• Spring 2024 meetup in Calgary - tentative date Saturday, April 20/2024. Other regions are also discussing meet ups. If you want one in your area get going on organizing it! discussion
  • We are having email/registration problems again. Diagnosis is underway. New users sorry if you are having trouble getting registered. We are exploring different options to get registered. Contact the forum via another member or on facebook if you're stuck. Update -> we think it is fixed. Let us know if not.
  • Spring meet up in Ontario, April 6/2024. NEW LOCATION See Post #31 Discussion NEW LOCATION

G3616 Conversion.

jcdammeyer

John
Premium Member
Nope. Unless they started opening on weekends again :-(
Not really a big deal. Unlike a servo that I can switch over to constant torque mode I have to be able to detect that the stepper motor has stalled with a specific power setting. If the stepper driver has resistors for setting current then it's pretty easy to use a relay to switch them based on the motor direction. 18 ft-lb for tighten. 20 ft-lb for loosen.

But most of the drivers I have use switches to set current. There is one where I think I can put a relay across the dip switch. But I also need to interface to a quadrature encoder so I can tell when it's stopped turning. Quadrature so if it's vibrating around the sensor edge that we also detect that.

I could use 1 PPR on the motor, which if stepping at say 10kHz with 2000 steps/rev I should see a pulse every 0.2 seconds. So if no pulse occurs after that time I will know it's stalled. Since it stalled at a higher speed and torque falls off on a stepper for higher speeds it's likely it will just buzz once stalled at that lower torque value. I could then step very slowly for another 1/4 turn (after reduction drive) to firmly tighten it.

To loosen, set higher current into motor, turn slowly for two turns. Then speed up if trying to remove R8.

This would be so much easier with a servo motor compared to a stepper.
 

jcdammeyer

John
Premium Member
Here are the two parts enlarged for the F03 specification size.
Planetary Gear output shaft size is 14mm with 5mm key. Screws are 5mm. Drive is 3/8" with a retainer pin hole that is 3mm diameter.
 

Attachments

  • GearHub.pdf
    131.8 KB · Views: 2
  • HubSocketDriver.pdf
    134.2 KB · Views: 3

jcdammeyer

John
Premium Member
no recess to use socket-head cap screws?
No. I don't have 5mm metric in stock. I have lots of Philips Pan head and the bolt diameter is large enough to clear the socket. Plus I might still mount a flat slotted encoder disk there instead of mounting an encoder on the back of the stepper motor.

I have a couple of US Digital encoders I pulled off my DC brushed servos because they were very unreliable. Switched to CUI for those motors. But likely for what I want on the draw bar it would be adequate. But making up an aluminum disk with slots using the slitting saw and indexing on the rotary table is pretty easy.
Or even just a disk with a ring of small holes which I can make LinuxCNC do without a rotary table.

Hence no socket head screw capability.
 

Attachments

  • EncoderDisk.jpg
    EncoderDisk.jpg
    53.2 KB · Views: 1

jcdammeyer

John
Premium Member
And just for general interest. The 3D parametric CAD drawing took less than 10 minutes. The CAM using AlibreCAM took about 25 minutes which includes generating and zipping the G-Code .ngc files.
I'm not suggesting anyone use the files but if you are interested in what I did.
How I would use them would be to:
1. 0.042" aluminum sheet a bit larger than the diameter placed onto scrap plywood or MDF and held in the corners. Ensure that the center point sits over a T-Slot.
2. Run the G-Code that cuts out the middle hole.
3. Hand drill through the plywood and with a large fender washer and a bolt anchor down the middle.
4. Next run the G-Code that does the center drilling for the mount holes and encoder holes.
5. Run the G-Code that does the 4 mounting holes.
6. Screw self tapping screws into the wood through the 5mm mount holes.
7. Run the G-Code to drill the encoder holes.
8. Finally run the G-Code to cut out the outer perimeter.
9. Remove the outer clamps, screws and bolt.
10. use a deburring tool to clean the perimeter if the edge is ragged. Same with the 14mm center hole.
11. Depending on the state of the encoder holes deburr from the rear with a hand tool.
12. Mount to adapter.

There. A 12 step program...
 

Attachments

  • EncoderHolesG-Code.jpg
    EncoderHolesG-Code.jpg
    329.4 KB · Views: 4
  • EncoderG-Code.zip
    4.1 KB · Views: 2

jcdammeyer

John
Premium Member
I found in the drawings/XAxis folder all sorts of drawings and photos I had done a few years ago. Working with ordering parts from AliExpress for the Robot Arm I realized I should get after getting ball screws. The thing that put that project #42 on hold was not knowing whether to use a 16mm, 20mm or replacement 2505 sized ball screw to replace the 1" ACME 0.2" pitch screw.

I had a photo of the screw nut and where it went when I had the table off a few years ago.
BronzeNut-1.jpg

X_Axis_Ways-2.jpg

But I really couldn't tell from those pictures whether the much larger ball nut and holder would fit. So first step is remove the RH side plate, wind the lead screw all the way over so the edge of the table lined up with the frame section in the photo above. Dangled a square over to measure to the screw center line.

EndViewRHS.jpg

Well lots of measurements. Sketches, then CAD drawings and finally taking the original ACME screw and nut assembly drawing I did in 2017 and adding it to the two parts I drew today.

Here's a perspective view.
XAxisRendering.jpg

So similar end view with the ACME screw centered on the X axis.
1676174657816.png

Looks like there will be plenty of room for a larger ball nut holder. Why not just go with a 16mm screw? Well the pulleys and bearings and spacers etc are all sized for the current 1" ACME screw. The cost of a 25mm over a16mm isn't that high and the rounded ends then match all my bearings and the pulley.

All I need to do is cast new end plates with the hub in line with the new ball screw center line. Easy to do since it was a two part pattern. Just move the hole up and use some wax to plug the old one.
EndPlatePattern.jpg

All for now. Time to download the 1" double ball nut step file and model it with the parts done here.
 

jcdammeyer

John
Premium Member
Took the model of the 25mm ball screw I did a couple of years ago. Now that I have the table assembly modeled I created a ball nut holder and with some bevels on the holder.

1676323148538.png
I will have to cast new end plates and come up with a better method of locating the holes. Unfortunately the Chinese builders didn't put them in symmetrical locations. On the RHS the set is even turned CW a tad so the top of the holder isn't level.

I suspect they had some sort of jig that held the ends of the ACME screws so it was parallel in the two dimensions to the table and then they just pressed the end plates and used a centering punch to set where the holes go. That a plate pivoted slightly wasn't important...

1676323554402.png
 
I found in the drawings/XAxis folder all sorts of drawings and photos I had done a few years ago. Working with ordering parts from AliExpress for the Robot Arm I realized I should get after getting ball screws. The thing that put that project #42 on hold was not knowing whether to use a 16mm, 20mm or replacement 2505 sized ball screw to replace the 1" ACME 0.2" pitch screw.

I had a photo of the screw nut and where it went when I had the table off a few years ago.
View attachment 30602

View attachment 30603

But I really couldn't tell from those pictures whether the much larger ball nut and holder would fit. So first step is remove the RH side plate, wind the lead screw all the way over so the edge of the table lined up with the frame section in the photo above. Dangled a square over to measure to the screw center line.

View attachment 30604

Well lots of measurements. Sketches, then CAD drawings and finally taking the original ACME screw and nut assembly drawing I did in 2017 and adding it to the two parts I drew today.

Here's a perspective view.
View attachment 30607

So similar end view with the ACME screw centered on the X axis.
View attachment 30605

Looks like there will be plenty of room for a larger ball nut holder. Why not just go with a 16mm screw? Well the pulleys and bearings and spacers etc are all sized for the current 1" ACME screw. The cost of a 25mm over a16mm isn't that high and the rounded ends then match all my bearings and the pulley.

All I need to do is cast new end plates with the hub in line with the new ball screw center line. Easy to do since it was a two part pattern. Just move the hole up and use some wax to plug the old one.
View attachment 30608

All for now. Time to download the 1" double ball nut step file and model it with the parts done here.
Going down this route soon.
 

jcdammeyer

John
Premium Member
Going down this route soon.
I'm finding it invaluable this time to do all the CAD up front. To make sure things fit. I know on the island if the mill is disabled there are a few people close by who can make some chips if I make a mistake. But I'd rather not have the mill that disabled.
 
I'm finding it invaluable this time to do all the CAD up front. To make sure things fit. I know on the island if the mill is disabled there are a few people close by who can make some chips if I make a mistake. But I'd rather not have the mill that disabled.
For me the plan is some maintenance to see what wear hard CNC is causing and plan mitigation. Measure up and design mounts for the ballscrews. Assemble and continue work will setting up for next retrofit (without glitches one hopes).
 

jcdammeyer

John
Premium Member
And just for general interest. The 3D parametric CAD drawing took less than 10 minutes. The CAM using AlibreCAM took about 25 minutes which includes generating and zipping the G-Code .ngc files.
I'm not suggesting anyone use the files but if you are interested in what I did.
How I would use them would be to:
1. 0.042" aluminum sheet a bit larger than the diameter placed onto scrap plywood or MDF and held in the corners. Ensure that the center point sits over a T-Slot.
2. Run the G-Code that cuts out the middle hole.
3. Hand drill through the plywood and with a large fender washer and a bolt anchor down the middle.
4. Next run the G-Code that does the center drilling for the mount holes and encoder holes.
5. Run the G-Code that does the 4 mounting holes.
6. Screw self tapping screws into the wood through the 5mm mount holes.
7. Run the G-Code to drill the encoder holes.
8. Finally run the G-Code to cut out the outer perimeter.
9. Remove the outer clamps, screws and bolt.
10. use a deburring tool to clean the perimeter if the edge is ragged. Same with the 14mm center hole.
11. Depending on the state of the encoder holes deburr from the rear with a hand tool.
12. Mount to adapter.

There. A 12 step program...
Spent some time thinking about how to mount the CUI or US Digital Encoder on the back of the stepper motor. Turns out the issue is mounting holes as there aren't any. Not like the SteppersOnline motor that I wish I had ordered.

Not sure I want to take the back off the motor in order to drill and tap them. The shaft isn't long enough for a mounting plate that is thick enough to be threaded.

On the LinuxCNC forum counter sunk screws epoxied in was one suggestion. Or 3D print an expansion plate that matches up to the existing encoder mounting plate and then use JB-Weld to attach. Then mount to the outer mounting holes.

Or just remove the back plate and leave the armature in place to avoid screwing the magnetism.

Anyway, hopefully today I will cast the new mounts for the power draw bar and then tracking revolutions will become important. The option to put in my own disk is still there but first I think I want the works all mounted on the new castings before I decide if I want to go that way.
 

Attachments

  • Encoders.jpg
    Encoders.jpg
    146.8 KB · Views: 6
  • StepperOnlineMotor.jpg
    StepperOnlineMotor.jpg
    31.1 KB · Views: 5
  • SocketDrive-2.jpg
    SocketDrive-2.jpg
    286.5 KB · Views: 5

jcdammeyer

John
Premium Member
I’m not sure how it all goes together but couldn’t you just make a new mounting plate for this?
View attachment 35924
That was the original idea. But a plate that can take either threads or countersunk for a screw or threaded insert raised the plate enough that the encoder disk locking screw is just on the edge of the shaft. Not the longest shaft. Same issue with the CUI holder. If I'd restarted this project 3 days earlier I could have added a StepperOnline motor with all the AR3 Robot Arm motors and reduction drives etc. But a single motor at $33 with the holes already there along with $32 shipping all US $ turns into about $100 Cdn.

I think I'll put a hold on this part until I cast the 3 draw bar plates and get them machined. It may well be that I can even run the stepper and planetary drive open loop with forwards for tightening at lower current than reverse for loosening. It will all depend on whether I can loosen what I've tightened. The desire for quadrature is so I can tell if it's stopped and not just vibrating against the stop.
 

jcdammeyer

John
Premium Member
I thought I'd posted a photo of the new draw bar assembly but I don't see it in the older messages. The top ring keeps the vertical posts aligned.
Now it's the middle one that moves up and down with the air cylinder body.

The 25:1 planetary drive takes the 0.8 8 ft-lb static torque and makes about 22 ft-lbs. I figure I'll tighten at about 18 ft-lbs and loosen at full current.

What would be really cool is to have something sense the flat edge of a the draw bar nut and then turn the stepper motor until the socket lines up so it fits over perfectly. No collisions trying to go down with a locked draw bar in the spindle and and the socket on a locked stepper motor system.
PTB_Assembly1.jpg
 

jcdammeyer

John
Premium Member
I have a problem and I'm not sure how to solve it at the moment. The knee shaft has the usual square end. The pulley is not keyed but has the two set screws on the shoulder.

1689890422662.png

I can't just run the set screws down to press on the flats because if there is some sort of over torque event the screw part that sticks out bends. I've been using some brass pins pushed down by the set screws but ultimately it's the same issue. Some sort of sudden load and the pulley turns a bit, the brass pins are bent a tad and my Z axis has changed position slightly.

1689890539838.png

If I make a 25mm bushing with a 14.14mm square hole and then slice it into key like pieces I could then jam them down with the two set screws as long as the screws went into a dimple in the insert.

1689891914969.png

Alternatively put a flange on the insert with the square hole. As long as the set screws enter a dimple in the bushing it won't twist. I suppose it would be better if the bushing were keyed to the pulley.

1689893531868.png

Suggestions?
 

DavidR8

Scrap maker
Administrator
Moderator
Premium Member
I have a problem and I'm not sure how to solve it at the moment. The knee shaft has the usual square end. The pulley is not keyed but has the two set screws on the shoulder.

View attachment 36692

I can't just run the set screws down to press on the flats because if there is some sort of over torque event the screw part that sticks out bends. I've been using some brass pins pushed down by the set screws but ultimately it's the same issue. Some sort of sudden load and the pulley turns a bit, the brass pins are bent a tad and my Z axis has changed position slightly.

View attachment 36693

If I make a 25mm bushing with a 14.14mm square hole and then slice it into key like pieces I could then jam them down with the two set screws as long as the screws went into a dimple in the insert.

View attachment 36694

Alternatively put a flange on the insert with the square hole. As long as the set screws enter a dimple in the bushing it won't twist. I suppose it would be better if the bushing were keyed to the pulley.

View attachment 36695

Suggestions?
Does it have to have the square hole?
Seems to me that a round bushing with bore that clears the corner-to-corner dimension would support the majority of the grub screw stickout.
What about increasing the grub screws diameter and adding two more on the two other flats?
 

jcdammeyer

John
Premium Member
Does it have to have the square hole?
Seems to me that a round bushing with bore that clears the corner-to-corner dimension would support the majority of the grub screw stickout.
What about increasing the grub screws diameter and adding two more on the two other fl

I think a square hole is a requirement. Although the entire pulley is pretty thick I could likely still broach a keyway in line with one of the set screws. Then make the square hole insert have a key way. If I can make the square be a press fit so it doesn't slop...
 
Top