My lathe originally had a taper attachment. The bar got lost at some point, but it still had the long slotted cross slide extension bolted on instead of a chip guard like your 820. I sold or gave away the extension and made an extended chip guard for mine which you can see at the 16:31 mark. Logan didn't spend a lot of time on cosmetics, which is why you can still see the milling marks on the back end of the cross slide. I stoned it a bit before attaching the bracket for the scale and the shininess made the milling marks stand out. It's all still going strong, and I just used it a couple of days ago. Good little lathe.

I didn't wire a remote pot, I use the one on the VFD. The braking resistor connection is really straightforward. You won't have any trouble figuring it out.

The documentation is at github.com/ccomito1223/ELS-Jon. I used a "Nema 24" motor from omc-stepperonline that was only available for a short time. It's a unicorn, with an 8mm shaft, Nema 23 mounting holes, 3.5Nm torque and was (I think) 110mm long. I've cut 8tpi Acme threads with it with no hiccups. I did get some occasional chatter with a 3Nm Nema 23 motor, which was why I upsized. You might get away with a Nema 24 4Nm motor if someone makes one shorter than 125mm, with a 6:1 20T/120T pulley combination.

If I was looking for a replacement for the motor I'm using now I would look at teknic.com/model-info/CPM-SDSK-2321S-RLN/?model_voltage=75VDC and get the 1/4" shaft option. It's a servo motor with step and direction input. I would have to go to a 75-volt power supply, too. Then I would look at timing pulley and belt options. The 15mm belt that I used is probably overkill, and 10mm would probably work fine (timing belts are amazingly strong for their size). I would stick with the 8:1 reduction on the pulleys. The Clearpath motor would probably outperform a stepper. Might be quieter, too. I think they have some cool software tools, too.

I just looked on eBay and found a couple of 8:1 timing pulley sets. I believe an HTD3M set with a 6.35mm (1/4") bore on the small pulley and 192.3mm wheelbase would work with the Clearpath motor. Be advised that the big pulley has to be bored, counterbored and a keyway cut to go on the leadscrew, if you haven't figured that out already. You're getting me going here. I can't get sucked back into this. I have too many other things going on these days :).

Bravo, Aaron. We just left San Diego this morning after spending Thanksgiving with our kids and grandkids. Now we're in Tucson for the night before heading the rest of the way home tomorrow. If the display is indicating the direction change and the leadscrew is turning and changing speed then the only thing left is the direction control connection between the Mega and the motor controller. Make sure DR+ is pulled up to 5V and DR- is connected to D7 on the Mega. -Jon

Thank you! I use it quite a bit, and have some ideas for enhancements. Maybe someday I'll revisit it and make an updated video. Maybe someday I'll make another video, period :).

It should be possible to do it using the "enable" input on the motor controller, but I haven't tried it.

The Arduino Mega gets the job done, but there's not a whole lot of CPU left to spare. It handles 1500rpm just fine, but going above 2000 would get into problems with the interrupt handler. Easiest way to resolve that would be to reduce the spindle encoder resolution from 800ppr to 600 or even 400. The feed resolution would still be sub-tenths.

There are several different ways to configure it. I don't know where my sketch is now, but I believe that I used the 3-wire mode with the E-Stop mushroom switch wired to terminal S1 and FWD/REV on S2 and S3. It doesn't have to be momentary switches on S1. Parameter 00-04 controls the mode. The FWD/REV switch is a modular 22mm 2-position rotary selector switch. If you're not familiar with such things you could just use a 2-way toggle switch. You could use an ON-OFF-ON 3-position toggle and run it in 2-wire mode without the E-stop in 2-wire mode. You could use two NC switch blocks on the E-stop switch and wire them in series with the FWD/REV lines in 2-wire mode. More than one way to do it. Please don't electrocute yourself.

@@jonrbryan I probably would use a toggle switch when I get everything setup which will be awhile. I’ll ask more questions when I get started, I’ve handled electric stuff a good bit so I’m pretty carful. I just can’t wrap my head around programming. I really appreciate you answering.

Thank you. They're surprisingly common machines, and I was lucky to get good ones. I once saw a Clausing 8520 in someone's basement on an episode of "This Old House".

Thanks, Matt. Yes, things like the encoder resolution can be easily changed. The Arduino has some practical limits, and if my lathe could go faster than 1500rpm I would probably drop to a lower resolution encoder so that the Arduino would still have some margin. It would still cut very good threads with a 400ppr encoder on the spindle while handling 3000rpm. There are people who manage to cut threads with 1ppm encoders. If I had a Monarch 10EE that could go up to 4000rpm I might start thinking about a faster microcontroller.

Thanks, Robert. I usually use jog to set the limits and leave the half nuts engaged. It resynchronizes with the thread whenever the feed direction is changed. Otherwise I use it exactly like a gearbox with the threading dial. I've toyed with the idea of implementing a rapid reverse, and maybe I'll get around to it one of these days.

I just spent a few minutes looking for it and couldn't come up with the right search terms. If I eventually find it I'll let you know.

@@jonrbryan Yeah, looking at the clutch assembly on the 10EE It's brilliant but definitely over engineered. I think I'm just gonna go with a set screw like yours.

@@jamiewilliams6763 The way I did it is kind of half-assed. Ideally, the dial would be part of the mechanism so that the positive stop aligned with "0" on the dial. My way "works", but requires that the thumb screw be tightened just enough to drag all the dogs into contact at the set point before tightening it firmly. I had to practice with it a bit to be consistent.

@@jonrbryan Right, I was concerned about binding. I was planning to having the two pins on the dial with a shoulder for the rings to sit on and have the lock screw on the other side locking to the hand wheel. Just gotta decide if I want to use alloy or O1 to make the rings. It's a shame that this is not a feature on any modern lathes.

I'm using an Arduino Mega, a spindle encoder with 800 counts per revolution, and the stepper is set to 400 steps per revolution and drives the 8tpi lead screw through an 8:1 reduction. It takes 25,600 steps to move the carriage one inch. The lathe maxes out at just under 1500rpm, but has enough margin to get close to 2000. Faster than that would require decreasing the encoder resolution. The real-time code is all interrupt driven, and doesn't lose counts or steps at reasonable feeds. I enhanced the user interface since I did this video and display the maximum spindle speed for the selected pitch/feed rate. I also warn the user if the spindle RPM is too high for the selected rate. It won't drive the carriage at 4tpi with the spindle running at 1500rpm, but you couldn't do that through a gear train either, and would likely strip the gears. Ultimately, the system is torque limited, and starting the spindle with the rate set too high just causes the stepper driver to error out because it doesn't have enough power to keep up. The code and design information are available at github.com/ccomito1223/ELS-Jon. We have a small user group at groups.io/g/AtomicELS.

@@jonrbryan Good to know! Thank you for sharing!

I'm using a 1/2hp Hallmark Industries MA0505E 3-phase motor from Amazon that I'm driving with a Fuji Frenic-Mini FRN0003C2S-6U VFD from Wolf Automation. I put the VFD and braking resistor in a proper enclosure after this video was made. It's worked great for everything that I've done with it so far, and I haven't felt any need for more horsepower.

Thank you very much for the information.

Hoping you do more videos on your 8520.

I got lucky and picked mine up from a retiring machinist. He had it set up to do finer work because it would hold closer tolerances than his Bridgeport clone (that I was actually there to look at). It's a good, solid machine if you can live with its limitations.

It's a Grizzly 5/8 with a JT3 taper. I believe the JT3 to MT2 arbor was also from them, but it's been too many years. It's been a good chuck.

@@jonrbryan Thank you for the info sir.

Thank you.

Thanks, Luc. I bought a couple of STM32 boards just in case, but the Mega is working fine. I've made some relatively minor modifications to both the hardware and code since I did this video. I was able to squeeze a slightly larger stepper in and also changed the stepper-to-lead screw ratio to 8:1, which gave better results cutting 8tpi Acme threads in bearing bronze. There is an "Atomic Leadscrew" forum on groups.io if you want to check it out.

Certainly. The fundamental components are the same. I used an 800-count spindle encoder, which works very well on my lathe. For an indexer I would want higher resolution, since any angular error is multiplied by the radius of the workpiece.

Well, the carriage won't move unless the half nuts are engaged. When they are left engaged I sync to the thread on direction changes, and have left and right programmable stops. Doesn't matter if it's English or Metric. It can also be used conventionally, engaging and disengaging the half nuts on a mark just like a regular geared setup. I've let this languish because it does exactly what I want it to do right now. There are some little things that I would like to add, and will get around to eventually. It's just that other things have required my attention for a while. There's an "Atomic Leadscrew" forum on groups.io where I've offered support for people who have implemented their own versions. The biggest stumbling block for people has been getting reliable encoder signals. Mine is rock-solid (of course :/).

I tried using the Arduino internal pull-ups at first, but switched to 2k external resistors because the slow rising edges were resulting in random counts in the wrong direction. I haven't had a problem since.

@@jonrbryan bingo. Thank you so much! Looks like I need to make a few changes to github.com/kachurovskiy/nanoels/tree/main/h1/pcb to support that :/ Hardware is hard! I think I managed to overcome this through software though: once spindle speeds past 10rpm, I just stop reading the second line assuming spindle can't change direction while turning :)

That's an excellent suggestion, Jason, and one I've thought about. I think I would need to add another setup screen to do it properly.

@@jonrbryan I was thinking the same thing. Maybe move the speeds table to an Information screen and use the free space on Settings to add the editors.

@@jonrbryan Another thought would be to use the encoder for the jog function. If you tap the Leadscrew textbox, the background changes color indicating that turning the encoder will jog the leadscrew. Tap it again and the color returns to normal. Maybe even support a single tap to jog slow and a double tap to jog fast.

I can't get the Arduino code to compile. Am I missing a library? Sorry, I'm new to Arduino and having a tough time figuring this part of the code out. AtomicELS:980:5: error: expected constructor, destructor, or type conversion before '(' token ISR(TIMER3_OVF_vect) { ^

@@jasonestes8954 Make sure your target processor is the Mega (2560).

You will need to change some #defines to scale things properly.

Take a look at the Teensy. Programming would be similar, with the Ardunio IDE. The Teensy is pretty small, very fast, and has hardware decoding of quadrature encoders. Wonderful project!

The forces on the tool are quite high. There are some equations in Machinery's Handbook that gave me numbers that were in line with what I observed cutting 8tpi Acme threads where I was getting some divots (momentary stalls) before I changed the motor and gear ratio. I have a couple of STM32 boards that I thought I might have to resort to, but the Arduino Mega is getting the job done at this point.

@@jonrbryan I'm playing around with my own implementation using an esp32 which would let me change parameters via a web interface. I have a lot of c++ to learn to get the Didge motion algorithm implemented though. I want a nice solid feeling lever control for the left/right feeding. Have you done any prototyping of physical controls?

@@JesseSchoch You can see the black box with the red toggle switch on top in this video. It's a MOM-OFF-MOM switch to select the left/right feed direction. Has to be momentary so the direction can be controlled from either the touch screen or the switch.

You said "I found a project called Didge which has a very very nice write up on how it does accurate step generation based on a modified Bresenham algorithm. It runs on a stm32 blue pill board" - I've looked everywhere for this - can you remember where you found it?

If you roll the chuck back and forth the lead screw and display arrows should change automatically. If they don't then one spindle encoder phase is not working and stuck either high or low.

Thank you for the kind words, Gene. I bought an STM32 board just in case, but the Mega has worked out well so far. As for going further, my goal was basically a programmable gearbox and I'm sticking to that. I added programmable stops because they were useful and easy, but otherwise the lathe works exactly like it would with a gearbox.

@@jonrbryan I understand. It does the job beautifully and is quite minimalist. Maybe the manual jogging (if you use it often), for which you implemented acceleration, could benefit from a joystick like an RC controller has (i.e. the further you push it away from center, the faster it goes). How easy/difficult would it be to adapt the code to work with another lathe? Besides the leadscrew pitch, is there anything else that would change? I think I might just go minimalist with my Hobbymat, and try doing it with your design. And with regards to the parts you used, do you think some changes would be needed? Hobbymat is a very little lathe :) By the way, did you do the improvements which you had in mind / mentioned in this video?

@@GenePavlovsky I did add a ON-OFF-ON momentary toggle switch to change direction, after I did this video, that attaches to the ways with magnet bars. A stick would be interesting to implement. I'm satisfied with the acceleration algorithm that I'm using for now, but it might be something to play with in the future. I have a "minion" in England who has adapted it to a Myford lathe. He had to use a two-step reduction (4 timing pulleys) to drive the lead screw and fit the available space, but it seems to work well for him once he scaled for the odd ratio that it came out to. Are you also in the UK? Another person in Australia has done a more involved adaptation for a very specific application. He's even changed to a smaller Nextion display. The size of the lathe has to be considered when selecting the stepper motor to get an appropriate amount of torque. I started with a Nema 23 motor and a 4:1 ratio, but changed to a Nema 24 motor and 8:1 to get more torque. On something like a Hobbymat you could downsize everything. I haven't watched the video in quite a while, so I forget what I talked about. I have changed the STOP screen quite a bit, and I have the feed synchronized when changing directions. I added British Association threads plus a couple of other oddball ones. The SETUP screen now displays only the belt positions/speeds recommended for the currently-selected feed rate. It's all available on github.com/ccomito1223/ELS-Jon if you want to give it a try. You can also join the Logan Lathe Users Group on groups.io where I hang out and help people with their implementations.

@@jonrbryan Thanks for the detailed reply. I've joined the Logan group as you suggested. What is meant by "minion" in this case? It's not the first time I encounter it's usage which I don't understand. The lead screw pitch also has an effect on the torque, right? E.g. on my lathe it's quite fine pitch, 1.0 mm. Given a same motor and belt reduction, the torque will be 3 times higher than on an 8 TPI leadscrew, at least that's my thinking :) If too much reduction is used, the max speed is going to get limited, right? On my lathe the leadscrew nut is solid, so I think I'm going to be using the powered traverse more often than you. By the way, if you move the carriage using the handwheel (half nuts engaged), does the ELS keep track of the position changes? How does that work? Another feature which could be nice to have with your setup, and this won't require any hardware changes, is spindle division. With an 800 PPR encoder, using both rising and falling edges of the pulses, you could have a resolution of 360 / (2*800) = 0.225 degrees. By the way, the Russian project I referred to earlier, is using a 1800 PPR encoder on the spindle, and counts both rising and falling edges for an effective resolution of 3600 steps per revolution. The spindle dividing feature has a resolution of 0.1 degrees. And it seems that Arduino Mega is still able to pull this off, although I don't know if it uses any tricks at high RPMs.

@@GenePavlovsky All evil geniuses have followers, Gene ;). A 1mm pitch lead screw would have a 3.175 mechanical advantage over an 8tpi leadscrew of the same pitch diameter, but friction has a large influence. I have a 1000-pound dial force indicator that I picked up on eBay to compare different motors and ratios, and torque/force isn't particularly linear. The only way to know is to try it. The ELS only knows how much it has turned the lead screw, so the half nuts have to remain engaged for the programmable stops to work. I haven't connected the driver alarm output back to the Arduino, either, so it doesn't "know" if you have stalled the motor. Since that is an unrecoverable condition requiring the power to be cycled I figured it would be pretty obvious. Several others have brought up spindle division, but I haven't felt strongly motivated to implement it. I would be more inclined to get a dividing head for my mill to do that sort of work, or use the bolt circle calculator on the mill DRO, which would provide much greater accuracy than the 800ppr encoder that I'm using on the ELS. At one point I had full quadrature decoding implemented, but it's cumbersome in software. To save clock cycles I switched to simply interrupting on the falling edge of one phase and looking at the high/low state of the other phase to determine direction, and it has been very stable.

Thanks for the kind words. I was using it yesterday to cut an internal 3/4-8tpi Acme thread. I suppose I should be making a video, eh?

@@jonrbryan yes you sude

That's one advantage gears have - they don't forget. I used incremental encoders. I've toyed with the idea of battery backup and/or saving the state to EEPROM, but decided to take my chances for now. There is a zeroing function.

@@jonrbryan Hi Jon, Thank you for replying. Would it be feasible to synchronise the chuck and lead screw before any screw cutting operations at all times . A possible way is to fit absolute encoders to the chuck and lead screw. When cutting a screw thread there will be a ratio between the two - say for argument for a particular thread the ratio is 6:1. As the chuck cannot be accurately positioned just read its position from its single turn absolute encoder -- say it is sitting at 37 degrees. Execute a synchronise command and the lead screw will rotate independently until it has reached the position of 37 x 6 = 222 degrees. When more than 360 just keep track of the turns or use a multi turn absolute encoder. The lead screw should then be synchronised with the chuck. As long as the work piece is not removed from the chuck. Would this work?

@@johnspathonis1078 Well, if I lost power while cutting a thread, I would probably just adjust the compound to reestablish the alignment and go from there.