#linuxcnc Logs
Jul 14 2022
#linuxcnc Calendar
12:33 AM randy: morning
01:22 AM -!- #linuxcnc mode set to +v by ChanServ
01:41 AM Deejay: moin
02:18 AM pere: hi
03:31 AM toluene3 is now known as toluene
04:42 AM Tom_L: morning
04:58 AM JT-Cave: morning
05:17 AM travis_farmer: Good Morning :) supposed to be a bit cooler today, so it sounds like a workshop day (or morning, anyway...)
05:30 AM toluene0 is now known as toluene
08:15 AM CaptHindsight[m]: only in the 70's here, but humid
08:17 AM CaptHindsight[m]: hazzy: I forgot Amada makes them. I usually just think lasers.
09:18 AM roycroft: we'll hit 30 here today, but for this time of year that's not too hot
09:23 AM CaptHindsight[m]: temps here have been close to normal, the heat has stayed to the south
09:27 AM hazzy[m]: CaptHindsight: you got that right! It’s been in the 100s down here
09:28 AM CaptHindsight[m]: hazzy: are parts getting any easier to get or is it still slow?
09:31 AM CaptHindsight[m]: https://libre.computer/products/s905x/ ~$50ea and has the Rpi pinouts
09:31 AM hazzy[m]: CaptHindsight: it depends. Some things have gotten better, but other parts are still next to impossible to get. We are shipping some machines without motors since we have no room to warehouse until they come in, and no expected delivery date. Problem is, we can’t demand payment until the machines are complete
09:33 AM CaptHindsight[m]: and the customer can't use them since they don't have motors :(
09:34 AM roycroft: our temps will be right about average for the next week at least
09:35 AM CaptHindsight[m]: I'd probably be retrofitting available motors and then swapping out later
09:35 AM hazzy[m]: Yep! But they have to store it, and then we have to pay to fly techs out and install the motors once we get them
09:36 AM CaptHindsight[m]: could be worse, cancelled orders due to no motors
09:36 AM hazzy[m]: CaptHindsight: unfortunately that’s hard to do with servo motors
09:37 AM CaptHindsight[m]: I never said it was going to be easy :)
09:40 AM rmu: CaptHindsight[m]: is RT PREEMPT kernel running on those "le potato"s with acceptable performance?
09:41 AM hazzy[m]: CaptHindsight: I actually got quotes on getting custom motors made in China to use temporarily, but the cost and lead time was almost as bad as the German made motors
09:41 AM CaptHindsight[m]: AML-S905X-CC might be, I don't have one yet, but I'd prefer something with an integrated micro for all-in-one board software steping
09:42 AM CaptHindsight[m]: hazzy: too bad that there are so few US servo shops
09:42 AM rmu: CaptHindsight[m]: i have one but it is running kodi
09:43 AM CaptHindsight[m]: since it runs u-boot there is a good chance that latency jitter may be kept low
09:44 AM CaptHindsight[m]: https://www.amazon.com/s?k=%22libre+computer%22+aml-s905x-cc&linkCode=sl2&linkId=6ef3196e613636af4f5bb7dfcd2d9935&tag=loverpi05-20&ref=as_li_ss_tl
09:46 AM CaptHindsight[m]: https://forum.linuxcnc.org/18-computer/46077-potential-all-in-one-single-board-computers list of potential ARM boards
09:48 AM CaptHindsight[m]: https://wiki.radxa.com/Rockpi4 needs a port of the LCNC for the Opi
09:49 AM CaptHindsight[m]: http://www.orangepi.org/html/hardWare/computerAndMicrocontrollers/details/orange-pi-4B.html
09:50 AM CaptHindsight[m]: has 2x microcontrollers integrated as well as PCIe
09:51 AM CaptHindsight[m]: https://www.amazon.com/dp/B09W8G67CV
09:52 AM CaptHindsight[m]: but getting into the x86 price range
09:57 AM rmu: CaptHindsight[m]: you will need a breakout board for CNCing anyway, I don't see an additional µC on such a break out board for step generation as a big cost item (if and when such things will become available again to mere mortals).
09:58 AM CaptHindsight[m]: in the future ....
10:01 AM CaptHindsight[m]: i want my dual core Ryzen embedded cpu with integrated GPU for $25
10:05 AM * roycroft is getting tired of cheap, crappy filament, and has decided that the cheap stuff is way to expensive
10:31 AM roguish[m]: roycroft: a poor man can only afford the very best..................pay me now, or pay me later...................
10:32 AM roycroft: indeed
10:33 AM roycroft: my only real disappointment is that there are no quantity discounts available for the good stuff
10:33 AM roycroft: the cheap filament gets even cheaper in bulk
10:33 AM roycroft: the stuff that i know works well is the same price whether i order 1kg or 100kg
10:33 AM roycroft: a slight quantity discount would be nice
10:35 AM roycroft: and on that note, gst3d just sent me an email
10:35 AM roycroft: "please send picture"
10:35 AM roycroft: do they want me to take a picture of their filament breaking inside the bowden tube?
10:35 AM roycroft: i wonder how i would do that
10:36 AM roycroft: would a picture of a partial print that never completed because the filament broke be helpful?
10:36 AM roycroft: why would i need a picture for that when i've already told them that happens?
10:37 AM roycroft: if they would tell me exactly what kind of picture they want i'll do my best to provide one
10:38 AM roycroft: but i'm not going to load their crap in my printer, waste time making a failed print, and then taking a picture that might not be what they want
10:39 AM CaptHindsight[m]: https://www.zyltech.com/zyltech-filament/
10:40 AM roycroft: have you personally used that filament, capthindsight?
10:41 AM CaptHindsight[m]: I rarely use FFF/FDM
10:41 AM CaptHindsight[m]: what everyone used to mention for low cost and US supply of PLA
10:42 AM roycroft: the price is ok
10:42 AM roycroft: it's double the bulk price of gst3d, but if it works it's way cheaper than gst3d
10:42 AM roycroft: i can live with $20/kg or so if it works
10:43 AM roycroft: atomic filament, which i really like and has never caused me any problems, is $30/kg
10:43 AM roycroft: and for day to day use, i don't mind that
10:43 AM roycroft: but for the 30kg or so that i need for these hardware storage boxes, i'd like to get closer to $20
10:44 AM roycroft: if shipping is not too much i'll try a kg or two of zyltech
10:45 AM roycroft: $10/kg for shipping
10:45 AM roycroft: so it costs more than atomic filament
10:47 AM CaptHindsight[m]: https://www.zyltech.com/5kg-texas-size-spool-abs-3d-printer-filament-1-75-mm-11-lbs/
10:48 AM CaptHindsight[m]: White $69.95
10:48 AM CaptHindsight[m]: $75 with tax + shipping for 5kg
10:50 AM roycroft: they don't have the colors i need in 5kg spools of pla
10:51 AM CaptHindsight[m]: white + can of Krylon
10:52 AM roycroft: atomic do have 3.5kg jumbo spools, but the price on those is still $28.57/kg
10:52 AM roycroft: not much of a savings
10:52 AM roycroft: and the larger spools would be inconvenient for me to handle
10:53 AM roycroft: but at the end of the day, that may be what i need to do - i'm getting really frustrated with wasting filament and, more importantly, not making steady progress on my project
10:54 AM CaptHindsight[m]: 25kg of sawdust and 5kg of PVA
10:55 AM roycroft: this stuff i have now is actually decent filament, but it's coiled horribly
10:55 AM roycroft: i have to babysit my prints to avoid breaking the filament due to jamming in the coil
10:56 AM roycroft: i've gone through almost the entire 1kg roll, and have six good boxes to show for it
10:56 AM roycroft: i should have about 30 boxes
11:00 AM CaptHindsight[m]: silicone mold + sawdust and Elmers polyester resin + BAPO = dual cure wood resin (UV and MEKP cure)
01:35 PM Kerr[m]: Hi! The engineering club I'm part of is considering building a CNC machine from scratch. We're considering LinuxCNC for the control... but there is so much *old* information we're concerned we would have trouble finding a hardware setup that work. Forum posts about writing your own drivers are scary lol
01:36 PM Kerr[m]: * Hi! The engineering club I'm part of is considering building a CNC machine from scratch. We're considering LinuxCNC for the control... but there is so much *old* information we're concerned we would have trouble finding a hardware setup that works. Forum posts about writing your own drivers are scary lol
01:36 PM Kerr[m]: Overall, are things smoother than old posts would suggest, or...
01:37 PM Kerr[m]: Hmm... I just saw that everyone here is a "bot" is this bridged to irc somewhere?
01:38 PM roycroft: no, not everyone here is a bot
01:38 PM roycroft: but there is a bridge to matrix
01:38 PM roycroft: the [m] users are on the matrix side
01:39 PM roycroft: i don't know what hardware you're talking about, but linuxcnc is under very active development, and there is a lot of current hardwar that is fully supported
01:40 PM jpa-: Kerr[m]: i built a custom CNC at our hackerspace; the linuxcnc part was mostly smooth and i didn't run into much outdated info - but it does have a lot of things to configure and choose and setup
01:41 PM roycroft: it is true that there's a fair amount of configuration work needed, but there is pretty decent documentation, a wiki, a forum site, and the irc channel
01:41 PM roycroft: folks are generaly quite willing to answer questions and otherwise help out
01:41 PM roycroft: when they're not busy talking about chickens or privacy issues and the like :)
01:43 PM Kerr[m]: Okay, that's encouraging. We're almost certainly just using stepper motors with step/dir interface on the drivers. Open loop. For a machine to run it on, I have an asrock n68 ucc motherboard with a LPT header on the board, and a bag of athlon II processors. Am I likely to run into issues using that?
01:44 PM Kerr[m]: s/machine/computer/
01:44 PM Kerr[m]: What sort of things should I be looking for?
01:44 PM jpa-: boot from the livecd and run the latency test
01:45 PM jpa-: one thing you may eventually get annoyed about is that such old system is too slow to run most CAD / CAM programs; it is quite convenient to be able to run them on the same machine
01:47 PM jpa-: i also started with stepper machine, but if possible, i would really recommend feedback - it avoids a lot of annoyance that lost steps due to mechanical issues can cause
01:47 PM roycroft: what jpa- said
01:47 PM roycroft: the first thing to do is boot the machine from the livecd and run the latency test
01:47 PM roycroft: if the latency is low enough you should be good to go
01:48 PM roycroft: if not, consider a different main system board with lower latency, or some kind of smart motion control hardware that you can offload to
01:48 PM Kerr[m]: What sort of sensors should we use for feedback?
01:49 PM roycroft: mesanet products are widely used for that, and the owner of mesanet is on this channel, as is a us distributor for those product (who has also written some nice configuration tools to help you get going faster)
01:49 PM Kerr[m]: Is it like... "hook it up to a pin on the lpt port & do some configuration" ?
01:50 PM roycroft: for step/gen you usually don't use position encoders - you rely on the number of steps to keep track of position
01:50 PM roycroft: so you just use limit switches for homing and the like
01:50 PM jpa-: on my latest machine, i'm using rotary encoders for feedback, mesa 6i25 PCIe card for interface and DC motors for drive; but it sure is a lot more complex than steppers
01:51 PM roycroft: and even if your latency is low, it is strongly recommended that you use a breakout board with optoisolators in between your lpt port and the drivers
01:51 PM roycroft: it's pretty easy to fry an lpt board
01:51 PM jpa-: (or a driver with integrated isolators)
01:52 PM roycroft: sure, but that doesn't protect against the other signals you need to monitor/manage
01:52 PM Kerr[m]: The drivers are ones the club designed last year, they have optoisolators on the inputs
01:54 PM jpa-: https://www.omc-stepperonline.com/e-series-nema-34-closed-loop-stepper-motor-4-8-nm-679-87oz-in-encoder-1000ppr-4000cpr-34he31-6004d-e1000 there are also stepper motors that come with rotary encoders already; linuxcnc works fine with stepper + encoder combo also - in practice it will compensate for small amounts of lost steps and signal error & stop if some axis gets fully stuck
01:56 PM Kerr[m]: oh that's pretty cheap too
01:57 PM jpa-: though if you are using only single parallel port, you may run out of signal lines quite quickly
01:57 PM Kerr[m]: hmm.
01:58 PM jpa-: if you have each axis with step/dir/encoderA/encoderB signals, that's already 12 pins which is the maximum amount of outputs for a parallel port
01:59 PM Kerr[m]: Would I be able to use the onboard parellel port in conjunction with a pcie parallel port?
01:59 PM Kerr[m]: s/parellel/parallel/
01:59 PM jpa-: should be
02:00 PM Kerr[m]: how does the latency test work - is there anything external (ie, with the lpt port) or it is just a test of the computer latency?
02:00 PM jpa-: it just tests the computer latency
02:01 PM Kerr[m]: Do I need to do a live cd, or can it be run from a flash drive?
02:01 PM jpa-: though the livecd also includes full linuxcnc system, so you can test generating stepper signals with http://linuxcnc.org/docs/html/config/stepconf.html
02:01 PM Kerr[m]: idk if I have a cd/dvd drive anywhere
02:01 PM jpa-: sure, USB stick works fine
02:02 PM jpa-: that's what most people will mean with "live cd" nowadays, it is just CD image that has been written onto USB stick
02:02 PM roycroft: "live image" would probably be a better term to use
02:06 PM Kerr[m]: using the stepper motor with encoder posted a moment ago as an example, which signals need to be connected? I see that one comes with a male 15 pin DB, of which there is EA+, VCC, GND, DB+, EB- EA- : Are those differential signals, or does it do something like emit a pulse indicating direction? (Apologies, I'm not familiar with encoders, electrically)
02:07 PM Kerr[m]: Also, VCC voltage isn't shown on the "full datasheet" lol
02:08 PM jpa-: that encoder has differential signals indeed; they handle noise better, but you'd need a differential receiver (e.g. some RS422 chip) to connect it to parallel port
02:12 PM jpa-: when you get to the stage of choosing a spindle, i'd recommend going with something relatively powerful like https://eur.vevor.com/other-c_11229/0-8kw-air-cooled-spindle-motor-and-2hp-1-5kw-variable-frequency-drive-vfd-p_010340991079 - in my experience a good spindle can make even a mediocre machine useful and even a good machine is annoying to use with bad spindle
02:16 PM Kerr[m]: This was the last one mentioned
02:16 PM Kerr[m]: https://www.ebay.com/itm/275088625379?epid=1972295114&hash=item400c8f4ee3:g:1MAAAOSwnz9h0wXB&amdata=enc%3AAQAHAAAA8DT1ptA6bzUOhY2btmw627fKySbPB6TfSfzmu%2FMNUNLj4l1WAS2gpUXs4JiLHbdwvZmnyIeY1rn2K3IF2x4r1wekTbtBk7qNW6oMPjPvwtlk9KN6lclfzw25WawXpUZ1G49omhBnKpJrx0ba5QieUCuAprqUgIKsha3NVF5%2FGaw430l0sSiUKbQRn%2FM7L2TFkVJkHnEvCaZddo442OqqankYHAGVXInkkydmg0zLeSQpuWTDpt%2BcvIehD26D74l2NUQdgR3OGGUulLEXo458oHwhXdRdqML85zvLDK9VVa0jOSFL5%2Bqt1fWyrFNwD18OP
02:16 PM Kerr[m]: g%3D%3D%7Ctkp%3ABFBMguWT4r9g
02:16 PM jpa-: yeah, probably similar stuff (a bit larger diameter)
02:21 PM Tom_L: anything after the ? in the link is not needed for the post
02:21 PM Tom_L: https://www.ebay.com/itm/275088625379
02:22 PM Tom_L: is just as effective
02:24 PM Tom_L: what sort of scratch built cnc btw? mill, lathe, laser cutter, router ?
02:26 PM Tom_L: asrock may not be the best choice but i'd still run a latency test to see
02:26 PM Kerr[m]: We're thinking a tabletop mill type thing. I want to do a cast iron base (~200lb) and make it a bit more rigid than most in that size
02:26 PM Kerr[m]: "tabletop" being a relative term. ~12"x15" working area maybe
02:27 PM Kerr[m]: Cost wise, debating buying pre-built ways from mcmastercarr, or maknig them myself
02:27 PM roycroft: the yuan-yang spindles are good spindles
02:27 PM Kerr[m]: Metal is so expensive right now 😦
02:27 PM roycroft: but two things:
02:27 PM Tom_L: gantry style or other?
02:27 PM roycroft: 1. there is a 2 bearing version and a 3 bearing version of almost all their spindles
02:28 PM roycroft: the 3 bearing version is slightly more expensive, but has much less runout and lasts a lot longer
02:28 PM Kerr[m]: thinking gantry style, one axis on the table, one on the gantry
02:28 PM Kerr[m]: then ofc z axis with the spindle
02:28 PM roycroft: 2. there are lot of huan-yang "clones" of inferior quality
02:28 PM Kerr[m]: gantry style just seems like a way more rigid design
02:28 PM roycroft: my recommendation is to go on amazon.com and buy the spindle from huan-yang directly to make sure you're getting the real thing
02:29 PM Tom_L: gantry seems the most popular
02:29 PM Tom_L: i went a different route on mine
02:29 PM roycroft: i'm building a gantry router to save space
02:29 PM XXCoder: gantry is bit easier to build I think
02:30 PM Tom_L: http://tom-itx.no-ip.biz:81/~webpage/cnc/Column_Mill_VMC.jpg
02:31 PM Tom_L: that was my design idea
02:31 PM Tom_L: completely scratch built
02:31 PM roycroft: i'm not seeing the air-cooled ones on amazon at the moment
02:31 PM roycroft: i have some water-cooled ones, and i prefer those anyway
02:32 PM roycroft: but that does require a sump and a pump and some additional periodic maintenance
02:32 PM roycroft: for a maker space it might not be a great idea
02:41 PM Kerr[m]: How do they do it the lower end of RPM
02:43 PM ZincBoy[CAON][m]: What do you consider lower end? Anything below 10k is not going to be great. Most of them are designed for wood routing and are not meant for low rpm operation.
02:43 PM Kerr[m]: That's a shame
02:43 PM XXCoder: I suppose can always do gearing but yeah
02:43 PM Kerr[m]: And he spend the recommendations for something more in the 5000 rpm range
02:44 PM Kerr[m]: More interested in cutting aluminum and a little bit of steel
02:45 PM ZincBoy[CAON][m]: A 24k spindle is fine for aluminum with the right tooling (1/4" cutters typically). You won't be running a 4" face mill but you won't have the HP or rigidity for that either.
02:46 PM ZincBoy[CAON][m]: Steel is technically possible again with the right cutters but is not usually practical. I have done it with a router and a 1/8" TiAlN coated endmill. It was not great.
02:47 PM Kerr[m]: Well on the topic of virginity that's why we're talking about doing a cast iron base and probably some heavy Box ways, Is not some 80:20 ant tubes and bushings
02:47 PM Kerr[m]: s/ant/and/
02:48 PM ZincBoy[CAON][m]: For what you are doing, something like this would be better: https://www.aliexpress.com/item/4000411342262.html
02:48 PM roycroft: i think you would grow impatient trying to make steel parts with a smallish gantry machine with an 80mm spindle
02:48 PM roycroft: you would have to use such tiny cutters it would take forever to make anything
02:50 PM roycroft: and btw, i've run those 24krpm spindles down to about 4krpm with a huan-yang vfd without stalling
02:50 PM roycroft: but i have not done so with an 8mm end mill cutting steel
02:50 PM ZincBoy[CAON][m]: That is just the first one that popped up, there are better ones on there. I have to say that a cast iron frame and box ways is ambitious. That is a ton of work. I would suggest just buying an older VMC with dead control and retrofit it.
02:51 PM ZincBoy[CAON][m]: I would be worried about overheating at that low an RPM. Plus your torque is going to suck. Might be fine for wood but not steel.
02:51 PM Kerr[m]: That's why I asked
02:51 PM roycroft: i use water-cooled spindles
02:52 PM roycroft: so the overheating thing is not an issue
02:52 PM Kerr[m]: If the spindles readily available aren't really able to do low rpm's, maybe the cast iron frame and box ways arent worth it
02:52 PM roycroft: and i did mention that i was not cutting steel with a big cutter at that speed
02:53 PM Kerr[m]: I don't want to also machine a spindle lol. The lathes at the makerspace suck
02:53 PM roycroft: i see fairly large cnc mills with dead fanuc controllers for $5k and less on cl regularly
02:53 PM roycroft: they tend to date back to the '90s or earlier
02:54 PM roycroft: but as was suggsted here a short while ago, something like that with a linuxcnc retrofit might be a good match for your project
02:54 PM roycroft: or should i say a linuxcnc upgrade? :)
02:54 PM ZincBoy[CAON][m]: Yes, I picked one up for $2k with a dead control. The total cost of the conversion was about $10k plus about $200k of my time 🙂
02:55 PM ZincBoy[CAON][m]: I was still faster and cheaper than trying to do a box way frame from scratch.
02:55 PM ZincBoy[CAON][m]: Plus I have a tool changer, coolant, and chip augers.
02:55 PM CaptHindsight[m]: Kerr: where are you located? +/- 100 miles or k
02:56 PM CaptHindsight[m]: km
02:56 PM Kerr[m]: Washington, usa
02:57 PM roycroft: where in washington?
02:57 PM * roycroft used to live in bellingham
02:57 PM CaptHindsight[m]: you're asking good questions, lots just hope and pray on their spindle and motor choices
03:00 PM roycroft: the problem with living in the pacific northwet is that we're far from the rust belt, and great deals on old, big iron are not common
03:01 PM Kerr[m]: Lets downgrade it from milling steel, just mill aluminium with one of those 18k spindles. What sort of options are there for prebuilt frames?
03:01 PM Kerr[m]: I'm inland northwest
03:01 PM XXCoder: most people living at wa state live to left of mountains lol
03:01 PM XXCoder: like me
03:01 PM XXCoder: (by left I mean west)
03:02 PM roycroft: what is the largest part you want to make?
03:02 PM roycroft: i'm not sure if you mentioned that already
03:02 PM roycroft: but that's a pretty important bit of information
03:03 PM ZincBoy[CAON][m]: What is your budget and are you looking for a primarily wood machine that can be used for Al or primarily for Al
03:03 PM roycroft: i would suggest as well that for a home shop, a combination wood/non-ferrous router is fine, but for a maker space, i would dedicate the machine to one material or the other
03:04 PM roycroft: you have to do extensive cleaning when you switch between wood and metal for best results
03:04 PM CloudEvil: It'd be nice from a number of perspectives if you could just run a shop with an oxygen free environment.
03:04 PM ZincBoy[CAON][m]: If this is a machine for Al, you will want coolant and some way to enclose the machine. It is possible to machine Al without coolant but it really, really sucks.
03:04 PM roycroft: and not everyone who works in a maker space can be relied on to do that kind of tidying up
03:05 PM Tom_L: funny how everybody jumps to the 4000000000 rpm spindles right off...
03:05 PM roycroft: yes, although one of those noga mist coolant systems would be fine, and would not require a sump to collect the wast coolant
03:05 PM roycroft: waste
03:05 PM CaptHindsight[m]: https://www.ebay.com/itm/234575703988
03:05 PM CloudEvil: roycroft: Or competent at cleaning up - it's quite possible to damage stuff incompetently cleaning it.
03:05 PM roycroft: but i'd still want to enclose the machine, even with mist coolant
03:06 PM roycroft: and locate it away from the rest of the machinery
03:06 PM roycroft: good point, cloudevil
03:06 PM ZincBoy[CAON][m]: Yes, it is even more important as the chips will blow everywhere.
03:07 PM roycroft: i don't think poorly of makerspaces, btw
03:07 PM roycroft: but i know that they tend to attract inexperienced makers
03:07 PM roycroft: and it can be difficult to supervise the users well enough to ensure bad things don't happen
03:08 PM CloudEvil: Working out who really needs a two hour (equivalent) video on how to do stuff... is hard
03:08 PM roycroft: my expectation is that tools at makerspaces tend to be in suboptimal condition, and expectations should be set appropriately
03:08 PM ZincBoy[CAON][m]: I am a member of a makerspace and there is a ton of training needed for most of the equipment and it is all accessed by your keyfob.
03:09 PM roycroft: my local makerspace has a bridgeport mill with a cnc conversion, and i was going to make parts for my own mill conversion there, but when i saw the condition of the machine and the eternal mess surrounding it, i knew that would likely not work out well for me
03:09 PM CaptHindsight[m]: https://www.ebay.com/itm/115450399231
03:09 PM roycroft: the better ones are like that, zincboy
03:09 PM CaptHindsight[m]: https://www.ebay.com/itm/203267825037
03:10 PM ZincBoy[CAON][m]: Agreed. Need to make sure things like door interlocks and the like work and are difficult to bypass.
03:10 PM roycroft: there's a woodworking guild in portland who are very particular about that
03:10 PM XXCoder: that okuma is exact model I ran before. I hated it
03:10 PM roycroft: there's a keyfob that is required to operate the machines, and you have train on each machine you want to use, and pass a test on it, before access to it is granted to your keyfob
03:10 PM CaptHindsight[m]: https://www.ebay.com/itm/275386233535
03:11 PM XXCoder: machine design is ok, its software side thats problemic. ie it would error out if you change from tool 1 to tool 1, leading to bad programming changes
03:11 PM XXCoder: making it unsafe, if you only use one tool for example
03:12 PM CaptHindsight[m]: looks like lots of used Bridgeport knock-offs in the WA area
03:15 PM Kerr[m]: Aluminium, no wood
03:16 PM Kerr[m]: Not going into a makerspace - the engineering club workshop
03:16 PM Kerr[m]: We just will have to go to a makerspace to use their bridgeport and grinder
03:16 PM XXCoder: whats your workspace needs like anyway
03:16 PM roycroft: my frustration with that wood shop, were i to consider becoming a member, is their approach to the question above - how to determine who has to go through the training before getting keyfob access
03:16 PM CaptHindsight[m]: Kerr: the tradeoffs are how good are your used part hunting skills vs budget and time, how much skill do you have properly assembling and aligning a machine?
03:17 PM Kerr[m]: Don't have the budget for a whole existing machine
03:17 PM roycroft: skip the training and take the test would be fine
03:17 PM roycroft: but their approach is that everyone trains on every machine, no matter how much experience and competence they demonstrate
03:17 PM Kerr[m]: The makerspace I go to has thrown away the training matrix like 4 times, requiring everyone to be retrained
03:17 PM ZincBoy[CAON][m]: Kerr One thing to consider is that if you want to do? Make parts or build a CNC machine? Very different requirements.
03:17 PM Kerr[m]: I spent 10 years as a machinist - in school for engineering now
03:18 PM Kerr[m]: Both, more of the latter zinc
03:18 PM XXCoder: wood cnc you can slap stuff together. alum cutting machine is bit harder to do
03:18 PM ZincBoy[CAON][m]: Okay, that helps, a conversion is not as much fun then 🙂 What is your timeline?
03:22 PM CaptHindsight[m]: Kerr: what is the budget?
03:25 PM CaptHindsight[m]: https://www.automation4less.com/store/hiwinlegacy.asp
03:25 PM roycroft: awesome
03:25 PM CaptHindsight[m]: https://www.automation-overstock.com/index.asp
03:26 PM roycroft: one of the credit unions i use have a "save to win" cd account, which is a high yield cd that i can contribute to monthly
03:26 PM roycroft: that alone is great, but they also give away money to random people
03:26 PM roycroft: two years ago i won $100
03:26 PM roycroft: last year i won $100
03:26 PM roycroft: i just got a call to tell me i just won $500
03:26 PM XXCoder: lol congats
03:26 PM roycroft: what are the odds that i would win three years in a row?
03:27 PM roycroft: yeah
03:27 PM roycroft: maybe i can afford some good filament now :)
03:28 PM roycroft: that sunlu stuff does work, btw, but i have to babysit it
03:28 PM roycroft: it gets tangled on the spool and then breaks
03:28 PM roycroft: so every five minutes or so i have to unspool a bunch of it
03:28 PM XXCoder: yeah sunlu was one I advised against
03:29 PM roycroft: when i first got it i was dubious, as the coils did not look very well organized
03:29 PM roycroft: yes, you did, but only after i had ordered it
03:29 PM roycroft: oh well, live and learn
03:29 PM roycroft: that $500 will get me a fair amount of atomic filament
03:29 PM XXCoder: yet I bought a few recently since its required to get 2 filament dryers very cheap
03:29 PM XXCoder: I paid $70 for 2 filament dryers and 7 pla rolls
03:29 PM roycroft: although the $500 is really more like $350 after taxes
03:30 PM XXCoder: even if I throw away all 7 rolls, its still half price for dryer.
03:30 PM roycroft: but still, i can get 10kg of atomic filament for that and have money left over for a decent dinner
03:30 PM XXCoder: the deal expired very, very fast. I barely snagged it
03:30 PM XXCoder: indeed
03:30 PM roycroft: the 3.5kg spools from atomic are looking nice
03:31 PM roycroft: they're only a tiny bit cheaper than getting 1kg at a time
03:31 PM roycroft: but i won't have as much end of spool waste
03:31 PM roycroft: or have to change spools as often
03:31 PM XXCoder: im kinda meh on that. not enough bulk discount
03:31 PM roycroft: yeah, i wish they had better discounts
03:31 PM roycroft: but honestly, the other brands i'm buying are ending up costing eve more, not even counting the time i'm wasting
03:32 PM roycroft: so many prints that i have to bin
03:32 PM XXCoder: most of their customers is businesses and very large bulks. hardly worth it to try attact smaller customers I guess?
03:32 PM roycroft: this spool i'm printing now cost me 1/3 less than atomic, but i ended up wasting about 80% of it
03:32 PM roycroft: that makes it very expensive
03:32 PM XXCoder: yeah
03:32 PM roycroft: with atomic my waste is just a wee bit at the end of the spool
03:33 PM roycroft: never a spoiled print due to the filament
03:33 PM roycroft: oh
03:33 PM roycroft: and another update from gst3d
03:33 PM roycroft: after my last response to them, asking them what they want pictures of
03:33 PM roycroft: "please send picture"
03:33 PM XXCoder: did you ask what picture they expens ah so you did lol
03:33 PM roycroft: the exact same response as the other two times
03:34 PM XXCoder: maybe just send picture of roll, with broken filament or something
03:34 PM roycroft: that's what i'm thinking
03:34 PM XXCoder: not nearly emopty one
03:34 PM roycroft: sometimes it breaks below the extruder
03:34 PM roycroft: if i can get it to do that then i can take a picture of the filament sticking out the bottom of the extruder, broken off
03:34 PM roycroft: i'm not going to stage it though
03:35 PM roycroft: and i'd rather not waste a bunch of time trying to print until it breaks in a way that's easy to document
03:35 PM roycroft: i keep asking them to check the qa logs for the batch they sent me, and i send them the batch number every time
03:35 PM roycroft: the have not acknowledged any of that
03:36 PM roycroft: so i'm assuming that "qa" means "we count how many spools were made during the shift"
03:36 PM XXCoder: if you have a roll thats extremely brittle just bend a little break few peices off, then show picture of roll with broken segements
03:36 PM XXCoder: seems so
03:36 PM roycroft: i can feel how uneven the filament is with my fingers
03:36 PM roycroft: and i can feel on part of it on the spool i checked a while ago that there's a "ridge" along the filament
03:37 PM roycroft: it's not round at all
03:37 PM roycroft: but i can't photograph that
03:37 PM roycroft: what i'd like to do is cut off a short piece and send it to them
03:37 PM roycroft: i asked about that and got "please send picture"
03:37 PM roycroft: that seems a good way for them to evaluate - it would cost less than a dollar to send them a sample
03:38 PM XXCoder: hmm what about pictures of micrometer measuring it and showing sizes beyond +- .04mm
03:38 PM roycroft: i asked them if they want that
03:38 PM XXCoder: or did they say +- .02mm? forgot
03:38 PM roycroft: "please send picture"
03:38 PM roycroft: it's like they don't read what i ask them at all
03:38 PM roycroft: which they probably don't, because they probably don't understand english
03:39 PM XXCoder: possibly
03:41 PM XXCoder: other company does that thing too. wanted pictures of package not arriving
03:42 PM XXCoder: guy just sent them bunch of pictures of hands, holding nothing
03:44 PM Kerr[m]: CaptainHindsight about $1k for everything.
03:45 PM Kerr[m]: But i have some 525 ozin steppers, and the aforementioned motherboard (&parts already)
03:45 PM Kerr[m]: And a set of 48v power supplies for the steppers
03:45 PM Kerr[m]: So that saves a bit
03:47 PM XXCoder: spindle and metal for frame is biggest expense
03:48 PM ZincBoy[CAON][m]: I don't think you are going to get a machine that will be good for milling aluminum for a $1k budget. Linear rails and ballscrews will pretty much consume that.
03:48 PM XXCoder: yeah.
03:49 PM XXCoder: wood, you can use double belt system and its vastly cheaper
03:49 PM XXCoder: spindle can be lot cheaper too
03:50 PM ZincBoy[CAON][m]: The bargain basement rails and ballscrews are not suitable for an aluminum focused machine. You need to step up to at least a decent import or Hiwin screw if you don't want to be chasing backlash and poor motion.
03:51 PM Kerr[m]: Well we were talking about making our own ways
03:51 PM roycroft: did we ever get the maximum part size for this project?
03:51 PM Kerr[m]: Do I need to do ballscrews, or would ground acme threaded rod work
03:52 PM Kerr[m]: 12x15 inch, ish. Doesnt have to be that big
03:52 PM ZincBoy[CAON][m]: You need ballscrews. ACME is for manual machines.
03:52 PM Kerr[m]: Work area
03:52 PM XXCoder: certainly possible with steel plates. like that old 3'x5' project I saw before cheap MGNs existed
03:52 PM roycroft: and the z height?
03:52 PM ZincBoy[CAON][m]: 12x15 inch work area is large for a hobby class machine.
03:53 PM roycroft: well 12x15 inch means it doesn't need to be a router
03:53 PM roycroft: it can be something that resembles a milling machine
03:53 PM roycroft: that's actually smaller than i guessed it would be
03:54 PM roycroft: with the 24krpm spindle and gantry design, i was envisioning a 24x48" router that could also gnawn on some metals
03:54 PM roycroft: gnaw
03:56 PM * xxcoder[m] uploaded an image: (1741KiB) < https://libera.ems.host/_matrix/media/r0/download/jauriarts.org/KsRzEMJJzTBpLuGvdMCNAKnU/unknown.png >
03:56 PM xxcoder[m]: example of rail using milled steel plate, block using bearings
03:57 PM xxcoder[m]: lemme find one of y axis
03:57 PM * xxcoder[m] uploaded an image: (1705KiB) < https://libera.ems.host/_matrix/media/r0/download/jauriarts.org/HndxjzCasyAhaGqAYzsxMAXp/unknown.png >
03:57 PM xxcoder[m]: its not best example, but something that might can be done cheaply if you have access to bridgeport
04:01 PM XXCoder: those extrusions would be too expensive for your budget unless you have access to some scrap cheap
04:01 PM ZincBoy[CAON][m]: I think these are false economy though. You have the cold roll steel for the rails and bearings to account for in the cost. Then you end up with something that is in every way worse than the cheapest linear rail for pretty much the same price.
04:01 PM XXCoder: yeah really deoends on access if nateruaks
04:01 PM XXCoder: materials
04:03 PM XXCoder: for example if can buy plates very cheaply at scrap metal for example, or has in stock
04:03 PM JT-Shop: roycroft, will titebond III adhere to a surface that's coated with T&T Varnish oil?
04:11 PM Kerr[m]: Where would I shop for guideways & ball screws
04:12 PM Kerr[m]: Also I guess I don't really understand what the issue with acme rod would be.
04:12 PM XXCoder: backlash usually
04:13 PM XXCoder: manual its not a problem since you can cut in only direction or adjust for it
04:13 PM XXCoder: cnc well its bit different
04:13 PM XXCoder: there is fairly cheap MGNs and ballscrews nowdays, c7 is actually affordable rated ballscrew too
04:14 PM ZincBoy[CAON][m]: The links that CaptHindsight posted are a good start.
04:14 PM XXCoder: MGNs well cheap ones you will have to do a rebuild right away
04:14 PM XXCoder: swap balls might be the case, though thankfully not expensive. since youre machinist you have mic to measure balls and swap bad ones
04:15 PM ZincBoy[CAON][m]: You want a P5 preloaded or better screw. C7 is transport grade and is not really great for a CNC.
04:15 PM XXCoder: agreed, but theres budget concerns
04:15 PM ZincBoy[CAON][m]: Again, this is for an aluminum CNC. C7 is fine for wood.
04:16 PM XXCoder: ok
04:16 PM JT-Shop: all thread and roller skates are fine if they meet your tolerance and expectations
04:16 PM ZincBoy[CAON][m]: If you can't afford to do it at least fairly well, you will end up with a machine that doesn't work and be out what ever you spent.
04:17 PM XXCoder: indeed. though not sure what guy needs are at also
04:18 PM ZincBoy[CAON][m]: Agreed. If you can live with 0.02" tolerance then that frees things up. If you are expecting 0.005" then what I am suggesting holds. If you want better than 0.005" then cost goes way up and you are better off with a used VMC.
04:19 PM XXCoder: MGNs, cheapish ones is still rebuildable, while I guess youre rigjt on ballscrews. less ajstable
04:20 PM JT-Shop: what are MGNs?
04:20 PM Kerr[m]: Those linear guide ways he posted are very affordable.
04:20 PM XXCoder: on phone be back in bit
04:20 PM Kerr[m]: A few dollars perinch
04:20 PM ZincBoy[CAON][m]: Yes, the ballscrews will limit you if not good to start. Linear rails are a little more tolerant.
04:22 PM ZincBoy[CAON][m]: Yes and those are generally well regarded. The NOS Hiwin are really good but a bit more pricy.
04:28 PM XXCoder: JT-Shop: its rail type
04:28 PM JT-Shop: oh it's a linear rail?
04:29 PM JT-Shop: I thought you were talking ball screws
04:29 PM * JT-Shop goes to make some chicken crack
04:29 PM roycroft: i just ordered a 3.5kg spool of atomic filament
04:29 PM roycroft: i'm feeling more relaxed already
04:34 PM Kerr[m]: So it looks like around $80/linear guide way + block, so $160/axis or $480. Then $100 per ballscrew if I did that...
04:35 PM Kerr[m]: Which year gets close to the budget on just those. The vertical axis is really short though, so if I were to make the ways for that, that would save a bit
04:35 PM XXCoder: machine that small you can use just one ballscrew for y axis, at center
04:39 PM Kerr[m]: https://www.automation-overstock.com/proddetail.asp?prod=LY150340ELC2B02P51
04:40 PM XXCoder: https://www.aliexpress.com/item/2251832620307321.html example of very cheap MGNs
04:40 PM XXCoder: JT-Shop: yeah, though was talking about both rail and screw
04:40 PM XXCoder: now, those aliexpress ones is very cheap. you'd want to replace many ball bearings in it, and do a rebuild
04:41 PM JT-Shop: ah ok
04:41 PM XXCoder: Kerr[m]: https://youtu.be/c53sa46C1qQ this video is HIGHLY recommanded
04:42 PM * JT-Shop just got his AMD Radeon RX6600... gotta pick an ATX case now
04:42 PM XXCoder: he reworks cheap aliexpress MGNs
04:42 PM XXCoder: jt nice. good that gpu market crashed recently.
04:42 PM XXCoder: wish rpi market would crash soon
04:44 PM JT-Shop: holy crap this video card is HUGE
04:45 PM roycroft: the high end graphics cards tend to be full length and need two slots and lots of power
04:46 PM JT-Shop: this one takes up the space of two slots but only has one connector
04:47 PM XXCoder: need lot of cooling space
04:47 PM roycroft: yeah, that's typical
04:56 PM * JT-Shop pulls out a gaming case with a handle on top storm trooper or something like that... lots of room and lots of fans
05:00 PM JT-Shop: going to be hard to stay focused and remember to mow the grass before dinner
05:03 PM roycroft: are you about, xxcoder?
05:04 PM XXCoder: phone still
05:04 PM roycroft: i just opened another spool of that crappy gst3d filament
05:04 PM roycroft: i think i can demonstrate the problem easily
05:05 PM XXCoder: thats good and bad
05:05 PM roycroft: what would you say if you got this? :
05:05 PM XXCoder: man wish I have spoken language, using asl to talk is making me seriously dizzy and sick
05:05 PM roycroft: roycroft.us/gst3d.jpeg
05:06 PM XXCoder: holy crap
05:06 PM roycroft: is there any wonder why it breaks and jams?
05:06 PM JT-Shop: wow
05:07 PM roycroft: and you all thought i was crazy for all the whining i've been doing :)
05:07 PM roycroft: i may well be crazy
05:07 PM roycroft: but my issue with that filament is real
05:11 PM roycroft: i just sent them a picture
05:11 PM roycroft: maybe that will get a reaction
05:20 PM XXCoder: damn phone calls finally done
05:20 PM XXCoder: feeling pretty sick
05:20 PM XXCoder: constant vision shaking for over a hour
05:21 PM XXCoder: had to be done though
05:21 PM roycroft: i'm sorry
05:21 PM roycroft: sounds like it was rather stressful
05:21 PM XXCoder: nah its just what im dealt with :)
05:21 PM XXCoder: anyway
05:21 PM XXCoder: your roll looks way way worse than what I got from them
05:21 PM XXCoder: mines just bit messy but not tangled to heck
05:21 PM roycroft: it's way worse than the first two spools i got
05:22 PM roycroft: i almost wonder if they pick the best stuff to send out as single orders
05:22 PM roycroft: and save all the crap for the bulk orders
05:22 PM XXCoder: that is possible, though we never ordered small orders. only 2x 20 filament when it was 8.99 usd per roll for bulk
05:22 PM roycroft: the tangles are a problem
05:22 PM roycroft: but the bends/kinks are the big problem
05:23 PM roycroft: when those bends hit the extruder roller it straightens them out instantly, causing a stress fracture
05:23 PM roycroft: and then it all goes downhill
05:24 PM XXCoder: you might try dry em and hope it retains some of flexability
05:24 PM XXCoder: doubt its soaked but maybe extract small amount would help
05:25 PM roycroft: that spool is fresh out of the vacuum bag
05:25 PM XXCoder: can I share your image elsewhere to warn?
05:25 PM roycroft: but the others i did dry after they started failing
05:25 PM roycroft: it did not help
05:25 PM roycroft: not yet
05:25 PM roycroft: i want to give gst3d a chance to respond first
05:26 PM XXCoder: true ok
05:26 PM XXCoder: is bags transparent like mine is? if so maybe give rest of boxes a look
05:26 PM XXCoder: no need to open any
05:26 PM roycroft: as disappointed as i am in their product, and espcially in their customer service so far, i want this to play out to the end before i form a final opinion of them
05:26 PM XXCoder: (any bags that is)
05:26 PM roycroft: yes, they are transparent bags
05:26 PM roycroft: and yes, i'll look at the rest when i get around to it
05:27 PM roycroft: even if they just send me a return shipping label, i'll inspect the others before i pack them up and ship them back
05:27 PM XXCoder: ok
05:27 PM XXCoder: hopefully you will recover money
05:27 PM roycroft: i already have initiated chargebacks with my cc company
05:28 PM roycroft: if it's not resolved by 16 august their conditional credits will be final
05:28 PM roycroft: and btw, gst3d did cancel my second order on 7 july, and informed me that a refund had been issued
05:28 PM roycroft: that refund has not shown up on my cc account a week later
05:29 PM roycroft: but i also have a chargeback on that
05:29 PM roycroft: so i know i'll get my money on that one
05:29 PM XXCoder: thats good
05:29 PM roycroft: it usually takes 3-4 days at most for a cc refund to post
05:29 PM roycroft: not a week +
05:30 PM roycroft: i'll leave that picture on my website, and after the issue is resolved, i'll let you kwow if it's ok to share it or not
05:30 PM XXCoder: no problem
05:30 PM roycroft: i can't really complain about the sunlu coils after seeing that :)
05:31 PM XXCoder: indeed lol
05:32 PM roycroft: ivana t**** has died
05:32 PM roycroft: ivanka's mom
05:33 PM XXCoder: dont really feel anything related to that
05:34 PM XXCoder: besides the usual for death of fellow human anyway
05:34 PM roycroft: it's news
05:34 PM roycroft: it doesn't affect my life
05:34 PM Kerr[m]: https://m.aliexpress.com/item/2255801116473795.html?spm=a2g0n.productlist.0.0.54961e13imqkHJ&browser_id=6cfb28a00e354fa094ae7cda003f4bbb&aff_trace_key=&aff_platform=msite&m_page_id=z9isgd5maikcawt1181fed7b1ed10cd575f9104615&gclid=&pdp_npi=2%40dis%21USD%21%2159.86%21%21%21%21%21%40210318b916578379593161115e67a9%2112000027406486319%21sea&algo_pvid=3617718f-4438-4d04-840c-8306176014e9
05:34 PM Kerr[m]: https://m.aliexpress.com/item/2255801116473795.html
05:35 PM roycroft: thank you for fixing that, kerr :)
05:35 PM XXCoder: c5. better but still not P rated I guess
05:35 PM XXCoder: and only half price? meh not sure if savings is worth qualty loss
05:36 PM XXCoder: ZincBoy[CAON][m]: what ya think
05:37 PM Kerr[m]: If its too expensive, we may downgrade to "engraving pcbs"
05:38 PM XXCoder: id go for wood
05:38 PM XXCoder: whatever can do wood also can engrave
05:38 PM TheMightyFozz: hi folks I installed LinuxCNC on top of Debian11 and Gmoccapy seems to bug out. but Axis works fine. is this fixable or just use Axis?
05:39 PM TheMightyFozz: i did use Mesa software to configure
05:40 PM roycroft: can you elaborate on what "seems to bug out" means in this context?
05:40 PM ZincBoy[CAON][m]: I am not a fan of the cheap aliexpress ballscrews. Really not worth the low price. They say C5, but that is really had to do in a rolled screw. Not likely at that price. Plus no mention of preload/backlash.
05:40 PM TheMightyFozz: err messages when using gmo, none when using axis
05:41 PM roycroft: the error messages probably provide you with clues as to what software/drivers are missing or are the wrong version for gmoccappy
05:42 PM ZincBoy[CAON][m]: Engraving PCBs is achievable on the budget you have set. Most of the cheap cnc machines on amazon will do that, just make sure you get one with a minimum 24k rpm spindle. I used to use a PCB mill years ago and it had a 50k rpm spindle and that was too slow.
05:42 PM roycroft: you might google those error messages or search for them on the linuxcnc forum site for some clues
05:44 PM Kerr[m]: That's a might
05:45 PM TheMightyFozz: i wish i could right click copy the error messages.
05:45 PM Kerr[m]: We'd still prefer to do aluminium milling
05:45 PM XXCoder: is text selectable? if so, just ctrl-c to copy
05:46 PM XXCoder: Kerr[m]: hard part is spindle. get good ballscrews from what zinc suggested, cheap MGNs that you rebuild, and framing isnt too hard
05:46 PM XXCoder: spindle though
05:47 PM JT-Shop: I think that is the new example of extreme vague statements
05:49 PM ZincBoy[CAON][m]: The challenge is your budget. I think it would be do-able for $3k with reasonable compromises. The other thing to factor in is the cost of tooling/workholding/CAM. General rule of thumb is to budget the cost of the machine for tooling.
05:50 PM TheMightyFozz: ok thanks, that was easy. i was pressing ctrl+shift+c . thanks (it's late here. LCNC is throwing up the usual message when the cards not connected not the one i need and it will wake up my partner if i did.)
05:51 PM roycroft: zincboy: if you don't mind, i'd like to throw my budget figure out and my overall design specs for a reality check
05:52 PM roycroft: i'm going to be building a roughly 2'x4' router, for wood, using a gantry design
05:52 PM roycroft: i'm thinking $5k should be a reasonble budget for that
05:52 PM roycroft: not including the computer
05:53 PM roycroft: but including the control systems with which the computer interfaces
05:54 PM roycroft: i want 6" of z, btw
05:54 PM roycroft: and i already have some of the components, but i'm counting what i have in the overall budget
05:54 PM ZincBoy[CAON][m]: $5k should be pretty good depending on how crazy you go with motion components. I would go with a steel frame if you have fabrication skills.
05:54 PM roycroft: i won't be doing that
05:54 PM roycroft: there are other criteria
05:55 PM roycroft: this will actually be what i call an "over/under" router
05:55 PM roycroft: i have an existing router table
05:55 PM roycroft: the conventional type with the spindle mounted underneath
05:55 PM roycroft: i don't have room to add a cnc router in my shop, so my plan is to build a low profile router that straddles the router table
05:55 PM roycroft: so i can use it with either the overhead or underside spindle
05:56 PM roycroft: the router table part will be made of sheet goods
05:56 PM roycroft: and i plan on using extrusions like 80/20 or similar for the y axis rails and the x axis rail
05:57 PM roycroft: the gantry sides will be custom fabricated
05:57 PM ZincBoy[CAON][m]: Cool idea. I like the dual use. I built a Joes2006 2'x4' router years ago out of MDF using a table saw and a drill press. Worked great.
05:57 PM roycroft: as well the z axis assembly
05:57 PM roycroft: i just don't have any room in my shop
05:57 PM roycroft: so i need to double up
05:57 PM roycroft: i don't plan on making parts with the cnc router
05:57 PM roycroft: i plan on making jigs/fixtures with it that i can use to make parts
05:58 PM ZincBoy[CAON][m]: That is why I sold my wood router. No room with the VMC.
05:59 PM roycroft: so these renders are months old, and quite preliminary
05:59 PM roycroft: i've made a lot of changes since then already, and am about to make a lot more
06:00 PM roycroft: https://roycroft.us/CNC-Router/Full-Assembly-Preliminary-3-Front-View.jpeg
06:00 PM ZincBoy[CAON][m]: I would go with a torsion box table with 80/20 to mount the rails and the rack. Go with a rack and pinion drive as that keeps the motion components compact. Mount the stepper/servo vertically to make it smaller.
06:00 PM roycroft: https://roycroft.us/CNC-Router/Full-Assembly-Preliminary-3.jpeg
06:01 PM roycroft: 0
06:01 PM roycroft: the frame that holds the y axis extrusions will be a torsion box
06:01 PM roycroft: the router table will be mounted in the front
06:03 PM roycroft: i'm just finishing up another project that consumed a lot of time, and am about to get back to the router design
06:03 PM ZincBoy[CAON][m]: I would mount the y axis rails on the side of the extrusion. Allows it to be more compact. The x/z seems to be a bit wider than needed. I would make the carriages square to free up travel or make the axis shorter.
06:03 PM roycroft: i've already redrawn the gantry sides to make them more compact
06:04 PM roycroft: i'm still wavering on what to do with the y rails
06:05 PM roycroft: i've been working on a design that adds a second set of rails to the bottom of the extrusion for more rigidity
06:05 PM roycroft: that would add a bit of cost, but might make it work better
06:05 PM roycroft: i was planning on using ball screws to move the gantry
06:06 PM ZincBoy[CAON][m]: You also want to add thickness to the gantry side plates. Looks like you have two plates in the design, space them out to increase stiffness. This is a weak point in many designs but if you look at commercial systems they almost always have a box section for the vertical.
06:06 PM roycroft: i'd either mount them on the outsides of the extrusions, or if i put the rails there i'd mount them on the bottom of the extrusions
06:06 PM roycroft: i think i already have the gantry sides redone as a torsion box
06:06 PM roycroft: as torsion boxes, rather
06:07 PM roycroft: these drawings date back to january of this year
06:07 PM roycroft: i did a bunch of work in april and may
06:07 PM roycroft: iirc the gantry sides are now 2 3/8" or 1/2" thick plates with 1/2" spacers in between them
06:07 PM ZincBoy[CAON][m]: Rack and pinion is much better for a wood router. It gets you all the accuracy needed for wood and also the travel speed without whipping issues found in ballscrew systems.
06:08 PM roycroft: i would be fine with rack and pinion for the y
06:08 PM roycroft: i think i was having trouble finding what i wanted for that
06:08 PM roycroft: i don't recall exactly what the issue was, but i'm happy to look into it again
06:09 PM roycroft: i am concerned about ~2m long ball screws whipping on rapids
06:09 PM ZincBoy[CAON][m]: This is unnecessary and just over constrains the system. You only need one rail per side for the y.
06:09 PM roycroft: for the x i think a ball screw would be fine, as well as for the z
06:10 PM roycroft: i'll consider all you've suggested
06:10 PM roycroft: but as a basic sanity check, the direction i'm going and my budget are ok?
06:10 PM ZincBoy[CAON][m]: Z use a ball screw for sure. X, can use a ballscrew but you will have issues with not being able to hit the same feed rates in X and Y. Wood routers really like high feeds.
06:10 PM roycroft: sure
06:11 PM roycroft: but keep in mind what i said earlier
06:11 PM roycroft: i won't be making parts with this
06:11 PM roycroft: i'll be making jigs/fixtures to make parts
06:11 PM roycroft: so maximum performance is not really necessary
06:11 PM roycroft: but having said that
06:11 PM ZincBoy[CAON][m]: Looks good to me. To be honest, I think your design would work and I am just making suggestions to improve things.
06:11 PM roycroft: as soon as it's built there's a good chance i'll want to make parts with it :)
06:12 PM roycroft: because that's how the world works
06:12 PM ZincBoy[CAON][m]: I think this is 100% 🙂
06:12 PM roycroft: rigidity issues are my primary concern right now
06:12 PM roycroft: which is why i'm doing the torsion box gantry sides
06:13 PM roycroft: and why i'm considering a second set of rails for the y
06:13 PM ZincBoy[CAON][m]: Yes, this is great. Puts you ahead of many designs.
06:13 PM roycroft: but if the gantry is really rigid, those second rails probably won't help much, if any
06:13 PM ZincBoy[CAON][m]: Will not help at all.
06:13 PM roycroft: well my goal is not to build cnc routers for a living
06:13 PM roycroft: my goal is to build one and use it for many years
06:13 PM roycroft: without constant mods
06:14 PM roycroft: so i'm taking my time with the design
06:14 PM ZincBoy[CAON][m]: Even if the gantry is a wet noodle, a second rail on y still won't help.
06:14 PM roycroft: the x rail is ginormous
06:14 PM roycroft: and the sides attach to it with 8 big bolts each
06:15 PM roycroft: i think that if i build rigid torsion box sides and bolt them tightly to the x rail, it's going to be pretty rigid
06:16 PM ZincBoy[CAON][m]: If you have fusion360, the FEA tool is quite helpful in optimizing designs. The basic deflection/stiffness simo runs local and doesn't need cloud credits.
06:16 PM roycroft: https://roycroft.us/CNC-Router/Left-Gantry-Side-Exploded.jpeg
06:16 PM roycroft: there's an early torsion box gantry side design
06:16 PM roycroft: i'm using sw 2018 pro
06:17 PM ZincBoy[CAON][m]: Ah then you should have FEA built in I think.
06:17 PM roycroft: yes
06:18 PM roycroft: i have the tools i need
06:18 PM roycroft: and much of the knowledge
06:18 PM ZincBoy[CAON][m]: Also, your y carriages are spaced a bit further apart than needed. You can shrink that down without sacrificing anything.
06:18 PM roycroft: i come here and other places for the rest of the knowledge :)
06:19 PM roycroft: i left room for ball screws
06:19 PM roycroft: but yeah, i can compact that
06:19 PM roycroft: especially if i go with rack and pinion
06:19 PM ZincBoy[CAON][m]: No, I meant along the y axis. the carriages can be closer together.
06:20 PM roycroft: i'm wondering if the reason i decided not to do rack and pinion had to do with how i would have to mount the steppers
06:21 PM roycroft: oh, i think i've tightened that up a bit already
06:21 PM ZincBoy[CAON][m]: Mount the rack on the side of the extrusion and the steppers facing up below the rail.
06:21 PM roycroft: i really need to get back to the project - i just checked it out from my gitlab server the other day
06:21 PM ZincBoy[CAON][m]: Also helps with chip clearance.
06:21 PM roycroft: i should generate some renderings of what i was doing in april/may
06:21 PM roycroft: and work from there
06:23 PM roycroft: if i'm really clever with this, i can use the overhead cnc part as a router fence for when i'm using the underside router table
06:23 PM roycroft: i could mount a fence across the front of the gantry and use the pendant to move it in and out in relation to the router table spindle
06:27 PM roycroft: so i cut off the gnarly mess at the beginning of that spool of filament, tossed the rest on the printer and tried to do a print
06:27 PM roycroft: the filament broke within the first couple minutes
06:27 PM roycroft: i just put a spool of atomic filament on the printer, tried the same print
06:27 PM roycroft: and it's going brilliantly
06:28 PM roycroft: this is my last roll of atomic that is the right color for my project, so hopefully the new stuff will ship soon
06:45 PM XXCoder: probably be quick
07:07 PM Centurion_Dan1 is now known as Centurion_Dan
08:21 PM roycroft: they usually ship next day or so
08:25 PM roycroft: and the first print with the new spool of atomic was completely uneventful and completed unattended
08:25 PM roycroft: just as it should be
08:25 PM XXCoder: indeed
08:28 PM roycroft: xxcoder: when this is all sorted, i'm still not going to ask you not to use that picture
08:28 PM roycroft: i just want you to have the whole story before you share the experience
08:29 PM roycroft: they sent me defective product and their customer service is horrible - that has been established and will not change
08:29 PM XXCoder: no problem
08:29 PM roycroft: but you should know if they issue a refund or i have to have my cc do a chargeback, as that might affect some folks' decision on whether to try it or not
08:30 PM roycroft: so far they have shown no signs of stepping up and taking ownership of the problem
08:30 PM XXCoder: yeah. man your roll is so bad, mine is just bit wavey but works fine
08:30 PM XXCoder: thats why I said it was not amazing quality
08:57 PM * roycroft is getting tired of the nag messages from nagios
08:57 PM roycroft: there's a power outage in a city that i monitor, and a dozen fiber nodes are down
08:57 PM roycroft: oh well, at least the clock is running (at after hours rates) while i'm fixing dinner
08:57 PM roycroft: i could use some extra overtime this month
08:58 PM roycroft: especially since i'm back to buying premium filament :)
09:10 PM XXCoder: yea
09:11 PM XXCoder: there should b many good $20/lb filament also
09:11 PM XXCoder: pla isnt hard. which makes gst3d and sunlu sad
09:13 PM XXCoder: never had problems with hatchbox for examople https://www.amazon.com/s?k=Hatchbox+PLA+Filament&tag=all3dp0c-20
09:13 PM XXCoder: reviews https://all3dp.com/1/best-pla-filament/
09:14 PM XXCoder: we have tried overture its good. matte ones is awesome
09:15 PM roycroft: sunlu is $20 filament
09:15 PM roycroft: i wasn't happy with the overture i bought
09:15 PM roycroft: but that was when i first got the printer, so it may have been more me than the filament
09:16 PM XXCoder: i had such problems printing with atomic, but it was me not filament
09:16 PM XXCoder: it was in 2016 with elcrapo
09:17 PM XXCoder: wow one company sells filament the color you want
09:17 PM XXCoder: bet its expensive lol (colorfabb)
09:18 PM XXCoder: 57 euro which is essentally 57 usd now ouch
09:24 PM roycroft: all i know is that i had the atomic filament page open today
09:24 PM roycroft: i was very close to just biting the bullet and buying the $30/kg stuff that i know works
09:25 PM roycroft: and then i got the phone call saying i had won $500
09:25 PM roycroft: i ordered the atomic stuff right after that call
09:25 PM roycroft: and i do like the idea of the 3.5kg jumbo spools
09:26 PM roycroft: which is what i ordered
09:33 PM XXCoder: yeah
09:36 PM XXCoder: havent found a way to report abusive amazon listing. I give up
09:56 PM mferraro89[m]: https://mattferraro.dev/posts/cnc-router
09:57 PM mferraro89[m]: roycroft, this might be helpful? IDK, it might be too basic for you
09:58 PM MikeM[m]: looks like a nice build
09:59 PM MikeM[m]: I designed and built my machine, had a lot of fun doing it too
10:09 PM roycroft: i'll read through that
10:09 PM roycroft: but i need a gantry machine
10:10 PM roycroft: i can't have the table move, as mine will be a combination cnc router/router table, with a spindle above and a spindle underneath the base
10:10 PM roycroft: but that's not to say i can't get any good ideas from that build blog
10:10 PM mferraro89[m]: yeah the post isn't about any one router design, it's just got a lot of information about general concepts and pros/cons of different designs
10:11 PM mferraro89[m]: there's stuff in there about work holding, spoilboards, dust collection, types of spindles, etc
10:12 PM XXCoder: just finished reading. not bad
10:14 PM mferraro89[m]: thanks! it really is just what I'd wish I knew before I designed and built my first one
10:15 PM mferraro89[m]: s/knew/known/
10:19 PM MikeM[m]: I have been considering buying one of the cheapo CNCs you can get on amazon and seeing if I can make it into something more robust.
10:19 PM roycroft: if you're not familiar with them, you might have a look at composite nailers to use a a hold down device
10:20 PM roycroft: they use plastic nails, so you can just nail your work down onto the spoil board and have at it without fear of damaging cutters
10:20 PM MikeM[m]: I have been using double sided tape mainly for work piece holdig
10:21 PM ZincBoy[CAON][m]: On error in the document: Ballscrews are backdriveable. They should not be used to support a Z axis 🙂
10:21 PM roycroft: https://www.amazon.com/dp/B08SMRGLPK
10:25 PM roycroft: i was thinking of it as something you might add to your document, not necessarily that you would use it yourself
10:26 PM XXCoder: roy interesting. though quite pricy ow
10:26 PM roycroft: they are
10:26 PM roycroft: but the time that can be saved can justify that cost pretty easily
10:27 PM roycroft: it would take seconds to attach the parts to the spoilboard, and seconds to remove them
10:27 PM roycroft: with no cleanup required
10:28 PM ZincBoy[CAON][m]: Excellent document! I have bookmarked it to send to people when they ask about building a CNC machine. Just the one minor error that I found but you hit everything else on the head.
10:28 PM roycroft: it's been on one of my wish lists for a long time
10:29 PM roycroft: probably as soon as i get my router built i'll get it
10:29 PM roycroft: or maybe while i'm doing the build
10:30 PM ZincBoy[CAON][m]: There is not much that is useable by the time you fix all the bad parts. Pretty much everything about them is either poorly designed, poor quality, or both. Maybe you could keep the steppers and controller?
10:31 PM mferraro89[m]: thank you! what is the error? I would be happy to make a fix to the text
10:31 PM MikeM[m]: yeah, that is what my thought would be, and I would replace the controller with a MESA one... so the steppers? lol
10:32 PM ZincBoy[CAON][m]: Ballscrews can be backdriven and should not be used to hold up the z axis.
10:33 PM MikeM[m]: My G/F wants a small/portable machine. Probably just as well to build one right from scratch.
10:33 PM XXCoder: yeah cheaper than buy cheapie then build whole thing
10:33 PM XXCoder: MikeM[m]: what materials and workspace?
10:34 PM ZincBoy[CAON][m]: Depending on your budget there are some small machines that are quite capable. Just not the $200 amazon CNC 🙂
10:35 PM roycroft: in your discussion of stepper motors you mention the disadvantage that that you can lose steps, which is correct
10:36 PM roycroft: but some folks add encoders with their stepper motors to track and adjust position
10:36 PM roycroft: as well, there are stepper motors with built-in encoders
10:36 PM roycroft: so while nothing there is wrong, there could be further discussion of steppers with encoders
10:36 PM XXCoder: yeah. there was that kickstarter for cheap stepper encoder board
10:36 PM ZincBoy[CAON][m]: Closed loop steppers are a midground between stepper and servo systems.
10:36 PM XXCoder: it uses hall effect and magnet on stepper rear (that indention)
10:37 PM roycroft: actually you say 'the only solution is to over spec your motors"
10:37 PM roycroft: so that might be a little bit wrong
10:37 PM ZincBoy[CAON][m]: They typically cost close to the same as a low end servo though. Better matched to a typical hobby system though.
10:38 PM roycroft: that's kind of nit-picky, but i think that's what you want for feedback, no?
10:40 PM roycroft: as far as using ball screws on the z axis, people do that all the time with stepper motors
10:40 PM MikeM[m]: I did with my build, no issues so far.
10:40 PM roycroft: i think the motor would help it hold position when unpowered
10:41 PM roycroft: but the axis would definitely need to be homed after a power cycle
10:42 PM roycroft: i have not started my build yet, but i do have a stepper motor and a ball screw for my z axis
10:42 PM roycroft: the ball screw wasn't that expensive, and i'm willing to be convinced it's a bad idea to use it
10:42 PM XXCoder: my machine is small enough that z ballscrew dont fall down unpowered
10:42 PM ZincBoy[CAON][m]: You are relying on the holding torque of the motor. Not the best idea. If you have a really light z axis it can work but ballscrew based systems should have a brake on the z.
10:43 PM roycroft: iirc my z axis is about 12kg
10:43 PM XXCoder: zinc whats interesting is old fadal z falls down when lose power. less interesting: all 8 inserts broken and part scrapped
10:43 PM roycroft: including spindle
10:43 PM XXCoder: setup rotated too so re-setup needed
10:44 PM ZincBoy[CAON][m]: lol. My z weighs somewhere around 500kg. It does the same thing if the brake were to fail 🙂
10:44 PM mferraro89[m]: thank you both for the feedback! I'll patch the text 🙂
10:46 PM roycroft: a brake that engages by default and disengages with something like a solenoid when the machine is powered up would be interesting
10:46 PM XXCoder: zinc a61 is bit weird. if it loses power, it will go down on Y axis (horzional axis) but only very short distance. doesnt help if it was milling surface on downwards direction
10:46 PM XXCoder: had it use big tapping tool when power loss happened. I couldnt do anything and just left it for day shift as I cnat fix that
10:53 PM roycroft: my design to date also has a timing belt to drive the z axis ball screw, with the motor pulley about 1/2 the size of the ball screw pulley, so that doubles the holding force of the motor
10:54 PM roycroft: https://roycroft.us/CNC-Router/Z-Axis-Motor.jpeg
10:55 PM XXCoder: I remember that yeah
10:55 PM XXCoder: i'm sorta tempted to do that also lol to increase torque at lower rpm speed
10:55 PM XXCoder: but then i want to mostly work on wood and 27k rpm is fine for that
10:55 PM roycroft: it both increases torque and gives finer granularity
10:56 PM roycroft: but this is to drive the ball screw
10:56 PM roycroft: not the spindle
10:56 PM XXCoder: doh yes I meant similiarm for spindle LOL
10:57 PM XXCoder: your design would certainly help, better control without going far in microstepping
10:57 PM roycroft: my mill has a similar setup for the spindle though
10:58 PM roycroft: https://roycroft.us/Mill-Motor-Mount/Mill-Motor-Mount-1.jpeg
10:58 PM XXCoder: it would be like 1/16 but you set it at 1/8 microstepping for example
10:58 PM XXCoder: or 1/32, 1/16 resectively
10:59 PM XXCoder: I do wonder whats limitions on pulley ratio though
11:00 PM roycroft: that's mostly a space constraint
11:00 PM roycroft: if your spindle pulley is 4x as large as the motor pulley you have to move the motor far away to make room for it
11:01 PM XXCoder: chaining would increase backlash I guess also
11:01 PM roycroft: yeah, you might lose some accuracy that way
11:03 PM XXCoder: I wonder if can offcenter stepper on y direction also for more room
11:03 PM XXCoder: though capped to sides of gantry unless it missed sides
11:03 PM roycroft: the pulley doesn't care
11:04 PM * roycroft is thinking about rack and pinion along the y again, because he's never liked the idea of 2m ball screws whipping around
11:04 PM * roycroft can't remember why he decided to not go with rack and pinion in the first place
11:04 PM ZincBoy[CAON][m]: This is the same on my mill. The reason is the axis brake takes time to engage. When the servo drive loses power or is disabled the output will turn off faster than the brake can engage. On mine, the Z will drop about 0.2" which normally isn't an issue but if you are cutting a the time bad things will happen.
11:05 PM XXCoder: indeed
11:06 PM roycroft: you could install a big capacitor to buffer the power to the mill and have the brake bypass the capacitor
11:07 PM ZincBoy[CAON][m]: With a 2.2kW Z drive that would be one big cap 🙂
11:07 PM XXCoder: battery would be better I think
11:07 PM roycroft: of course the best thing is to ensure the the mains never fail :)
11:07 PM XXCoder: battery only need to last a minute
11:07 PM roycroft: not even a minute
11:07 PM roycroft: a few milliseconds, i should think
11:07 PM XXCoder: yeah, though I like to design in overkill
11:08 PM XXCoder: it could use mroe time to save exact positions, gcode line, etc etc
11:08 PM roycroft: if it takes 10ms to engage the brake, then 20ms is overkill :)
11:08 PM ZincBoy[CAON][m]: You can do fine with just having an input that indicates mains failure. There is enough holdup time in all of the DC link caps to shut things down. Note to self: implement this feature.
11:10 PM roycroft: i'm happy to nag at your conscience :)
11:11 PM ZincBoy[CAON][m]: Damn you! Now I have another project!
11:11 PM roycroft: it's the kind of thing that would bug me
11:11 PM XXCoder: my machines way too small to be concerned with power failures
11:11 PM XXCoder: worse it can do is ruin wood peice
11:12 PM roycroft: no machine is too small to disregard power failure scenarios
11:12 PM roycroft: at the VERY least you need to ensure that when power is restored it will not start up again on its own
11:12 PM XXCoder: if I sum up all ponental failurs up to 10 years, it wont cover 1% of savings of cost
11:13 PM ZincBoy[CAON][m]: This should be handled by the main power contactor. I agree that this is a very high priority.
11:13 PM roycroft: improper handling of power transitions is a safety issue, not a lost profit issue
11:14 PM XXCoder: sure though risk is very low in my case. its quite small machine
11:14 PM XXCoder: if blackout happens I just hit e-stop
11:14 PM XXCoder: and z wont even fall
11:14 PM roycroft: for some reason, a lot of woodworking machinery does not come with a magnetic switch
11:14 PM roycroft: when i purchase such a machine, the first thing i do is install a magnetic switch along with the hour meter that also is rarely found on a woodworking machine
11:15 PM XXCoder: hour meter?
11:15 PM roycroft: yes
11:15 PM roycroft: to keep track of how many hours it's been operated
11:15 PM roycroft: an important thing for maintenance schedules
11:15 PM XXCoder: ahh interesting
11:15 PM roycroft: they are cheap and i don't understand why every machine does not have one
11:16 PM roycroft: "lube the bearings once/year"
11:16 PM roycroft: once per year if i use the machine four times during the year, or once per year if i'm running it 14 hours/day 5 days/week?
11:17 PM roycroft: how about "lube the bearings after every 100 hours of use"
11:17 PM roycroft: now the hour meter tells me when to do the maintenance
11:17 PM XXCoder: there have to be time comonent since oils lubes do decay
11:17 PM ZincBoy[CAON][m]: Yes, I do the same thing. Magnetic contactors are cheap safety. The hour meter comes in handy for PM as you say.
11:17 PM roycroft: sure
11:18 PM roycroft: but you know that a machine that sits mostly idle needs maintenance at shorter intervals that one that's in regular use
11:18 PM XXCoder: how do magnetic contactors work anyway. I guess switch on it uses some current to stay on
11:18 PM roycroft: yes
11:18 PM XXCoder: otherwise it deteches in blackout and wont turn on
11:18 PM roycroft: it's a self-latching circuit
11:19 PM roycroft: i believe the eu recently mandated mag switches on all power tools, hand-held or stationary
11:19 PM XXCoder: cool
11:19 PM XXCoder: I wish usa would make all extensions have fuses too
11:20 PM XXCoder: along with high current tools
11:20 PM roycroft: there's talk that one of the festool routers that recently had an update and a new model number was to comply with the new regs
11:20 PM ZincBoy[CAON][m]: Hmmm. Misumi S87A-12N looks to be a good power loss detector. 10/20ms detection time with an open collector output. Feed that into an HAL input and trigger the brake as soon as it is seen. Think I will give it a try.
11:21 PM roycroft: sounds like that might not be too much work
11:21 PM roycroft: buy me a beer the next time you save a tool/part on a power loss :)
11:21 PM ZincBoy[CAON][m]: Yes, pretty trivial if it works.
11:22 PM skunkworks: https://www.youtube.com/watch?v=pzk0wM1HERY
11:22 PM ZincBoy[CAON][m]: Sure! That is cheap in comparison 🙂
11:22 PM XXCoder: roy should have said "every time" for income guantee for years ;)
11:22 PM roycroft: i'm not greedy
11:22 PM XXCoder: lol
11:22 PM roycroft: it certainly did not cost me a beer to make the suggestion
11:23 PM roycroft: so if i get one in return i am already ahead
11:23 PM * roycroft does not suffer from monty burns syndrome at all
11:49 PM XXCoder: roy yea