I think this is exceptionally produced content and it's evident that you have the spirit of a researcher/teacher. Just the ability to verbally connect with your audience in a clear-concise manner puts you into a class of your own. Showing mistakes, errors and oversight helps keep the content humanly interesting. Good job. (virtual back patting)
@@Clough42 Mistakes are the best learning tool!! When someone asks me what my education is, I tell them I have a PhD from the School of Hard Knocks... It usually gets a laugh, but it's more true than most folks realize!!
I’m stunned your subscriber count hasn’t skyrocketed in the last couple months. You’re videos are gold. They’re so professionally produced and enjoyable to watch. Just the right amount of detail and never boring. Thanks for all the time you invest in these.
You are welcome. I did have a big bump from 1K to now nearly 10K over the last few months, but things have slowed. Summer is upon us, and things usually slow down.
13:46 "looks alright to Me" proceeds to max out the rpm instantly 😂 man i just love this series and the channel in general! Joking aside this is really a top notch channel.
I'm replacing the gear on the spindle with a timing pulley. The main drive pulley came off quite easily. Just need to put a big bore and keyway in the timing pulley. Your layout and presentation of content is awesome. Very nice work.
This is what home cnc is all about. Stuff that would be difficult manually is relatively easy with even a low power cnc machine. It’s going to be interesting to see if you can cnc the feed screw.
SDP/SI! That place has saved my bacon on multiple occasions. Fast (kinda pricey) shipping, reasonable prices on parts, and just about everything under the sun available.
As always SUPERB! Some of the finest instructional videos I have EVER seen. I too can't wait to see this come to fruition. Keep up the good work kind Sir! And may Jesus continue to gift you with unique talents. My lady watches EVERY video with ernest as I do.
James, thank you again for another well thought out (along the way) project. I commend you for not hiding goofs and gaffs and looking human in the process. Someday, I hope to be as competent as you. I am going to skip back and watch from the beginning. This looks like a great process for solving my gears kicking loose under tension. I want to see how you calibrate all the speed settings for proper threading without having the change gears, and simplify it for my old man head. My equipment is similar, an old Asian lathe that looks like the Grizzly G4000 and a mill the same as your G0704. All mechanical still.
I do not have a lathe but I am so excited to see all of this work. I am building a shock dyno on the same principle i got the same encoder and digital readout and the BSS 138 converter. I also got the F280049C encoder. Great job. Thank you! I am using the same VFT you are using to run my 3 HP 3PH motor.
Gates can't help the Chinese part numbering debacle... but the technical resources on the Gates website has a really cool application to size belts for pulley sizing and spacing. It's simple to use and very flexible. Our high school robotics team uses it. I can't wait to see this whole thing up and running.
Also would be cool to have a readout showing degrees of rotation of the spindle to allow using it as a low accuracy dividing head for drilling flanges/ holes around a PCD etc with a toolpost drill for those of us without a CNC milling machine.
This was the first video I saw of yours where you were running the Grizzly-based CNC, and was totally impressed by it running. That and the fact you just buzzed up a couple pulleys and another bracket on the 3D printer really makes me want to try one, lol!
If you wanted to mount the leadscrew stepper direct drive, you'd have the option of doing so by mounting the stepper to the tail stock end of the leadscrew and making up a mounting bracket. That way you get to leave the gear box cover on the lathe and move the stepper wiring away from the lathe controls.
About Chinese belts misunderstanding - I think that this type of numbering of belts is common at least here in Eastern Europe. I always order belts by their length in mm, not by number of teeth.
Damn! Just binged on this build and got to the end! Seriously looking forward to seeing things spin! I love the idea, I'd like to do a similar thing as a part of a gradual CNC conversion on my Colchester lathe.
Great update. I've been following a few of these ELS projects and eventually decided to give one a go. It was a commercial one, but quite cheap and he sold it as a complete kit with (recommended) stepper motors. I already had a couple of servo's so I decided to go with them instead. I direct drive the lead screw thru the gearbox, as you do and drive the cross side with a 1:1 belt drive, as this it what the creator suggested. It works fine but I have noticed that the servo(DMM) on the cross side has little holding torque and can 'back out' a little on heavy cuts sometimes(I guess the lead screw one might be the same). A stepper would have better holding torque to stop this. I think if you are going to use a 1:1 or low ratio drive, steppers would be fine and if you are going to use a belt drive and higher ratio servos would be fine. Either way I'm having fun and look forward to seeing where you take yours. I do have a Mazak Rex 610-1500 20Hp lathe with no change gears and am interested in giving yours a go on this one, if I can work out sizing and find servo's large(and cheap!) enough to drive it!
Yes was powered, it's a Canadian DMM servo so I can connect the driver to the laptop and see if there is something I can change in the settings, but I didn't think servos had the same holding torque as steppers? Steppers are full power all the time, servo's power as required(I thought?). Didn't move much and probably no more that if I was using it manually, but I was going to put a ballscrew on the cross slide to remove backlash and this could make it worse as the ballscrew would have far less mechanical drag/ resistance. Would be interesting to do a test with stepper/(true )servo connected to a ballscrew/leadscrew with a pushing load applied to the screw to see if there was any movement. I of course maybe completely wrong, as I often am, but am happy to proved so for the greater good :-).
@clough42 You can always part off on the round part of the stock. Cut your flats just deep enough. Then part on the very edge of the round stock. A little metrology, and you can have a precise part....
Just turn it, part it off, and put in the mill vise. The flange is plenty big enough to hold it, and the pip will register for height. The mistake is thinking the flats have to be perfectly parallel - they only need to be close, and it can be done by eye.
Another excellent video! A quick-ish fix for the over sized pulley would have been to bore it out larger, press fit a plug it, then bore to the correct size. Or....order a new one. 😬
I thought about it. It probably would have worked fine, but I'd have to hold it in the 4-jaw, with some kind of fixture to prevent damaging the aluminum teeth. Also I didn't want to put the change gears back on the lathe or use it without feed. So...there were enough excuses it seemed easier to just 3D-print one.
Where there’s a mill there’s a way. 😬A couple hold down clamps, interpolate a hole, bobs your uncle. I actually appreciate the 3D printed part, I like seeing something that’s gonna take some stress being printed. Sure makes for some quick and easy prototypes.
Gates also measure their belts by the pitch length rather than by the number of teeth. For instance a 5mm pitch Powergrip GT3 (300-5MGT) is 300mm long with a 5mm pitch tooth and has 60 teeth.
So glad I ran across your channel. Just about everything I am interested in and trying to learn is right here. I have binge watched every one of your videos and anxiously await more. I especially enjoy the process as you explain it from the need/idea to planning and design in Fusion 360, prototyping, problem solving, machining (3D and CNC), mistakes, solutions, everything! I find your content inspirational and motivational in just the right ratio. Congrats on the weight loss and thanks for taking the time to produce such excellent content for those of us just starting our journey (however long it took us to get started:)... P.S. Are you planning anything special for the 10K subscribers milestone which is right around the corner?
Hi James. What software are you using to design your 3d printed pulleys? You've got an amazing project here. It's so instructive to see your entire process. You epitomize what education is really about. Thanks many times over.
with this asesories you havent backless and you can have a high presicion lathe you have good ideas in your head , to that in all you live , succes mister
I'm curious, why not go for optical sensing of spindle speed? It seems like you could've milled one or more slots in the 60 tooth gear, shined an IR LED through it at whatever sensor you prefer, and there's your data.
This is actually an optical encoder, and a very high-resolution one. Milling slots in the gear isn't something I had considered. For that matter, I could probably read the teeth of the gear optically. It *might* even be possible to get 240 counts/rev with some kind of quadrature setup. The software approach I'm taking at the moment depends on the high resolution, though, so I'll probably continue with that for now.
@@Clough42 Yeah, I suppose for thread synchronization you need not just RPM but high accuracy spindle position at high sample rate. Still, you could read the teeth like you mentioned, or attach a slotted rotor sheet to the gear with as many slits as you like for high resolution optical sensing. It just seems to me that doing the sensing with a mechanically linked encoder is inviting complication, error, and wear.
@calastigro If you would really want to read the gear, that would be better served by an inductive / Hall effect sensor: it's what ABS in cars actually does; you really wouldn't want to muck around shining light through the cogs. But as noted, the resolution would be drastically worse, and that's not a good thing when you're trying to cut a thread instead of just figuring out whether a wheel is slipping or not...
Hi James, you said the rotary encoder 3Dprinted gear is mod 1in the video at 4:45minutes. I have the same lathe type but my gears are mod 1.5. Mod 1 is for very small loads.
There are two gear types on my lathe. The first 60T plastic gear fed from the 40T plastic gear on the spindle is about 90mm in diameter, so it's likely mod 1.5. The rest of the cast iron gears appear to be mod 1. I just measured the 60T cast iron change gear, and it's approximately 60mm in diameter. The leadscrew gear train carries relatively small loads.
@@Clough42 Thank for the mod clarification. Overall gear ratio from spindle, rotary encoder up to the leadscrew you prefer to keep as integer 1, 2 or 4 and in combination with the lathe selector ABC and in the CCSsoftware code microsteps variable the total ratio should be 1?
This is my favorite TH-cam channel. I have 2 questions: what motors are you using for your table top mill, steppers or servos? Also i never hear your air compressor kick on and off. Where do you have your air compressor located?
Direct drive is the way to go. I've got NEMA24 560oz. in. closed loop stepper motor which easily drives Z axis head stock with 1605 ball screw on my Sieg mill. So for this application it will do the job just hands down. Otherwise PU toothed belt with steel cord is going to be fine as well.
great build so far, use only premium gates belts ,worth it over imports , I also like belt drives with steppers I believe they do a good job at damping. on expensive VMC's you will also see belts used on drives and encoders.
Has anyone given thought to using the existing RPM readout on many of the current VS lathes as an encoder? curious if there is a way to tap into that instead of introducing an additional point of failure into the mix. I have all the parts to do the conversion but there may be an opportunity to simplify here.
I think the main issue with the existing sensors would be resolution. It is possible to thread with one or ten pulses per revolution, but you have to make a lot of assumptions about constant speed operation and it would be slow to respond to speed or direction changes.
@@Clough42 Fair enough. I actually came to a similar conclusion after a friend said his kept RPM with only 4 magnets around the shaft. I can definitely see the lack of resolution really negating the whole project. Thanks for the sanity check.
I wanted to say I really appreciate this, and all your hard work. I urge other people to stop buying from Grizzly. They offer what seems to be a good product at a good price, but this is the sort of issue that comes from them. They will continue to sell sub-standard equipment as long as people keep choosing them over other dealers who sell equipment that is not missing basic functions.
@@Clough42 Gearbox, reverse tumbler, etc. I can't remember all the things, I just remember being very frustrated about how close they came to being a good product. Wound up buying both a lathe and mill from Precision Mathews. They both cost more, but end up cheaper in the end because you don't have to modify them
@@joshua43214 Grizzly has a bunch of different offerings with different price points and feature sets. The G0602 is a particularly good deal if you're doing a CNC conversion and throwing all the extra stuff away. IIRC, the PM lathes have power cross feed. That's a winner.
Bore the pulleys out to 20-22mm, then make bushing that press fit in with a 14mm center hole and keyway. You could make them from steel which would wear better than the aluminum.
Hi James. I think it was this video where you mentioned you might eventually try and mount the Servo toward the back of your lathe. Just wondering if its still on your todo list? I have the same lathe as you and use my ELS all the time. And I think of the door being open often too. Just curious if you've given any more thought. Thanks. I'm just assembling my new box and ELS faceplate that just arrived in the mail. Thanks for all the great content.
Excellent series James, I'm buying up the bits and pieces to begin this build (already bought your kits) on my Grizzly G4000. I'm down in the Caribbean, so everything ordered on-line takes forever to get here, especially from China. Very frustrating when working on a project as you would appreciate, and Lord forbid the wrong thing arrives! :) Two questions: 1) The ebay link for the Servo motor is dead, can you suggest an alternate source? 2) Do you have recommended source(s) for acceptable toothed belts and pulleys?
In the lands Metric, M series belts are ordered by Series ~ Length ~ Width. In Australia I don't think any belts are classified via tooth count, the books however have the tooth count listed also. You may find that you could run into some belt tooth climb with the 3m series belts in practice. On a 2hp ball screw drive a 5m is normally used, the 3 m may need to be tight beyond the capability of the nose bearing of the motor? I don't think that will be an issue for you until you get up into the 2mm per rev range of diameter turning. Say a 2 X 2 mm cut on steel. The risk of this problem is greatly reduced the bigger the drive pulley diameter becomes. Cheers
Hi James. By replacing the leadscrew by a ball screw you can turn that lathe into a CNC machine much more capable than just cutting threads without change gears. Installing the servo straight to the leadscrew on the tail stock end is not an option? less cluttered space. Great job and keep on. Subscribed and following
Hi James, your lathe is very similar to mine and I'm very interested in this. To repeat what I've said earlier, please bear in mind that anyone else like me is not au fait with coding software and we (or maybe just me!) will need help with compiling the code. That's the aspect that most concerns me with replicating your exceptional work. A pre-burnt EPROM could be a big selling point. BobUK.
I missed it and tried to zoom in but still couldn't see it. What size gt2 belt did you use on the encoder? I also have a G0602 and I'm currently doing both your VFD conversion and the lead screw. These videos are gold!! Thanks for your explanations
That one is 127 grooves, part number "A 6R51M127060" from SDP-SI. Be warned that Gates belts are sized by number of grooves, but many import belts are sized by belt length, in mm. SDP-SI has a center distance calculator that comes in really handy for selecting belts.
Great, excellently produced videos. I'm in the process of installing encoders on my lathe and mill. I too used 3D-printed pulleys to show proof of concept. What 3D modeling software did you use to model the pulleys? Where did you find the information necessary to model the teeth on the timing pulley? I found a parametric timing gear model on Thingiverse that did a great job modeling the timing pulleys but it is in OpenSCAD. But, not being conversant in OpenSCAD I was not able to make all the modifications I wanted to make to the pulley. I'd like to be able to model a timing pulley in Fusion 360, my 3D modeling software of choice. I understand that the Gates GT series of pulleys tooth profile is proprietary and not available. I purchased my aluminum timing pulleys and belts from SDP/SI. Very good company to work with. They have a lot of technical information online.
Just an idea .... when you put an optical trigger disc on the lathe spindel ..you have no backlash problems ..of corse you have some encoding/decoding ussues but with your knowledge of electronics i think it would be feaseble with more accuracy.
I don't think the backlash is a problem except for the noise it makes with the plastic gears. The angular error it's causing is isn't going to accumulate. A synchronous drive directly off the spindle would solve the same issue and remove the hazard of the belt getting contaminated with gear lube. IIRC, he's using a 4096 count/rev encoder. Trying to fabricate an equivalent resolution encoder that fits directly on the spindle would be a lot more difficult (and prone to contamination itself) than pressing the output gear off the spindle and replacing it with a synchronous belt sheave. I'll agree that it would be an interesting design challenge though.
Have you looked at adding a pendant to the mill so you have an interface you can use without leaning over to the keyboard when doing manual operations?
I actually have a ShuttleXpress on the machine, and I use it when edge-finding. For manual operations, though, I usually set the feed rate and use the cursor keys. Just habit, I guess.
@@Clough42 I've been putting off one for my CNC router since I use the keyboard the same way. Thinking of doing a CNC conversion on my Lagun mill, but I think I'll want a pendant on it.
This is a great project! I wonder if the software can be easily adapted to a closed loop control driver modified stepper? BIGTREETECH S42B V2.0 Upgrade 42 Stepper Motor, Closed Loop Driver Control Board, 3D Printer Parts for SKR V1.4/V1.4 Turbo/V1.3, MKS Gen L (with OLEP Display) Thanks Doug
I am also waiting for the next video :-) very good project ... and when you are ready and it is running (successful) I will adapt it on my EMCO Emcomat 8.6 *smile*
I will. There's some basic info in the GitHub project (link in description) and more is coming. If I can make it work, I'd really like to make PC boards available.
I'm really enjoying the video content. I used to work on CNC lathes, programming /setting/operating. And I look forward to seeing the finished driver. Have you thought about a patent for your work?
I have a question about you lathe stand. What is it made of? Because it looks like wood? What do you do to prevent/dampen vibrations that cause tool chatter?
The lathe is bolted to my wooden workbench. The top is a grid of half-lapped 2x4s, a layer of 3/4" particle board and a layer of hardboard. It's supported on a 3/4" plywood cabinet carcass and bolted to the wall.
Really excited about doing this to my9x20 Harbor Freight lathe. I am somewhat chanlanged on the programming and rratio of the steeper motor pulleys and my lead screw TPI.. I have purchased the same hybrid stepper you used and I have a 16 TPI lead screw. Do I go with a 3 to 1 pulleys like you used on your stepper motor ?
@@sethshook I completed mine work great. Can't push it to hard or the stepper motor kicks out. I had to go in to the program and adjust timing. It's been almost several year so I don't remember what I changed Send me an email address and I will forward photos
I take them directly from the end mill manufacturer's (YG-1 Alu-Power) data sheet and then lighten up just a little if I think it's a heavy cut for the rigidity of my machine. In this case, it's a quarter-inch 3-flute tool in Aluminum, slotting 1/8" deep (0.5D). The sheet calls for 10,000RPM and 56.7IPM. That's .0019" per tooth. I think I ran it at .0015" per tooth. Of course, my feed rate is slower since I only have a 2200RPM spindle.
Hey James, did you ever mount the stepper inside the cabinet, above the machine's motor( under the head)? Or is it still outside attached to the spider?
Great intro -- I've muttered that flat, disgusted "Perfect!" more than once myself. Usually preceded by or followed by :-) This project is both entertaining and educational, and is also something I hope to emulate. Clyde
I have a question about the ratio of the spindle to encoder. Forgive me if I'm not seeing this correctly...You were originally going to drive the encoder via the teeth on the idler, but are now driving from the idler shaft. This (I think) introduces a ratio that is no longer 1 to 1. I know you can compensate for this in your programming for your particular spindle to idler ratio. Will there be a way for those of us who have different lathes to adjust for this?
There are numerous methods for connecting a rotary encoder to the spindle, and devices other than rotary encoders that could be used. Please do not consider my suggestions as criticisms of the methods you chose. I am just offering alternatives that you may not have considered. You can eliminate the idler gear by replacing the gear on the lathe spindle with a tooth belt pulley. This would eliminate the noise and some maintenance issues associated with the gears. I believe that a better option is to use a reluctor wheel mounted on the lathe spindle with an abs sensor. It will not provide the counts/revolution that your rotary encoder produces, but it is capable of generating more than adequate resolution for this application. ABS sensors are dirt cheap and incredibly durable and reliable. The same can not be said for the rotary encoder your using. In regards to the reluctor wheel, your existing gear could be used or a higher count gear for a little better resolution. Another advantage to going this route is that the reduced pulse count would allow the use of a cheaper and more convenient CPU.
I actually considered most of those options. Replacing the existing gear on the spindle requires cracking open the spindle stack, which requires re-setting the bearing preload. I didn't really want to take that on. The ABS brake sensor is an interesting idea. The approach I'm taking depends on the high pulse count to avoid having to make any assumptions about timing or speed. With this much resolution, I never have to estimate how fast anything is moving and make any steps in anticipation of continued motion. I can make everything reactive and avoid any guessing. So far, I'm getting results I like. We'll see how the rest of the project goes.
@@Clough42 I have a similar lathe. It's not a big deal to take the pulley and gear off. You'll have to at some point to replace the grease anyway and, because the Chinese use such high quality stuff, this is something that should be done sooner rather than latter:) A simple method to bore out those aluminum pulleys is to use shim stock/spacers with a four jaw chuck. Use a DTI on the existing bore to center. Between swarf and lubricants, a rotary encoder is going to have a very tough life in this application. In contrast, this is nothing compared with what a typical ABS sensor was designed to endure. In regards to the reduction in resolution, I would have approached this matter from a different direction. Rather than taking samples based upon a time interval, I would calculate how many servo steps are required in relation to spindle counts. This could be easily accomplished with an Arduino or Raspberry pi, and there's an abundance of supplemental hardware that works with these platforms. There's no need to reinvent the wheel.
@@ChristophPech There's a physical limit on how many magnets could be distributed on lathe spindle. They have to be far enough apart for a hall effect sensor to function properly and there's limited space on a lathe for a wheel large enough that would provide an adequate diameter for acceptable counts/revolution. There is also another issue with this method; swarf on the magnets would be an annoying problem. In contrast, ABS sensors are basically very sensitive proximity detectors that can detect individual gear teeth. Install a 120 tooth gear and that's how many counts the senor would provide per revolution (one pulse every 3 degrees). That should be adequate for this project.
I started with a 2N.m servo, and it's just barely enough for this lathe. I'm now running an inexpensive 180w servo with 3:1 belt reduction, and I'm much happier. The biggest issue is that the hybrid stepper loses torque at high speeds, and this will vary from motor to motor, based on inductance.
Just a thought and maybe not a valid one; if all your feed selections are programmable, wouldn't it be more "mechanically positive" to directly drive the other end of the lead screw rather than go through the gear change box?
Perhaps. I like the idea of keeping everything under the cover, and I might someday end up wanting to use the extra gears. There's another 2:1 reduction in there if I need it.
Good lord that's beautiful. So when you manufacture these for everyone I'm thinking the total cost for all the machined parts should be somewhere in the range of $ .42 and $3.60. Either way, I'd like 50 sets.
It occurs to me that this project is sufficient justification to build a desktop lathe around it, and possibly a lathe/mill combination tool ... or to license the technology to an existing manufacturer (preferable domestic) to build a small desktop CNC machine tool around the threading tool ...
Uhh that would be the banjo on your lathe not the spider. Easy to make a common mistake while dealing with the videoing and work being done though. As always great informative video.
Could you cut a slug of metal, weld it into the oversized bore after you've back gouged the join, the drill and bore that to size? Assuming you have a welder at this stage. I do know you have a lathe :P
I don't have a welder. I could press something in or use loc-tite. The bigger concern is holding the pulley in the lathe, getting it centered properly, and not damaging the aluminum profile. I think the replacement pulley was about $6. Plus three weeks' time for delivery.
I think this is exceptionally produced content and it's evident that you have the spirit of a researcher/teacher. Just the ability to verbally connect with your audience in a clear-concise manner puts you into a class of your own. Showing mistakes, errors and oversight helps keep the content humanly interesting. Good job. (virtual back patting)
Very well stated!!
Thanks! Mistakes are my specialty.
@@Clough42 Mistakes are the best learning tool!! When someone asks me what my education is, I tell them I have a PhD from the School of Hard Knocks... It usually gets a laugh, but it's more true than most folks realize!!
I’m stunned your subscriber count hasn’t skyrocketed in the last couple months. You’re videos are gold. They’re so professionally produced and enjoyable to watch. Just the right amount of detail and never boring. Thanks for all the time you invest in these.
You are welcome. I did have a big bump from 1K to now nearly 10K over the last few months, but things have slowed. Summer is upon us, and things usually slow down.
Every episode I watch, I can’t praise enough your workmanship. Fabulous! Thank you for allowing us to tag along. FYI, the suspense is killing me.
Yeah, me too!
Good to see 3D printing can help out with practical prototypings like these. Looking forward to seeing the whole thing work!
I don't know how I got by without 3D printers. I'm always printing little parts and tools that I need for various things.
13:46 "looks alright to Me" proceeds to max out the rpm instantly 😂 man i just love this series and the channel in general! Joking aside this is really a top notch channel.
Thanks!
I don't own my G0602 lathe any longer - I upgraded to a G4003G lathe, but I am thoroughly enjoying this series! Thank you for sharing your journey.
I enjoy watching you work, you have a lot of talent and a knack for finding solutions to conundrums.
I recently subscribed and want to thank you for your hard work and willingness to share.
I'm replacing the gear on the spindle with a timing pulley. The main drive pulley came off quite easily. Just need to put a big bore and keyway in the timing pulley. Your layout and presentation of content is awesome. Very nice work.
Your work, presentation, and production is excellent. I dont understand why there arent more likes!
Thanks!
This is what home cnc is all about. Stuff that would be difficult manually is relatively easy with even a low power cnc machine. It’s going to be interesting to see if you can cnc the feed screw.
SDP/SI! That place has saved my bacon on multiple occasions. Fast (kinda pricey) shipping, reasonable prices on parts, and just about everything under the sun available.
Yeah, SDP-SI and Automation Direct are good stateside premium sources.
As always SUPERB! Some of the finest instructional videos I have EVER seen. I too can't wait to see this come to fruition.
Keep up the good work kind Sir! And may Jesus continue to gift you with unique talents.
My lady watches EVERY video with ernest as I do.
Very cool. Glad you could join us.
James, thank you again for another well thought out (along the way) project. I commend you for not hiding goofs and gaffs and looking human in the process. Someday, I hope to be as competent as you. I am going to skip back and watch from the beginning. This looks like a great process for solving my gears kicking loose under tension. I want to see how you calibrate all the speed settings for proper threading without having the change gears, and simplify it for my old man head. My equipment is similar, an old Asian lathe that looks like the Grizzly G4000 and a mill the same as your G0704. All mechanical still.
After 2 years I saw your video I will start the same project. I will try to do it with Arduino and all 3D printed parts. Thanks for sharing 🙂
I do not have a lathe but I am so excited to see all of this work. I am building a shock dyno on the same principle i got the same encoder and digital readout and the BSS 138 converter. I also got the F280049C encoder. Great job. Thank you! I am using the same VFT you are using to run my 3 HP 3PH motor.
Nice!
Gates can't help the Chinese part numbering debacle... but the technical resources on the Gates website has a really cool application to size belts for pulley sizing and spacing. It's simple to use and very flexible. Our high school robotics team uses it.
I can't wait to see this whole thing up and running.
I usually use the "Center Distance Designer" on the SDP-SI website. It's very helpful.
Awesome progress and I'm really enjoying this series. Thank you once again.
Excelent work. I am very anxious to see how it works. Thanks for sharing.
So am I. I have started working on it, and it's looking good so far.
Also would be cool to have a readout showing degrees of rotation of the spindle to allow using it as a low accuracy dividing head for drilling flanges/ holes around a PCD etc with a toolpost drill for those of us without a CNC milling machine.
This was the first video I saw of yours where you were running the Grizzly-based CNC, and was totally impressed by it running. That and the fact you just buzzed up a couple pulleys and another bracket on the 3D printer really makes me want to try one, lol!
If you wanted to mount the leadscrew stepper direct drive, you'd have the option of doing so by mounting the stepper to the tail stock end of the leadscrew and making up a mounting bracket. That way you get to leave the gear box cover on the lathe and move the stepper wiring away from the lathe controls.
About Chinese belts misunderstanding - I think that this type of numbering of belts is common at least here in Eastern Europe. I always order belts by their length in mm, not by number of teeth.
That type of numbering is probably common for all metric users
This is a wonderful series. Can’t wait to see if the motors actually do the job.
You and me both. The good news is that steppers and servos have standard interfaces, so it's easy to scale up or down for different applications.
Damn! Just binged on this build and got to the end! Seriously looking forward to seeing things spin! I love the idea, I'd like to do a similar thing as a part of a gradual CNC conversion on my Colchester lathe.
Interestingly what you refer to as a "spider" here in the UK we call it a banjo.
Great update. I've been following a few of these ELS projects and eventually decided to give one a go. It was a commercial one, but quite cheap and he sold it as a complete kit with (recommended) stepper motors. I already had a couple of servo's so I decided to go with them instead. I direct drive the lead screw thru the gearbox, as you do and drive the cross side with a 1:1 belt drive, as this it what the creator suggested. It works fine but I have noticed that the servo(DMM) on the cross side has little holding torque and can 'back out' a little on heavy cuts sometimes(I guess the lead screw one might be the same). A stepper would have better holding torque to stop this. I think if you are going to use a 1:1 or low ratio drive, steppers would be fine and if you are going to use a belt drive and higher ratio servos would be fine. Either way I'm having fun and look forward to seeing where you take yours.
I do have a Mazak Rex 610-1500 20Hp lathe with no change gears and am interested in giving yours a go on this one, if I can work out sizing and find servo's large(and cheap!) enough to drive it!
Was the servo on the cross-slide powered when it was backing out? Or was it disconnected? Ballscrews, or just lead screws?
Yes was powered, it's a Canadian DMM servo so I can connect the driver to the laptop and see if there is something I can change in the settings, but I didn't think servos had the same holding torque as steppers? Steppers are full power all the time, servo's power as required(I thought?). Didn't move much and probably no more that if I was using it manually, but I was going to put a ballscrew on the cross slide to remove backlash and this could make it worse as the ballscrew would have far less mechanical drag/ resistance.
Would be interesting to do a test with stepper/(true )servo connected to a ballscrew/leadscrew with a pushing load applied to the screw to see if there was any movement. I of course maybe completely wrong, as I often am, but am happy to proved so for the greater good :-).
awesome.. ive printed XL 5mm pulleys before and they come out great and hold up to surprisingly high torque..
My only concern is wear on the keyway. It's possible the plastic pulley will be fine, but I'll probably make a metal one anyway.
@clough42 You can always part off on the round part of the stock. Cut your flats just deep enough. Then part on the very edge of the round stock. A little metrology, and you can have a precise part....
Just turn it, part it off, and put in the mill vise.
The flange is plenty big enough to hold it, and the pip will register for height.
The mistake is thinking the flats have to be perfectly parallel - they only need to be close, and it can be done by eye.
This is great. I have really enjoyed the excellent work on these videos.
Awesome! Really looking forward to see the next video.
Thanks!
Another excellent video! A quick-ish fix for the over sized pulley would have been to bore it out larger, press fit a plug it, then bore to the correct size. Or....order a new one. 😬
I thought about it. It probably would have worked fine, but I'd have to hold it in the 4-jaw, with some kind of fixture to prevent damaging the aluminum teeth. Also I didn't want to put the change gears back on the lathe or use it without feed. So...there were enough excuses it seemed easier to just 3D-print one.
Where there’s a mill there’s a way. 😬A couple hold down clamps, interpolate a hole, bobs your uncle. I actually appreciate the 3D printed part, I like seeing something that’s gonna take some stress being printed. Sure makes for some quick and easy prototypes.
Thank you for producing such great content.
Great job. Super Interesting to watch how you solved some of the problems. Thank you for the insight into your solutions. Best wishes.
Thanks!
Gates also measure their belts by the pitch length rather than by the number of teeth. For instance a 5mm pitch Powergrip GT3 (300-5MGT) is 300mm long with a 5mm pitch tooth and has 60 teeth.
So glad I ran across your channel. Just about everything I am interested in and trying to learn is right here. I have binge watched every one of your videos and anxiously await more. I especially enjoy the process as you explain it from the need/idea to planning and design in Fusion 360, prototyping, problem solving, machining (3D and CNC), mistakes, solutions, everything! I find your content inspirational and motivational in just the right ratio. Congrats on the weight loss and thanks for taking the time to produce such excellent content for those of us just starting our journey (however long it took us to get started:)...
P.S. Are you planning anything special for the 10K subscribers milestone which is right around the corner?
Hi James. What software are you using to design your 3d printed pulleys? You've got an amazing project here. It's so instructive to see your entire process. You epitomize what education is really about. Thanks many times over.
Very cool video! Can't wait to watch the progress
with this asesories you havent backless and you can have a high presicion lathe you have good ideas in your head , to that in all you live , succes mister
I'm curious, why not go for optical sensing of spindle speed? It seems like you could've milled one or more slots in the 60 tooth gear, shined an IR LED through it at whatever sensor you prefer, and there's your data.
This is actually an optical encoder, and a very high-resolution one. Milling slots in the gear isn't something I had considered. For that matter, I could probably read the teeth of the gear optically. It *might* even be possible to get 240 counts/rev with some kind of quadrature setup. The software approach I'm taking at the moment depends on the high resolution, though, so I'll probably continue with that for now.
@@Clough42 Yeah, I suppose for thread synchronization you need not just RPM but high accuracy spindle position at high sample rate. Still, you could read the teeth like you mentioned, or attach a slotted rotor sheet to the gear with as many slits as you like for high resolution optical sensing. It just seems to me that doing the sensing with a mechanically linked encoder is inviting complication, error, and wear.
@calastigro If you would really want to read the gear, that would be better served by an inductive / Hall effect sensor: it's what ABS in cars actually does; you really wouldn't want to muck around shining light through the cogs. But as noted, the resolution would be drastically worse, and that's not a good thing when you're trying to cut a thread instead of just figuring out whether a wheel is slipping or not...
Hi James, you said the rotary encoder 3Dprinted gear is mod 1in the video at 4:45minutes. I have the same lathe type but my gears are mod 1.5.
Mod 1 is for very small loads.
There are two gear types on my lathe. The first 60T plastic gear fed from the 40T plastic gear on the spindle is about 90mm in diameter, so it's likely mod 1.5. The rest of the cast iron gears appear to be mod 1. I just measured the 60T cast iron change gear, and it's approximately 60mm in diameter. The leadscrew gear train carries relatively small loads.
@@Clough42 Thank for the mod clarification.
Overall gear ratio from spindle, rotary encoder up to the leadscrew you prefer to keep as integer 1, 2 or 4 and in combination with the lathe selector ABC and in the CCSsoftware code microsteps variable the total ratio should be 1?
This is my favorite TH-cam channel. I have 2 questions: what motors are you using for your table top mill, steppers or servos? Also i never hear your air compressor kick on and off. Where do you have your air compressor located?
Direct drive is the way to go. I've got NEMA24 560oz. in. closed loop stepper motor which easily drives Z axis head stock with 1605 ball screw on my Sieg mill. So for this application it will do the job just hands down. Otherwise PU toothed belt with steel cord is going to be fine as well.
great build so far, use only premium gates belts ,worth it over imports , I also like belt drives with steppers I believe they do a good job at damping. on expensive VMC's you will also see belts used on drives and encoders.
Has anyone given thought to using the existing RPM readout on many of the current VS lathes as an encoder? curious if there is a way to tap into that instead of introducing an additional point of failure into the mix. I have all the parts to do the conversion but there may be an opportunity to simplify here.
I think the main issue with the existing sensors would be resolution. It is possible to thread with one or ten pulses per revolution, but you have to make a lot of assumptions about constant speed operation and it would be slow to respond to speed or direction changes.
@@Clough42 Fair enough. I actually came to a similar conclusion after a friend said his kept RPM with only 4 magnets around the shaft. I can definitely see the lack of resolution really negating the whole project. Thanks for the sanity check.
Outstanding job
I wanted to say I really appreciate this, and all your hard work.
I urge other people to stop buying from Grizzly.
They offer what seems to be a good product at a good price, but this is the sort of issue that comes from them. They will continue to sell sub-standard equipment as long as people keep choosing them over other dealers who sell equipment that is not missing basic functions.
You mean the lack of a threading gearbox?
@@Clough42 Gearbox, reverse tumbler, etc. I can't remember all the things, I just remember being very frustrated about how close they came to being a good product.
Wound up buying both a lathe and mill from Precision Mathews. They both cost more, but end up cheaper in the end because you don't have to modify them
@@joshua43214 Grizzly has a bunch of different offerings with different price points and feature sets. The G0602 is a particularly good deal if you're doing a CNC conversion and throwing all the extra stuff away. IIRC, the PM lathes have power cross feed. That's a winner.
Bore the pulleys out to 20-22mm, then make bushing that press fit in with a 14mm center hole and keyway. You could make them from steel which would wear better than the aluminum.
Have you thought about adding Diameter pitch to cut gear worms???
Hi James. I think it was this video where you mentioned you might eventually try and mount the Servo toward the back of your lathe. Just wondering if its still on your todo list? I have the same lathe as you and use my ELS all the time. And I think of the door being open often too. Just curious if you've given any more thought. Thanks. I'm just assembling my new box and ELS faceplate that just arrived in the mail. Thanks for all the great content.
Excellent video. Can't wait for the next one!
Excellent series James, I'm buying up the bits and pieces to begin this build (already bought your kits) on my Grizzly G4000. I'm down in the Caribbean, so everything ordered on-line takes forever to get here, especially from China. Very frustrating when working on a project as you would appreciate, and Lord forbid the wrong thing arrives! :)
Two questions:
1) The ebay link for the Servo motor is dead, can you suggest an alternate source?
2) Do you have recommended source(s) for acceptable toothed belts and pulleys?
Hi why I cannot turn the hand crank on a lathe after I install the stepper motor thanks Eli
In the lands Metric, M series belts are ordered by Series ~ Length ~ Width. In Australia I don't think any belts are classified via tooth count, the books however have the tooth count listed also.
You may find that you could run into some belt tooth climb with the 3m series belts in practice. On a 2hp ball screw drive a 5m is normally used, the 3 m may need to be tight beyond the capability of the nose bearing of the motor? I don't think that will be an issue for you until you get up into the 2mm per rev range of diameter turning. Say a 2 X 2 mm cut on steel. The risk of this problem is greatly reduced the bigger the drive pulley diameter becomes.
Cheers
I doubt you'd want to take such a heavy cut on a machine of his calibre.
Hi James. By replacing the leadscrew by a ball screw you can turn that lathe into a CNC machine much more capable than just cutting threads without change gears. Installing the servo straight to the leadscrew on the tail stock end is not an option? less cluttered space. Great job and keep on. Subscribed and following
Will you do the final assemble video on the Diamond Grinder?
Probably someday. It's sitting on the floor next to my bench right now, still waiting for my attention. :)
Hi James, your lathe is very similar to mine and I'm very interested in this. To repeat what I've said earlier, please bear in mind that anyone else like me is not au fait with coding software and we (or maybe just me!) will need help with compiling the code. That's the aspect that most concerns me with replicating your exceptional work. A pre-burnt EPROM could be a big selling point. BobUK.
Where's a good place to find a decent selection of timing pulleys?
I missed it and tried to zoom in but still couldn't see it. What size gt2 belt did you use on the encoder? I also have a G0602 and I'm currently doing both your VFD conversion and the lead screw. These videos are gold!! Thanks for your explanations
That one is 127 grooves, part number "A 6R51M127060" from SDP-SI. Be warned that Gates belts are sized by number of grooves, but many import belts are sized by belt length, in mm. SDP-SI has a center distance calculator that comes in really handy for selecting belts.
@@Clough42 Thank you so much! I really appreciate you trailblazing this incredible project! ❤
I see you used the slitting saw under digital control. One digit, specifically... :)
Great, excellently produced videos. I'm in the process of installing encoders on my lathe and mill. I too used 3D-printed pulleys to show proof of concept. What 3D modeling software did you use to model the pulleys? Where did you find the information necessary to model the teeth on the timing pulley? I found a parametric timing gear model on Thingiverse that did a great job modeling the timing pulleys but it is in OpenSCAD. But, not being conversant in OpenSCAD I was not able to make all the modifications I wanted to make to the pulley. I'd like to be able to model a timing pulley in Fusion 360, my 3D modeling software of choice. I understand that the Gates GT series of pulleys tooth profile is proprietary and not available. I purchased my aluminum timing pulleys and belts from SDP/SI. Very good company to work with. They have a lot of technical information online.
Just an idea .... when you put an optical trigger disc on the lathe spindel ..you have no backlash problems ..of corse you have some encoding/decoding ussues but with your knowledge of electronics i think it would be feaseble with more accuracy.
I don't think the backlash is a problem except for the noise it makes with the plastic gears. The angular error it's causing is isn't going to accumulate. A synchronous drive directly off the spindle would solve the same issue and remove the hazard of the belt getting contaminated with gear lube.
IIRC, he's using a 4096 count/rev encoder. Trying to fabricate an equivalent resolution encoder that fits directly on the spindle would be a lot more difficult (and prone to contamination itself) than pressing the output gear off the spindle and replacing it with a synchronous belt sheave.
I'll agree that it would be an interesting design challenge though.
Have you looked at adding a pendant to the mill so you have an interface you can use without leaning over to the keyboard when doing manual operations?
I actually have a ShuttleXpress on the machine, and I use it when edge-finding. For manual operations, though, I usually set the feed rate and use the cursor keys. Just habit, I guess.
@@Clough42 I've been putting off one for my CNC router since I use the keyboard the same way. Thinking of doing a CNC conversion on my Lagun mill, but I think I'll want a pendant on it.
Washers as standoffs between the mounting plate and the servo?
This is a great series!
This is a great project! I wonder if the software can be easily adapted to a closed loop control driver modified stepper? BIGTREETECH S42B V2.0 Upgrade 42 Stepper Motor, Closed Loop Driver Control Board, 3D Printer Parts for SKR V1.4/V1.4 Turbo/V1.3, MKS Gen L (with OLEP Display) Thanks Doug
I am also waiting for the next video :-) very good project ... and when you are ready and it is running (successful) I will adapt it on my EMCO Emcomat 8.6 *smile*
That's the plan. My goal is to make this as adaptable as I can.
James,
How many teeth / long is that GT2 belt you used? Ordering some parts and could not see where you showed or mentioned details on that belt.
Clough42. What kind of mill do you have? It’s sweet.
Paul
Excellent job!! I'm hoping you publish the electronics details so i can make my own unit, the mechanical side will be no problem, here's hoping.
I will. There's some basic info in the GitHub project (link in description) and more is coming. If I can make it work, I'd really like to make PC boards available.
@@Clough42 Thanks for your consideration, great channel by the way!
I'm really enjoying the video content.
I used to work on CNC lathes, programming /setting/operating. And I look forward to seeing the finished driver.
Have you thought about a patent for your work?
Your open to the video made me laugh. Thank you
You're welcome. Thanks for stopping by.
Is there not an optical sensor that is fast enough to read a painted dot on the lathe drive, instead of a mechanical connection to a separate encoder?
It's not just the RPM, it's the position. The encoder has ~4k(?) measurements per revolution - that would be a lot of painted dots...
Correct depth flats interrupt interrupted cuts.
I have a question about you lathe stand. What is it made of? Because it looks like wood? What do you do to prevent/dampen vibrations that cause tool chatter?
The lathe is bolted to my wooden workbench. The top is a grid of half-lapped 2x4s, a layer of 3/4" particle board and a layer of hardboard. It's supported on a 3/4" plywood cabinet carcass and bolted to the wall.
Really excited about doing this to my9x20 Harbor Freight lathe. I am somewhat chanlanged on the programming and rratio of the steeper motor pulleys and my lead screw TPI.. I have purchased the same hybrid stepper you used and I have a 16 TPI lead screw. Do I go with a 3 to 1 pulleys like you used on your stepper motor ?
Im just starting on mine (ive got a HF 9x20 as well) and trying to figure out the gearing. what did you end up doing if you dont mid me asking?
@@sethshook I completed mine work great. Can't push it to hard or the stepper motor kicks out. I had to go in to the program and adjust timing. It's been almost several year so I don't remember what I changed Send me an email address and I will forward photos
James Iam trying to duplicate your ELS setup but don't know where to purchase the high torque drive motor m3 pulleys, HELP, Thanks Al
We are proud of you....💕💖
Great work
What are your typical aluminum slotting recipes?
I take them directly from the end mill manufacturer's (YG-1 Alu-Power) data sheet and then lighten up just a little if I think it's a heavy cut for the rigidity of my machine. In this case, it's a quarter-inch 3-flute tool in Aluminum, slotting 1/8" deep (0.5D). The sheet calls for 10,000RPM and 56.7IPM. That's .0019" per tooth. I think I ran it at .0015" per tooth. Of course, my feed rate is slower since I only have a 2200RPM spindle.
Hey James, did you ever mount the stepper inside the cabinet, above the machine's motor( under the head)? Or is it still outside attached to the spider?
nice work
Grrr... really want to see it go. Did you convert your mill on your own? Do you have details on the conversion?
Great intro -- I've muttered that flat, disgusted "Perfect!" more than once myself. Usually preceded by or followed by :-) This project is both entertaining and educational, and is also something I hope to emulate. Clyde
Thanks! I thought it was appropriate. It was either that or just pretend I don't make mistakes. :)
I have a question about the ratio of the spindle to encoder. Forgive me if I'm not seeing this correctly...You were originally going to drive the encoder via the teeth on the idler, but are now driving from the idler shaft. This (I think) introduces a ratio that is no longer 1 to 1. I know you can compensate for this in your programming for your particular spindle to idler ratio. Will there be a way for those of us who have different lathes to adjust for this?
Okay...I re-watched from the beginning and see you're using encoder pulleys with a 60:40 ratio.
Thanks for doing this project!
You SURE have enough "toys" in your shop!
Are you using Code Composer Studio and programming in C? I've just started playing with my board and that IDE. Great work!
Yes. I'm using Code Composer Studio 9. The video description has a link to my code in GitHub.
Yea besides your lathe is undergoing surgery. Love it!
There are numerous methods for connecting a rotary encoder to the spindle, and devices other than rotary encoders that could be used. Please do not consider my suggestions as criticisms of the methods you chose. I am just offering alternatives that you may not have considered.
You can eliminate the idler gear by replacing the gear on the lathe spindle with a tooth belt pulley. This would eliminate the noise and some maintenance issues associated with the gears.
I believe that a better option is to use a reluctor wheel mounted on the lathe spindle with an abs sensor. It will not provide the counts/revolution that your rotary encoder produces, but it is capable of generating more than adequate resolution for this application. ABS sensors are dirt cheap and incredibly durable and reliable. The same can not be said for the rotary encoder your using. In regards to the reluctor wheel, your existing gear could be used or a higher count gear for a little better resolution. Another advantage to going this route is that the reduced pulse count would allow the use of a cheaper and more convenient CPU.
I actually considered most of those options. Replacing the existing gear on the spindle requires cracking open the spindle stack, which requires re-setting the bearing preload. I didn't really want to take that on. The ABS brake sensor is an interesting idea. The approach I'm taking depends on the high pulse count to avoid having to make any assumptions about timing or speed. With this much resolution, I never have to estimate how fast anything is moving and make any steps in anticipation of continued motion. I can make everything reactive and avoid any guessing. So far, I'm getting results I like. We'll see how the rest of the project goes.
@@Clough42 I have a similar lathe. It's not a big deal to take the pulley and gear off. You'll have to at some point to replace the grease anyway and, because the Chinese use such high quality stuff, this is something that should be done sooner rather than latter:)
A simple method to bore out those aluminum pulleys is to use shim stock/spacers with a four jaw chuck. Use a DTI on the existing bore to center.
Between swarf and lubricants, a rotary encoder is going to have a very tough life in this application. In contrast, this is nothing compared with what a typical ABS sensor was designed to endure.
In regards to the reduction in resolution, I would have approached this matter from a different direction. Rather than taking samples based upon a time interval, I would calculate how many servo steps are required in relation to spindle counts. This could be easily accomplished with an Arduino or Raspberry pi, and there's an abundance of supplemental hardware that works with these platforms. There's no need to reinvent the wheel.
What about just gluing a bunch of magnets on the spindle gear and then put a hall effect sensor over it?
@@ChristophPech There's a physical limit on how many magnets could be distributed on lathe spindle. They have to be far enough apart for a hall effect sensor to function properly and there's limited space on a lathe for a wheel large enough that would provide an adequate diameter for acceptable counts/revolution. There is also another issue with this method; swarf on the magnets would be an annoying problem.
In contrast, ABS sensors are basically very sensitive proximity detectors that can detect individual gear teeth. Install a 120 tooth gear and that's how many counts the senor would provide per revolution (one pulse every 3 degrees). That should be adequate for this project.
Whats the spec on the servo 2 or 3nm
I started with a 2N.m servo, and it's just barely enough for this lathe. I'm now running an inexpensive 180w servo with 3:1 belt reduction, and I'm much happier. The biggest issue is that the hybrid stepper loses torque at high speeds, and this will vary from motor to motor, based on inductance.
Lookin' good.
Just a thought and maybe not a valid one; if all your feed selections are programmable, wouldn't it be more "mechanically positive" to directly drive the other end of the lead screw rather than go through the gear change box?
Perhaps. I like the idea of keeping everything under the cover, and I might someday end up wanting to use the extra gears. There's another 2:1 reduction in there if I need it.
nice!!! cant wat wait!
Looking good still seems over kill for ppr encoder
Overkill is underrated. :). The high resolution encoder allows me to create smooth motion without interpolating.
Thanks
Good lord that's beautiful. So when you manufacture these for everyone I'm thinking the total cost for all the machined parts should be somewhere in the range of $ .42 and $3.60. Either way, I'd like 50 sets.
Thanks!
It occurs to me that this project is sufficient justification to build a desktop lathe around it, and possibly a lathe/mill combination tool ... or to license the technology to an existing manufacturer (preferable domestic) to build a small desktop CNC machine tool around the threading tool ...
Those little ones are the belts my vacuum cleaner uses..
Uhh that would be the banjo on your lathe not the spider. Easy to make a common mistake while dealing with the videoing and work being done though. As always great informative video.
Could you cut a slug of metal, weld it into the oversized bore after you've back gouged the join, the drill and bore that to size?
Assuming you have a welder at this stage. I do know you have a lathe :P
I don't have a welder. I could press something in or use loc-tite. The bigger concern is holding the pulley in the lathe, getting it centered properly, and not damaging the aluminum profile. I think the replacement pulley was about $6. Plus three weeks' time for delivery.