You should always use a granite surface plate desk, even in the kitchen! I sleep on a granite surface plate to make sure I get the most precise dreams ever, to within 3 microns.
@@reps EVERYTHING IS CROOKED! REALITY IS POISON! *LAMBS TO THE COSMIC SLAUGHTER*! I've seen that scene at least 5 times this week because it keeps popping up in TH-cam for me.
Hey Marco, the Steinmeyer ball screws are induction hardened as well (only the threads up to a few mm depth into the material). They also start as rolled part but then they get precision ground to fit to the pitch within tolerance (as low as a few µm). Of course the abrasive grinding will heat up the thread but they balance the cutting depth to control the soft-annealing. So no worries - they will hold up fine.
@@operator8014 Your in luck I happen to be in charge of the cnc department of Infinity Drain. We have brand new AMS Loader from Amada paired to our 4kw Laser. The Loader uses just that to tell where the pallet traverser is and to also measure sheet material thickness.
@@operator8014 already done. We use those at my work. Picometer accuracy. It's a pain though, as in Australia this accuracy encoder systems requires defence approval. We had one system fail, and I had to arrange approval from ministry of defence to be able to send it back to the manufacturer for repair.
@@brianjensen5200 Does everything run in a temperature controlled oil bath? I saw a video from Mitutoyo and that's how they keep their master leadscrew at a fixed temperature when screw cutting.
@@evildrome no our motion stages are in free air. We keep the technical hall under 1K temperature deviation and encoder feedback loop controls the rest
bearing blocks on linear rails must have a certain amount of preload to give you accuracy and rigidity under load. that's why a new linear blocks don't slide freely and have some amount of friction. (and that used rails from industrial machines slides freely just because they are WORN OUT and lost their preload. that's why they was removed and replaced with new ones)
Sounds all correct! The banggood ones however have an uneven friction and seem to get caught on some high spots. And since hobby stuff is the only area where you would consider used rails anyway I think even the most worn-out ones will be easily good enough? Maybe I can find a way to actually measure my worn-out Rexroths ...
@@reps I've recently bought some used miniature Bosch rexroth linear guides (size 15, 4 bearings, 2x330mm rails) from ebay, alongside 500 steel bearing balls. Had to replace the old balls and even replaced the new ones multiple times until I found a "good batch" of those cheap balls that were running smoothly with slight preload. All in all this costs me 120€ compared to new ones from Bosch for over 500€
@@michaelkelly3158 the original Bosch rails are very similar to hiwin rails. You'd have to compare lifetime, friction or other technical properties in a specific use case. Just saying "these are better" is impossible. As long as you get new brand name rails (hiwin, Bosch, SKF, Schneeberger,...) you should be good to go for any hobby project.
I have some cheap Chinese linear rails in a 20mm size. They run kind of rough but don't seem like it's specific high spots. Do you think taking them apart and cleaning would help smooth out the motion? I kind of suspect it's small particles in side the carriages are causing it to stick
I know it's been a while but at my job we have what is essentially a large pair of calipers to calibrate cinema cameras. In order to discover and correct for long distance accuracy we use the distance traveled and revolutions of the motor. It uses a belt and pulley along with a very nice absolute encoder. We note the error (difference) of where it ended up vs the target and use the encoder resolution as well as the revolutions of the pulley. So if the machine was 0.015" off we would use that distance multiplied by the ticks/mm of the encoder (a constant), divided by the number of revolutions made during the move. This would give us a long range error correction value that the machine can compensate with. Essentially the number of encoder ticks it should increase or decrease by with every revolution of the pulley. I believe a similar calibration technique could be used with a ball screw mechanism.
6:04 When a ballscrew shaft is ground, the metal can already be very hard prior to grinding and the resulting screw can be far superior than a ballscrew that gets some hardening in its roll forming process.
I think it's important to note that the final hardening comes just before grinding (which happens after rolling in the case of rolled ball screws) and so the effect of surface hardening is minimal compared to the higher precision and therefore tolerances of purely ground ball screws.
I'm working with high precision CNC mills and we use 30mm diameter ballscrews with 16mm pitch. But the bearing blocks are something different. They cost around 1500€/piece and weight around 6-8kg I think. But we use much more beefier motors. 2.2kW per ballscrew. That crackling noise in that nut hurted very badly.
Here in Vietnam we can get those linear rail and ball screw second hand for dirt cheap. They are often disassembled from old machines. Just got a pair of unused IKO LWA15 rails and NSK 1602 ball screw for like $40. $200 can get you a nice 400w servo with driver from Japanese brand like Yakaswa.
Always interesting ! In my experience, too much precision is useless, because, at the end, the total deflexion of the structure, leadscrews, the motor head, the tool, etc. is very significant, but, of course, "too much precision" is better than "meh!". I'm saving money for the parts of my next CNC router. Thanks for sharing !
Depends a lot on the application, but most companies will replace ball screws when it starts causing problems. If it's in a machine that NEEDS to run without a hitch, consumables will generally get replaced or at least checked for wear on a predetermined schedule. I used to work for a company that would always have a maintenance crew scheduled for holidays so they could do that sort of thing without creating excessive downtime.
The standard operating procedure in North America is: A more and more frequent motor overload error, operators get frustrated enough to complain, so maintenance come and grease it, same thing happens a few more times, maintenance get frustrated with greasing it every damn shift, and then they tear it open to find out it’s completely fucked, ball bearings all over the inside of the machine, parts are ordered in a hurry, often resorting to physically sending a guy to the distributor to get the parts faster, because the line is completely down and hemorrhaging money. You guys have a better system in Germany?
FYI: You should look into Klipper if you are interested in interfacing a regular threaded OS with a realtime OS. Klipper is a 3d printer software that does all the heavy lifting on a computer (usually a Pi) and then uses one or more microcontrollers to schedule and execute commands.
@@reps I honestly think it's the way forward for 3d printing as opposed to using STM32 based boards with webUI addons like the Duet. Plus it can be infinitely scalable. Want to have two Y-axis gauntries with two X-axis carriages with two independent extruders each? Slap a few cheap reprap boards together and hook them up to a Pi and install Klipper on everything.
The qualifying test for becoming a marine artificer in the royal navy in the 1970s was to build a miniature metal lathe. By hand from scratch, to a maximum accuracy of so many thou. Apparently creating those jack screws and whatever they made to run along it is tricky.
Yep, turning the nut is tricky. It takes a while turning it back and forth to be ‘happy’ again. For a while it was trying to force out the plastic seal. At one end the plastic seal is held in by two tiny grub screws.
I know this is off topic, but what is your planned controller setup? I know you have some LinuxCNC experience, but have you looked at the MachineKit? It’s a pretty cool fork, far more modern in architecture and much more modular ecosystem. You can find out more at machinekit.io. I have no connection to the project, just believe that is superior and worth a look.
Finally, a new video!! I was getting really tired of watching reruns and can't seem to get my fix anywhere else! I just got my MK3 Extended last week and already started my collection of cheap Chinese knock-offs….I'm in too deep already and obsession/fascination with Micro Reps isn't going away..one video every couple months just won't suffice!
A simple laser interferometer can be used to measure accuracy over long distances. All you have to do is to program maybe an arduino to count how many times the interference pattern cycles. The accuracy may not be great because you can't really properly calibrate it but it should be super precise, as long as you don't loose count.
To measure long distances you could attach levers to the encoder. It's the same technique that people used back in the day to convert large movements into small ones for precise machining. You'd be trading micrometer precision for mm or cm precision though.
Your drift is likely caused by thermal expansion. You'll get about 10um/m change if your ball screw changes temperature by 1 degree. You can confirm this by heating the ball screw with your hand.
Nice to see some SMC stuff in the video. I have never used anything from them other than pneumatics but i got a tour around one factorie and they do make very high quality stuff....
Hello, it's my first comment here.. First of all love your channel...So I am writing aboute the "el chepo chinese linear rails" i've bought some more then a year ago.. Frustrated over the laggy performance, etc... Solution disassemble the carriage submerge in alcohol (to clean some wierd oils inside) reassemble, greas up... It worked for me... Still going strong after a year or so of modernet to heavy use..
That's some quality hardware! I'm only worried a bit about this Z axis, it looks going to be 2-3 times longer than spacing between rails, this will amplify load on the rails, and then amplify deflection at the tip of an end mill even more. I might be overthinking it but i'd be tempted to make something like Okuma double column machining center.
Edit: Gets addressed in the 'Shaft' video. Don't CNC machines usually control their travel using DRO? If built that way, the lead precision of a ball screw doesn't matter as much as the backlash. (I think "Umkehrspiel" called backlash in English.) Probably because the ball screw, because it is warmed by friction and cannot be protected as good, is much more susceptible to thermal expansion than a glass scale. Possible that a good controller even uses both to achieve sub micron positioning.
You might want to consider adding a DRO to your CNC machine. You can get them cheaply on ebay up to ~2m+ lengths and its probably easier to make the glass ruler in the DRO nearly perfect than it is to make a perfect ball screw. You could use the DRO for absolute position in a feedback loop with the servo motor.
Anxiety when using expensive ball screws? We use them in a foundry enviroment where there's sand everywhere, they trash out about once a year. Also, you know when linear guides are replaced? When the carriages explode from all the dirt and grease inside them and the rails look like an apprentice filed them out of one solid bar and got bored halfway through.
you can get away with using thinner screws by tensioning them slightly, either between 2 angular contact bearings (use roller ones, preloading is hard on bearings that are mostly stationary), or rigid and rotating the nut. I like the chasing micrometers, but harmonised stiffness, adequate stiffness and vibration dampening is where it's at with diy cnc methinks. still, i wouldn't say no to those Steinmeyers either:)
For measuring longer travel accuracy, a fairly cheap method would be to get one of the cheap 12" (~240mm) digital calipers. Note that it has to be one of the cheap ones, but for some reason nearly all of the cheaper digital calipers have serial output, and there are tons of tutorials for interfacing with them. Resolution isn't spectacular (+-.05"), but they are much cheaper than something like a lathe bed axis DRO and would likely provide some good data (at least for repeatability testing). At least in the US, they are normally avaliable for ~15USD, I would imagine they aren't much different in the EU. Plus, then you don't have to feel bad about hacking a nice starrett or mitutoyo instrument!
What's the amount of the slow drift? Steel has a coefficient of linear expansion of ~10-12 microns/meter/degrees c, so temperature expansion vs the desk should certainly shown up on your micron scale.
I am also German but have been living in the US for over 40yrs. I am looking to build a reasonably priced desktop machine to route/mill parts for my hobby. Why is it that I am also obsessed with precision and worry about thermal expansion of materials? I guess it’s in my genes? 🤪
Hello Marco. At 3:38 you show us a voice coil similar in construction with voice coil from hard disk. Please explain me how it remain fixed on the position. On hard disk if you apply +and- the voice coil goes to the end. If you reverse polarity - and+ it goes to the oposite end. The question is what kind of electricity makes this voice coil to remain fixed on a track on any other position betwen right end and left end ? This is interesting to put an osciloscope on the voice coil on this device and show us how can be stand still with incredible acuracy. Nobody know to answer me this question for several years. Thank you.
Hallo Marco, Ich habe mir auch die BK20 gekauft auf aliexpress, und sehe da sind ganz normale lager verbaut. Denkst du das die irgendwie ok isnd für den konstanten axialen druck auf zwei seiten? Ich brauche es fuur meine Fräsmaschine (machinable area = 2440 x 1220). Mein setup hat zwei 2525 ball screw spindles für die lange Y achsee, und eine 1616 für die X Achse. Denkst du es würde sich lohnen gleich jetzt vor der montage um zu rüsten, oder erst mal schauen was die billigteile hergeben? Danke für einen Tip!
Cannot imagine that those ground screws are not first rolled to the general shape, then cut. Would seem to be the best way to do it for many reasons. You gain the extra free hardness and also wear your grinding machine less.
Maybe you could try Schneeberger MiniScale Plus rails. They have a linear encoder integrated into the rail so would let you do a lot of tests with no worries about aligning the encoder to the rail. The resolution is very good at 0.1 μm and the encoders have good length accuracy at +/- 5 μm over 1m.
Tiny fpga? It's time that you discovered the Parallax Propeller. The new P2 is mind blowing. Any/all of its "smart pins" can decode quadrature signals at sys clock rate (180 to 300+ MHz) . Eight independent processors can be programmed in C/BASIC/FORTH/ASM and Python is on the way.
Less vibration since a long screw would flex while rotating. a nut that rotate around a static screw would not make the screw wobble. You could use thinner screw and improve over longer distance
I'd recommend using a double ball set up not, a single.. take those plastic seals out & degrease the bearing balls and you'll find a good amount of play. The grease & plastic seals mask the ball screws axial play.
I love ball screws, but everytime I see one my mind flashes back to the $500 Bosch-Rexroth paperweight on my desk resulting from the spindle spinning off the shaft and onto a dirty floor.
Have you ever thought about about a flashlight or a headlight with power leds and minimal heatsink and Peltier Cells? A cool but power hungry light. Full of Watts and coolness...
"These are the polar opposite to the Chinese ones" Aaaaand of course they are german. Germany, the place where a a simple thread can be overengineered so it's super expensive but still being worth it lol.
If I pick a handful of TH-camrs about 3D printing (you being one) and I can get us all all in a conference call with video if possible so “talking with hands to help with visuals” will help. .... And in this live group call we form a “company/Organization” for my truly revolutionary design of a printer to be sold as a product. I’m not asking for money, I am a over the road semi truck driver, I make plenty of money for everything needed, I just don’t have the time to do anything other than talk really. And I live out of my Simi Truck. So I don’t have build space. Most of my design is set in stone, but there are multiple ways to do different factors of my design, can we all discuss it and maybe add tweaks to make it better. But what we all discuss as a collective does not leave the project and stays private until it is ready for sale, can we all agree on this and talk about
At my old job they did replace them after operating hours. And yes this is done by a lot of companies as they said. It's cheaper to replace a rails + carriages before it fails rather than failing during a cnc job which will cost a lot more like failed material and downtime hours which can cost thousands of dollars.
@@Underp4ntz_Gaming_Channel It's also usually done when "you're already in there." If you've got a machine torn apart, and it's halfway to the replacement time, it's easier and cheaper to do it all at once.
You should always use a granite surface plate desk, even in the kitchen! I sleep on a granite surface plate to make sure I get the most precise dreams ever, to within 3 microns.
< Insert Rick & Morty True Level Scene here >
It also straightens out the back
@@DeDeNoM ...Or make a "C" with your column, if you sleep over your side. Cheers !
@@DeDeNoM Raises hell with the "roids" tho!
@@reps EVERYTHING IS CROOKED! REALITY IS POISON! *LAMBS TO THE COSMIC SLAUGHTER*!
I've seen that scene at least 5 times this week because it keeps popping up in TH-cam for me.
Hey Marco, the Steinmeyer ball screws are induction hardened as well (only the threads up to a few mm depth into the material). They also start as rolled part but then they get precision ground to fit to the pitch within tolerance (as low as a few µm). Of course the abrasive grinding will heat up the thread but they balance the cutting depth to control the soft-annealing. So no worries - they will hold up fine.
When you go below 1um accuracy you start getting weird effects like the heat from your body causing thermal expansion.
I want to see someone use laser interferometry to measure the position to kick off the next generation of hyper accurate making.
@@operator8014 Your in luck I happen to be in charge of the cnc department of Infinity Drain. We have brand new AMS Loader from Amada paired to our 4kw Laser. The Loader uses just that to tell where the pallet traverser is and to also measure sheet material thickness.
@@operator8014 already done. We use those at my work. Picometer accuracy. It's a pain though, as in Australia this accuracy encoder systems requires defence approval. We had one system fail, and I had to arrange approval from ministry of defence to be able to send it back to the manufacturer for repair.
@@brianjensen5200 Does everything run in a temperature controlled oil bath? I saw a video from Mitutoyo and that's how they keep their master leadscrew at a fixed temperature when screw cutting.
@@evildrome no our motion stages are in free air. We keep the technical hall under 1K temperature deviation and encoder feedback loop controls the rest
bearing blocks on linear rails must have a certain amount of preload to give you accuracy and rigidity under load.
that's why a new linear blocks don't slide freely and have some amount of friction.
(and that used rails from industrial machines slides freely just because they are WORN OUT and lost their preload. that's why they was removed and replaced with new ones)
Sounds all correct! The banggood ones however have an uneven friction and seem to get caught on some high spots. And since hobby stuff is the only area where you would consider used rails anyway I think even the most worn-out ones will be easily good enough? Maybe I can find a way to actually measure my worn-out Rexroths ...
@@reps I've recently bought some used miniature Bosch rexroth linear guides (size 15, 4 bearings, 2x330mm rails) from ebay, alongside 500 steel bearing balls. Had to replace the old balls and even replaced the new ones multiple times until I found a "good batch" of those cheap balls that were running smoothly with slight preload. All in all this costs me 120€ compared to new ones from Bosch for over 500€
@@mariusb6035 How do these Bosch rails compare to Hiwin rails? Only ever heard of Hiwin rails being used in high end CNC applications.
@@michaelkelly3158 the original Bosch rails are very similar to hiwin rails. You'd have to compare lifetime, friction or other technical properties in a specific use case. Just saying "these are better" is impossible. As long as you get new brand name rails (hiwin, Bosch, SKF, Schneeberger,...) you should be good to go for any hobby project.
I have some cheap Chinese linear rails in a 20mm size. They run kind of rough but don't seem like it's specific high spots. Do you think taking them apart and cleaning would help smooth out the motion? I kind of suspect it's small particles in side the carriages are causing it to stick
That's going to be one hell of a PCB etcher
the worlds most expensive dickbutt plotter
Can’t wait to see the 1:100000 scale city that he etches.
I know it's been a while but at my job we have what is essentially a large pair of calipers to calibrate cinema cameras. In order to discover and correct for long distance accuracy we use the distance traveled and revolutions of the motor. It uses a belt and pulley along with a very nice absolute encoder. We note the error (difference) of where it ended up vs the target and use the encoder resolution as well as the revolutions of the pulley. So if the machine was 0.015" off we would use that distance multiplied by the ticks/mm of the encoder (a constant), divided by the number of revolutions made during the move. This would give us a long range error correction value that the machine can compensate with. Essentially the number of encoder ticks it should increase or decrease by with every revolution of the pulley. I believe a similar calibration technique could be used with a ball screw mechanism.
6:04
When a ballscrew shaft is ground, the metal can already be very hard prior to grinding and the resulting screw can be far superior than a ballscrew that gets some hardening in its roll forming process.
Marco we see so much stuff in your videos we cant find nowhere else..TY.
And cant afford haha!
I just love how you worked that ad right in there. Smooth as butter i didn’t see it coming and then BAMM like a freight train.
I understand NOTHING but I can't stop watching.
Precision engineering with German precision dry humour. God I love this guy!
I think it's important to note that the final hardening comes just before grinding (which happens after rolling in the case of rolled ball screws) and so the effect of surface hardening is minimal compared to the higher precision and therefore tolerances of purely ground ball screws.
Made with carbon metal not a trace of cabon wood LMAO
I'm working with high precision CNC mills and we use 30mm diameter ballscrews with 16mm pitch. But the bearing blocks are something different. They cost around 1500€/piece and weight around 6-8kg I think. But we use much more beefier motors. 2.2kW per ballscrew.
That crackling noise in that nut hurted very badly.
Just gonna change the ballscrews at a customers machine tomorrow
Here in Vietnam we can get those linear rail and ball screw second hand for dirt cheap. They are often disassembled from old machines. Just got a pair of unused IKO LWA15 rails and NSK 1602 ball screw for like $40. $200 can get you a nice 400w servo with driver from Japanese brand like Yakaswa.
Always interesting ! In my experience, too much precision is useless, because, at the end, the total deflexion of the structure, leadscrews, the motor head, the tool, etc. is very significant, but, of course, "too much precision" is better than "meh!". I'm saving money for the parts of my next CNC router. Thanks for sharing !
Oh finally a new video. 😀
Just found this channel... Been binge watching since : )
Your videos are usually pretty funny, but that joke at the end about rails getting replaced preemptively in industrial settings was hilarious.
Bigger companies do this all the time. Regular maintenance is important when your Nestle and bottling 10's of thousands of water bottles a day.
@@adisharr
I am not sure if you have ever replaced a ball screw, but it doesn't sound like it.
@@SystemsPlanet Most industrial machines (outside Germany at least) get run to death, rarely get the love or greasing they should.
Depends a lot on the application, but most companies will replace ball screws when it starts causing problems. If it's in a machine that NEEDS to run without a hitch, consumables will generally get replaced or at least checked for wear on a predetermined schedule. I used to work for a company that would always have a maintenance crew scheduled for holidays so they could do that sort of thing without creating excessive downtime.
The standard operating procedure in North America is: A more and more frequent motor overload error, operators get frustrated enough to complain, so maintenance come and grease it, same thing happens a few more times, maintenance get frustrated with greasing it every damn shift, and then they tear it open to find out it’s completely fucked, ball bearings all over the inside of the machine, parts are ordered in a hurry, often resorting to physically sending a guy to the distributor to get the parts faster, because the line is completely down and hemorrhaging money. You guys have a better system in Germany?
FYI: You should look into Klipper if you are interested in interfacing a regular threaded OS with a realtime OS. Klipper is a 3d printer software that does all the heavy lifting on a computer (usually a Pi) and then uses one or more microcontrollers to schedule and execute commands.
Yeah, cool project. I think Cetus is also based on that principle
@@reps I honestly think it's the way forward for 3d printing as opposed to using STM32 based boards with webUI addons like the Duet. Plus it can be infinitely scalable. Want to have two Y-axis gauntries with two X-axis carriages with two independent extruders each? Slap a few cheap reprap boards together and hook them up to a Pi and install Klipper on everything.
Thank you so much. You saved me a lot of time. Going to replace my end bearings and get that Tiny FPGA kit now! Excellent.
Had no idea FPGAs could be so adorable. I need to dust off some micro-controller unworthy projects and exchange some financial details with Elektor.
Elektor: "only 0 left"
Me: "oh great... fuck my life."
My wallet: *quiet sigh of relief*
I can confirm that in 2025, FPGAs will go beying Kittens in the minds of serial-video-watchers.
The qualifying test for becoming a marine artificer in the royal navy in the 1970s was to build a miniature metal lathe.
By hand from scratch, to a maximum accuracy of so many thou.
Apparently creating those jack screws and whatever they made to run along it is tricky.
Yep, turning the nut is tricky. It takes a while turning it back and forth to be ‘happy’ again. For a while it was trying to force out the plastic seal. At one end the plastic seal is held in by two tiny grub screws.
FYI: micrometre = μm, micrometer = an instrument for measuring small distances
Thanks
nerd :)
@@reps fair, fair
I know this is off topic, but what is your planned controller setup? I know you have some LinuxCNC experience, but have you looked at the MachineKit? It’s a pretty cool fork, far more modern in architecture and much more modular ecosystem. You can find out more at machinekit.io.
I have no connection to the project, just believe that is superior and worth a look.
Finally, a new video!! I was getting really tired of watching reruns and can't seem to get my fix anywhere else! I just got my MK3 Extended last week and already started my collection of cheap Chinese knock-offs….I'm in too deep already and obsession/fascination with Micro Reps isn't going away..one video every couple months just won't suffice!
Came for the precision, subscribed for the humour! Nice and dry, just how I like it.
best video ever. inspired me to study engineering and will build a machine.
Linear rails have different amount of preload and various combinations of seals that will affect how easy they slide.
I got my first two 12mm ballscrews today from china, 600mm long with a straight nut, I hope they are not bent
Just because linear rails don't move freely, does not mean they are bad. They could just be preloaded, wich can be beneficial in some applications.
German precision oozing out of this video. I had a German accent for an hour after watching this. Love this channel! Subscribed!
Precision wood block table
A simple laser interferometer can be used to measure accuracy over long distances. All you have to do is to program maybe an arduino to count how many times the interference pattern cycles. The accuracy may not be great because you can't really properly calibrate it but it should be super precise, as long as you don't loose count.
That tiny FPGA thing... I didn't knew such a thing existed! That's the size of an Arduino nano!
thank you for the lesson in angular contact bearings
YESSS! I freaking LOVE your videos!!!
To measure long distances you could attach levers to the encoder. It's the same technique that people used back in the day to convert large movements into small ones for precise machining. You'd be trading micrometer precision for mm or cm precision though.
Your drift is likely caused by thermal expansion. You'll get about 10um/m change if your ball screw changes temperature by 1 degree. You can confirm this by heating the ball screw with your hand.
Your videos are awesome, I can't believe I hadn't discovered these sooner, bravo.
For testing longer range you could Mount an standing plate and Drive the carige against it..
I'm really glad I'm subscribed to this channel. I feel like the humor here is within the same vein as This Old Tony.
You are such a knowledgeable and funny guy.
Nice to see some SMC stuff in the video. I have never used anything from them other than pneumatics but i got a tour around one factorie and they do make very high quality stuff....
Any "Stuff Made Here" fam?
Looking up what screw balls are? 😂 Same
@@creativeoutlet3148 Ayyyy
@@creativeoutlet3148 lol same
No
That's why I'm here lol
Hello, it's my first comment here.. First of all love your channel...So I am writing aboute the "el chepo chinese linear rails" i've bought some more then a year ago.. Frustrated over the laggy performance, etc... Solution disassemble the carriage submerge in alcohol (to clean some wierd oils inside) reassemble, greas up... It worked for me... Still going strong after a year or so of modernet to heavy use..
That's some quality hardware! I'm only worried a bit about this Z axis, it looks going to be 2-3 times longer than spacing between rails, this will amplify load on the rails, and then amplify deflection at the tip of an end mill even more. I might be overthinking it but i'd be tempted to make something like Okuma double column machining center.
Valid point!
Edit: Gets addressed in the 'Shaft' video.
Don't CNC machines usually control their travel using DRO? If built that way, the lead precision of a ball screw doesn't matter as much as the backlash. (I think "Umkehrspiel" called backlash in English.) Probably because the ball screw, because it is warmed by friction and cannot be protected as good, is much more susceptible to thermal expansion than a glass scale. Possible that a good controller even uses both to achieve sub micron positioning.
I could listen to you say "steinmeyer ballscrews" over and over on repeat endlessly. I wish I had your voice, fantastic vid
That's how we Germans say it ... 😀
You might want to consider adding a DRO to your CNC machine. You can get them cheaply on ebay up to ~2m+ lengths and its probably easier to make the glass ruler in the DRO nearly perfect than it is to make a perfect ball screw. You could use the DRO for absolute position in a feedback loop with the servo motor.
Very good. I like your sense of humor too.
I love the shot at 3:37 where you can see a die with all the traces allong it.
There is a recent video how to clean and "upgrade" cheap linear bearing, like you are using at 7:45 the difference is outstanding
Anxiety when using expensive ball screws? We use them in a foundry enviroment where there's sand everywhere, they trash out about once a year.
Also, you know when linear guides are replaced? When the carriages explode from all the dirt and grease inside them and the rails look like an apprentice filed them out of one solid bar and got bored halfway through.
Funny , ohh the humor , most excellent content. This is exactly the info I was looking for as I'm currently trying to repair a TL-1 Lathe : )
The lubrication sounds are very inspiring. Fruits & vegetables are good for you, Marco!
well the good thing about carbon wood is that its actuall pretty good at absorbing vibration
I actually shuddered when you rotated the ballscrew with abrasive in it
I used to work with these in laser trimmers and C02 lasers. The C02 had a 200 pound polished slab of granite that everything was mounted on
Nice, I just replaced a Z-axis ball screw on a VM-3. Fun times.
Oh, it seems I got lost in the CNC TH-cam rabbit hole. That said, now THIS is some serious DIY CNC shit.
you can get away with using thinner screws by tensioning them slightly, either between 2 angular contact bearings (use roller ones, preloading is hard on bearings that are mostly stationary), or rigid and rotating the nut. I like the chasing micrometers, but harmonised stiffness, adequate stiffness and vibration dampening is where it's at with diy cnc methinks. still, i wouldn't say no to those Steinmeyers either:)
Finally a new video! I'm so happy to see your CNC machine coming together😍 can't wait to see those beauty in action
For measuring longer travel accuracy, a fairly cheap method would be to get one of the cheap 12" (~240mm) digital calipers. Note that it has to be one of the cheap ones, but for some reason nearly all of the cheaper digital calipers have serial output, and there are tons of tutorials for interfacing with them. Resolution isn't spectacular (+-.05"), but they are much cheaper than something like a lathe bed axis DRO and would likely provide some good data (at least for repeatability testing). At least in the US, they are normally avaliable for ~15USD, I would imagine they aren't much different in the EU. Plus, then you don't have to feel bad about hacking a nice starrett or mitutoyo instrument!
dunno about the tiny fpga, but you can use the blue pill stm32 fr a cheaper/faster upgrade to the arduino
Yes, the $1 STM32F103C8T6 with true USB 2.0 at 72mhz ought to be enough. Maybe?
What's the amount of the slow drift? Steel has a coefficient of linear expansion of ~10-12 microns/meter/degrees c, so temperature expansion vs the desk should certainly shown up on your micron scale.
I am also German but have been living in the US for over 40yrs. I am looking to build a reasonably priced desktop machine to route/mill parts for my hobby. Why is it that I am also obsessed with precision and worry about thermal expansion of materials? I guess it’s in my genes? 🤪
One way I could think of measuring long distance movement, would be measuring rotation shaft movement vs fixed point via laser distance sensor.
Thank you for the unfiltered sharing of information. It is quite refreshing.
Classic case of a "spec creep"
Jep
Know that feeling
@@johnsnow8186 en.wikipedia.org/wiki/Scope_creep
3:10 when your hardware is so precise that your nut has bolts inside it
A new video from Mr Reps makes my day.
Good humour and technical knowledge. Subbed.
So from Steinmeyer and Hiwin to Skillshare, you must have made loads in royalties. Then TH-cam too. Fantastic, share some my way.
Excellent video about a thing that gets ignored quite often.
More, please sir! One of my favorite channels.
Genuine Hiwin rails take force to move because the fit is tight - I have some and they move just like those.
Hello Marco. At 3:38 you show us a voice coil similar in construction with voice coil from hard disk. Please explain me how it remain fixed on the position. On hard disk if you apply +and- the voice coil goes to the end. If you reverse polarity - and+ it goes to the oposite end. The question is what kind of electricity makes this voice coil to remain fixed on a track on any other position betwen right end and left end ? This is interesting to put an osciloscope on the voice coil on this device and show us how can be stand still with incredible acuracy. Nobody know to answer me this question for several years. Thank you.
Hallo Marco, Ich habe mir auch die BK20 gekauft auf aliexpress, und sehe da sind ganz normale lager verbaut. Denkst du das die irgendwie ok isnd für den konstanten axialen druck auf zwei seiten? Ich brauche es fuur meine Fräsmaschine (machinable area = 2440 x 1220). Mein setup hat zwei 2525 ball screw spindles für die lange Y achsee, und eine 1616 für die X Achse. Denkst du es würde sich lohnen gleich jetzt vor der montage um zu rüsten, oder erst mal schauen was die billigteile hergeben? Danke für einen Tip!
Cannot imagine that those ground screws are not first rolled to the general shape, then cut. Would seem to be the best way to do it for many reasons. You gain the extra free hardness and also wear your grinding machine less.
Excellent breakdown and explanation with valuable tips and tricks for the aspiring CNC machine builder. Thank you. Subscribed.
Well I'm screwed with my M6 threaded rod and 5V stepper motors lmao
Maybe you could try Schneeberger MiniScale Plus rails. They have a linear encoder integrated into the rail so would let you do a lot of tests with no worries about aligning the encoder to the rail. The resolution is very good at 0.1 μm and the encoders have good length accuracy at +/- 5 μm over 1m.
Yes please! :)
Tiny fpga? It's time that you discovered the Parallax Propeller. The new P2 is mind blowing. Any/all of its "smart pins" can decode quadrature signals at sys clock rate (180 to 300+ MHz) . Eight independent processors can be programmed in C/BASIC/FORTH/ASM and Python is on the way.
I did not know that Germans knew about humour LOL
ONE OF THE BEST TECHNICAL CHANNELS ON YT (and funny too)
Buy the cheap ones, install the standing replacement bearings that are made with rolling truncated cones.
A German tech nerd channel with a sense of humor?
Subbed!
It's not that unusual, check out bitluni.
Initially I thought with heck this guy is doing with less precise ball screws for VMC. I have used THK several Times.
Marco.. You should check out rotating nut instead of rotating screw
Why?
Less vibration since a long screw would flex while rotating. a nut that rotate around a static screw would not make the screw wobble. You could use thinner screw and improve over longer distance
th-cam.com/video/4unEngqU09A/w-d-xo.html
th-cam.com/video/WXv5breXGLw/w-d-xo.html
@@IBICO74 mkay, interesting! but I am already too far on the 'normal' route ... maybe later
I'd recommend using a double ball set up not, a single.. take those plastic seals out & degrease the bearing balls and you'll find a good amount of play. The grease & plastic seals mask the ball screws axial play.
You need to give your narrator a raise in wages. What a great voice.
You worry about the possible error induced by the table? Are you building a gravity wave detector? o.0
I love ball screws, but everytime I see one my mind flashes back to the $500 Bosch-Rexroth paperweight on my desk resulting from the spindle spinning off the shaft and onto a dirty floor.
Have you ever thought about about a flashlight or a headlight with power leds and minimal heatsink and Peltier Cells? A cool but power hungry light. Full of Watts and coolness...
Whoa that SMC card motor is just so cool!
Liquid nitrogen on an ATMega was the funniest niche shit ever. Thank you sir.
"These are the polar opposite to the Chinese ones"
Aaaaand of course they are german. Germany, the place where a a simple thread can be overengineered so it's super expensive but still being worth it lol.
Clearly you have never visited Japan...
If I pick a handful of TH-camrs about 3D printing (you being one) and I can get us all all in a conference call with video if possible so “talking with hands to help with visuals” will help. ....
And in this live group call we form a “company/Organization” for my truly revolutionary design of a printer to be sold as a product. I’m not asking for money, I am a over the road semi truck driver, I make plenty of money for everything needed, I just don’t have the time to do anything other than talk really. And I live out of my Simi Truck. So I don’t have build space.
Most of my design is set in stone, but there are multiple ways to do different factors of my design, can we all discuss it and maybe add tweaks to make it better. But what we all discuss as a collective does not leave the project and stays private until it is ready for sale, can we all agree on this and talk about
7:48 "they get replaced based on milage or time" soooo, i work as cnc lathe operator, and no they do not :D
That depends on the shop, where I work they get replaced BEFORE they fail, since they´re cheaper to replace than most of the products being machined!
At my old job they did replace them after operating hours. And yes this is done by a lot of companies as they said. It's cheaper to replace a rails + carriages before it fails rather than failing during a cnc job which will cost a lot more like failed material and downtime hours which can cost thousands of dollars.
@@Underp4ntz_Gaming_Channel It's also usually done when "you're already in there." If you've got a machine torn apart, and it's halfway to the replacement time, it's easier and cheaper to do it all at once.
to get bearings in easley use vaseline they stick to the bearing or magnetize them
That tiny fpga looks f*n cool !
Only you could make ball bearings and vertical axis travel sound humorous.
Nicely done.
You can use your transducer you bought to shake parts from plastics. 4:22 :D