Welcome to Our Community

Some features disabled for guests. Register Today.

Escape from (VFD) noise?

Discussion in 'General Talk' started by Batcrave, Mar 29, 2020.

  1. Rob Taylor

    Rob Taylor Master
    Builder

    Joined:
    Dec 15, 2013
    Messages:
    1,470
    Likes Received:
    746
    Yeah, the TIG got it, the mill will probably get it, and in all likeliness machine #4 will also get it. It's kind of a problem, at a certain point, though right now "real life" kinda precludes too much attempted overlap.

    I think it can drop-in, especially since I removed my compound and have the rigidity to work with. I'd also looked at flex-shafts, there's one on Amazon for like $100 but I don't know if it's worth the flutter vs a spindle with more bearings, or actually building a spindle and belt driving it with a "spindle". You see why I've never actually attempted the project...

    Do you need one? Really all you need is to fix a big hole in your casting. If there's a mm or two gap between the fill and the liner that gets filled with oil, what does that matter? Other than it all falling on you when you next remove the liner.

    That was my theory. As long as you properly lube the liner (even if you don't plan on removing it again for a few years) it should be fine.

    Yeah, I need a pressure pot too, probably more like 3 gal though, depending on the dimensions. Of course first I need a vacuum chamber because you can't cast resin into non-vac'd silicone, but anyway. Just another one on the list.

    It's doable. I wouldn't turn wood on a machine lathe, but that's just me. Many people do, presumably just fine. I'd try turning down a PVC pipe, possibly a PVC pipe covered in Bondo (for diameter), before I tried that. Or a steel tube of some kind. Maybe I'd use some silicone too, for easy removal. Assuming Smooth-On are currently operating.

    Though, I'd argue that you have the exact same risk and margin of error on both methods, but in one method you've glued a piece of wood in your spindle and in the other you've made your spindle liner non-removable until the next time you feel like blowtorching and banging on it. One of these results in a usable mill...

    Exactly. And I did just move my lathe away from the wall for exactly this sort of hare-brained rear-access scheme.

    Hardening, annealing, hell, even SMD soldering if I build it right. Needing to chuck stuff in the oven is another one up there with needing to surface grind something.

    That's my guess. Though I also have additional ceramic caps on the lines, which I may make more permanent along with the NC-ing.

    Hard to fail closed, yeah. Harder, in terms of signal:noise ratio, to overpower a solid +5VDC (or 12, or 24) voltage, than kick up what seems to essentially function as a floating ground.

    From what I've read, most mechanical switches are more accurate than we think- even cheap microswitches. I'm sure they're comfortably hitting the 0.05-0.1mm range or under, which is more than enough for an extrusion machine without a toolchanger or other G53-based operations. I have optical switches that were originally going to go on the laser, but I decided fast and simple was better than making a bunch of little flags.

    Oh yeah, that was the fourth thing I forgot. Those things are pretty cool!

    Kinda, yeah. :p

    I don't know if it'll happen, but apparently there have been Discussions.

    That was my assumption, tbh.

    Incredible how eager they are to privatize something so apparently useless, inefficient and wildly unprofitable.

    I too, like this plan...

    Mostly for the flag too, actually.

    The door can have collet holders in it.
     
  2. Semper Why

    Builder

    Joined:
    May 16, 2018
    Messages:
    32
    Likes Received:
    12
    Yeah, I once had this thought that ran along the lines of "You know, a CNC router would allow me to make so many things and it is a hobby I can enjoy at home." That one has cost me quite a bit.

    Thanks. I'm good at "odd". Still working on the brilliant part.

    I can think of a couple of hacks to possibly get around this. I dunno how much EFI escapes from gaps, but if you could limit the gaps towards the read of the VFD, it might help and/or be sufficient. You may consider figuring out how much depth you need and then slicing off the back of the mailbox minus an inch or so. Cut inch-long slits in the ragged edge every 2" or so and then bend the metal outward to create mounting flanges. Attach those to a small piece of sheet steel or whatever as a back mounting plate. Mount the VFD to the sheet steel, mount the plate to the wall. Getting the cords into the mailbox I'll leave as an exercise.

    As a bonus, this method would keep the little flag intact.

    Oooh, I like it.
     
    #32 Semper Why, Apr 6, 2020
    Last edited: Apr 6, 2020
  3. Batcrave

    Batcrave Journeyman
    Builder

    Joined:
    Apr 20, 2018
    Messages:
    361
    Likes Received:
    165
    Oooo... right! I could plug in a WELDER! Where's that damned drooling smiley?

    "Real life" seems to be precluding quite a lot of things for people lately. Who'd ever think a little thing like a plague would cause that much disruption?

    If I was looking for stability & low runout with the flexy option, I'd probably be tempted to grab a cheap import flex shaft that's compatible with a genuine Foredom handpiece. Some of those things are pretty seriously built.

    Or I suppose there's the option of something fun like starting with a dental handpiece with the 200k rpm air turbines, although I'm not sure quite how to fit it with anything as heavy-duty as an 1/8" bit.

    I'm starting to think we've gone so far off-topic that even Amazon's creepy Ring network couldn't spot it, so, while it would be more in-character of me to continue a lifetime spent doubling down on my missteps, I'm going to split this off into a new thread over in General. Probably named something awful like The Overflow from Bats' Lubricated Spindle.

    You and me both. But when you find a beautiful shiny 10gal stainless pot (plus, for no good reason, a bazooka-sized pneumatic epoxy gun & scads of tips) for $35 and the alternative is paying well over $100 for someone's battered & filthy old 3gal... well... it's hard to say no.

    The vac chamber's easy - it's just a pot with a thick slab o' plastic on top with a plumbing fitting and rubber gasket. Or you can be lazy and grab something off Amazon being sold for totally-not-weed-extracts (which is what I did... and then promptly ruined the clear lid with solvent spatters). But it's the pump that'll get you. I eventually found a little lab grade unit that (after a week or three of cleaning) I use for resin stabilizing wood & degassing epoxy (sometimes more successfully than others), but it was a long search to find something affordable that wasn't just a venturi.

    Although I suppose we may have very different budgets.

    I've used a little countertop convection oven for rough annealing, and it would probably even do the job for SMD work, as long as your temperature tolerances/times aren't too tight. Hardening is obviously another matter entirely, though.

    I guess I sort of assumed the little cap was just for debouncing... your caps on the lines could definitely be helping, though.


    *nod* And I think the Openbuilds ones are operating at 24V.

    I'm sure it's possible for the mechanical switches to be accurate... but the ones I had on before definitely weren't. Some in the batch were flaky or flickery, others seemed to have a surprisingly large variation (1-3mm) in where they tripped. Of course, I suspect the even the cheap OB ones would be of better quality.

    I've got an assortment of IR sensors pulled from... something or another (printers, maybe?), but never even got around to testing them, never mind deciding how to mount them & the flags to trip them.

    I've never done much of any sheet metal work, but, yeah, an English wheel always seemed absolutely bursting with potential.

    Maybe not quite as much potential as it takes up space... but an awful lot.

    I'm not sure it would be quite as frequently handy & utilitarian as the brake, though.

    What I ended up reading (not very extensively), it sounded less like they were going to shut down the USPS due to the disease, and more like politicians were threatening that if it didn't get more money, it would have no choice but to shut down. So probably just a bargaining chip, since I suspect no one wants to see it shut down.

    Well, except for UPS. And FedEx. And maybe Amazon. And... ok, maybe a lot of people want to see it shut down.

    I guess we'll see which of those people own the most politicians.


    "We'll just offload it and save the government the money. I own lots of stock in a great company that would be willing to make the sacrifice. But we may have to offer them some... additional incentives. Say, maybe a ten year tax holiday and a promise to subsidize 50% of all pre-tax expenses. Just as a show of good faith, mind you."


    The flag is really what makes it. Although I'm not sure whether it should be a variable signal, or should just pop straight upright when the spindle starts.

    On the signal side, I assume an arduino (I've still got a spare Uno and at least one Digispark kicking around) could eat signals from one of the VFD's assorted outputs (I think it's got a serial or RS-485 in there somewhere).

    I'm pretty sure I've still got some servos left over from an RC glider I built waaaay too many years ago (probably under Bush I). Not sure whether they've got the muscle for flipping up a (presumably kinda stiff?) flag like that, though. Or how to interface with them. Or if they've been eaten by rats. I've also got a big sack of steppers, but driving them is (ironically) a bit too much of a headache for something like this.

    Yes!

    Although I'm not immediately sure how to manage a collet holder on a flip-down door that won't dump them out in one position or another - that may need a little more thought. Worst case, they could go in a rack that just sits inside.

    If the drive were shoved all the way to the back, the interior could also potentially be used for storing cutting tools too (with or without racks) - although I'm worried about shoving stuff in too deep & losing track of it, or not being able to see it in the dark.

    It'd probably also need a little exhaust fan punched through the back wall, but I'd need to do that for any enclosure.

    I'm still not sure where I'd put it... and I am pretty sure that it's a terrible idea... and I'm absolutely positive that mactec wouldn't approve at all... but now I CAN'T GET IT OUT OF MY HEAD.

    This cannot go unpunished, Semper!


    -Bats
    *Trying to remind myself that I've still got a backlog of Christmas gifts to make & shouldn't prolong this repair* ... *Failing*
     
  4. Batcrave

    Batcrave Journeyman
    Builder

    Joined:
    Apr 20, 2018
    Messages:
    361
    Likes Received:
    165
    How about "CNC would let me make things out of wood without requiring a basement full of woodworking tools"?

    *looks at basement full of machine tools*

    No worries. It was probably just a momentary flash, anyhow. No need to worry that it's a chronic condition yet.

    I seem to remember the answer is "lots", but I don't remember the details , and know it depends a lot on the wavelengths you're trying to block (which is why microwave ovens can have all those holes in the window & still block any emissions).

    That's a thought... although I'm getting increasingly attached to the idea of a post sticking up from the CNC bench with a mailbox sitting on top of it. If I can figure out where/how to do it, the potential for a little extra storage might even be the push to swing me fully into trying this nonsense.

    That's the easy part - just a step drill with some grommets in the holes. Or, I suppose I've been told those should be shielded glands, to bond with the cable shields. The fan takes slightly more effort, but I've already got a bimetal hole saw that's sized for an 80mm fan (if I go for something smaller, I may need a new one... although I'd been thinking about getting one for 60mm fans for another project anyhow).

    The flag's a must - even if it's not motorized. I suppose cutting the front would've required something messy, like welding the whole circumference back together behind the flag, then grinding it down & re-painting the whole thing. Which has absolutely no appeal.

    Especially because I can't weld.

    Or paint, for that matter.

    But, yeah, no flag, no mailbox.


    -Bats
    (No flag, no mailbox, no wife, no horse, no mustache)
     
  5. Semper Why

    Builder

    Joined:
    May 16, 2018
    Messages:
    32
    Likes Received:
    12
    At this point, I think you're chasing the concept of "perfect" and ignoring the concept of "better".

    Nuts, there I go again.

    Interesting and amusing. It would, however, put the VFD closer to your electronics than you might wish.

    Right. As a proponent of the "good enough" theory of shop design, I would suggest just soldering the stuff together if you must. Perhaps a two part epoxy to get it in place and then solder the edges to ensure a metal seal. Maybe. I dunno. I'm making this up as I go. My theory is that even a cheap mailbox with small gaps in the back edge near the rear of the VFD is going to be an improvement over a wall-mounted VFD open to your shop.
     
  6. Batcrave

    Batcrave Journeyman
    Builder

    Joined:
    Apr 20, 2018
    Messages:
    361
    Likes Received:
    165
    Well, it matters in broad terms. If it's mostly very narrow wavelengths, it might have to be practically airtight to make much of a difference.

    Based on this (which has lots of fancy numbers I don't quite understand, but has a quick rule of thumb and quicker chart), blocking out 5GHz signals (like wifi, or the crap the microwave spits out), you'd need the slots (since they ours probably wouldn't be nice round holes) to be less than .01". Something with a couple .1" gaps will be of very minimal usefulness at best.

    On the other hand, if the problem is mostly lower frequencies - say, 800kHz, like the detuned radio I was using to locate noise sources - then the gaps can apparently be as large as... err... over ten inches??? So, yeah, I don't really know where most of the noise (or the most troublesome of the noise) lies, so I can't pretend to guess at the effectiveness of any given enclosure (except that "no holes" will probably be more effective than "big holes"). The front actually looks like it should be a good thing - radiation hates right angles, so that overlapping door is great. And if I can figure out somewhere to stick it on a post, then the whole worry about creating gaps at the back end goes away.

    The even better solution would be to roll the lip of the mailbox and new back piece together... but I think we're going to need those fancy sheet-forming tools that Rob was wishing for to do that (I'm going to learn my lesson from him and probably not what pick this for my first big hammer-forming project)

    Knock that off! Look at all the trouble your first thought is still causing!

    True, but at several bucks a foot for the continuous flex-rated power cable, I don't think I'll be able to afford to get it far enough away for the distance to be a great protection in itself (I'd tried that before, and look where it landed me) - even if the long cables didn't turn into broadcast antennas - and I'll still want it close enough to see the readout (at least until I move on to some software that has some sort of RS-485/serial/etc support to transmit speeds back). That said, if it's properly shielded ("if"), then all those problems should go away and distance shouldn't be necessary.

    I've never done any sheet metal soldering - I imagine it conducts heat far too fast for most irons or guns. But I suppose any excuse to swing around a torch is a pretty good excuse.

    What about a metal-filled epoxy like JB-Weld - is that conducti... no, apparently not. Oh well. So, yeah, probably some sort of solder fill - and/or maybe conductive tape - if I go that route.

    You and me both. I wasn't even sure whether I was taking the idea entirely seriously until... well... nowish, I guess. :p

    I'm not sure the back-vs-front distinction would make much of a difference - except that the overlap of the door should already make a good seal... and the all-important flag, of course - but, yeah, it would definitely make an improvement over no enclosure. What I'm trying to decide, though, is whether it'll be close enough to a proper electrical enclosure like this one.

    If I keep the existing back and do a passable job making my knockout holes, then, yeah. While it's not likely to thrill the sort of guys who wander around looking for code violations to post angry replies about, I can't see any reason (barring loose seams) that it wouldn't be effective. If I have to shorten it, though, that's going to be an open question (and at least partially one of how competent I turn out to be).


    -Bats
    ( "Black paint isn't code! Only grey powdercoat!" "And what's that flag? Flags definitely aren't code!" )
     
  7. Batcrave

    Batcrave Journeyman
    Builder

    Joined:
    Apr 20, 2018
    Messages:
    361
    Likes Received:
    165
    Update:

    So, despairing of the alternatives, I finally went ahead and ordered another Huanyang VFD on Monday - supposedly it should arrive some time next week. It'll be interesting to see whether the troubles resurface if it's hooked up identically to the old one.

    I've also been looking into mailboxes (blame it on the flag). The one out front matched the dimensions of the cheap Home Depot medium (translation: "smallest") size, but it turns out the interior dimensions are quite different from the exterior. Not just the rolled sheet edges, but the interior floor is also significantly raised from the exterior lower edge - so while the VFD fits the numbers with room to spare, it doesn't fit in the box (not by a wide margin, but too much to easily hack off with a dremel). And I'm sure any watching neighbors would've been all sorts of curious about why I was measuring and trying to cram a large electronic thingy that was almost certainly a bioweapon dispersal device into the mailbox.

    I bugged support asking for some more useful figures, so we'll see what they come up with. Otherwise I'll probably try the Large. The XL is... yikes... accurately named - I realized I've got shelves that're smaller - so that's probably not an option even post-mounted with new lighting underneath (to fill the newly generated shadow), that fills up a huge volume of airspace.

    Although... hrm... maybe I could use it to install a small 3D printer the the leftover space...

    Or a coffeemaker?

    Another worry with the smaller sizes is if it'll be too much of a pain in the bats to mount the filter & any other support electronics and route the wires through the small opening... although depending on how it's joined, I suppose I could try to split the seams and then replace them with nuts & bolts, or maybe thumbscrews. I figure an external power switch probably wouldn't be a bad idea either, although I'm not quite sure where I'd put it. It sorta crowds the look of the control panel in the door.

    On the flag front, I managed to dig out the old Futaba "New" Attack-R ("new circa '90ish) box. It looks like everything still works, but the servos only make a +/- 1/8 turn from idle. The limit may be the controller, though - I don't have much recent experience with servos and haven't tried wiring anything up to drive them independently yet. If I can't get more travel out of them, I can probably do it with a linkage of some sort instead.

    I was initially thinking about a 4/4 as a post (maybe with a nice, decorative, wrought irony bracket underneath... maybe from some nice, decorative, non-wrought aluminum), but it doesn't look like Homebound Desporation likes to deliver big chunks of wood, and I share a house with a couple "at risk" types who've gotten very touchy about me going out, so I'm looking at alternatives or interim solutions. There's a hemlock scheduled to come down out back next week, but it's big enough to clock Socrates for a loop, and I don't know whether it'll generate anything small enough to haul up & toss on the bandsaw (there's also a fair bit of last year's downed hemlocks lying around, but I don't know what sort of shape it's in after a winter on the ground, and I already used up my only Socrates joke). The other alternative is bending something from the big pile of angle iron that's been sitting around collecting rust for the past few years. Not very elegant or thematic, but eminently functional.

    It's too bad I don't have the High-Z mod to engrave a house number (or bat) on the side... although if Rob's right, it might have to read " In loving memory of the United States Postal Service, RIP: "Neither snow nor rain nor really pretty gorgeous weather stayed these couriers from dropping your packages in a puddle and running them over with a truck" ".


    -Bats
    (how long can I get away with calling this idle speculation in a thread before someone makes me move it to a build?)
     
  8. Semper Why

    Builder

    Joined:
    May 16, 2018
    Messages:
    32
    Likes Received:
    12
    Aha. One more dumb idea: Instead of having the flag go up & down to tell you the status of the VFD, why not rig up the flag to be the switch for the VFD? Raise the flag when you want the VFD to be active.
     
    Batcrave likes this.
  9. Batcrave

    Batcrave Journeyman
    Builder

    Joined:
    Apr 20, 2018
    Messages:
    361
    Likes Received:
    165
    I hope you realize that if you don't quit with the dumb ideas, this project is likely to end up turning into something practical and affordable. I'm sure you understand what a disaster that would be.

    Hrm... While I really like the idea of watching the flag pop up, I suppose manually flipping the flag to tell something the mailbox has work to do would be more thematically appropriate. It might also be easier to implement... although I suppose I'd be trading the arduino/servo complexity for some sort of mains-voltage relay, which I'm even less familiar with.

    This definitely bears consideration, though.

    [edit: it looks like I've got a 4-way 3-32VDC to 120/240VAC relay in the parts bin that would do the job... although I feel like it would be a bit of a waste not to use it for something that takes advantage of all four channels - maybe a Arduino-controlled 4-outlet smart power strip? There's also a Digispark relay shield, but at 5A that's going to be switching a really tiny VFD]

    -Bats
    (I mean, no! No it doesn't! Terrible, practical, stupid, awful... sensible, even!)
     
    #39 Batcrave, Apr 9, 2020
    Last edited: Apr 10, 2020
  10. Rob Taylor

    Rob Taylor Master
    Builder

    Joined:
    Dec 15, 2013
    Messages:
    1,470
    Likes Received:
    746
    That's an interesting idea. I'll look into the Foredom bearing setup at some point, though I think speccing the bearings and boring my own spindle would be more fun, overall. Might be a good combo CNC/lathe project.

    There are a few used air spindles on eBay here and there that I've found (whenever I decide to start looking for 60,000rpm spindles), but I think I want more in the 5000-15000rpm range with 1/4" shanks for most rotary grinding applications.

    Chamber's easy because it's 15psi of negative pressure. It's pressure pots that'll get you. But yeah, most decent vac pumps seem to be in the ~$200 region, which doesn't seem unreasonable to me, I just need more molding and casting projects to justify it. However, those are the very projects the mill is intended to make obsolete.

    When you run the numbers a DIY programmable heat treat oven really isn't that difficult or expensive. Just a case of figuring out your power levels and volume and stuff.

    Speaking of DIY, I just discovered that people sell lathe beds on eBay. No carriage, no headstock, just a nice ol' pair of V-ways... Perfect for a DIY surface grinder base.

    This is a very important lesson.

    I don't remember what this thread's about, but it's sure making me want to do some projects I really don't need to start just yet!
     
  11. Batcrave

    Batcrave Journeyman
    Builder

    Joined:
    Apr 20, 2018
    Messages:
    361
    Likes Received:
    165
    They may be beefier than I thought... just saw an Instructable where a guy carved(!) a dragon bracelet out of stainless with one.

    I don't find much fun in the speccing part. But the boring? Oh, definitely.

    Yeah. That's how I ended up with the 10 gallon stainless behemoth sitting outside my door, and... hey, wasn't that also how we got onto this topic in the first place?

    Part of my problem may have been that the ~$200 region was well out of reach for me at the time, so I may not even have been flagging those in my searches. The fact that I did eventually find a <1micron pump in the <$75 range is a little surprising in itself (never mind the fact that I found it long before seeing anything cruder - venturi excepted - turn up)

    Well clearly the only solution is to mill a vacuum pump!

    Or you can justify it by milling your form, then molding & making a cast of it.

    I've actually been curious about the potential for rewiring the convection oven with some sort of PID circuit in place of whatever thermostat it has right now. Obviously it still wouldn't get hot enough for things like hardening, but even having tightly controlled temperature up to 500-550ish would be nice for a lot of applications.

    Funny, I'd just noticed a couple people on Craigslist in the past week or two advertising "lathes" that turned out to be missing just about everything other than the bed...

    Of course, I guess if I could fit another lathe, I would've had no trouble finding somewhere for a dainty little grinder...

    What's even more important is that I got someone else to learn it for me!

    Mail delivery, I think?

    True. Better keep a few big projects saved up for when they decide you're not too essential to be replaced by a robot & you wind up stuck indoors like everybody else.

    Although maybe it's not too soon to start stockpiling parts...


    -Bats
    (then again, if I got rid of all the parts I'd stockpiled, I might have room for the grinder and another lathe)
     
  12. Batcrave

    Batcrave Journeyman
    Builder

    Joined:
    Apr 20, 2018
    Messages:
    361
    Likes Received:
    165
    Update:

    The new VFD showed up several days early, and, for testing purposes, I set up everything identically to the old one.

    As it turned out... that didn't work. The spindle would slowly crawl up to ~4-500 rpm, and then the drive would overload & die. The settings Mactec was so adamant about over on The Zone just made it worse.The cure seemed to be lowering the V/F curve's "Intermediate voltage" setting a little - at which point it ended up accelerating far more smoothly than the old one ever did. It looks like either there's significant variation between units, or else one of the two is defective (or differently defective than it first seemed). I'm curious to see if Mac can explain exactly what's happening. Sort of.

    After getting that worked out & everything wired back up, I let the spindle run for a couple hours without seeing any sign of noise in the logs. However improbable, so far all signs point toward the old VFD - not the Blackbox, poor shielding, poor shield grounding, misrouted wiring, lack of enclosure, lack of filters, the furnace, the solar panels, the electrical grid, Chinese secret agents, US secret agents, lack of faith, insufficient filial piety, or any other nasty little environmental gremlins just dying to get their teeth into my poor little Gcode.

    In the interest of time, money, and appeasing the wrathful and jealous gods of combined shipping, I ended up (at least for the moment) passing up both mailboxes :cry: and oversized electrical boxes and getting a small electrical enclosure from Mouser along with the EMI filter (which was sold out with a 2-5mo lead time everywhere I looked, except in a DIN rail-mounted flavor - hoping that won't be too much trouble to work around) and shielded cable glands. The box should have enough space for everything (barely) (although I couldn't find a model or dimensions for the filter) - I'm just hoping it isn't too miserably cramped working inside it. I came to the conclusion (well-founded or otherwise remains to be seen) that the manual's airspace requirements were A) primarily for still air rather than active cooling, and B) designed to cover their 380V 55kW behemoths as well as my delicate little flower, so hopefully a cramped little box with a fan will do the job.

    I also grabbed a 20A double-pole light switch, an 80mm AC fan, and a longer ribbon cable for the VFD panel, and should still have a bunch of terminal strips & wire nuts left over from when I had to do this for the lathe. I think the only thing I haven't ordered is the breaker, but I figure that can wait, since it looks like it'll have to be external to the box anyhow. I'm also still toying with the idea of an external switched outlet, to control the pump, or something really useful like a string of Christmas lights.


    -Bats
    (or maybe a tape deck playing an endless loop of "Ride of the Valkyries")
     
  13. Semper Why

    Builder

    Joined:
    May 16, 2018
    Messages:
    32
    Likes Received:
    12
    At the very least, I want pictures when you're done with this whole ordeal. Or if not done, at least to the point where you can carve your name without tripping a limit switch.
     
    Peter Van Der Walt likes this.
  14. Peter Van Der Walt

    Peter Van Der Walt OpenBuilds Team
    Staff Member Moderator Builder Resident Builder

    Joined:
    Mar 1, 2017
    Messages:
    13,751
    Likes Received:
    4,070
    (; BlackBox wins!
     
    Batcrave likes this.
  15. Semper Why

    Builder

    Joined:
    May 16, 2018
    Messages:
    32
    Likes Received:
    12
    Ugh. I tried to get some carving done last night and the VFD noise got progressively worse. Lots of hard limit stop messages. First pass got about 30% through the file, the second made it to 80% and then it got worse until I couldn't even set the spindle to spin up directly from the console. Moving the router further away from the VFD didn't help. I had to disable the hard limits in the firmware and watch it like a hawk until it completed. Then I re-enabled them.

    I ordered the new limit switches today. I suppose I should check the ground on the spindle while I wait for them to arrive. So frustrating.
     
  16. Batcrave

    Batcrave Journeyman
    Builder

    Joined:
    Apr 20, 2018
    Messages:
    361
    Likes Received:
    165
    That's worrying, since the progression sounds an awful lot like what happened on mine. Hopefully your control electronics won't be the next to start complaining.

    What kind of switches were you running before? And what sort did you order?

    I've got the fancy blinky Openbuilds sort, which - in addition to being NO (contrary to the advice of exactly 1.523 million safety-obsessed internet machinists) - are supposed to have onboard noise filtration of some flavor or another. I'm wondering if one or both of those saved me from ever seeing the noise show up in my switches.

    I wish I hadn't killed my noisy VFD, so I could give some sort of report on how effective the EMI filter & enclosure (which I'm still waiting for the last couple pieces and a 13/16" step drill to finish up) end up being. As it is, all I'm going to be able to prove is a negative - that the noise didn't start appearing. Nor did any tigers.

    One takeaway from the exceedingly frustrating thread over on CNCZone was the apparent claim that "noise saturation" damages the VFD's components over time - which is what causes the problem to appear and then worsen with time. Of course, that gets a little harder to understand when the VFD is also the source of the noise... but "hard to understand" could be applied to the entire conversation.

    Apparently BlackBox and fancyswitches win.


    -Bats
    (now if only they could win enough to overcome the VFD aggressively losing)
     
  17. Peter Van Der Walt

    Peter Van Der Walt OpenBuilds Team
    Staff Member Moderator Builder Resident Builder

    Joined:
    Mar 1, 2017
    Messages:
    13,751
    Likes Received:
    4,070
    Apparently BlackBox and fancyswitches win.[/QUOTE]
    No no, VFDs suck :)

    ... can confirm VFD noise eats a lot of their work time
     
  18. Batcrave

    Batcrave Journeyman
    Builder

    Joined:
    Apr 20, 2018
    Messages:
    361
    Likes Received:
    165
    I can't exactly disagree. Especially cheap Chinese ones designed for circuits that don't exist.
    (I can hear mactec objecting already)

    Unfortunately I have yet to find anything other than a VFD that'll run a [ beautiful | shiny | silent | ER16 | not-so-damned-blowy ] 3-phase spindle off a residential circuit... so sucky VFDs it is. :banghead:

    (technically I suppose a rotary phase converter might also do the job... if I didn't mind giving up speed control, could figure out the 60/400hz bit, and, most challenging of all, had room to put it)

    (this is probably Rob's cue to point out how all three could be accomplished in a package no larger than the average mailbox)

    ...had way too much time on their hands to begin with.


    -Bats
    ( New Thread: "Why Your Machine Is Unsafe And You're A Terrible Machinist Because You Disagree With Me, Vol. XXIII" )
     
  19. Peter Van Der Walt

    Peter Van Der Walt OpenBuilds Team
    Staff Member Moderator Builder Resident Builder

    Joined:
    Mar 1, 2017
    Messages:
    13,751
    Likes Received:
    4,070
    Agreed, my point being, blame the right "bad" component (;
     
  20. Batcrave

    Batcrave Journeyman
    Builder

    Joined:
    Apr 20, 2018
    Messages:
    361
    Likes Received:
    165
    Hey, you're the one trying to deprive the BlackBox & switches of their rightful glory :p

    But, seriously, I've gotten so utterly spoiled by this spindle that it hurts just to think of going back to trim routers - even after all the headaches this VFD has caused.

    And I was more than half joking when I mentioned the rotary converter, but it would give me a single source of 3-phase for the spindle, my South Bend, and the Bridgeport I've always dreamed of, meaning a total of three less VFDs.


    -Bats
    (except that then I'd end up installing three 3PH-3PH VFDs just to get speed control back)
     
  21. Semper Why

    Builder

    Joined:
    May 16, 2018
    Messages:
    32
    Likes Received:
    12
    Ugh. That would suck rocks. I've already thrown a lot of money at this thing.

    The ones that came with the OpenBuilds parts store Workbee kit. I just ordered the new special blinky ones. Hopefully they will help with my problem.

    I have a friend who was welding up a box for my VFD, but that was before this quarantine mess started. I haven't seen him since.

    I'm running the previous controller board, the XPro v4. Like the blinky limit switches, the BlackBox wasn't available when I purchased my kit.

    I really do not want to mount a router on this thing. I usually end up running this thing later at night and I don't want to disturb the neighbors too much.
     
  22. Batcrave

    Batcrave Journeyman
    Builder

    Joined:
    Apr 20, 2018
    Messages:
    361
    Likes Received:
    165
    Believe me, I'm keeping my fingers crossed that yours is strictly limited to the switches & not an early sign of the mess I ended up dealing with. Having your spindle effectively strangle your controller any time it turns on is no fun at all - and frustratingly difficult to work around.

    If the problem is limited to over-sensitive switches, then, yeah, they should definitely do the trick. NO-Is-Bad safety arguments aside (which aren't entirely unfounded, but more than a little overblown on our class of machines), it's a lot harder for noise to make an open connection appear closed than a closed connection appear open - and that's not even accounting for whatever filtration it is that they use.

    Also, they blink!

    (of course, blinking wasn't enough to save my first couple from being destroyed when I accidentally wired them to the wrong axes and the gantry kept moving while the poor little switches blinked away in silent agony)

    Oh no... maybe you should check up on him. He might've had a tragic accident and welded himself inside the box!

    Mine either. I started off running on a Gecko G540 (which I really, really liked, but eventually ran into too damned many problems with the whole PC stepgen end of things), and only recently ended up giving the BlackBox + CONTROL setup a shot... which Peter has no doubt come to regret already :p.

    (Joking aside, I had some misgivings about the idea, but it really does seem to be an awfully nice little piece of hardware. Before everything with the VFD went to hell, I was enjoying the novel experience of realizing that nearly all of my CNC problems were turning out to be my own damned fault)

    I know the feeling - and my problems may be even more immediate. I'm in a partially-finished basement, while my "shop" is in the partially-unfinished part - meaning that when I'm machining at 3am there's not much to keep the noise from coming up through the floor overhead... or angry thumps and yells from coming back down.


    -Bats
    ("But, Bats!" you say, "The solution seems obvious. Just replace the people overhead with more machines"! And to this I have no answer)
     
  23. Governor

    Governor New
    Builder

    Joined:
    Jan 29, 2016
    Messages:
    30
    Likes Received:
    6
    First I have really enjoyed this thread. It has been very funny at times yet informative. I seem to be having the the same problems and have been for the past 4 years, sad to say and embarrassed about it as well. I get really frustrated with it and step away for several months at a time. I can no longer go on with this not working. So here's the question. Did I make a BIG newbie mistake by buying a 2.2 kw spindle and VFD for this build and trying to run it with an Arduino Uno and V3 sheild. There seems to be massive EMI going on but don't know how to check for it. I will mainly be cutting aluminum is the reason for such a big spindle. When I try cutting a simple pocket in multiple dept of cuts the machine acts a fool after the first step down. Looking at the spindle position on the DRO it reads right but the spindle will actually step down 2-3 time more than what the DRO is reading. At first I thought it was the Nema 23 loosing steps so I upgraded to a 425 in. oz. with no luck. I am about to pull the trigger on the Blackbox motion control system and upgrade the limit switches from Open Builds today but I don't want to just throw money at this and don't know what the he!! I am doing. I also noticed that you kept mentioning putting the VFD in a metal box. Is that also something I need to do? Please any help is GREATLY APPRECIATED!!!!!! I feel your pain and thank you for the uplifting thread. I so needed to laugh while in this nightmare.
     
  24. Rob Taylor

    Rob Taylor Master
    Builder

    Joined:
    Dec 15, 2013
    Messages:
    1,470
    Likes Received:
    746
    Kill the V3 shield, it can't drive those motors, the drivers are too low power. You need to tap the Arduino STP/DIR/ENA outputs and send them to real drivers like DM542s that'll push 2-3A or more comfortably, or go with the Blackbox which also has higher power drivers inside.

    A 2.2kW spindle is way overkill for an aluminum extrusion machine. Even 1.5kW is a little on the high side, but probably doable with C-Beam based machines. 800W (real watts remember, not peak (fake) router watts) is probably more of a sweet spot for most general machines. You don't necessarily need to get rid of it, but make sure your gantry isn't flexing under its weight and make sure your Z axis motor is beefy enough to drive it up and down without losing steps (once you have the right drivers). You won't come close to maxing out its power and bogging down though, so at least that's nice.

    Once you've taken care of those two main issues, worry about EMI from the VFD, if the upgraded limits actually give any problems.
     
    Governor and Peter Van Der Walt like this.
  25. Batcrave

    Batcrave Journeyman
    Builder

    Joined:
    Apr 20, 2018
    Messages:
    361
    Likes Received:
    165
    I take absolutely no responsibility for any informative content anyone may have infected this thread with!

    A giganormous spindle paired with a teensy little Arduino (not to be confused with a little Teensy *duino) certainly makes for a funny-looking combination, but I don't think there's any inherent incompatibility. Unlike a Gecko/parallel-port/software stepgen system, it doesn't take much muscle at all to feed a GRBL board (I recently picked up an old CNC'd Taig Microlathe, and I'm running it with an xPro v3 controlled by a 7 year old android tablet - after running into frustration with an old Raspberry Pi model B and failing to convert an old Macbook Pro to linux), and the VFD doesn't know or care what it gets its on/off or PWM signals from.

    The one thing that might be an issue - if your problem is really EMI - is that, unless you've got them in an appropriate enclosure, neither the xPro nor the Uno has anything in the way of shielding to protect it. That said, I'm not sure whether your problem really is EMI yet- at least without more information.

    That's both really easy and really... not easy.

    It's easy to identify prolific noise sources by waving around a portable AM radio tuned to dead areas. In my case, noise from the VFD, steppers, and AC wiring all tended to be most observable - read: screechiest - around 880kHz. But that doesn't tell you much, when you're dealing with a bunch of devices (like, say, a VFD, steppers, and AC wiring) that you already know spit out lots of noise.

    What's more challenging is isolating which of those noise producers are actually causing problems - especially when they're all part of the same system.

    Usually a more productive angle is to look for symptoms that are especially noise-y to begin with. Semper was seeing a lot of limit switches triggering when nothing was pressing them, which is an extremely common problem when there's noise involved. In my case, I was getting literal noise turning up in in the console/GRBL logs.

    (ok, so maybe it wasn't technically noise so much as corrupted GRBL status messages... but what's noise anyhow, if not corrupted signal?)

    (fine, so maybe it's technically "irrelevant or meaningless data or output occurring along with desired information"... but that's close enough)

    Based on my experience, even my 1.5kW is bottlenecked by the machine's (lack of) rigidity - I strongly suspect someone with an 800W on a similar aluminum extrusion machine would get about the same performance as either of us... but I still like mine for the bigger collets it offers. And, besides, we use what we've got, not what might theoretically be a bit more optimal under certain uncertainly certain metrics.

    That size spindle does introduce some problems of its own, though - even moreso if it's one of the really monstrous 110/120V versions (I seem to remember there being some moderately slimmer 220/240V versions).

    And that's where it starts sounding like one of those problems unique to those of us with spindle. My first guess was the same as yours - that it was just plain too damned heavy for the Z motor to handle & was being dragged down... but the heavier stepper should have addressed that.

    The next things I'd want to look into before spending any money on new electronics (Sorry Mark, Peter - I'm not trying to cost you guys sales! Honest!) are also tied to potential weight issues - although if you've been fighting this for years, you may well have ruled at least some of these out already:

    First, just to have a solid starting point, I'd double-check all the hardware on the Z axis (yes, again. sometimes I don't notice the whole stepper is rattling loose until my third check) - make sure all the screws (especially those holding on the end caps & steppers, as well as the setscrews on the leadscrew couplings & collars - I've had similar problems when the stepper shaft/leadscrew was able to spin in the coupling) are tight and there isn't any significant backlash (you don't need to pull out the indicators here - just make sure you don't feel a lot of shifting when you try to lift & press down the spindle/axis manually).

    You might also want to make sure the leadnut isn't stripped and/or jumping threads, although that's more of a theoretical "what if" - it's not something I can remember ever seeing anyone complain about on Openbuilds hardware.

    Once you've eliminated any potential hardware issues, I'd try turning the Z axis speed and acceleration way down (and, if that doesn't change anything, try turning it waaaaay down). This should eliminate most inertia-related problems, while if the problem is noise-related, you should still see the same behavior.

    If you're still seeing it, the next step (or maybe it should be the first step - I always get those two confused) would be to try running the same toolpath (zeroed at the same height, just to maintain consistency, but without a workpiece) both with and without the spindle running. That might not completely eliminate every weird possibility (vibration? acoustic noise? leaking water? I'm kinda stretching here) but it should give a pretty strong indication of whether or not the VFD/spindle is really the culprit.

    Holding off on the electronics purchase for the moment is the right idea. I can absolutely recommend both the BlackBox and blinkyswitches in their own right (see, Mark? see, Peter? please don't hate me!), but I don't know that they're likely to do anything to solve your problem.

    The fancy switches were the right way to go for Semper, because the problem was with the existing switches picking up noise (as reflected by a lot of spurious hard limit messages). The BlackBox was a bit of a different matter - it didn't solve anything in itself (well, it did - it solved a fair bit - but those problems were utterly unrelated) - it just won by virtue of being cleanly exonerated when the fierce but fickle finger of justice finally came down pointing squarely (what's an "f" word for "squarel... firmly!"?) pointing firmly at the habeas corpse of my by-then-deceased VFD.

    Then you're in good hands. Neither do I! :confused:

    That depends. A solid enclosure, shielded cables between the VFD & spindle, proper wire routing, and an EMI filter before the mains line goes into the VFD are all generally recommended (as are a wide variety of other things that most of us also neglect)... but let's hold off a little longer on jumping into that. If everything does still point to the VFD, then we can get into the really fun stuff (read: the really, really, not fun stuff), but first let's try to rule out all the easier possibilities before trying to go to war with the really nasty gremlins. The invisible species always put up the biggest fights.

    What good are nightmares you can't laugh at?

    -Bats
    (laughing at nightmares since last Tuesday)
     
  26. Batcrave

    Batcrave Journeyman
    Builder

    Joined:
    Apr 20, 2018
    Messages:
    361
    Likes Received:
    165
    When Rob & I give the opposite advice, always defer to the one who isn't named after a species known for battiness and belfries.**

    I didn't even think about the issue of driving the steppers - and inadequate driver power definitely could contribute to a heavy spindle dropping.

    At the same time, if the problem predated the move to larger steppers (which is what the description made it sound like), that's probably not the source of it, and so replacing the driver might not be a solution in itself.

    Funny, I think I said exactly the same thing immediately before/after (or possibly even during) your post.

    Probably means we're both wrong.

    Governor didn't mention what sort of machine he(?) was running, either. Four years would predate the Lead, but I don't remember when the earlier C-Beam Machines were first offered.

    That's the other thing I forgot to ask about (which is surprising, since it's usually the first thing I'm curious about when anyone else mentions running a big spindle), was how much flex the gantry was showing. I have trouble imagining that being responsible for the problem described, though (although it could obviously cause headaches of its own)

    That's an understatement. I usually worry that I'm going to damage the machine before I bog down the spindle :p

    Now this I'm going to have to (partially) disagree with. I agree (or at least I hope) that chasing EMI should probably be low on the list of priorities - since I'm not convinced that's the real problem - and obviously big steppers will need bigger drivers, but I'm not sure diving right in to throwing money at that end of the problem (especially without knowing exactly where the problem he's asking about lies) is really the right way to start off the process. I'd feel awful (and I'm sure he'd be frustrated) if I suggested a big cash/time investment in the electronics only to discover the core problem was rooted elsewhere & was still around. Sure, it would've solved other problems that he'd probably be running into eventually, but they wouldn't have solved The Problem that brought him here.

    It also may turn out that throwing the money in a different direction would turn out to be more sensible. If he's got a machine with a flexy gantry, it's possible it would make more sense to downgrade to a smaller spindle/drive and sidestep the stepper/driver problem that way.

    So new electronics may well be the way to go (and if money isn't tight, the BB is a massive upgrade just on style alone), but if there's a budget to consider, I'd be inclined to dig a little deeper before deciding what gets the green paper bandage first.


    -Bats
    ** not to be confused with the Flapsalottos Robitaylori bat.
     
  27. Governor

    Governor New
    Builder

    Joined:
    Jan 29, 2016
    Messages:
    30
    Likes Received:
    6
    Thanks Rob I will be ordering a Blackbox limit switches and cables in the morning and start there.
     
    Peter Van Der Walt likes this.
  28. Governor

    Governor New
    Builder

    Joined:
    Jan 29, 2016
    Messages:
    30
    Likes Received:
    6
    Batcrave I'm with you on all of your suggestions which makes me feel a little better. My thoughts were the same and I've tried all of the above that you have mentioned. Several times now latest being two days ago. I can say that I have noticed some deflection but it is not from the gantry. It seems to be coming from the Z actuator do to the heavy spindle. The deflection I can live with for now because it's not much at all but I do have a fix for that when I can get rid of this ghost. I must say that I don't mind spending the money on a Blackbox set up if it's a step in the right direction. There are not many with this style set up therefore, not much help. You guys seem to VERY knowledgeable and willing to help. I do take my hat off to you all.
     
  29. Governor

    Governor New
    Builder

    Joined:
    Jan 29, 2016
    Messages:
    30
    Likes Received:
    6
    The one thing I have not tried lately is totally removing the spindle from the gantry. I'll be back in a half hour. going to try that now and let you know.
     
  30. Governor

    Governor New
    Builder

    Joined:
    Jan 29, 2016
    Messages:
    30
    Likes Received:
    6
    Well who wants to know how that test without the heavy spindle went. :( Not well at all. Amazingly it did exactly the same thing. No power to the VFD just to the controller. I am so baffled right now. thought for sure it would at least do something different. By the way the limit switches are not hooked to the Arduino at all.:banghead:
     
    #60 Governor, Apr 30, 2020
    Last edited: Apr 30, 2020

Share This Page

  1. This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
    By continuing to use this site, you are consenting to our use of cookies.
    Dismiss Notice