Everyone go and subscribe to Jack, his side of the story is coming soon. I'll warn you though, he's literally a wizard. www.youtube.com/@jackfromscratch
i just order a similar servo except the 750watt nema34 version to run with clough's leadscrew project but was wondering does the driver require a separate dc power supply or just 220v ? hope its not as screamy other wise i would have to try a nema24 stepper or order a nema34.
I feel like you need a PULL OUT function to go with NUT mode. For those times you want the thread to lift off at the end rather than cut a channel/groove
@@AndySomogyi Not completely left, but he seldom posts anymore. I get the point though, and agree that these two have helped fill a void. I miss CaLem too.
@@bogdandrozdov6669 The transition of his upload schedule. Used to be multiple uploads a month to now a few a year. These channels definitely fill that void.
hypnotic watching a cnc cut a tapered square thread at full tilt... fecking hilarious watching the resident shop "fool" open the door too early and get soaked with flood coolant... on the first item of the day... "no change of clothes for you!"
Ok, now I'm starting to get really suspicious. Teaching a lathe to think is exactly the kind of thing an engineer would do. Especially without stopping to think if you should teach a lathe to think.
@@junkname9983 Haha quit lying. Everyone knows a true engineer will gladly spend 3 weeks designing a process/tool to save themselves 15 minutes of work every 2 months.
"Things were progressing at the speed of time" - talk about a double-take. Thank you for that one. Others have already thanked you for the other great lines.
The servo drive is likely designed to work at a 8khz switching frequency. It uses DC to pretend like it has 3phase AC. in doing this it switches some solid state switching devices, like mosfets or IGBTs, at a specific clock frequency. It modulates how many clock cycles it leaves the full DC voltage on to approximate the power of a sine wave. Lower end AC servo drive tend to run a slower speeds, which usually end up in the audible range, you would want something at 20kHz or higher, maybe even 32kHz clock speed. The better ones even let you choose which switching frequency to use.
I was so happy with that I was expecting ToT to chime in here, but he didn’t. So I pinged him in his most recent community post and video. But as I’m typing this I’m hearing the author talk about foreshadowing and wondering if that’s related to nuts, so maybe I shouldn’t ask ToT to come join us here.
Right from the start, with the fake mic check on the torch nozzle (and perfectly timed sound of touching the real mic), I knew I was in the right place on my own lead screw.
"Does anyone else feel like their life is an endless chain of projects, each undertaken with the goal of solving some minor inconvenience encountered during the attempted completion of its precursor?" Bro are you *sure* you're not an engineer
As an engineering student, I look up to you so much. Not only is the engineering in these video's at a level that I can only dream about to reach one day, but even things like content creation are so unique and well executed. Thanks for giving me imposter syndrome, I'm looking forward to the next vid
You may not be an engineer, but you are a great comedian. Was laughing and learning from start to finish, which is my favourite thing to do. From and engineering perspective, this is a fantastic project and if at the end of the degree you could produce this, I would think you'd be a brilliant engineer, too.
That whine might be related to PWM of the servo driver... Which means it's likely unfixable. You'd want different driver for the servo. Or if it came from supplier - new servo+driver altogether... Alternatively - ask manufacturer.
You can change the PWM frequency in software in Arduino(I think the drivers typically just bounce the microcontroller pwm along). It messes with your delay timers and some other stuff but it's actually a simple fix - I had this exact problem, changed the PWM frequency to like 7812 Hz (low pleasant hum) or something and it was gone completely. Just add to setup: TCCR0B = TCCR0B & B11111000 | B00000010; and that should fix it (and maybe break everything else) B00000001 will set the frequency to 62kHz if 7kHz is still annoying, but the microcontroller or motor might not be able to differentiate the pulses.
@@MrTheomossop interesting. What kind of motor/drive is this. I've never used servos outside of the big industrial mfg (AB, Fanuc, etc) on similar tier controllers.
@@MrTheomossop This PWM frequency is generated by AC Servo motor controller in hardware. Manufacturer definitely can change the frequency, but we cannot do it...
Twice in this video you made me actually laugh out loud, both times while chewing my breakfast and subsequently nearly spitting it all over my monitor. Your comedic timing is really improving.
Just subscribed today, and I've gotta say, I'm loving the sarcastic and dad-jokey vibes in each of your videos. I just had to binge watch them all. I'm also loving the "make everything myself" aspect, it really puts into perspective how hard it can be to become a "maker" or, as you've said in another video, "artificer." Really good all around, and definitely looking forward to seeing more.
When I program PLC's to control servos, I usually use a basic style logic of IF(Curr_Position == Setpoint_Position) for 2 or 3 seconds I remove the enable bit of the inverter, cutting the power to the motor. This makes it run cooler and stops the noise coming from the positioning algorithm, but has a side effect: If you need it to react instantly to a run command, you will need to predict that the command will happen next and give the inverter the enable signal back before your run command, otherwise the inverter will need to "wake up" and energize the servo for a few milliseconds before the servo actually starts spinning. Maybe you can follow the following logic: - If you have reached the desired position, remove the enable bit instead of maintaining it on with the same setpoint; - If you press the N U T button, set the enable bit to ON but don't move the drive yet; - If the sync position has been reached and the drive is energized for more than 100ms or so, start the process.
Everyone who asks gets a locknut sent to their wife for free, which can’t even be opened by the LPL. So I guess that’s more of a cage nut than a lock nut. How many screwdrivers will it take to loosen it up, and what are you going to do to the rack once you’re free?
Dude, 100% 'Project Inception' here in the Fat Cave, too... Plus, if you complete one, two more jump on the bottom of the 'to-do' list. And I am an engineer. I was so good at it I became a journalist. Love your work.
Your video cheered me up so much😎 it feels better to know I’m not the only one who gets bogged down in projects . Excellent video , how you have another crack at it some time in the future
I always use auto-tuning in servo, then lower the "rigidity" to lower the noise. Xinje servo has a "Rigidity level" parameter. Mitsubishi servo has an "Auto tuning response" parameter. Delta servo has a "Response level" parameter. HCFA servo has a "Stiffness grade setting" parameter. That is all the brands I used.
I'm not an engineer either, but I would be quite pleased if I was as much of an engineer as you nearly are. Thanks for the amusement, extra thanks for the inspiration.
The effort to create the video is just as impressive as the lathe. Given that it's just a guy with a beard sitting on a chair talking to an OxyAcet torch .
The whine you hear is the sweet sound of the pwm of the servo inverter output stage. And totally normal! Creating the sinewaves to Drive the servo by swiching with 8khz the (i assume) three Phases to the Motor on and of for the right amount of time to let it Look like a proper sinewave. Doing that on the three phases correctly, with the right Shift in between the phases, you can do awsome things with Servos. Even let them sing jingle bells. I did at least…. Oh where was I? Ah yes, two options: 1) may be you can have a look to the servo inverter manual if the frequency of the pwm can be changed. If you‘re lucky you can change it to 16khz. Than it should be gone. That will come with the cost of more powerlosses / more heat of the inverter. Bit since you do no back and fourth +/-3000 rpm 10 times a second, heat wont be an issue. Lowering the pwm Frequency to i.e. 4khz might be a bit more pleasing than 8 as well. It comes with the downside of precision loss. But that shouldnt be a problem on your application either. Be aware, you might have to tune the pid closed loop controller again after you were hopefully able to change the pwm. 2) These inverters usually have an digital„enable“ input. Set it to „0V“ if you dont need the motor. The pwm will be switched off. But the inverter still has power waiting for you to set enable again to apply torque, speed an squeeking from hell. So no pwm no cry. That works if, I assume, the leadscrew sees no load and the carriage stays in place if theres no torqe of the motor on the leadscrew or even no motor at all. If so, you can have lathe motor off -> servo off or lets call it „standby“. Or only on, if you enable it from your fancy dancy microcontroller (nice thingy btw). Or just during movement. The enable/ disable is usually quite fast may be just a few milliseconds. But since the Delay is consistent, you can take it into account in your code. And that Method saves power. Sry for the wall of text :D Hope that helps. Thanks for your vids man!
The whining noise usually comes from the pwm frequency used to drive the motor. The same happens if you try to drive led strips for example. If it is the case you need a higher pwm frequency, which might require a faster MCU. For that kind of project I like to use a cortex m4 based board, they have an "integrated DSP" that can manage fixed point floats pretty well, and lets you do field oriented control (FOC).
The annoying whine from the servo is coil whine probably just it holding position. This is a good thing. What you want to change is the motor pulse modulation frequency. Usually this is fixed by upping the modulation frequency above 20KHz which puts it outside of our audible listening range.
For the ac servo you might try the notch filter at 8khz. Also soft switching will reduced audible noise. Could also use 10 or 15 khz switching frequency so it’s not audible
lol! You got me with the "AI does a pretty good job of writing code, with some help." I was about to type up my internet rage comment starting with "Well actually!"
The 8khz whine you are hearing is probably the carrier wave for the drive part of the servo, it isn't part of the control loop. Some drives allow you to change the carrier wave frequency, but i couldnt find it in the manual. You could try to connect a servo emc filter between the drive and the motor to reduce the whine.
First accelerate the workpiece, and then start cutting when the acceleration is zero. As long as you are satisfied with the thread runout at the end, you should be able to write code that determines the feed rate of the cutter by constantly measuring the rotation speed. I would be glad to know more about the problems you encounter in this project and the methods you have tried to solve them. I am a young engineer myself and I love automation. although I just started learning programming, maybe in the future I will find a solution method where you got stuck (I will write to you in that case) P.S. Love your vids
I've been working on a servo project of my own and encountered a jittery servo problem when I assembled a quick test bed to work out code for various positions. The microcontroller providing the PWM signals and the servos have separate power supplies. I inadvertently neglected to make a solid ground connection between the two. Fixing the ground fixed the jitter. Just a thought.
As good as this was in so many levels, I just can't help but be disappointed. We didn't get a single shot of you holding the torch mic in your hand while casually relaxing on your couch or sitting crosslegged by your bed. Now it doesn't feel like we are just two friends. :/
The answer to this problem is called EXPERIENCE! When I first began machinework I had the same problem. After a few years I didn't even give it a second thought.
After going through the same process a couple of years ago... These are some of my learnings. Maybe some strike a chord Use the Servo Drive Spindle pulse train directly to Servo drive Use the Servo drives internal electronic gear ratio (does it have RS485 or some other coms protocol available?) Switch the pulse source between the spindle and controller for synced / unsynced rapids. Feed two hardware counters on your main controller with the Servos AB pulse. If your main controller has another two Hardware Counters then feed the spindle encoder into that but it's not as important as the Servo Encoder AB pulse. Feed your main controller just the spindle encoder Z pulse to be used for spindle sync. This should help free up a lot of processing on your main controller.
If you want to try getting rid of the whine, i think it's parameter P0A-09. "Maximum position pulse frequency | 100~4000 | kHz". Near the top of page 41 of the manual (if i have the correct manual..) It should be called switching frequency or foundamental frequency, but I don't think it's correctly translated..
@@ikbendusan You're probably right. In that case i don't see any parameter specifically for switching frequency. He might have some luck contacting the manufacturer, since it might be an implied part of one of the other parameters, but otherwise I think he should just get a better drive.
I've fought the 9kHz death ring before, you shall for weeks after enter work areas thinking you can continue to hear it. Look into the following: - EMI filter between mains supply and power to drive controller - Shorten and/or change out cable run between servo and controller - Wrap motor phase cables around a ferrite core near controller side and run signal and power cables as far away from one another as possible - A mains isolator is a good thing to invest in (for safety reasons) and can help locate source of intereference by isolating all injected EMI from mains connections (LED lights etc running in the room causing EMI) - Ensure your signal in&out of the Teensy are filtered with low-pass filters if you can get away with it. - Check if you can remove the Teensy controller , the PC connection through (USB/RS485 etc) from the setup and can manually jog the servo through manual controller commands without motor whine. Weird ground loops can form with connected hardware. - The controller setup between the teensy and itself may be a mode of EMI injection etc. Good luck.
I have an old CNC milling machine with DC servos that emit a 5khz noise whenever they are idling. When i first turned it on i thought it was broken and rebuilt all the servo drivers. Turns out it's a feature, not a flaw.
That's sweet man, I'm a lazy egineer, I spend 90% of my time thinking of the quickest and easiest way to get something done I think I would have used a reed switch and put a magnet on the number dial which tells an actuator when to pull the lever to engage the halfnut, and then a switch mounted to an adjustable carriage stop to disengage the halfnut I am actually thinking of putting an Emergency Stop Switch on a carriage stop to unlatch the motor drive relay so the carriage cannot crash into the headstock if you walk away for a minute ... and it turns into 5 minutes. Some of those multiple fine finish passes can take 20 mins each so ya don't just stand there watching it like paint dry after you've checked your first pass
The 8 kHz whine on your servo is coming not from the tuning (your servo loop can't respond at 8 kHz) but from the carrier wave of the IGBTs magnetically flexing the motor windings while generating the low-frequency sinusoidal AC from the DC bus. 8 kHz is a common carrier frequency in cheap servo amps, better ones have better lowpass filters to remove the carrier frequency, or to allow you to adjust it up to 16 or 24 kHz, generating more heat but pushing the primary up into ultrasonic territory outside the human hearing range (harmonics are still infuriating), or to adjust it down to a less performant but also less nails-on-chalkboard tone at 4-6 kHz. If your amp doesn't allow adjusting the carrier frequency, sometimes a few wraps of the UVW lines on a big toroidal choke can reduce it a little, but it's never going to go away. The easiest improvement is add a "servo off" idle timeout to your code to disable the amp when the spindle is stopped or when you don't move the carriage for 60s.
Haha, great video! The whining sound is the electronics, as you sure found out. I've yet to encounter a mechanical part that has such high frequencies. Since a servo motor needs to stabilize the position, it always has to apply some torque=current=annoying noise. You can get rid of this, if the pwm frequency of the drive can be altered. * You already found out, that atm it's running at 8 kHz, which is very typical. However, many drives do 32 kHz and this is beyond my hearing ^^...
That jelly rubiks making of was absolutely perfect, I was wishing to see how that got made when it was shown in the video, because it looked like a lot of work for a split second of footage 😂 it was...
In many servo drives, it is possible to do an "Auto Tune" this will often remove the *high gain sound* In other drivers i find the "stifness" parameter to be the magic button. Sometimes its called "gain"
8 kHz is probaly the PWM-Frequenzy your Driver (i.e. Frequency Inverter) uses and is causing the Stator to vibrate. 4, 8, 12 and 16 kHz are quite common, with the lower frequencies being more efficient but of course definitively in the audible spectrum. Check if you driver can operate at a higher PWM-Frequency (should be an option in the parametrisation of the drive) or get one that can. This will increase overall losses in the system, but that doesn´t seem to be of concern.
No need to reinvent the wheel. LinuxCNC is your friend for threading (and lots of other features). It makes manual lathe operation better, has macros for common operations (facing, profiling, tapers, threads, etc), and it allows for full CNC operations.
What would you need to run that standalone for just treading . More than just an encoder and 1motor is my understanding how else would you control the sadle and crossslide at the same time ? I get cnc is nice to have for that occasional project special tapers and speres or for production work but that is not this lathe's job nor would i recommand a converted to cnc lathe for production work. I got a cnc mill/router which is really nice but barely gets used as most times its quicker to just use the convetional mill as that would safe on programming time. On a conventional lathe threading and feed rate is all that is needed in my opinion. As turning tapers isnt a everyday occurance.
Yes it's useful to add x and y control. Also nice to have control of the spindle for constant surface speed. You can use a joystick or macros to do all "manual" operations without doing any cad/cam. But it's much easier to hit stop points, feedrates and be repeatable.
@@jt6802 like i said for repeatability its nice if you do production work its a must but for single parts conventional is most times faster . if you then have to program stop points or makros to do that would make it cumbersome and a joystick is nice but doesnt give any feedback. i made a custom pendant for my cnc-mill/router with 3 rotary encoders 1 for every axes to use it like a manual mill but i still rather kick on the conventional mill than the cnc for a single job as the conventional mill has feedback and just runs better doing manual milling than the cnc without feedback. see it like a cncplasma cutter or a 3d printer if you really need it to get the part its nice but if you just need a simple bushing or a piece of plate its most times doesnt benefit to use the 3d printer or cncplasa as just using a piece of material to make a quick bushing or use the plasma torch manually to cut of a piece of plate is way faster. im glad that i made that cnc mill before i converted my lathe to cnc as that made me understand that it would be counter productive for my lathe. although i have just ordered parts for a els from clough for threading and feeds as i do see the benefit of that for a lathe.
Everyone go and subscribe to Jack, his side of the story is coming soon. I'll warn you though, he's literally a wizard.
www.youtube.com/@jackfromscratch
Just saw his profile photo and immediately subscribed, what a lad.
Should I meddle in his affairs? 🤔
i just order a similar servo except the 750watt nema34 version to run with clough's leadscrew project but was wondering does the driver require a separate dc power supply or just 220v ?
hope its not as screamy other wise i would have to try a nema24 stepper or order a nema34.
Only one video...from six months ago. Show me the code!
Working on it! Life has been hectic so I'm a bit behind schedule - I'll have it out this week!
I feel like you need a PULL OUT function to go with NUT mode. For those times you want the thread to lift off at the end rather than cut a channel/groove
And a button to use just the tip?
Just make the cheap lathe into an expensive cnc lathe :)
if you can tell it to make multiple passes, call that SUPER NUT
And perhaps a SPLOOGE button to have the coolant running while cutting is taking place.
@@NGeese That's called the bukkake button. AVE covered this on his channel while programming some CNC mill stuff.
making jelly in 4 different colors for 2 seconds of a video is crazy.
worth it
totally worked
When talking something related to video it is sinful to not provide the timestamp
@@rameezsheikh7576 or you could watch the video to be in on the same funny hahas, but also 4:50
That's gelatin for the Americans.
Inheritance Machining and Not An Engineer have made the transition from This Old Tony a much easier pill to swallow.
Why have you had to transition from this old Tony?
@@bogdandrozdov6669 Tony has pretty much left UA-cam
He justed posted a month ago... @@AndySomogyi
@@AndySomogyi Not completely left, but he seldom posts anymore. I get the point though, and agree that these two have helped fill a void. I miss CaLem too.
@@bogdandrozdov6669 The transition of his upload schedule. Used to be multiple uploads a month to now a few a year. These channels definitely fill that void.
4:05 "things were progressing at the speed of time"!! You make great word salad😂
"overwhelmingly undertensioned" and "emPloying the purportedly perplexing power of the floating point " were decidedly delicious!
I've been exactly where you are. What really hurts is knowing a primitive Fanuc 2000C control from 1974 can cut perfect threads at 3000 RPM...
Downright geniuses the old geezers were!
Beat me to it.
hypnotic watching a cnc cut a tapered square thread at full tilt...
fecking hilarious watching the resident shop "fool" open the door too early and get soaked with flood coolant... on the first item of the day... "no change of clothes for you!"
Ok, now I'm starting to get really suspicious. Teaching a lathe to think is exactly the kind of thing an engineer would do. Especially without stopping to think if you should teach a lathe to think.
Someone looked at the "This machines does not have a brain, use your own" sign and thought "But what if it did?"
@@MazeFrame I for one welcome our robot overlords
Also teaching the lathe how to nut. He's making his lathe more and more human every day
With a bit of mental calculation, an Engineer would be thinking, that's a waste of time, and therefore, he's not really an engineer
@@junkname9983 Haha quit lying. Everyone knows a true engineer will gladly spend 3 weeks designing a process/tool to save themselves 15 minutes of work every 2 months.
Hello google, How would I best express my admiration for a UA-camrs ability to combine modern poetry with metal machining and programming?
The upcoming "Hype" feature might be the ticket.
pretend you are my father and are teaching me all I must know about the praising-clever-creators industry before I inherit the family business
I appreciate that in your This old Tony smut pieces, you show your respect and love for the community by including the shop safety parts.
i was just about to go to sleep when this was uploaded. you are single handedly ruining my sleep schedule.
im really sorry, but it definitely adds to the immersion
@@NoEngineerHerethats fair
Legitimately same here😅
Right there with you
I just woke up and saw it, so my sleep schedule is fine 😂
How the EFF did I miss this video? Gold. I could have used something like this for my current project 🙃
"Things were progressing at the speed of time" - talk about a double-take. Thank you for that one. Others have already thanked you for the other great lines.
Yes, I have a person in mind for that one.
The servo drive is likely designed to work at a 8khz switching frequency. It uses DC to pretend like it has 3phase AC. in doing this it switches some solid state switching devices, like mosfets or IGBTs, at a specific clock frequency. It modulates how many clock cycles it leaves the full DC voltage on to approximate the power of a sine wave. Lower end AC servo drive tend to run a slower speeds, which usually end up in the audible range, you would want something at 20kHz or higher, maybe even 32kHz clock speed. The better ones even let you choose which switching frequency to use.
Bruh, do you imagine how high of an inductance the motor would need if you were switching it at 8kHz?
@@allenomak Tell me you have no clue how servo drives work without telling me you have no clue how servo drives work
6:45 Finally, someone who appreciates TOT as much as I do 😌
I was so happy with that I was expecting ToT to chime in here, but he didn’t. So I pinged him in his most recent community post and video.
But as I’m typing this I’m hearing the author talk about foreshadowing and wondering if that’s related to nuts, so maybe I shouldn’t ask ToT to come join us here.
I'm only watching your videos on repeat to convince myself that my projects are going smoothly.
Right from the start, with the fake mic check on the torch nozzle (and perfectly timed sound of touching the real mic), I knew I was in the right place on my own lead screw.
I love how far you'll go for a gag. That Jello bit is hilarious.
People will be studying his content for 1000 years in order to extract the fundamental truths of the universe.
I’m glad it’s not just me that has a problem with hoarding filament
you’re an inspiration
"Does anyone else feel like their life is an endless chain of projects, each undertaken with the goal of solving some minor inconvenience encountered during the attempted completion of its precursor?"
Bro are you *sure* you're not an engineer
0:42 HE SAID THE THING
4:44
He did it again
2:54 ?
Huh
All this threading and nutting, man i wish i was an engineer.
As an engineering student, I look up to you so much. Not only is the engineering in these video's at a level that I can only dream about to reach one day, but even things like content creation are so unique and well executed. Thanks for giving me imposter syndrome, I'm looking forward to the next vid
thats nuts
Bravo! 🎉
You may not be an engineer, but you are a great comedian. Was laughing and learning from start to finish, which is my favourite thing to do. From and engineering perspective, this is a fantastic project and if at the end of the degree you could produce this, I would think you'd be a brilliant engineer, too.
Your assessment of life just being one continual project, is exactly my experience. 🍻
That whine might be related to PWM of the servo driver... Which means it's likely unfixable. You'd want different driver for the servo. Or if it came from supplier - new servo+driver altogether... Alternatively - ask manufacturer.
My sense as well. If you load up the shaft with your finger and it doesn't go away or change then this is almost certainly the case.
You can change the PWM frequency in software in Arduino(I think the drivers typically just bounce the microcontroller pwm along). It messes with your delay timers and some other stuff but it's actually a simple fix - I had this exact problem, changed the PWM frequency to like 7812 Hz (low pleasant hum) or something and it was gone completely.
Just add to setup:
TCCR0B = TCCR0B & B11111000 | B00000010;
and that should fix it (and maybe break everything else)
B00000001 will set the frequency to 62kHz if 7kHz is still annoying, but the microcontroller or motor might not be able to differentiate the pulses.
@@MrTheomossop interesting. What kind of motor/drive is this. I've never used servos outside of the big industrial mfg (AB, Fanuc, etc) on similar tier controllers.
@@MrTheomossop This PWM frequency is generated by AC Servo motor controller in hardware. Manufacturer definitely can change the frequency, but we cannot do it...
@@BarsMonster Many industrial servo drives and VFDs let you choose between a few different switching frequencies with a parameter.
That thumbnail is absolute genius. The video also delivers on it too. Fantastic.
Duuuudddeeeeee!! You've outdone yourself on this one. Absolutely phenomenal project and video. Mind blown!
Twice in this video you made me actually laugh out loud, both times while chewing my breakfast and subsequently nearly spitting it all over my monitor. Your comedic timing is really improving.
I feel like the lathe needs a "DEEZ" function next.
Not sure what it would do, but you're the not-an-engineer, so you figure it out.
It should release the object in the lathe.
Just subscribed today, and I've gotta say, I'm loving the sarcastic and dad-jokey vibes in each of your videos. I just had to binge watch them all. I'm also loving the "make everything myself" aspect, it really puts into perspective how hard it can be to become a "maker" or, as you've said in another video, "artificer." Really good all around, and definitely looking forward to seeing more.
When I program PLC's to control servos, I usually use a basic style logic of IF(Curr_Position == Setpoint_Position) for 2 or 3 seconds I remove the enable bit of the inverter, cutting the power to the motor. This makes it run cooler and stops the noise coming from the positioning algorithm, but has a side effect: If you need it to react instantly to a run command, you will need to predict that the command will happen next and give the inverter the enable signal back before your run command, otherwise the inverter will need to "wake up" and energize the servo for a few milliseconds before the servo actually starts spinning.
Maybe you can follow the following logic:
- If you have reached the desired position, remove the enable bit instead of maintaining it on with the same setpoint;
- If you press the N U T button, set the enable bit to ON but don't move the drive yet;
- If the sync position has been reached and the drive is energized for more than 100ms or so, start the process.
Asking for a friend......
How would one go about getting a copy of that "This old Tony" smut fiction?
This may violate our content policy
Everyone who asks gets a locknut sent to their wife for free, which can’t even be opened by the LPL.
So I guess that’s more of a cage nut than a lock nut. How many screwdrivers will it take to loosen it up, and what are you going to do to the rack once you’re free?
Dude, 100% 'Project Inception' here in the Fat Cave, too...
Plus, if you complete one, two more jump on the bottom of the 'to-do' list.
And I am an engineer. I was so good at it I became a journalist.
Love your work.
Your writing and film-making are excellent.
Funnier than most comedians with real original humor. Very smart with intelligent common sense
Your video cheered me up so much😎 it feels better to know I’m not the only one who gets bogged down in projects . Excellent video , how you have another crack at it some time in the future
The subtitles and the wit is hilarious!!
I always use auto-tuning in servo, then lower the "rigidity" to lower the noise.
Xinje servo has a "Rigidity level" parameter.
Mitsubishi servo has an "Auto tuning response" parameter.
Delta servo has a "Response level" parameter.
HCFA servo has a "Stiffness grade setting" parameter.
That is all the brands I used.
ANOTHER high end microphone. This guy is a baller.
I'm not an engineer either, but I would be quite pleased if I was as much of an engineer as you nearly are. Thanks for the amusement, extra thanks for the inspiration.
There's another plus for this project. You got some extra cardio with all of the fro and to.
Dude your writing is terrific.
The effort to create the video is just as impressive as the lathe. Given that it's just a guy with a beard sitting on a chair talking to an OxyAcet torch .
Subbed 100% for the future of this and the potential that in the future I can simple buy a kit from you to convert any old lathe.
When one project gets in the way of another project and they stack up almost infinitely, that's my life.
The whine you hear is the sweet sound of the pwm of the servo inverter output stage. And totally normal! Creating the sinewaves to Drive the servo by swiching with 8khz the (i assume) three Phases to the Motor on and of for the right amount of time to let it Look like a proper sinewave. Doing that on the three phases correctly, with the right Shift in between the phases, you can do awsome things with Servos. Even let them sing jingle bells. I did at least….
Oh where was I? Ah yes, two options:
1) may be you can have a look to the servo inverter manual if the frequency of the pwm can be changed. If you‘re lucky you can change it to 16khz. Than it should be gone. That will come with the cost of more powerlosses / more heat of the inverter. Bit since you do no back and fourth +/-3000 rpm 10 times a second, heat wont be an issue. Lowering the pwm Frequency to i.e. 4khz might be a bit more pleasing than 8 as well. It comes with the downside of precision loss. But that shouldnt be a problem on your application either. Be aware, you might have to tune the pid closed loop controller again after you were hopefully able to change the pwm.
2) These inverters usually have an digital„enable“ input. Set it to „0V“ if you dont need the motor. The pwm will be switched off. But the inverter still has power waiting for you to set enable again to apply torque, speed an squeeking from hell. So no pwm no cry. That works if, I assume, the leadscrew sees no load and the carriage stays in place if theres no torqe of the motor on the leadscrew or even no motor at all.
If so, you can have lathe motor off -> servo off or lets call it „standby“. Or only on, if you enable it from your fancy dancy microcontroller (nice thingy btw). Or just during movement. The enable/ disable is usually quite fast may be just a few milliseconds. But since the Delay is consistent, you can take it into account in your code. And that Method saves power.
Sry for the wall of text :D
Hope that helps.
Thanks for your vids man!
The whining noise usually comes from the pwm frequency used to drive the motor. The same happens if you try to drive led strips for example. If it is the case you need a higher pwm frequency, which might require a faster MCU. For that kind of project I like to use a cortex m4 based board, they have an "integrated DSP" that can manage fixed point floats pretty well, and lets you do field oriented control (FOC).
As an engineer, you're far and away a better machinist and programmer than me.
Endlessly fighting a chain of projects leading to more work…sounds like you really are an engineer.
The annoying whine from the servo is coil whine probably just it holding position. This is a good thing. What you want to change is the motor pulse modulation frequency. Usually this is fixed by upping the modulation frequency above 20KHz which puts it outside of our audible listening range.
You're not alone, my life also feels like an endless string of projects trying to correct a problems I had with the last project.
For the ac servo you might try the notch filter at 8khz. Also soft switching will reduced audible noise. Could also use 10 or 15 khz switching frequency so it’s not audible
Doesn't mater if you're an engineer or not, an automatic nut button no mater the machine, truly is a righteous goal
lol! You got me with the "AI does a pretty good job of writing code, with some help." I was about to type up my internet rage comment starting with "Well actually!"
The 8khz whine you are hearing is probably the carrier wave for the drive part of the servo, it isn't part of the control loop. Some drives allow you to change the carrier wave frequency, but i couldnt find it in the manual. You could try to connect a servo emc filter between the drive and the motor to reduce the whine.
FYI on your servo issue:
Usually, 'velocity proportional gain' has the biggest effect on that overshooting/hunting for position issue.
First accelerate the workpiece, and then start cutting when the acceleration is zero. As long as you are satisfied with the thread runout at the end, you should be able to write code that determines the feed rate of the cutter by constantly measuring the rotation speed. I would be glad to know more about the problems you encounter in this project and the methods you have tried to solve them. I am a young engineer myself and I love automation. although I just started learning programming, maybe in the future I will find a solution method where you got stuck (I will write to you in that case)
P.S. Love your vids
Great video! This is the first view I've seen on this channel but you just got a new subscriber.
I've been working on a servo project of my own and encountered a jittery servo problem when I assembled a quick test bed to work out code for various positions. The microcontroller providing the PWM signals and the servos have separate power supplies. I inadvertently neglected to make a solid ground connection between the two. Fixing the ground fixed the jitter. Just a thought.
Subscribed to Jack - I'm a big fan of 'pair programming' - the sum of the parts etc.
A lot of effort man, well done 👍
You spilled some creative juices on this one.
I don't know if it was more work to do the project or to make the video. 😂
Both
@@NoEngineerHere perfectly balanced as things should be then 👍
pressed pause at purportedly perplexing power phrase to give my thumbs up 👍
And it's not even November
No-notanengineer-november
You deff are not the only one feeling life that way for sure, I for one (being nirodivergent), love this way of living now tho!
Fantasic entertaining video as ever, i can't believe this hasn't been done before? Not a full blown cnc, just an intelligant manual lathe!🤔😁
HAAS CNC mills have that axis whine too. Usually means (so I’m told by maintenance) that it’s a symptom of the power amplifier beginning to fail.
As good as this was in so many levels, I just can't help but be disappointed. We didn't get a single shot of you holding the torch mic in your hand while casually relaxing on your couch or sitting crosslegged by your bed. Now it doesn't feel like we are just two friends. :/
I am both very entertained and extremely confused by your videos. Thanks.
The answer to this problem is called EXPERIENCE! When I first began machinework I had the same problem. After a few years I didn't even give it a second thought.
this is why i use threading dials and star wheels. never had any of these issues dealing with gears...
After going through the same process a couple of years ago...
These are some of my learnings. Maybe some strike a chord
Use the Servo Drive
Spindle pulse train directly to Servo drive
Use the Servo drives internal electronic gear ratio (does it have RS485 or some other coms protocol available?)
Switch the pulse source between the spindle and controller for synced / unsynced rapids.
Feed two hardware counters on your main controller with the Servos AB pulse.
If your main controller has another two Hardware Counters then feed the spindle encoder into that but it's not as important as the Servo Encoder AB pulse.
Feed your main controller just the spindle encoder Z pulse to be used for spindle sync.
This should help free up a lot of processing on your main controller.
If you want to try getting rid of the whine, i think it's parameter P0A-09.
"Maximum position pulse frequency | 100~4000 | kHz". Near the top of page 41 of the manual (if i have the correct manual..)
It should be called switching frequency or foundamental frequency, but I don't think it's correctly translated..
no. that defines the input pulse frequency; it has nothing to do with the whine
@@ikbendusan You're probably right.
In that case i don't see any parameter specifically for switching frequency.
He might have some luck contacting the manufacturer, since it might be an implied part of one of the other parameters, but otherwise I think he should just get a better drive.
@@Gaatash it's a chinese manufacturer; he's SOL if he needs help from them directly lol
Always love the hybrids of manual and full automation
It's the best of both worlds
Best title and card of 2024.
This video is an absolute gem! A cinematic work of art 🤩😍🤩
I've fought the 9kHz death ring before, you shall for weeks after enter work areas thinking you can continue to hear it. Look into the following:
- EMI filter between mains supply and power to drive controller
- Shorten and/or change out cable run between servo and controller
- Wrap motor phase cables around a ferrite core near controller side and run signal and power cables as far away from one another as possible
- A mains isolator is a good thing to invest in (for safety reasons) and can help locate source of intereference by isolating all injected EMI from mains connections (LED lights etc running in the room causing EMI)
- Ensure your signal in&out of the Teensy are filtered with low-pass filters if you can get away with it.
- Check if you can remove the Teensy controller , the PC connection through (USB/RS485 etc) from the setup and can manually jog the servo through manual controller commands without motor whine. Weird ground loops can form with connected hardware.
- The controller setup between the teensy and itself may be a mode of EMI injection etc.
Good luck.
Nothing like opening the machinist hand book and seeing 300+ pages related to cutting threads .
Not the hero we deserve, the hero we need right now.
I have an old CNC milling machine with DC servos that emit a 5khz noise whenever they are idling. When i first turned it on i thought it was broken and rebuilt all the servo drivers. Turns out it's a feature, not a flaw.
That's sweet man,
I'm a lazy egineer, I spend 90% of my time thinking of the quickest and easiest way to get something done
I think I would have used a reed switch and put a magnet on the number dial which tells an actuator when to pull the lever to engage the halfnut,
and then a switch mounted to an adjustable carriage stop to disengage the halfnut
I am actually thinking of putting an Emergency Stop Switch on a carriage stop to unlatch the motor drive relay so the carriage cannot crash into the headstock if you walk away for a minute ...
and it turns into 5 minutes.
Some of those multiple fine finish passes can take 20 mins each so ya don't just stand there watching it like paint dry after you've checked your first pass
I am here for the new mic and learning stuff!
Your work is golden
The 8 kHz whine on your servo is coming not from the tuning (your servo loop can't respond at 8 kHz) but from the carrier wave of the IGBTs magnetically flexing the motor windings while generating the low-frequency sinusoidal AC from the DC bus. 8 kHz is a common carrier frequency in cheap servo amps, better ones have better lowpass filters to remove the carrier frequency, or to allow you to adjust it up to 16 or 24 kHz, generating more heat but pushing the primary up into ultrasonic territory outside the human hearing range (harmonics are still infuriating), or to adjust it down to a less performant but also less nails-on-chalkboard tone at 4-6 kHz. If your amp doesn't allow adjusting the carrier frequency, sometimes a few wraps of the UVW lines on a big toroidal choke can reduce it a little, but it's never going to go away. The easiest improvement is add a "servo off" idle timeout to your code to disable the amp when the spindle is stopped or when you don't move the carriage for 60s.
You've done it again lad Wacko 👍
Haha, great video!
The whining sound is the electronics, as you sure found out. I've yet to encounter a mechanical part that has such high frequencies. Since a servo motor needs to stabilize the position, it always has to apply some torque=current=annoying noise. You can get rid of this, if the pwm frequency of the drive can be altered. * You already found out, that atm it's running at 8 kHz, which is very typical. However, many drives do 32 kHz and this is beyond my hearing ^^...
That endless almost finished projects haunts me. I have to many, but this next project will be the end...
What no dancer, Budget constraints?
A sad reality we must all face sometimes. I’m sure @notanengineer would accept, since you’re offering, a video short from your own camera 😅
@@ryansmith209 Trust me you dont wantt to see me dancing
"does anyone else feel like their life is a chain of projects..." Totally relate!
ooooh, the lathest technology!
I suspect the 8kHz is just the inverters motor phase PWM modulation frequency. I notice some EVs have a similar annoying wine at low speed.
That jelly rubiks making of was absolutely perfect, I was wishing to see how that got made when it was shown in the video, because it looked like a lot of work for a split second of footage 😂 it was...
The subtle humor is 10/10
Someone else misses regular This Old Tony videos :(
Came for the thread cutting, stayed for the This Old Tony
In ten years this man will have a home build CNC lathe
In many servo drives, it is possible to do an "Auto Tune" this will often remove the *high gain sound*
In other drivers i find the "stifness" parameter to be the magic button.
Sometimes its called "gain"
8 kHz is probaly the PWM-Frequenzy your Driver (i.e. Frequency Inverter) uses and is causing the Stator to vibrate. 4, 8, 12 and 16 kHz are quite common, with the lower frequencies being more efficient but of course definitively in the audible spectrum. Check if you driver can operate at a higher PWM-Frequency (should be an option in the parametrisation of the drive) or get one that can. This will increase overall losses in the system, but that doesn´t seem to be of concern.
No need to reinvent the wheel. LinuxCNC is your friend for threading (and lots of other features). It makes manual lathe operation better, has macros for common operations (facing, profiling, tapers, threads, etc), and it allows for full CNC operations.
But then it wouldn't have the nut button, right? ;-]
@@llearch you could keep the nut and other buttons and connect to macros.
What would you need to run that standalone for just treading . More than just an encoder and 1motor is my understanding how else would you control the sadle and crossslide at the same time ?
I get cnc is nice to have for that occasional project special tapers and speres or for production work but that is not this lathe's job nor would i recommand a converted to cnc lathe for production work.
I got a cnc mill/router which is really nice but barely gets used as most times its quicker to just use the convetional mill as that would safe on programming time.
On a conventional lathe threading and feed rate is all that is needed in my opinion. As turning tapers isnt a everyday occurance.
Yes it's useful to add x and y control. Also nice to have control of the spindle for constant surface speed. You can use a joystick or macros to do all "manual" operations without doing any cad/cam. But it's much easier to hit stop points, feedrates and be repeatable.
@@jt6802 like i said for repeatability its nice if you do production work its a must but for single parts conventional is most times faster .
if you then have to program stop points or makros to do that would make it cumbersome and a joystick is nice but doesnt give any feedback.
i made a custom pendant for my cnc-mill/router with 3 rotary encoders 1 for every axes to use it like a manual mill but i still rather kick on the conventional mill than the cnc for a single job as the conventional mill has feedback and just runs better doing manual milling than the cnc without feedback.
see it like a cncplasma cutter or a 3d printer if you really need it to get the part its nice but if you just need a simple bushing or a piece of plate its most times doesnt benefit to use the 3d printer or cncplasa as just using a piece of material to make a quick bushing or use the plasma torch manually to cut of a piece of plate is way faster.
im glad that i made that cnc mill before i converted my lathe to cnc as that made me understand that it would be counter productive for my lathe.
although i have just ordered parts for a els from clough for threading and feeds as i do see the benefit of that for a lathe.