I really appreciate how you looked at the sound in audacity. We are often going by sound and gut in these situations, but I haven't seen many people try to use instrumentation to inspect the sound. Pretty clever, really.
Single flute end mills are awesome for aluminum. Nearly impossible to chip weld. The surface finish when roughing is pretty good, but a finish pass with a 3 flute certainly doesn't hurt.
I appreciate your videos! I got the Avid 4848 CNC pro with dual Z’s. The plasma worked fantastically but I haven’t had much luck with the spindle yet. I’m still figuring out speeds and feeds. I already beefed up the table by adding 125 pound T5 plate and adding 3 more steel crossmembers. It made a world of difference!! The table doesn’t move anymore. I’m almost done with all the holes then I’ll square my vise up and try aluminum again. Keep your videos comin’!!
I have a similar CNCRouterparts router (from before they were AVID) and I've thought about upgrading to the S30C for toolchanging. Excited to see your progress here, it seems like you are driving directly toward my dream router. Subscribed. Thanks for posting.
:high-five: CNCRP buddy :) I'll make sure to post any tips I find along the way! Already have a bunch about wiring and programming the spindle into the Avid system, and I'm assuming I'll run into more gotchas when programming up the tool change macros. My initial impression is that Mach4's documentation is uhh... sparse in places :)
@@BreakingTaps Awesome, can't wait. I have to say, after using PathPilot on my Tormach, I never want to go back to using Mach again. Maybe they'll get it right someday.
@@ViceChief I agree - pathpilot is seriously nice to use. Mach seems ancient in comparison. Of all of the 'hobby' controller interfaces, it's by far taken the cake for me.
@@TheSeaShadow clearpaths are sweet. Their silence is what scares me the most. Rather than the soothing whirr of steppers, it's just wind noise from your machine zipping around lol.
Your videos are so good but when your displaying the speeds and feeds in the edit it would be amazing if your could include metric I think your views would go up
Ah that's a great idea, I'll do that in the future! Thanks for the suggestion. It's funny, I do all my modeling in metric, but fall back to imperial for feeds/speeds. :)
Nice setup. When you're speed limited using lower flute counts can make a big difference. I find that Onsrud 0Flute aluminum end mills are magical in light machines like this. Lets you keep your chip load up without having to run at very high cutting speeds, and the extra relief on the back of the flute helps avoid the chip welding when slotting like you ran into just before the endmill started to pack. I also tend to run slotting very conservatively, but it works well to make it more reliable. I run shallow passes, 2mm DOC at a time, two passes with a cleanup. So then you're doing a 2mm deep slot, a .5mm WOC clean up pass, this leaves just that little extra bit of room for chip evacuation at the next slotting depth and makes it that little bit easier to get air deep into the cut. Also check out high helix endmills for finishing cuts. The high helix mill that Maritool sells gives me the best sidewall finish from any of the ones I've tried. Much less of the vertical chatter marks and closer to a mirror finish. I suspect it might be because it helps keep the Z axis engaged better so it doesn't bounce around as much
Thanks, will try these tips out! I'm a sucker for tooling so will definitely go hunt down some of those Onsrud 0flute and Maritool high helix end mills.
Idea - get piezo microphone (yeah, that thin disc), place it somewhere on spindle or frame and try to record sound from it. I tried this at work for one project and for example knocking on metal handrail sounds totally different. I'd be really interested to hear all sounds it could catch up and compare it to "real life" sounds that usual microphone picks up. Maybe you could detect chattering soon enough and even somehow tell CNC to act accordingly.
Some of it was the tool/collet as well. In a later video I realized I was loading the ER collets incorrectly (oops) and so the tools were not held nearly as firmly as they should. I suspect that contributed a lot of deflection and chatter seen in this video (on top of the limitations of the machine itself)
I'm at about 4:30 and feel compelled to let you know that IPT and IPM change together if all other variables are held constant. Double IPM would lead to double IPT.
From the first spectrogram it looks sort of like your mic, or the front end electronics, were picking up something extremely high frequency, and aliasing it into the signal. (Spindle commutation RF? 20kRPM Bearing vibration? beats me) If you set up a fancy contact microphone thing, you should consider balanced audio, and then low pass filter it before bringing it into your ADC. Very neat to see that done, I was just thinking about rigging something up myself.
Oh interesting, yeah I think you're right! Hadn't noticed that before! I'm working on a more robust setup right now... going to take a while to get fully operational but I think it'll give some interesting data: four electret mics spread around the enclosure, two piezo contact mics (one on spindle, one on z-axis) and threeaccelerometers (spindle, z-axis, bed). All feeding into some high bandwidth ADCs on an STMF32 and then up the chain to something like an RPi or laptop for visualization. No firm plans yet, but I think your suggestion for better analog processing before it gets digitized is probably a good idea.
What kind of tolerances have you found you are able to hold in aluminum? The product page says +- 1 thou accuracy, with +- .5 of repeatability. You think that’s accurate in aluminum? Also is there any kind of calibration you need to do to get the machine to the +- 1 thou level?
To be honest, there aren't many choices at a reasonable price point :( The import Jianken ATC spindles are generally considered to be pretty good, and a bit cheaper. Product page here (www.jian-ken.com/automatic-pneumatic-tool-change-spindle-motor/jgl-80-1-5kw-2-2kw-2-5kw-24000rpm-40000rpm.html), and they have an official Aliexpress store somewhere if you go looking
you said that you think one of the main limitaitons of this machine is its weight, and by extension the weight and stiffness of the table it is on - do you think a ooverbuilt table made of 2x4s and similar lumber with some cinderblocks to add mass would be suitable for this machine? furthermore, is it nessecery for the feet of the machine to be removed in favor of bolting the machine down to the surface it occupies, or are the feet passible?
Looks good, a few tips. Chip removal is important for tool life, not all carbide is the same, I have used some that didn't last as long as HSS. Polished carbide is great on aluminum, much better than any coatings I have used.
Cannot agree enough! A well polished carbide cutter combined with a single flute design - the finish, reduced chatter, and lack of chip welding offers distinct advantages.
If the machine is too lightweight... have you considered augmentation with epoxy-granite? (it’s like concrete, except it’s got epoxy instead of portland cement as the binder, because epoxy won’t shrink, warp, and crack).
I've considered it! Or something similar like sand/lead, just to add mass. Most of the aluminum extrusions are hollow so it'd be possible to fill with something, or replace the bed with something a bit more massive. I have a few big pieces of steel that I'm going to turn into fixture plates, and that'll add another ~100lbs. Figured I'd see how much that helps and go from there.
Draino :) Aluminum will dissolve in both dilute acids and bases, but I find NaOH a lot more pleasant to work with than 30% hardware store HCl. Let it sit for ~30-60min and the bond between the end mill and chips should have weakened enough you can remove it with a pick.
how would you compare this machine to a step craft Q204 with the 2.2kw spindle that takes iso30 er32 tool holders, very curious as i saw you machined steel with it.
No personal experience, so I don't want to speculate too much on a comparison. I personally avoided it because of the round rail/bearing motion system, instead of proper linear rails. I don't care what the marketing material says, it's not better than linear rails otherwise real VMCs would use them too ;) Otherwise it has many similarities (size, ballscrews, mostly-aluminum extrusion, spindle stats, etc). Does have a steel gantry support which is maybe a perk compared to the Avid (which is aluminum), but the Avid gantry support is much thicker so it might even out. The Avid is a touch smaller in Y direction which probably makes it a little more stout. Price ends up being similar.
@@BreakingTaps thank you very much, im looking to buy in the next week and although rated for aluminum im sure if the feeds and speeds are more on the conservative side, more robust routers such as these can cut steel accurately with good finish.
@@willacademic Yeah that's probably true, I think with most machines you can get away with steel if you dial in the feeds/speeds. Winston Moy's channel has a ton of steel machining on his Shapeoko which is even lighter, and he has some great surface finishes. Just need to find the right combo that works for the machine. :) Goodluck!
Thank you for the video! I’m wondering how you stabilized the table wobble you experienced after the first run? I’ve got the Benchtop Pro 2424 on a 30x60 workbench with standard metal legs. Table wobble is bad when the machine changes directions.
you may be able to change acceleration profiles, not sure about your machine, but that might do it. Otherwise reinforce your table with diagonal legs, and/or cross diagonal legs, to absorb/distribute shockwaves through the table. When 3d printers were new they did this too, i would attach bungee cords at the anti-nodes and hook them up to a wall.
I'm using "TRIM Mist", it's a synthetic coolant designed for mist application. I'm... mostly happy with it. I find I have to mix it at a higher strength than recommended to keep from welding aluminum frequently, and if some of it does get into the air it's definitely not great to breathe. But otherwise it seems to work well, and a gallon will last forever (1-5% dilution). I'm thinking about trying Koolmist next.
@Breaking Taps you said that you think one of the main limitations of this machine is its weight, and by extension the weight and stiffness of the table it is on - do you think a overbuilt table made of 2x4s and similar lumber with some cinderblocks to add mass would be suitable for this machine? furthermore, is it nessecery for the feet of the machine to be removed in favor of bolting the machine down to the surface it occupies, or are the feet passible? Thanks a bunch in advance!
Yep, a big heavy table would help. Once I stiffened up my table I saw an immediate improvement. I'm still using the feet though, so there's a limit to how much it can help. If the axis get moving fast enough and then sudden stop (at the end of a toolpath or something), there's enough mass in the moving axis to overcome the friction of the feet and cause the machine to move a little on the table. It's not enough to really get it "walking", but it's enough to not be fully coupled to the table. For best results it should bolted down... but I've been too lazy to do that :) It works well enough for me and I just keep the speeds reasonable.
@@ratherrelatablemantisshrim2409 No problem! Mostly steel, I welded up a table out of scrap and used a thick piece of MDF as the top surface. It wasn't quite sturdy enough after I assembled the machine, so I welded more braces on afterwards. It's a rather frankenstein table at this point :) A big chunky lumber table should do fine I would think, I just had the right metal scrap at the time to do it instead of lumber. Pretty much the entire rest of my shop are 2x4 tables and benches.
Conicidence? My Avid CNC Benchtop Pro 2x3 is due to arrive tomorrow! I’m looking forward to the conversions you made. I have a 2-3 year old CNC Router Parts 4x8 machine that just went through too many upgrades at once: S30C spindle upgrade, Wine Rack ATC holder extension to table length, conversion from the NEMA 43 steppers to ClearPath servos, and finally MACH 3 to MACH 4. I’n now willing to admit that I bit off more than I could chew and the totally new electroncs box and MACH 4 configuration is still in a state of, “development”! I needed a turnkey system from Avid so I could move forward so I got the Pro 2x3 and intended to initially power it with the old CNC Router Parts spindle and NEMA 43 electroincs from the original machine. in the mean time, I’ve been designing a table. I suspected much of what you’ve found - the machine itself needs to be bolted down to some higher mass table in order to keep things from dancing around. I plan on using a table built from 80/20 parts with a 1.5 x 48 x 60” butcher block maple top. Wasn’t sure about bolting the Pro 2x3 down to the butcher block top but after watching your video, I’m now convinced! Looking forward to your solutions for the other parts integration. Might help me get myself un-stuck.
Good luck, keep me updated! Definitely know the feeling of juggling too many balls in the air on a project and changing too many things at once. :) How'd you like those ClearPath servos though? I've been eyeing them as a potential upgrade in the future. The Pro 2x3 mounting system are basically M12 studs with rubber feet, so it would be relatively straightforward to grab longer M12 bolts and through-bolt it to the surface. I need to make my table a bit more robust first (cobbled together welded frame with MDF top, would like to bolster the frame and then get a more solid top) but that's my plan next. I'm also considering adding a big plate of 6061 or mild steel, either on top of the existing t-slots or maybe replacing them. Will give me a surface I can deck of nice and flat, and add some more mass (a 24x36x1" plate of 6061 is ~80lbs, and around 250lbs for A36).
The ClearPath servos have definite advantages in controlling acceleration and jerk. = faster moves without the table shaking. Having a direct substitute that can take data levels for step/direction (or quadrature) inputs makes wiring easier. I suspect that, like Tormach, it won't be long before Avid offers this as an upgraded driver option.
Gotcha, thanks for the details! Better control of shaking seems very useful (in addition to servos generally, like not missing steps) given the size of the Pro2x3. And a direct substitute makes it very appealing for sure. Might have to bump that up on my upgrade list :)
Good vid. Tip 1. For spiral down it looks to me that you go to slow. Machine should be able to handle tad more. Tip 2. Larger stock for finishing in unstable machined is preffered, more than usual. This way you will tension the tool more agsinst matetial and prevent some of that horrible high pitch chattet. Tip 3. Sometimes is bettet to increase F intead of Ae while keeping Ap to the max, again more tension on tool. And please metric ad well as imperial ;)
Thanks very much for the great video. I have this machine but am yet to upgrade the spindle on it. Would like to know if adding weight somehow would reduce shakiness and thus make the cuts appear smoother. They are plenty smooth but I am still keen to get them as smooth as an industrial CNC machine.
Adding weight should, theoretically, help. I have a large plate of steel I'm going to use as a fixture plate (~100 extra pounds) but haven't had a chance to work on it yet. But generally more mass should help here, since the machine is pretty light relative to it's horsepower. Also note: a lot of shakiness in this video was ultimately caused by the table it is sitting on. I reinforced the table with more welded braces and it improved dramatically, so that's a big factor too.
Great results either way for a desktop CNC. What spindle motor do you have? I looked through Avid CNC website and didn't see the spindle motor with tool change capability
Thanks! I have the S30C from CNCDepot (cncdepot.net). It was a little pricey but I really wanted ATC, didn't want to try and get a chinese model working (although I've heard good things about Jianken), and wanted local support. It's been great so far, and Alex from CNCDepot has been very responsive to questions.
Not yet, but it's on my todo list! I do have a reddit post explaining some of my thought proces though: www.reddit.com/r/hobbycnc/comments/eooasq/first_chips_on_avid_cnc_benchtop_pro_aluminum/feeo3oq/ You'll see that I was mostly considering small VMCs rather than desktop routers, to the Benchtop Pro was actually the smallest machine I was considering (in terms of weight at least, it has a larger work volume than the other machines). So I wasn't actively considering other desktop routers like MPCNC, Shapeoko, import 6040s, etc.
Have a new setup with the same 2'x3' Avid, different ATC spindle, Clearpath servos and dual Fogbuster. I have added a 12 mm aluminium plate to the T-slot table. If you want to cooperate (in Fusion 360) on a table + cover, please let me know.
Oh nice! Were you planning to drill/tap holes in your plate? Cut t-slots? It's definitely on my todo list, I'll drop you a line when I get around to working on it. What model of Clearpath servos did you end up with? I was looking at the specs over the weekend and had to force myself to close the tab so I didnt purchase any :)
@@BreakingTaps CPM-SDSK-2310S-ELN was the closest I could find to the NEMA 34 that is included in the Avid offer. Works great but I had a smaller issue running the calibration program for the Z axis. I have M6 holes in the plate. 100 x 100 mm grid. Should have done a 50 x 50 mm array, so that's on the to-do list.
Nice, thanks for the servo model! Will check those out. Did you notice the machine feeling more rigid/less lightweight after adding the plate? It's obviously useful as a fixture plate, but I was also hoping it would firm up the machine a bit too :)
0:28 @Breaking Taps , are you planning to make this video any time soon? I'm in the initial initial stages of choosing a machine myself, and feel like your thoughts on this one could be really helpful to me and other doing the same.
Afraid I haven't, and probably won't for a while (or ever? dunno). I did write up some of my thoughts on reddit which might be helpful. Thread starts here: www.reddit.com/r/hobbycnc/comments/eooasq/first_chips_on_avid_cnc_benchtop_pro_aluminum/feeo3oq/ Considering changes to fusion360 in last year (no more toolchanges in free version) i might not have gotten the ATC since that hamstrings it pretty well, and not sure I want to spend on another yearly subscription. And I might have splashed out for servos instead of steppers and built the electronics myself. But otherwise the breakdown and reasoning mostly still applies. I didn't quite appreciate how limiting the small Z height would be though. It's not a _huge_ deal, but it can be annoying once you add in a vise. Not so much because of needing it for machining, but I didn't consider how the part has to clear the gantry itself which can be very irritating, even if the spindle is otherwise positioned fine. I think I'd probably still go down the same route though, since the overall work volume suits my needs and diverse projects.
@@BreakingTaps I think I saw on your trying mild steel reddit post that you are able to hold abt + - 2 thou tolerances in mild steel - how does that compare to what this machine can manage in aluminum?
@@BreakingTaps I'd also be interested to know how long the whole setup process was for you - both the physical assembly, and the electrical + software tuning and setup.
@@ratherrelatablemantisshrim2409 Hmm, think I got it assembled over the course of a few evenings. So probably 3-4 hours max? Was pretty easy build, mostly just watching the Avid assembly video and following along. Wiring took _a lot_ longer because I was using a third party spindle (with ATC), so I had to do my own work on the VFD, pneumatics, connector breakout, etc. That was probably a week on/off of fiddling. Mach4 was pretty plug'n'play since Avid provides their own screenset, which is great for stock experience. Tweaking it to use my own probe and ATC was a bit of a battle, because Mach documentation is terrible :(
How that you’ve had this machine for a while, what tooling would you suggest someone buying the same machine buy to get started? What tooling do you use the most?
Most of the work I do on the machine ends up being aluminum, so I've found myself using single flute cutters extensively. Seem to hit the sweet spot in terms of feeds/speeds, don't clog easily and evacuate chips well (even with an underpowered fogbuster). Single flute 1/4" is my main tool when I just want to get work done. Two flute works ok but not amazing, three flutes tend to weld chips quickly. For occasional steel, variable flute really seems to help keep chatter down. Haven't had luck with corncobs, they really make a rucus and chatter a ton. Exception being Datron's fancy cross-cutter tool, which worked a treat. Expensive though. Otherwise, a big vise makes life a lot easier to setup but really hurts in terms of z-height, so I really need to invest in some more clamps and work off the table more. Threadmilling works surprisingly well, albeit very high clench factor while it's running 😉 I don't drill on the machine much, high spindle speed makes that hard. But when I have to, stub carbide drills work alright if you take it conservatively.
I have a Kobalt "QUIET TECH" 26-Gallon quiet compressor, plus an 8gal tank to add some more capacity to the system. It's OK.... a little underpowered if the fogbuster is going hard, probably 60% duty cycle. But it's reasonably quiet and I don't always have the fogbuster turned up to max so it mostly works :)
Work volume and table size mostly. I kept running into things I'd like to do on a CNC... but would be impossible (or very complicated, with multiple setups) to do on the Taig. My Taig has 9.5x5.5x6" travels and 15.5x3.5" table. Compared to 24x36x9 on the Benchtop Pro, or around 16x12x16'ish travel and 30x15" table for the various mini VMC competitors (Syil, Skyfire, Tormach, Haas MiniMill, etc). Part of the reason I opted for the BenchtopPro over the VMCs is that I may want to do larger sheet goods in plastic too, which really needs a larger work volume. The horsepower/rigidity of the Taig is obviously a lot smaller too, but that can mostly be accomodated by correspondingly smaller feeds/speeds, so that didn't concern me as much (as a hobbyist and prototyper, slow speeds are fine). Finally, ability to do probing and auto tool changes was high on my list, not necessarily because it's required but just because I want to automate that. I really hate touching off parts and measuring tools :) The Taig is still a great little machine for what it is -- especially at the price I got it -- and I'll be keeping it around for smaller parts or working in plastic, etc. And I do really like the coolant that I installed on it, works great :)
I don't think so actually, or at least not without other modifications. I rarely run into situations where the spindle bogs down, typically keep it very high RPM and high feed rates. So it's usually the steppers that are struggling to keep up. If I were to do it again, I'd order the NEMA 34 steppers, or just get some servos. With servos installed, I suspect the frame itself will be the limiting factor and so more horsepower would just go to waste.
Why on Earth would someone program a CNC router to cut through the material quickly, yet the rapid to return to start the next cut goes at 1/3 speed? Am I missing something?
When running a program for the first time, the most likely time you'll find an error is when moving to the next position. Moving to next hole to bore, or the next feature to profile, etc etc. Keeping rapid speed low is a good way to sanity check the program while it's running, since you can quickly mash the emergency stop if it looks like you've messed something up. It's saved my butt a few times, like if the z-height is incorrect (starts to plunge lower than expected) or dimensions were off and it's starting to head towards the vise jaws or another feature on the part. You _can_ run slower when cutting too, but that's more difficult since the material needs to be cut at a certain speed. So there is less room for slowing things down. But slow rapid speed is an easy safety measure and doesn't add much time. In a production environment, once the program is watched and it looks good, you'd turn the speeds back up to 100% and walk away, knowing it would do the right thing without being monitored :)
I had a hard time deciding between this and the MillRight mega v plasma/cnc I ended up going with the MillRight (still waiting for it!) and the money I saved I’m cnc’ing my Bridgeport I have all the brackets done now I’m doing the electronics. I went with centroid acorn for the Bridgeport. I’m curious how well your machine does machining on aluminum and the tolerances you can get out of it!
Awesome, goodluck! I've heard good things about the Acorn, I think I would use that myself if I were doing a full DIY. Not a huge fan of Mach 4 and the rest of the ecosystem. I'll be doing more tests in Aluminum soon, including tolerance tests. Want to finish up some accessory work first (tool changer, probes, etc). Cheers!
That looks like an awesome spindle! Monitoring/recording its current should give you a better idea of how hard you're pushing it. It looks like you have the "passively cooled" (no fan or liquid) version, right? How warm/hot does it get?
Good point! The VFD can definitely output the amperage being used, and with a bit of tinkering I think I could route that back through the Avid system for display in Mach. Would be really nice to have a "spindle load" meter on the display like an industrial system. The spindle has full case pressurization (~40 PSI flowing through the case) which I imagine helps cool it a bit, but yeah there's no fan or water cooling. I haven't noticed it getting very hot but I also haven't had it running for hours at a time yet either. From what I've read elsewhere, folks that use it in production for hours at a time haven't had issues with temperature either.
Why aren't you running the spindle at 24000 RPM to minimize cutting and machine forces? I'm embarrassed to say that I've been using Audacity for decades and wasn't aware of the spectrogram display mode - thanks for pointing that out!
Mainly going off what HSMAdvisor was recommending. I don't yet have a good sense how "off" feed/speed calculators are for the machine, so trying to start around the recommended settings and play from there. Been fiddling with steel (videos to come!) and am finding the recommendations are farther off than aluminum. Definitely going to take a while to figure out the best set of recipes.
@@BreakingTaps I believe that HSMAdvisor is really intended for spindle power limited CNC mills rather than machine force limited CNC routers. CNC routers generally have far less cutting force capability than mills (generally more than an order of magnitude less). Since spindle power (i.e. material removal rate) is proportional to the product of spindle speed and cutter torque (i.e. cutting force), using the highest speed your spindle and endmill support should help you get the most out of your machine. You can force HSMAdvisor to use that speed to estimate cutting forces for varying cutting parameters (WOC, DOC, and feed-rate), but since that calculation is based on the assumed unit horsepower for the workpiece, it's probably not very accurate. Measuring/monitoring the increase in spindle current (which is proportional to torque) when milling would provide a much more accurate, real time estimate of both cutting force and power
Yeah I think you're right, that all lines up with the various tests I've done so far. I suspect aluminum "mostly worked" because it's aluminum, and anything fast and not welding will work well enough. But in mild steel the assumptions of HSMAdvisor you mentioned are definitely starting to show. Two questions for you, if you have a sec! The torque/power curve for my spindle looks like this (roughly, i'm told the chart is just an estimate, but it tracks with other similar motors from what I've seen): imgur.com/jLRe01V So does that mean the "sweet spot" is around 18k RPM, since torque starts to drop past that point? Or is it that anything >=18k is good because we're in the constant power range, despite torque dropping as speed increases? The other question has to do with IPT. Presumably I'll need to increase feeds to keep up with increased RPM, otherwise I'll start rubbing the end mills? E.g. need to keep IPT constant while increasing RPM?
Dude Im ultra curious why you went this route over a more substantial vertical mill, you said you were going to talk about that in a later video, Id love to know that part
Was mostly space and power constraints at the time, plus spouse approval :) Price too. There is basically no used market in my area, so I'd either have to drive/fly far to get a used machine or buy new and have it freighted in. Both looked to be expensive or a hassle, so getting something like the Avid made sense. I also assumed it would be good enough for what I wanted/needed. It _was_ good enough, but of course you quickly realize you can put more machine to use if you had it. 🙂 It was a good stepping stone for me but I'd like to have a proper VMC some day for sure!
@@BreakingTaps For sure, thanks for that explanation, I have a shapeoko and a diy pm728vt conversion but Im looking to up the anti a little bit with something more rigid, if not more capable than what I have now.
Came here from Hackaday, did you think about running your analysis again with a contact microphone mounted on the spindle? Should get you a much cleaner signal.
Agreed! And definitely thinking about it for future experiments. Ditto to accelerometers and force sensors on the spindle/frame in various places. The external mic was a sort of a proof-of-concept to see if it'd work at all, being entirely non-invasive (and already recorded for the video). But now that it seems to work, I would definitely like to explore better methods of collecting the data. I have a much nicer condenser mic I'll probably drag out to the garage to capture better data, rather than the dynamic shotgun mic on my camera. I also want to toy with better data processing. E.g. The spindle, fogbuster and background compressor sounds could be captured in isolation and subtracted from the "cutting" signal to get a lower noise floor.
It will get finished, don't worry! I'm super curious to know how it will work myself, and if the hobby motors will hold up to the abuse. Just temporarily stalled because I hit a silly issue that was irritating me (and I kept needing the mill for other things and didn't want to change it over). Basically, I wired it all up in an enclosure... but the enclosure is too small. :'( I need to design/print a new enclosure and disassemble all the electronics to put in the new enclosure. It was just demoralizing having to redo all that work so I temporarily shelved it out of frustration :)
NEMA 23. I asked Avid their opinion, since I was considering the 34 instead. They recommended the 23 because A) they didn't think the enhanced speed/torque would be useful if the main goal was metal (compared to cutting wood on their larger routers, where more torque at 350+ IPM is useful) and B) ballscrew whip would start to be an issue at higher feeds on the benchtop machines.
Where was the chatter analysis? I watched the whole video but didn't see any analysis. You showed the load vs no load speed of the spindle and that was about it. Edit: I don't mind a little clickbait if the title is in the content, but sadly it is not in this video.
The audio analysis is very cool. Are you a musician or audio engineer by any chance? Or maybe a mechanical engineer? Some of the words you used made me wonder. When I saw Avid (and audio analysis) on Hackaday, my mind went to Avid Pro Tools DAW :) Peace
Software engineer :) I've done a fair amount of tangentially related work though (I work on search/analytics software and some of my past major projects have been in the area of time-series analysis work). I've also gone down the rabbit hole of reading "smart spindle" and realtime chatter detection research papers, so some of the terms have rubbed off on me :)
@@BreakingTaps Very cool! I'm an analytical software engineer as well. Plant growth and health most recently, from images. Once the sensors do their job, it's all just data. Glad to meet you :)
Oh awesome, great to meet you too! Is your work with IR/hyper/multi-spectral imaging of plants? Don't know much about it, but from what I've seen it looks like a really neat field. And hard, so much data to process! Or is it more individual plant level, like leaf gas-exchange sensors? Funnily enough, I'm an ex-biologist (mammalian, not plant) so crossover between tech and bio always excites me :)
@@BreakingTaps Wow, that's cool! My only exposure to animals was helping in a shark lab with molecular phylogenetics. Now, I'm building instruments for imaging plants, from basic digital cameras to fancy hyperspectral and infrared cameras. Much of the hardness comes from researchers not having a well defined trait to measure. I'm working on the motion of growing plants, interaction with soil microorganisms, and making a phenotyping appliance that's affordable. Please drop me an email if you think there might be something we could collaborate on, living or metallic :)
Neat stuff! Makes sense re: researchers not having a well-definied trait. Biology is messy, and researchers are fickle :) You have a really interesting channel, subbed and will drop you an email!
Wow thanks for doing this video. I have a build series going on the same machine. Your video will definitely be helpful. Maybe my build videos will help you to. Cheers. Ken
Insurance against accidents :) Most (user) errors happen during rapids from one cut to the next, so it makes sense to rapid slowly so there is time to hit the e-stop. Once you know a program is good and you've already run it, you can dial up the rapids. Feed rate can't change much because you need to cut the material at a minimum speed/feed for it to cut well ;)
I enjoyed your test here. If you would like help on anything send me a message. I currently run a small Aluminum machine shop in Utah. Just a heads up your facing finish would be much better with a all to center fluted endmill. i like the YG alupower HPC for this.
Might take up up on that offer in the future, thanks! I picked up a YG-1 Alupower recently and plan to test it in the near future, I've read good things about it. :)
Yeah agreed, I noticed that while re-watching the footage (especially deeper in the pocket at certain angles). I'm tempted to try the fogbuster with just air and no coolant. The coolant seemed to encourage chips sticking to side walls when they might have otherwise blown out. I did dial up the air pressure, but the fogbuster is only rated for so much.
I really appreciate how you looked at the sound in audacity. We are often going by sound and gut in these situations, but I haven't seen many people try to use instrumentation to inspect the sound. Pretty clever, really.
Single flute end mills are awesome for aluminum. Nearly impossible to chip weld. The surface finish when roughing is pretty good, but a finish pass with a 3 flute certainly doesn't hurt.
I appreciate your videos! I got the Avid 4848 CNC pro with dual Z’s. The plasma worked fantastically but I haven’t had much luck with the spindle yet. I’m still figuring out speeds and feeds. I already beefed up the table by adding 125 pound T5 plate and adding 3 more steel crossmembers. It made a world of difference!! The table doesn’t move anymore. I’m almost done with all the holes then I’ll square my vise up and try aluminum again.
Keep your videos comin’!!
Nice! I was considering adding a big plate of 6061 or A36 on top (or replacing) the t-slots, good to hear something like that helps!
I have a similar CNCRouterparts router (from before they were AVID) and I've thought about upgrading to the S30C for toolchanging. Excited to see your progress here, it seems like you are driving directly toward my dream router. Subscribed. Thanks for posting.
:high-five: CNCRP buddy :) I'll make sure to post any tips I find along the way! Already have a bunch about wiring and programming the spindle into the Avid system, and I'm assuming I'll run into more gotchas when programming up the tool change macros. My initial impression is that Mach4's documentation is uhh... sparse in places :)
@@BreakingTaps Awesome, can't wait. I have to say, after using PathPilot on my Tormach, I never want to go back to using Mach again. Maybe they'll get it right someday.
Ditto, I'm in the process of getting a nearly identical setup, except trying out clear path servos instead of steppers.
@@ViceChief I agree - pathpilot is seriously nice to use. Mach seems ancient in comparison. Of all of the 'hobby' controller interfaces, it's by far taken the cake for me.
@@TheSeaShadow clearpaths are sweet. Their silence is what scares me the most. Rather than the soothing whirr of steppers, it's just wind noise from your machine zipping around lol.
Alupower endmils with the coating are amazing. I used them a ton in aluminum and they last forever before they dull out
Your videos are so good but when your displaying the speeds and feeds in the edit it would be amazing if your could include metric I think your views would go up
Ah that's a great idea, I'll do that in the future! Thanks for the suggestion. It's funny, I do all my modeling in metric, but fall back to imperial for feeds/speeds. :)
@@BreakingTaps +1 for metric. I had to pause and use my calculator a lot :p
@BreakingTaps
16:20 What did you pour on the endmill to remove the welded aluminum?
Nice setup. When you're speed limited using lower flute counts can make a big difference. I find that Onsrud 0Flute aluminum end mills are magical in light machines like this. Lets you keep your chip load up without having to run at very high cutting speeds, and the extra relief on the back of the flute helps avoid the chip welding when slotting like you ran into just before the endmill started to pack.
I also tend to run slotting very conservatively, but it works well to make it more reliable. I run shallow passes, 2mm DOC at a time, two passes with a cleanup. So then you're doing a 2mm deep slot, a .5mm WOC clean up pass, this leaves just that little extra bit of room for chip evacuation at the next slotting depth and makes it that little bit easier to get air deep into the cut.
Also check out high helix endmills for finishing cuts. The high helix mill that Maritool sells gives me the best sidewall finish from any of the ones I've tried. Much less of the vertical chatter marks and closer to a mirror finish. I suspect it might be because it helps keep the Z axis engaged better so it doesn't bounce around as much
Thanks, will try these tips out! I'm a sucker for tooling so will definitely go hunt down some of those Onsrud 0flute and Maritool high helix end mills.
Idea - get piezo microphone (yeah, that thin disc), place it somewhere on spindle or frame and try to record sound from it. I tried this at work for one project and for example knocking on metal handrail sounds totally different. I'd be really interested to hear all sounds it could catch up and compare it to "real life" sounds that usual microphone picks up.
Maybe you could detect chattering soon enough and even somehow tell CNC to act accordingly.
You have some serious flex in the Z axis or the Gantry at the end of the pass where the cutter cuts back the edge of the block
Some of it was the tool/collet as well. In a later video I realized I was loading the ER collets incorrectly (oops) and so the tools were not held nearly as firmly as they should. I suspect that contributed a lot of deflection and chatter seen in this video (on top of the limitations of the machine itself)
Fascinating audio analysis!
I'm at about 4:30 and feel compelled to let you know that IPT and IPM change together if all other variables are held constant. Double IPM would lead to double IPT.
From the first spectrogram it looks sort of like your mic, or the front end electronics, were picking up something extremely high frequency, and aliasing it into the signal. (Spindle commutation RF? 20kRPM Bearing vibration? beats me) If you set up a fancy contact microphone thing, you should consider balanced audio, and then low pass filter it before bringing it into your ADC. Very neat to see that done, I was just thinking about rigging something up myself.
Oh interesting, yeah I think you're right! Hadn't noticed that before! I'm working on a more robust setup right now... going to take a while to get fully operational but I think it'll give some interesting data: four electret mics spread around the enclosure, two piezo contact mics (one on spindle, one on z-axis) and threeaccelerometers (spindle, z-axis, bed). All feeding into some high bandwidth ADCs on an STMF32 and then up the chain to something like an RPi or laptop for visualization. No firm plans yet, but I think your suggestion for better analog processing before it gets digitized is probably a good idea.
Thanks for the video! It would be interesting to hear more about why you choose that machine in particular.
Great video!!!👍👍👍
What kind of tolerances have you found you are able to hold in aluminum? The product page says +- 1 thou accuracy, with +- .5 of repeatability. You think that’s accurate in aluminum? Also is there any kind of calibration you need to do to get the machine to the +- 1 thou level?
The YG-1 alu power is perfect finish pass
Thank you for info packed video. Just wondering if there were other ATC spindles on the list that were short-listed.
To be honest, there aren't many choices at a reasonable price point :( The import Jianken ATC spindles are generally considered to be pretty good, and a bit cheaper. Product page here (www.jian-ken.com/automatic-pneumatic-tool-change-spindle-motor/jgl-80-1-5kw-2-2kw-2-5kw-24000rpm-40000rpm.html), and they have an official Aliexpress store somewhere if you go looking
you said that you think one of the main limitaitons of this machine is its weight, and by extension the weight and stiffness of the table it is on - do you think a ooverbuilt table made of 2x4s and similar lumber with some cinderblocks to add mass would be suitable for this machine? furthermore, is it nessecery for the feet of the machine to be removed in favor of bolting the machine down to the surface it occupies, or are the feet passible?
Looks good, a few tips. Chip removal is important for tool life, not all carbide is the same, I have used some that didn't last as long as HSS. Polished carbide is great on aluminum, much better than any coatings I have used.
Cannot agree enough! A well polished carbide cutter combined with a single flute design - the finish, reduced chatter, and lack of chip welding offers distinct advantages.
If the machine is too lightweight... have you considered augmentation with epoxy-granite? (it’s like concrete, except it’s got epoxy instead of portland cement as the binder, because epoxy won’t shrink, warp, and crack).
I've considered it! Or something similar like sand/lead, just to add mass. Most of the aluminum extrusions are hollow so it'd be possible to fill with something, or replace the bed with something a bit more massive. I have a few big pieces of steel that I'm going to turn into fixture plates, and that'll add another ~100lbs. Figured I'd see how much that helps and go from there.
What did you use in order to clean the cutter at 16:19?
Draino :) Aluminum will dissolve in both dilute acids and bases, but I find NaOH a lot more pleasant to work with than 30% hardware store HCl. Let it sit for ~30-60min and the bond between the end mill and chips should have weakened enough you can remove it with a pick.
how would you compare this machine to a step craft Q204 with the 2.2kw spindle that takes iso30 er32 tool holders, very curious as i saw you machined steel with it.
No personal experience, so I don't want to speculate too much on a comparison. I personally avoided it because of the round rail/bearing motion system, instead of proper linear rails. I don't care what the marketing material says, it's not better than linear rails otherwise real VMCs would use them too ;) Otherwise it has many similarities (size, ballscrews, mostly-aluminum extrusion, spindle stats, etc). Does have a steel gantry support which is maybe a perk compared to the Avid (which is aluminum), but the Avid gantry support is much thicker so it might even out. The Avid is a touch smaller in Y direction which probably makes it a little more stout. Price ends up being similar.
@@BreakingTaps thank you very much, im looking to buy in the next week and although rated for aluminum im sure if the feeds and speeds are more on the conservative side, more robust routers such as these can cut steel accurately with good finish.
@@willacademic Yeah that's probably true, I think with most machines you can get away with steel if you dial in the feeds/speeds. Winston Moy's channel has a ton of steel machining on his Shapeoko which is even lighter, and he has some great surface finishes. Just need to find the right combo that works for the machine. :) Goodluck!
Thank you for the video! I’m wondering how you stabilized the table wobble you experienced after the first run? I’ve got the Benchtop Pro 2424 on a 30x60 workbench with standard metal legs. Table wobble is bad when the machine changes directions.
you may be able to change acceleration profiles, not sure about your machine, but that might do it.
Otherwise reinforce your table with diagonal legs, and/or cross diagonal legs, to absorb/distribute shockwaves through the table. When 3d printers were new they did this too, i would attach bungee cords at the anti-nodes and hook them up to a wall.
Awesome video.
very cool
Thank you or posting. May I ask what kind of fluid you have in the fogbuster?
I'm using "TRIM Mist", it's a synthetic coolant designed for mist application. I'm... mostly happy with it. I find I have to mix it at a higher strength than recommended to keep from welding aluminum frequently, and if some of it does get into the air it's definitely not great to breathe. But otherwise it seems to work well, and a gallon will last forever (1-5% dilution). I'm thinking about trying Koolmist next.
@Breaking Taps you said that you think one of the main limitations of this machine is its weight, and by extension the weight and stiffness of the table it is on - do you think a overbuilt table made of 2x4s and similar lumber with some cinderblocks to add mass would be suitable for this machine? furthermore, is it nessecery for the feet of the machine to be removed in favor of bolting the machine down to the surface it occupies, or are the feet passible? Thanks a bunch in advance!
Yep, a big heavy table would help. Once I stiffened up my table I saw an immediate improvement. I'm still using the feet though, so there's a limit to how much it can help. If the axis get moving fast enough and then sudden stop (at the end of a toolpath or something), there's enough mass in the moving axis to overcome the friction of the feet and cause the machine to move a little on the table. It's not enough to really get it "walking", but it's enough to not be fully coupled to the table. For best results it should bolted down... but I've been too lazy to do that :) It works well enough for me and I just keep the speeds reasonable.
@@BreakingTaps that's very helpful thank you! Out of curiosity, is your table made of lumber or steel?
@@ratherrelatablemantisshrim2409 No problem! Mostly steel, I welded up a table out of scrap and used a thick piece of MDF as the top surface. It wasn't quite sturdy enough after I assembled the machine, so I welded more braces on afterwards. It's a rather frankenstein table at this point :) A big chunky lumber table should do fine I would think, I just had the right metal scrap at the time to do it instead of lumber. Pretty much the entire rest of my shop are 2x4 tables and benches.
Conicidence? My Avid CNC Benchtop Pro 2x3 is due to arrive tomorrow! I’m looking forward to the conversions you made. I have a 2-3 year old CNC Router Parts 4x8 machine that just went through too many upgrades at once: S30C spindle upgrade, Wine Rack ATC holder extension to table length, conversion from the NEMA 43 steppers to ClearPath servos, and finally MACH 3 to MACH 4. I’n now willing to admit that I bit off more than I could chew and the totally new electroncs box and MACH 4 configuration is still in a state of, “development”! I needed a turnkey system from Avid so I could move forward so I got the Pro 2x3 and intended to initially power it with the old CNC Router Parts spindle and NEMA 43 electroincs from the original machine. in the mean time, I’ve been designing a table. I suspected much of what you’ve found - the machine itself needs to be bolted down to some higher mass table in order to keep things from dancing around. I plan on using a table built from 80/20 parts with a 1.5 x 48 x 60” butcher block maple top. Wasn’t sure about bolting the Pro 2x3 down to the butcher block top but after watching your video, I’m now convinced! Looking forward to your solutions for the other parts integration. Might help me get myself un-stuck.
Good luck, keep me updated! Definitely know the feeling of juggling too many balls in the air on a project and changing too many things at once. :) How'd you like those ClearPath servos though? I've been eyeing them as a potential upgrade in the future.
The Pro 2x3 mounting system are basically M12 studs with rubber feet, so it would be relatively straightforward to grab longer M12 bolts and through-bolt it to the surface. I need to make my table a bit more robust first (cobbled together welded frame with MDF top, would like to bolster the frame and then get a more solid top) but that's my plan next.
I'm also considering adding a big plate of 6061 or mild steel, either on top of the existing t-slots or maybe replacing them. Will give me a surface I can deck of nice and flat, and add some more mass (a 24x36x1" plate of 6061 is ~80lbs, and around 250lbs for A36).
The ClearPath servos have definite advantages in controlling acceleration and jerk. = faster moves without the table shaking. Having a direct substitute that can take data levels for step/direction (or quadrature) inputs makes wiring easier. I suspect that, like Tormach, it won't be long before Avid offers this as an upgraded driver option.
Gotcha, thanks for the details! Better control of shaking seems very useful (in addition to servos generally, like not missing steps) given the size of the Pro2x3. And a direct substitute makes it very appealing for sure. Might have to bump that up on my upgrade list :)
I ditched Mach a while back in favor of UCCNC software, no regrets.
Good vid.
Tip 1. For spiral down it looks to me that you go to slow. Machine should be able to handle tad more.
Tip 2. Larger stock for finishing in unstable machined is preffered, more than usual. This way you will tension the tool more agsinst matetial and prevent some of that horrible high pitch chattet.
Tip 3. Sometimes is bettet to increase F intead of Ae while keeping Ap to the max, again more tension on tool.
And please metric ad well as imperial ;)
@Breaking Taps what height did you set the table to? is it the medium or lowest setting?
Hi there, really like your vid! How did you record the audio into audacity?
Cheers!
Thanks! Just put a shotgun microphone near the machine, pointed at the spindle. Took the recording and dropped it into Audacity after filming :)
What compressor do you use to run the spindle and air blast for this machine? Do you run the case air on the spindle?
Thanks very much for the great video. I have this machine but am yet to upgrade the spindle on it. Would like to know if adding weight somehow would reduce shakiness and thus make the cuts appear smoother. They are plenty smooth but I am still keen to get them as smooth as an industrial CNC machine.
Adding weight should, theoretically, help. I have a large plate of steel I'm going to use as a fixture plate (~100 extra pounds) but haven't had a chance to work on it yet. But generally more mass should help here, since the machine is pretty light relative to it's horsepower. Also note: a lot of shakiness in this video was ultimately caused by the table it is sitting on. I reinforced the table with more welded braces and it improved dramatically, so that's a big factor too.
AFTER A DECADE OF INTEREST IN CNC...
INCLUDING ABOUT $1K INVESTMENT....
I LIKE YOUR CHOICES,... AND ATTITUDE !
'SCRIBED !!
What coolant do you like the most for aluminum and steel on the fogbuster?
Ok... so why DID you choose this over a Tormach or a Haas mini mill or a Syil, or a skyfire (or a QueenBee)?
Great results either way for a desktop CNC.
What spindle motor do you have?
I looked through Avid CNC website and didn't see the spindle motor with tool change capability
Thanks! I have the S30C from CNCDepot (cncdepot.net). It was a little pricey but I really wanted ATC, didn't want to try and get a chinese model working (although I've heard good things about Jianken), and wanted local support. It's been great so far, and Alex from CNCDepot has been very responsive to questions.
@@BreakingTaps Thank you, I have heard of the Jainken so how do you feel this compares?
Do you have a video explaining your choice of router ?
Not yet, but it's on my todo list! I do have a reddit post explaining some of my thought proces though: www.reddit.com/r/hobbycnc/comments/eooasq/first_chips_on_avid_cnc_benchtop_pro_aluminum/feeo3oq/ You'll see that I was mostly considering small VMCs rather than desktop routers, to the Benchtop Pro was actually the smallest machine I was considering (in terms of weight at least, it has a larger work volume than the other machines). So I wasn't actively considering other desktop routers like MPCNC, Shapeoko, import 6040s, etc.
Have a new setup with the same 2'x3' Avid, different ATC spindle, Clearpath servos and dual Fogbuster. I have added a 12 mm aluminium plate to the T-slot table. If you want to cooperate (in Fusion 360) on a table + cover, please let me know.
Oh nice! Were you planning to drill/tap holes in your plate? Cut t-slots? It's definitely on my todo list, I'll drop you a line when I get around to working on it.
What model of Clearpath servos did you end up with? I was looking at the specs over the weekend and had to force myself to close the tab so I didnt purchase any :)
@@BreakingTaps CPM-SDSK-2310S-ELN was the closest I could find to the NEMA 34 that is included in the Avid offer. Works great but I had a smaller issue running the calibration program for the Z axis.
I have M6 holes in the plate. 100 x 100 mm grid. Should have done a 50 x 50 mm array, so that's on the to-do list.
Nice, thanks for the servo model! Will check those out. Did you notice the machine feeling more rigid/less lightweight after adding the plate? It's obviously useful as a fixture plate, but I was also hoping it would firm up the machine a bit too :)
@@BreakingTaps The first part that I made was the plate so I really can't compare before and after :)
0:28 @Breaking Taps , are you planning to make this video any time soon? I'm in the initial initial stages of choosing a machine myself, and feel like your thoughts on this one could be really helpful to me and other doing the same.
Afraid I haven't, and probably won't for a while (or ever? dunno). I did write up some of my thoughts on reddit which might be helpful. Thread starts here: www.reddit.com/r/hobbycnc/comments/eooasq/first_chips_on_avid_cnc_benchtop_pro_aluminum/feeo3oq/
Considering changes to fusion360 in last year (no more toolchanges in free version) i might not have gotten the ATC since that hamstrings it pretty well, and not sure I want to spend on another yearly subscription. And I might have splashed out for servos instead of steppers and built the electronics myself. But otherwise the breakdown and reasoning mostly still applies. I didn't quite appreciate how limiting the small Z height would be though. It's not a _huge_ deal, but it can be annoying once you add in a vise. Not so much because of needing it for machining, but I didn't consider how the part has to clear the gantry itself which can be very irritating, even if the spindle is otherwise positioned fine. I think I'd probably still go down the same route though, since the overall work volume suits my needs and diverse projects.
@@BreakingTaps I think I saw on your trying mild steel reddit post that you are able to hold abt + - 2 thou tolerances in mild steel - how does that compare to what this machine can manage in aluminum?
@@BreakingTaps I'd also be interested to know how long the whole setup process was for you - both the physical assembly, and the electrical + software tuning and setup.
@@ratherrelatablemantisshrim2409 Hmm, think I got it assembled over the course of a few evenings. So probably 3-4 hours max? Was pretty easy build, mostly just watching the Avid assembly video and following along. Wiring took _a lot_ longer because I was using a third party spindle (with ATC), so I had to do my own work on the VFD, pneumatics, connector breakout, etc. That was probably a week on/off of fiddling. Mach4 was pretty plug'n'play since Avid provides their own screenset, which is great for stock experience. Tweaking it to use my own probe and ATC was a bit of a battle, because Mach documentation is terrible :(
How that you’ve had this machine for a while, what tooling would you suggest someone buying the same machine buy to get started? What tooling do you use the most?
Most of the work I do on the machine ends up being aluminum, so I've found myself using single flute cutters extensively. Seem to hit the sweet spot in terms of feeds/speeds, don't clog easily and evacuate chips well (even with an underpowered fogbuster). Single flute 1/4" is my main tool when I just want to get work done. Two flute works ok but not amazing, three flutes tend to weld chips quickly. For occasional steel, variable flute really seems to help keep chatter down.
Haven't had luck with corncobs, they really make a rucus and chatter a ton. Exception being Datron's fancy cross-cutter tool, which worked a treat. Expensive though.
Otherwise, a big vise makes life a lot easier to setup but really hurts in terms of z-height, so I really need to invest in some more clamps and work off the table more. Threadmilling works surprisingly well, albeit very high clench factor while it's running 😉 I don't drill on the machine much, high spindle speed makes that hard. But when I have to, stub carbide drills work alright if you take it conservatively.
@@BreakingTaps wow thats great information thanks a bunch!
I need to try this on my machine.
A quick question; what compressor do you use for the ATC and the fog buster?
I have a Kobalt "QUIET TECH" 26-Gallon quiet compressor, plus an 8gal tank to add some more capacity to the system. It's OK.... a little underpowered if the fogbuster is going hard, probably 60% duty cycle. But it's reasonably quiet and I don't always have the fogbuster turned up to max so it mostly works :)
@@BreakingTaps Thank you! I'm in Australia and am in the process of setting mine up - same machine and the ATC spindle... Not cheap!
I am curious what wasn't working for you with the Taig?
Work volume and table size mostly. I kept running into things I'd like to do on a CNC... but would be impossible (or very complicated, with multiple setups) to do on the Taig. My Taig has 9.5x5.5x6" travels and 15.5x3.5" table. Compared to 24x36x9 on the Benchtop Pro, or around 16x12x16'ish travel and 30x15" table for the various mini VMC competitors (Syil, Skyfire, Tormach, Haas MiniMill, etc). Part of the reason I opted for the BenchtopPro over the VMCs is that I may want to do larger sheet goods in plastic too, which really needs a larger work volume.
The horsepower/rigidity of the Taig is obviously a lot smaller too, but that can mostly be accomodated by correspondingly smaller feeds/speeds, so that didn't concern me as much (as a hobbyist and prototyper, slow speeds are fine).
Finally, ability to do probing and auto tool changes was high on my list, not necessarily because it's required but just because I want to automate that. I really hate touching off parts and measuring tools :)
The Taig is still a great little machine for what it is -- especially at the price I got it -- and I'll be keeping it around for smaller parts or working in plastic, etc. And I do really like the coolant that I installed on it, works great :)
do you think this machine would benefit from having a spindle with more horsepower? (assuming you got that machine bolted down to a heavy table.)
I don't think so actually, or at least not without other modifications. I rarely run into situations where the spindle bogs down, typically keep it very high RPM and high feed rates. So it's usually the steppers that are struggling to keep up. If I were to do it again, I'd order the NEMA 34 steppers, or just get some servos. With servos installed, I suspect the frame itself will be the limiting factor and so more horsepower would just go to waste.
what cleaner u use for cleaning that endmill ?
Love it! :D
Why on Earth would someone program a CNC router to cut through the material quickly, yet the rapid to return to start the next cut goes at 1/3 speed? Am I missing something?
When running a program for the first time, the most likely time you'll find an error is when moving to the next position. Moving to next hole to bore, or the next feature to profile, etc etc. Keeping rapid speed low is a good way to sanity check the program while it's running, since you can quickly mash the emergency stop if it looks like you've messed something up. It's saved my butt a few times, like if the z-height is incorrect (starts to plunge lower than expected) or dimensions were off and it's starting to head towards the vise jaws or another feature on the part.
You _can_ run slower when cutting too, but that's more difficult since the material needs to be cut at a certain speed. So there is less room for slowing things down. But slow rapid speed is an easy safety measure and doesn't add much time.
In a production environment, once the program is watched and it looks good, you'd turn the speeds back up to 100% and walk away, knowing it would do the right thing without being monitored :)
I had a hard time deciding between this and the MillRight mega v plasma/cnc I ended up going with the MillRight (still waiting for it!) and the money I saved I’m cnc’ing my Bridgeport I have all the brackets done now I’m doing the electronics. I went with centroid acorn for the Bridgeport. I’m curious how well your machine does machining on aluminum and the tolerances you can get out of it!
Awesome, goodluck! I've heard good things about the Acorn, I think I would use that myself if I were doing a full DIY. Not a huge fan of Mach 4 and the rest of the ecosystem. I'll be doing more tests in Aluminum soon, including tolerance tests. Want to finish up some accessory work first (tool changer, probes, etc). Cheers!
That looks like an awesome spindle! Monitoring/recording its current should give you a better idea of how hard you're pushing it. It looks like you have the "passively cooled" (no fan or liquid) version, right? How warm/hot does it get?
Good point! The VFD can definitely output the amperage being used, and with a bit of tinkering I think I could route that back through the Avid system for display in Mach. Would be really nice to have a "spindle load" meter on the display like an industrial system.
The spindle has full case pressurization (~40 PSI flowing through the case) which I imagine helps cool it a bit, but yeah there's no fan or water cooling. I haven't noticed it getting very hot but I also haven't had it running for hours at a time yet either. From what I've read elsewhere, folks that use it in production for hours at a time haven't had issues with temperature either.
Why aren't you running the spindle at 24000 RPM to minimize cutting and machine forces? I'm embarrassed to say that I've been using Audacity for decades and wasn't aware of the spectrogram display mode - thanks for pointing that out!
Mainly going off what HSMAdvisor was recommending. I don't yet have a good sense how "off" feed/speed calculators are for the machine, so trying to start around the recommended settings and play from there. Been fiddling with steel (videos to come!) and am finding the recommendations are farther off than aluminum. Definitely going to take a while to figure out the best set of recipes.
@@BreakingTaps I believe that HSMAdvisor is really intended for spindle power limited CNC mills rather than machine force limited CNC routers. CNC routers generally have far less cutting force capability than mills (generally more than an order of magnitude less). Since spindle power (i.e. material removal rate) is proportional to the product of spindle speed and cutter torque (i.e. cutting force), using the highest speed your spindle and endmill support should help you get the most out of your machine.
You can force HSMAdvisor to use that speed to estimate cutting forces for varying cutting parameters (WOC, DOC, and feed-rate), but since that calculation is based on the assumed unit horsepower for the workpiece, it's probably not very accurate. Measuring/monitoring the increase in spindle current (which is proportional to torque) when milling would provide a much more accurate, real time estimate of both cutting force and power
Yeah I think you're right, that all lines up with the various tests I've done so far. I suspect aluminum "mostly worked" because it's aluminum, and anything fast and not welding will work well enough. But in mild steel the assumptions of HSMAdvisor you mentioned are definitely starting to show.
Two questions for you, if you have a sec! The torque/power curve for my spindle looks like this (roughly, i'm told the chart is just an estimate, but it tracks with other similar motors from what I've seen): imgur.com/jLRe01V So does that mean the "sweet spot" is around 18k RPM, since torque starts to drop past that point? Or is it that anything >=18k is good because we're in the constant power range, despite torque dropping as speed increases?
The other question has to do with IPT. Presumably I'll need to increase feeds to keep up with increased RPM, otherwise I'll start rubbing the end mills? E.g. need to keep IPT constant while increasing RPM?
Dude Im ultra curious why you went this route over a more substantial vertical mill, you said you were going to talk about that in a later video, Id love to know that part
Was mostly space and power constraints at the time, plus spouse approval :) Price too. There is basically no used market in my area, so I'd either have to drive/fly far to get a used machine or buy new and have it freighted in. Both looked to be expensive or a hassle, so getting something like the Avid made sense.
I also assumed it would be good enough for what I wanted/needed. It _was_ good enough, but of course you quickly realize you can put more machine to use if you had it. 🙂 It was a good stepping stone for me but I'd like to have a proper VMC some day for sure!
@@BreakingTaps For sure, thanks for that explanation, I have a shapeoko and a diy pm728vt conversion but Im looking to up the anti a little bit with something more rigid, if not more capable than what I have now.
Slow down the acceleration/deceleration and you will be able to run high speeds without the machine jumping off the table.
Oh, doh! That's a really good point. I'll go fiddle with the settings today and see if I can keep it from walking off the table :) Thanks!
Came here from Hackaday, did you think about running your analysis again with a contact microphone mounted on the spindle? Should get you a much cleaner signal.
Agreed! And definitely thinking about it for future experiments. Ditto to accelerometers and force sensors on the spindle/frame in various places. The external mic was a sort of a proof-of-concept to see if it'd work at all, being entirely non-invasive (and already recorded for the video). But now that it seems to work, I would definitely like to explore better methods of collecting the data. I have a much nicer condenser mic I'll probably drag out to the garage to capture better data, rather than the dynamic shotgun mic on my camera.
I also want to toy with better data processing. E.g. The spindle, fogbuster and background compressor sounds could be captured in isolation and subtracted from the "cutting" signal to get a lower noise floor.
Mo
So, does this mean that the motor upgrade on the Rockwell won't get finished???
I was looking forward to seeing how that would work.
It will get finished, don't worry! I'm super curious to know how it will work myself, and if the hobby motors will hold up to the abuse. Just temporarily stalled because I hit a silly issue that was irritating me (and I kept needing the mill for other things and didn't want to change it over). Basically, I wired it all up in an enclosure... but the enclosure is too small. :'( I need to design/print a new enclosure and disassemble all the electronics to put in the new enclosure. It was just demoralizing having to redo all that work so I temporarily shelved it out of frustration :)
Are you using NEMA 23 or NEMA 34?
NEMA 23. I asked Avid their opinion, since I was considering the 34 instead. They recommended the 23 because A) they didn't think the enhanced speed/torque would be useful if the main goal was metal (compared to cutting wood on their larger routers, where more torque at 350+ IPM is useful) and B) ballscrew whip would start to be an issue at higher feeds on the benchtop machines.
@@BreakingTaps Thanks!
So this is about a $9000 setup, yes? That is within the ballpark of a tormach.
Where was the chatter analysis?
I watched the whole video but didn't see any analysis. You showed the load vs no load speed of the spindle and that was about it.
Edit: I don't mind a little clickbait if the title is in the content, but sadly it is not in this video.
The audio analysis is very cool. Are you a musician or audio engineer by any chance? Or maybe a mechanical engineer? Some of the words you used made me wonder. When I saw Avid (and audio analysis) on Hackaday, my mind went to Avid Pro Tools DAW :) Peace
Software engineer :) I've done a fair amount of tangentially related work though (I work on search/analytics software and some of my past major projects have been in the area of time-series analysis work). I've also gone down the rabbit hole of reading "smart spindle" and realtime chatter detection research papers, so some of the terms have rubbed off on me :)
@@BreakingTaps Very cool! I'm an analytical software engineer as well. Plant growth and health most recently, from images. Once the sensors do their job, it's all just data. Glad to meet you :)
Oh awesome, great to meet you too! Is your work with IR/hyper/multi-spectral imaging of plants? Don't know much about it, but from what I've seen it looks like a really neat field. And hard, so much data to process! Or is it more individual plant level, like leaf gas-exchange sensors? Funnily enough, I'm an ex-biologist (mammalian, not plant) so crossover between tech and bio always excites me :)
@@BreakingTaps Wow, that's cool! My only exposure to animals was helping in a shark lab with molecular phylogenetics. Now, I'm building instruments for imaging plants, from basic digital cameras to fancy hyperspectral and infrared cameras. Much of the hardness comes from researchers not having a well defined trait to measure. I'm working on the motion of growing plants, interaction with soil microorganisms, and making a phenotyping appliance that's affordable. Please drop me an email if you think there might be something we could collaborate on, living or metallic :)
Neat stuff! Makes sense re: researchers not having a well-definied trait. Biology is messy, and researchers are fickle :) You have a really interesting channel, subbed and will drop you an email!
Wow thanks for doing this video. I have a build series going on the same machine. Your video will definitely be helpful. Maybe my build videos will help you to.
Cheers.
Ken
Why is your feed rate faster then your linking moves lol.
Insurance against accidents :) Most (user) errors happen during rapids from one cut to the next, so it makes sense to rapid slowly so there is time to hit the e-stop. Once you know a program is good and you've already run it, you can dial up the rapids. Feed rate can't change much because you need to cut the material at a minimum speed/feed for it to cut well ;)
👏👏👏👏
It looks like they've changed to rack and pinion on the X and Y axes! :-(
Just depends on model! The "Pro" versions have ballscrews, the non-pro (I think) are all rack/pinion.
@@BreakingTaps Yup - thanks! I was looking at the Pro rather than the Benchtop Pro.
I enjoyed your test here. If you would like help on anything send me a message. I currently run a small Aluminum machine shop in Utah. Just a heads up your facing finish would be much better with a all to center fluted endmill. i like the YG alupower HPC for this.
Might take up up on that offer in the future, thanks! I picked up a YG-1 Alupower recently and plan to test it in the near future, I've read good things about it. :)
Need to stop recutting chips. That is killing your surface finish. Get the chips out of the path of the cutter!
Yeah agreed, I noticed that while re-watching the footage (especially deeper in the pocket at certain angles). I'm tempted to try the fogbuster with just air and no coolant. The coolant seemed to encourage chips sticking to side walls when they might have otherwise blown out. I did dial up the air pressure, but the fogbuster is only rated for so much.
Pretty shit machine tbh. But then again I'm used to 6 figure machines.
Why did you choose the S30C over S30 or S30F?