this specific approach is: "we believe that the servo press design approach has the potential to offer the automation/labour-saving and tight manufacturing quality control benefits of commercial stamping press machines at a tiny fraction of the cost and shop footprint, with the compromise of offering inferior workpiece envelope/applied force/working speed; if acquiring any sort of real production press whatsoever is out of the question, the above tradeoffs don't apply, so why not" Ambrose Burnside fucked around with this message at 23:46 on Apr 11, 2019 |
|
#
?
Apr 11, 2019 17:08
|
|
|
Right now I use a hydraulic press to do sheet metal artistic work using acrylic dies and urethane rubber pads; i'm lucky enough to have a pressure gauge but even then no two pieces are the same, I get a lot of failed parts due to drawing too fast, etc. Without the gauge you basically guess as to where you're at, the only definite indicator being the snap of the sheet shearing because you, icarus-like, flew close to the 20-ton sun. Servo presses, though? You can program an operating cycle juuuuuust so; complex deep-drawn forms that would traditionally go through 5 separate dies can be done in a single die with a complex press cycle that may repeatedly apply pressure, dwell, back off etc. to do that five-die draw in just one. I don't expect to be able to just do that with an experimental press made from Aliexpress CNC router parts or whatever like it's nothin, but even a rudimentary ability to tailor a press cycle to the needs of a given part like this isn't normally something you can really do for less than "a low six figures" or so. If we can do it on the cheap we don't have much to lose, and even if it's a bust it's a learning experience (that I don't have to pay for out of pocket eyyy) e: it looks like there has been a fair bit of hobby work on CNC *press brakes*, focused on stepper-driven multiaxis back gages that handle all the finicky indexing work of doing a bunch of bends at varying depths, i e https://www.youtube.com/watch?v=T5R7O8GPSSg but they generally ignore the actual pressing aspect of it, one example had a solenoid controlling the operation of a pneumatic bottle jack but only in an on/off capacity. good to know, at any rate the closest off-the-shelf solution to what i'm imagining is prolly the potter electric-hydraulic press, the envy of urethane dieformers everywhere https://www.youtube.com/watch?v=D8bqavPnmYk which is fantastic for a benchtop artistic press, but it's still fundamentally manually-operated albeit with the hard labour of pumping the jack automated. i'm also dubious about a CNC-controlled hydraulic pump n powertrain being a good fit here- I know that hydraulic actuation used to be common back in the old NC days but it doesn't sound like it's a particularly good way to do things given modern motor options and very poorly-suited to being done on the cheap https://www.cnczone.com/forums/linear-and-rotary-motion/129886-cnc.html Ambrose Burnside fucked around with this message at 22:16 on Apr 11, 2019 |
|
#
?
Apr 11, 2019 17:18
|
|
|
Ambrose Burnside posted:Right now I use a hydraulic press to do sheet metal artistic work using acrylic dies and urethane rubber pads; i'm lucky enough to have a pressure gauge but even then no two pieces are the same, I get a lot of failed parts due to drawing too fast, etc. Without the gauge you basically guess as to where you're at, the only definite indicator being the snap of the sheet shearing because you, icarus-like, flew close to the 20-ton sun. You can get an electric log splitter, a load cell, amplifier set and some control electronics for like $5000. Get a VFD hooked up to the electric motor that takes 0-10v for the speed, hook the load cell up in line with the ram, and drive the whole thing electronically. The load cell gives you true force applied to the part, and can chase down a part that tends to yield or flow and keep applying more force until it levels out at some programed level, and the VFD lets you do a much more finely tuned job in terms of force over time and creep up vs. a simple relay that starts/stops the motor. Programming it all will be annoying, but an arduino or raspberry pi would be able to do the input/outputs without issue. If you wanted to be extra fancy, replace the poo poo tier induction motor driving the splitter with a used 2kw drive servo off some old CNC machine so you can start/stop/creep with it without the motor catching fire after you use it for an hour.
|
|
#
?
Apr 11, 2019 22:57
|
|
|
Just to comment, PID loops on hydraulics are really annoying even with faster acting proportional valves vs varying flow rate by changing motor speed.
|
|
#
?
Apr 12, 2019 00:48
|
|
|
Methylethylaldehyde posted:You can get an electric log splitter, a load cell, amplifier set and some control electronics for like $5000. Get a VFD hooked up to the electric motor that takes 0-10v for the speed, hook the load cell up in line with the ram, and drive the whole thing electronically. The load cell gives you true force applied to the part, and can chase down a part that tends to yield or flow and keep applying more force until it levels out at some programed level, and the VFD lets you do a much more finely tuned job in terms of force over time and creep up vs. a simple relay that starts/stops the motor. yeah using load cells like that is a huge selling point of doing things this way, making the best use of em i can is definitely the plan you are right that a log splitter is probably the easiest ready-to-go electrically-powered jack assembly i can get my hands on; i was dubious about zincboy's suggestion because run as-is they don't offer any clear advantage over the sort of conventional press you could get for the same money, but there's no reason you couldn't modify one like that, yeah. if i had that kind of budget, though- knock a zero off the end for now- I think it'd make sense to design a press properly, as in settling on firm operating reqs and picking parts to suit. At the tonnages I want out this it seems like I'd end up paying for a lot of log splitter I don't intend to ever make use of, i.e. i'll never need all that stroke length so could get away with a shorter ram, smaller reservoir etc. mekilljoydammit posted:Just to comment, PID loops on hydraulics are really annoying even with faster acting proportional valves vs varying flow rate by changing motor speed. yeah, ive done some research and nobody has anything encouraging to say about trying to do something with fine control and repeatability with hydraulics, and that's people working with the right equipment for the job, not the cheapest equipment that can still achieve the job like i prolly will. hydraulics are a mediocre solution in every aspect except for that pesky thing where it's probably the only way I can get the tonnage needed within our budget. i'd love to do sth purely mechanical and make things easy but the easy ways to achieve that seem to max out at a couple tons Ambrose Burnside fucked around with this message at 01:23 on Apr 12, 2019 |
|
#
?
Apr 12, 2019 01:13
|
|
|
Joking aside, you can hydraulic systems to do some amazing stuff ... it's just annoying and the good valves are expensive. Not envisioning exactly what you need vis-a-vis pressure vs time (or displacement or whatever) but it seems like you could probably do it with proportional pressure valves and an LVDT feedback on the press ram, and if you had the valve right at the ram the response time could be pretty good.
|
|
#
?
Apr 12, 2019 01:29
|
|
|
Ambrose Burnside posted:yeah using load cells like that is a huge selling point of doing things this way, making the best use of em i can is definitely the plan Your comments about why you want the servo drive make a lot of sense. I haven't done much press work that would require that type of control so couldn't see why it was needed. Thanks for explaining. I think you could do this for a lot less that 5k depending. The log splitter would be a few hundred. A load cell and amplifier is less than a hundred off aliexpress. The cost would be the servo motor/drive but you can get a 2kw unit for ~$400 off the same site. Quality may be suspect. Should be able to do it for 1k all said and done. The servo loop tuning would be very annoying I am sure. One thing I saw when I took a quick look at servo presses is that they use a toggle style system for mechanical advantage. With that you would want a longer stroke cylinder. The machining and fabrication of the toggle would be more involved and may be much more costly depending on what tools you have available. I think it would be practical to use a leadscrew to drive the toggle and get a direct servo drive. The manufacturing complexity would be a bit higher.
|
|
#
?
Apr 12, 2019 01:32
|
|
|
Ambrose Burnside posted:yeah, ive done some research and nobody has anything encouraging to say about trying to do something with fine control and repeatability with hydraulics, and that's people working with the right equipment for the job, not the cheapest equipment that can still achieve the job like i prolly will. hydraulics are a mediocre solution in every aspect except for that pesky thing where it's probably the only way I can get the tonnage needed within our budget. i'd love to do sth purely mechanical and make things easy but the easy ways to achieve that seem to max out at a couple tons How much tonnage do you actually need though? How large a part needs to fit within the ram assembly? What's the actual total stroke? You can do some fun stuff with a servo, reducing gearbox and an arbor press, but that tends to top out at like 3-5 tons. An electric screw jack would also work, and would be easy enough to retrofit a servo into. And yes, hydraulic PID loops are at best 'close counts' type systems without fancy valves and accumulators and other fun bits and gubbins.
|
|
#
?
Apr 12, 2019 01:33
|
|
|
Methylethylaldehyde posted:How much tonnage do you actually need though? How large a part needs to fit within the ram assembly? What's the actual total stroke? You can do some fun stuff with a servo, reducing gearbox and an arbor press, but that tends to top out at like 3-5 tons. An electric screw jack would also work, and would be easy enough to retrofit a servo into. 20 tons minimum, 25 or 30 would be nice, but really we'll go as high as we reasonably can, money and press frame strength permitting; workpieces/dies top out at around 8x8" (but are usually significantly smaller, say 4x4") and shouldn't ever take up more than 6" of height; 12-16" of stroke is likely more than enough and lets us get weird with bolt-on tooling. It's the tonnage that makes things tricky (or at least restrictive on a budget); the other requirements are trivial. A proof of concept could probably be ~5 ton capacity and still be useful for small work/press-fitting/blanking and the like but isn't really what we want it for so I'm inclined to not settle for less than 20. A screw jack's another option I thought of, yeah. There are combined ram/gearbox/motor units that would make things very simple, although they're definitely out of budget right now. They're basically the only purely-mechanical option that approaches the tonnage we need and isn't a servo press itself that I've run into. Ambrose Burnside fucked around with this message at 02:55 on Apr 12, 2019 |
|
#
?
Apr 12, 2019 02:51
|
|
|
Ambrose Burnside posted:20 tons minimum, 25 or 30 would be nice, but really we'll go as high as we reasonably can, money and press frame strength permitting; workpieces/dies top out at around 8x8" (but are usually significantly smaller, say 4x4") and shouldn't ever take up more than 6" of height; 12-16" of stroke is likely more than enough and lets us get weird with bolt-on tooling. It's the tonnage that makes things tricky (or at least restrictive on a budget); the other requirements are trivial. A proof of concept could probably be ~5 ton capacity and still be useful for small work/press-fitting/blanking and the like but isn't really what we want it for so I'm inclined to not settle for less than 20. Then you need to think about how fine a line you need between 27.5 tons and 28 tons. A load cell and a servo driven log splitter can get you 90% of the way there, and as long as your PID loop is stretched out over long enough, you can creep up on whatever value you need. The issue then becomes cycle time and dealing with all the fiddly bits involved in the hydraulic control loop. If going from 27.5 and overshooting to 28.5 over 10 seconds won't cause issues, then that would work quite well. Take a log splitter ram and mount it on a 30 ton pipe bender frame, set up the controls and go ham on it.
|
|
#
?
Apr 12, 2019 08:13
|
|
|
The bootstrapping continues; I’ve got the power feed servo working. It’s fine when going one direction but seems to back off going in reverse. Is there a backlash nut in the servo coupling?
|
|
#
?
Apr 12, 2019 21:51
|
|
|
Methylethylaldehyde posted:Then you need to think about how fine a line you need between 27.5 tons and 28 tons. A load cell and a servo driven log splitter can get you 90% of the way there, and as long as your PID loop is stretched out over long enough, you can creep up on whatever value you need. The issue then becomes cycle time and dealing with all the fiddly bits involved in the hydraulic control loop. If going from 27.5 and overshooting to 28.5 over 10 seconds won't cause issues, then that would work quite well. Yeah, that's what'll probably end up happening if we don't work up to it w/ a lighter mechanical prototype. We may very well end up doing that first depending on the sticker shock for tentative BOMs for various approaches, or on the feedback from a hydraulics pro once we've got a more developed design and more concrete requirements irt how responsive and precise a press we want. An alternate purely-mechanical approach someone suggested that may or may not be a waste of time to pursue is replacing one big, beefy motor/leadscrew/nut block of the needed specs with several, always operated in unison. On paper it's got some strong benefits- a leadscrew was always the ideal solution to this given our requirements, but off-the-shelf solutions don't seem to exist for the tonnages we really want, and would definitely be prohibitively-expensive unless we design and fabricate it all in-house (not very likely). 3/4/6 etc much smaller motor-leadscrew sets, though? We can source all that easy-peasy, budget-tier options become possible, we can stick to stuff with engineering data available so we can tailor the system to whatever the weak link is (iirc the nut blocks we looked at were the weak links in terms of what would fail first). It's got obvious pitfalls- extra complexity, motors missing steps or not running quite in synch making for unequal forces, that sort of thing- but if we're planning on lots of load cell feedback it'd be trivial to use enough sensors to measure force distribution across the press platens and trip the ol e-stop if a screw stops turning or forces start getting dangerously uneven. Maybe you could even use those multiple independent screws to compensate for unequal loading while the press is applying pressure, although that's on the programming end of things that I'm weak with. e: looking around and multi-screw commercial presses have at minimum been experimented with, which is good to know: http://www.j-mst.org/On_line/admin/files/18-J2011-339_2323-2334_.pdf they also did work on how to distribute and share the force of multiple independent leadscrew drives. which yeah may be necessary for this approach. in any case this seems to be a viable option, albeit a particularly experimental one Ambrose Burnside fucked around with this message at 01:59 on Apr 14, 2019 |
|
#
?
Apr 14, 2019 01:48
|
|
|
The International Journal of Advanced Manufacturing Technology October 2013, Volume 68, Issue 9–12, pp 2371–2381 Small mechanical press with double-axis servo system for forming of small metal products Abstract Multi-axis servo systems have been applied to many machine tools in recent decades, but they are not commonly used for press machines. They are particularly uncommon for small presses because it is still believed that the precision of a small press is dependent on the precision of its components and parts. However, high-precision components and parts are difficult to manufacture and assemble, so the costs of small precision presses are roughly equal to or higher than those of large press machines. Moreover, the accuracy of conventional small presses worsens over long operation times, particularly when dealing with off-center loading frequency. Therefore, we propose a small mechanical press with an adjustable driving system, using a slide and double-axis servo system, for manufacturing small products. Through this proposed design, we expect higher-quality products to be produced for lesser cost than conventional small presses.   so i guess this is almost exactly what i'm thinking of, down to using multiple independent servo drives assembled with (relatively) cheap low-precision components and then compensating with dynamic motor control adjustments to achieve the results of much larger and more expensive designs that conventionally do not scale down well. main difference is that it uses multiple cranks instead of leadscrews and has a stroke length of 18mm, and only a few tons of pressing force, but i wouldnt be reinventing the wheel here. also their prototype runs at up to 125 cycles per minute and in tests was blanking aluminum sheet parts with only +2 μm variation from spec, which is pretty wild for a tiny benchtop machine first iteration.
|
|
#
?
Apr 14, 2019 19:04
|
|
|
I just want to comment that this is all pretty cool.
|
|
#
?
Apr 14, 2019 19:19
|
|
|
I work in much bigger presses and I keep wanting to butt in but you guys are teaching me things so please carry on.
|
|
#
?
Apr 14, 2019 19:38
|
|
|
cakesmith handyman posted:I work in much bigger presses and I keep wanting to butt in but you guys are teaching me things so please carry on. well hey, feedback is extremely welcome, the only press i've ever built was a 20-ton no-weld engineer-approved design operated with a harbor freight special hydraulic bottle jack, and nobody else involved has any experience with press design specifically; and knowing of a couple tentative academic experiments that worked out well does not mean that our own, significantly more resource-constrained plans are also gonna work out. beyond an informal sounds-like-a-smart-approach chat prior to locking a design in, we won't even have an actual engineer involved until we need our math irt loadings and safety checked or we otherwise get intractably stumped with some design aspect. so if there's anything good, bad or misguided that sticks out to you, it can only help us~ Ambrose Burnside fucked around with this message at 21:07 on Apr 14, 2019 |
|
#
?
Apr 14, 2019 21:04
|
|
|
At my last job we did a lot of wire crimping and most of the machines were a single crank style with an AC motor. A team tried to make one with a linear motor but they designed it so it would all fall down when lost power then everyone just complained and cancelled it. I wasn't on the team dealing with them directly but I can share my impressions: The crank style worked fine. They were in production so we just had to tune and maintain one setup. I remember the tooling needing the most maintenance and the presses felt bulletproof. They only broke a few times and it was because of the user. Some measurements from memory: our crank radius was about 2-4 inches and the crimp stroke was about 200 thou. I understand the motor advantage changes throughout the rotation. We ran fast enough that the press cycle was less than a second. In your case, you might need to do some speed reduction to get advantage for a deep press. The paperwork of the 4 servo design is interesting. I wonder why they have ball socket connections like that. I'd keep the crank motion 2D so you can use rollers or bushings instead of balls so your contact area is higher and stresses lower.
|
|
#
?
Apr 15, 2019 19:54
|
|
|
The ball and socket join allows you to tweak slide parallelism, which is very important for repeatable setups and load balancing. Ambrose Burnside posted:well hey, feedback is extremely welcome, the only press i've ever built was a 20-ton no-weld engineer-approved design operated with a harbor freight special hydraulic bottle jack, and nobody else involved has any experience with press design specifically; and knowing of a couple tentative academic experiments that worked out well does not mean that our own, significantly more resource-constrained plans are also gonna work out. beyond an informal sounds-like-a-smart-approach chat prior to locking a design in, we won't even have an actual engineer involved until we need our math irt loadings and safety checked or we otherwise get intractably stumped with some design aspect. so if there's anything good, bad or misguided that sticks out to you, it can only help us~ Will do, I maintain and work with 2000t and 4600t electro mechanical presses. We'd love an upgrade to servo but £10m isn't falling out of the sky. The system you linked is a reasonable way of making a small servo press, ideally you'd use a single larger servo but cost for the drive becomes a problem, chaining the two together and coding them as a pair is an acceptable middle ground. It looks like they're using linear slide blocks as the jib slides, they don't allow adjustment in a way I'd like, screw jacks in the slide to jack/adjust phosphor bronze bushes would be preferable. Rigidity in the frame becomes your next problem, just go overkill, at the 20t scale that's affordable. The biggest threaded rod you can find as tie rods will help that. Finally you'll get the slide jammed at some point, it's inevitable, make provisions for the servo not being able to back the slide off under power, the two ways I know are either complicated pneumatic overload protectors (they resist a shitload of pressure them when they get squashed just enough they fire upwards about 100mm and e-stop the system) the second way is to heat up the uprights so they stretch slightly, some machines have heaters built into the frame for this purpose.
|
|
#
?
Apr 15, 2019 20:36
|
|
|
bred posted:At my last job we did a lot of wire crimping and most of the machines were a single crank style with an AC motor. A team tried to make one with a linear motor but they designed it so it would all fall down when lost power then everyone just complained and cancelled it. I wasn't on the team dealing with them directly but I can share my impressions: The crank design is ubiquitous in other press styles so it's clearly a smart way to go about it, but we have reservations about the stroke length- that example I posted only has 18mm of travel, and the height of potential tooling (as well as the total stroke length needed for deep drawing and stuff) varies by way more than that. Ours doesn't have to be 18 but it's gonna a limitation of that kind of eccentricity-derived force. With hydraulic presses you just get the ram out of the way like its nothing, worst-case scenario you do some build-up under the tooling so contact tweenst ram and tool happens where the ram has the most strength. With a very short stroke we'd need to think hard about how to accomodate existing work with that design such that everything works happily with limited travel, or have some other facility for adjusting the height of the lower platen (which sounds like a huge pain). i do know that we'd much rather have a generous possible stroke length before a very fast-cycling press; we're not expecting to do real production work with it to the point that that fast cycle matters to us working profitably. just being able to program whole stamping cycles and run them by pressing a button is the real labour-saving aspect here, v2 can iterate stuff like working faster.We could live with a ram travel of, I dunno, 10"/min. or something else glacially-paced, at least for this iteration anyways. Ambrose Burnside fucked around with this message at 22:27 on Apr 15, 2019 |
|
#
?
Apr 15, 2019 21:38
|
|
|
cakesmith handyman posted:The ball and socket join allows you to tweak slide parallelism, which is very important for repeatable setups and load balancing. What generally causes the slide to get jammed, does it rack a little under uneven loading and wedge or sth like that? I know my 20-ton press has jammed due to racking or weird alignment of that one slightly-bent threaded rod, but it's never been a game-stopper if I have a crowbar handy. It's like wrenching two new die shoes apart before they're worked in. We're definitely planning on a dead-simple "two big steel plates with holes we'll run bigass threaded rods through" load-bearing frame because 1) its the only simple way to do it from scratch while being confident in it meeting the on-paper engineering numbers (without a ticketed welder in our pocket, anyways) and 2) it lets us reconfigure the press or break it down for transport/storage fairly painlessly. for now we'll probably have the slide as a third plate that slides on seamless tubing passed over the exposed threaded rod between the outer plates; it's not gonna provide jack-poo poo in the way of adjustability by itself, but it's a very well-demonstrated slide design for small presses that's accessible and constrains the slide well (youd be surprised how often welded DIY press designs using steel channel to guide the slide are not good at that function when physics decides youve had a boring week just as the pressure gauge passes 30 tons on a hella unequal loading) i suspect we'll want a proof of concept of a programmable n force-measuring electric press design meeting the tonnage req before we bog down in stuff like adjustable guides or optimizing parallelism, not that those things arent important but rather that this'll be a fun experiment that didn't really go anywhere if it doesn't make The Guy Who Makes At Least Twice My Salary clap his hands and gurgle excitedly like a baby as we initiate the 'Pizzeria-Size 10-Gal. Drum Of Whole Canned Roma Tomatoes' Precision Crushing Cycle for his review Ambrose Burnside fucked around with this message at 22:30 on Apr 15, 2019 |
|
#
?
Apr 15, 2019 22:25
|
|
|
Ambrose Burnside posted:What generally causes the slide to get jammed, does it rack a little under uneven loading and wedge or sth like that? I know my 20-ton press has jammed due to racking or weird alignment of that one slightly-bent threaded rod, but it's never been a game-stopper if I have a crowbar handy. It's like wrenching two new die shoes apart before they're worked in. Wrong tool height set, tool heating up over excessive usage, accidentally stamping 2 pieces, wrench or die grinder left in tool after rework. quote:We're definitely planning on a dead-simple "two big steel plates with holes we'll run bigass threaded rods through" load-bearing frame because 1) its the only simple way to do it from scratch while being confident in it meeting the on-paper engineering numbers (without a ticketed welder in our pocket, anyways) and 2) it lets us reconfigure the press or break it down for transport/storage fairly painlessly. That design is a good start but doesn't prevent racking from uneven loading. Add in a fabricated steel u-shaped frame (sides and back of the press) and bolt through/to that. Welder needed obviously but dead simple to design and still let's you use the seamless slides you mention.
|
|
#
?
Apr 16, 2019 15:01
|
|
|
|
| # ? May 10, 2024 16:46 |
|
|
This is a metalworking question about a woodworking machine but I promise not to get sawdust in your coolant. This thing showed up today- I had hoped it would be delivered via a lift gate, but the gods decided my day would be better if it was delivered by falling off the end of the lift gate. It weighs about 2500#, and so it needed a push to get on the lift gate, but the driver I guess have it just a little too much push, and once it started moving it kept moving. Luckily, nobody was hurt. The mortiser thankfully landed on its back, and an inch deep dent in the asphalt tells me on concrete this could have gone very differently. With some fun with chainfalls and ratchet straps I got it back vertical.  Most of these machines have been painted all over and used hard and put up wet. I’ve been looking around for one for a while and have never seen one with polished tables.  The body is a single solid casting 5/8”+ thick, and it landed on a just as thick access door that did crack-what is the best way to repair this?  Probably not worth worrying about, but the machine is in such great shape I’d like to take care of if I can. I have a little Lincoln buzzbox stick welder but I don’t think that can weld cast iron? In addition, any way to check for any other cracks or fractures in the body of the machine without stripping the paint? These machines (Wysong and miles 284) were built like tanks to run in furniture factories 2 shifts a day for 30 years, and this one has had a very easy life (and will continue to), but is there anything else I should check considering it plunged 4 feet with a very hard landing? Thirty second test run did fine, but I’m no gearhead, unfortunately.
|
|
#
?
Apr 22, 2019 22:42
|
|
