# Tube-O-Matic

Discussion in 'Workshop Tips and Secrets / Tools' started by FritzW, Oct 10, 2019.

### Help Support HomeBuiltAirplanes Forum by donating:

1. Oct 12, 2019

### Hephaestus

#### Well-Known Member

Joined:
Jun 25, 2014
Messages:
1,171
Likes Received:
253
Location:
YMM
I get that, and think 3dimensionally - the Chuck could be turned around and only the passthrough be above the bed, have the Chuck extend below off the bed.

You could tweak that printed Chuck to have a very large passthrough, incorporate the bearing, mount, drive gear into a nice compact package.

2. Oct 12, 2019

### proppastie

#### Well-Known Member

Joined:
Feb 19, 2012
Messages:
3,779
Likes Received:
1,094
Location:
NJ
I have to wonder if Fritz ever sleeps......pretty amazing stuff

Pops and akwrencher like this.
3. Oct 12, 2019

### pictsidhe

#### Well-Known Member

Joined:
Jul 15, 2014
Messages:
6,835
Likes Received:
1,832
Location:
North Carolina
I'd go about this differently. Partly because I only have a baby CNC...

Spindle is fixed to the bed using it's standard gantry clamp. The tube stepper is attached to the gantry. Sliding steady rests support the tube as required. If you have a 10' tube, rests can be elsewhere in the workshop... For indexing tubes. A printed clamp is attached to the tube somewhere it won't be in the way of anything, a level is then used on the clamp to index the tube prior to each end being cut. Length, a tape measure? Stops could be set up for multiple identical tubes. This scheme isn't great for production, but will work with very long tubes on even small tabletop CNCs. As Scott suggests, I may well use off the shelf hardware.

4. Oct 12, 2019

### Hot Wings

#### Well-Known MemberHBA Supporter

Joined:
Nov 14, 2009
Messages:
6,352
Likes Received:
2,282
Location:
Rocky Mountains
If you have to resort to this forget CNC and just get a drill press and a tube notcher. The idea of CNC is to poke a "run" button and then go do something else productive.

5. Oct 12, 2019

### Hephaestus

#### Well-Known Member

Joined:
Jun 25, 2014
Messages:
1,171
Likes Received:
253
Location:
YMM
Yes and no...

If you have the Y axis length, you could do full complete cuts. But just the end is fairly helpful - using a single tool you'll use multiple times on numerous processes - vs a 250$dedicated tool (yes there are cheaper, but the better accurate ones are 250$+)

6. Oct 12, 2019

### FritzW

#### Well-Known Member

Joined:
Jan 31, 2011
Messages:
3,624
Likes Received:
3,251
Location:
Las Cruces, NM
Very few tubes are going to be long (8 foot'ish), most are going to be two or three feet long. The way I drew it post #1 was how it would be setup to cut 3 or 4 shorter tubes out of an 8 footer.

The way I see it happening for a typical homebuilder:

Slide the steady rest rings and the collet on to the tube and set the tube in the mounts. The first steady rest ring gets a hose clamp to stop the X motion. Line up the cog belt pulley on the collet and tighten the hand wheel.

The first tool (drill) drills all the holes, the second tool (1/8" end mill) cuts the tube end profiles starting at the far end of the machine. There's no need for a tube feeder.

It would only be a 5 or 10 minute operation to cut 3 or 4 rivet ready tubes. If you don't over think it, you could -literally- be making parts faster than you could rivet them together.

If the tube is longer than the X travel of your machine just draw an index line on the ends of the tubes. Mark a zero index on the collet. Cut one end, flip the tube around and cut the other end.

Post #48 on this thread shows a paper template version of what I'm talking about, just replace the template with a CNC machine.

7. Oct 21, 2019

### PiperCruisin

#### Well-Known Member

Joined:
Jan 18, 2017
Messages:
118
Likes Received:
39
Location:
Idaho
8. Oct 21, 2019

### GeeZee

#### Active MemberHBA Supporter

Joined:
Feb 25, 2019
Messages:
42
Likes Received:
15
Location:
Indianapolis, IN
While checking out that 3 jaw chuck on Thingverse I saw this:
Rotary 4th axis for CNC (MPCNC) https://www.thingiverse.com/thing:1877336
There are a couple of videos that are pretty cool. The design would need to be modified a bit to allow the tube to pass through the chuck.

I don’t yet understand all I need about profiling the tubes but does this comment mean he’s got S/W to control the 4th axis (in the way that we need to profile a tube?). “I've modified Scorchwork's awesome G-code ripper software to wrap rotary toolpaths on Marlin firmware based CNCs. I'll be posting the software on the Vicious 1 website soon.”

9. Oct 21, 2019

### 12notes

#### Well-Known MemberLifetime Supporter

Joined:
Aug 27, 2014
Messages:
836
Likes Received:
511
Location:
Louisville, KY
I'm not sure how stout the bearing blocks are, so I would think mounting the pulley side of the belt drive in a bearing block would be better, or at least a bearing block on either side of the pulley. If you've got enough tension to move and hold a tube during cutting, it might bend a thin tube. Perhaps you've already calculated this and it doesn't matter, just doing some backseat TLAR engineering.

10. Oct 21, 2019

### FritzW

#### Well-Known Member

Joined:
Jan 31, 2011
Messages:
3,624
Likes Received:
3,251
Location:
Las Cruces, NM
I'm playing with the wrapper plugin for CamBam to wrap the profile. It's completely not necessary but makes the hardware a one time setup instead of having to change it every time you want to cut a tube.

This thing doesn't have to be industrial grade. For a home hobbyist that wants to cut out one airplane I think the setup will be fine. You're cutting thin wall 6061 or 4130 tube with a very shallow depth of cut and moving the tool slowly (not healthy for your cutting tool, you'd plan to go through \$20 worth of ebay end mills to cut out a kit, ...no biggy)

I've cut and ground more than my share of tubes. It's not rocket science but it does get painfully tedious if you don't enjoy that sort of thing.

The whole point of this thing is: it's a 3D printed 4th axis that goes on a mostly 3D printed CNC machine so you can cut out a simple T&G airplane so you can go have some fun. It's a tool for the tube version of the "airplane on a thumb drive" idea.

GeeZee likes this.
11. Oct 23, 2019

### FritzW

#### Well-Known Member

Joined:
Jan 31, 2011
Messages:
3,624
Likes Received:
3,251
Location:
Las Cruces, NM
I started printing the 4th axis GeeZee found on thingiverse. It looks pretty good but it's not a pass-through and there are a few things I'd like to change.

Based on that one, I came up with a much cleaner version of mine. ...the first project will be SuperVeep Volksplane rudder tubes

GeeZee and Hephaestus like this.
12. Oct 23, 2019

### Hephaestus

#### Well-Known Member

Joined:
Jun 25, 2014
Messages:
1,171
Likes Received:
253
Location:
YMM
How well will the collet setup hold up do you think? Layer lines would bind them up as you tighten...

13. Oct 23, 2019

### FritzW

#### Well-Known Member

Joined:
Jan 31, 2011
Messages:
3,624
Likes Received:
3,251
Location:
Las Cruces, NM
I don't know but I suspect it'll do what it's designed to do. It's worth a few hours of printing to find out. If one of the parts isn't up to the task I'll just tweak it a little and print another one. It only has to cut a handful of tubes (maybe two or three handfuls for a more complicated airplane), it's not a tool to pass down to your grand kids.

The height of the bumps in the layer lines are less than a thousandths of an inch and they would have to line up to within an arc second or two to mesh. And there would have to be less flex in the system (flex in the collet and deflection in the plastic) than the height of the bumps. ...I'm not too worried about it.

Typical layer lines on medium quality print. Even if they did bind, just give it a quick wipe with a little sandpaper.

litespeed likes this.
14. Oct 23, 2019

### addicted2climbing

#### Well-Known Member

Joined:
Jan 27, 2012
Messages:
957
Likes Received:
270
Location:
Glendale, CA
Looks great Fritz. My only concern is on the collet that the infil may do odd things to a flexure.

15. Oct 23, 2019

### ScaleBirdsScott

#### Well-Known Member

Joined:
Feb 10, 2015
Messages:
1,033
Likes Received:
656
Location:
Uncasville, CT
If I was printing the collets it would be 100% infill.

Also yeah you'd probably need one of those 1/2" diameter sanding drums for a drill and run that a few times to smooth out the interior. Or make some kind of way to mount some sandpaper to a rod and run it through a few times.

I've honed out 3D printed tubes to be viable short paintball barrels, so smooth enough to grab a piece of tube should not be tough.

GeeZee likes this.
16. Oct 23, 2019

### FritzW

#### Well-Known Member

Joined:
Jan 31, 2011
Messages:
3,624
Likes Received:
3,251
Location:
Las Cruces, NM

You can't think of this thing like you do the collet on a 3 ton Bridgeport mill. Think of it more like a Dremel tool. Like I said in post #5, it might take 3 or 4 passes to get through .049 wall tube. It's not for turning crankshafts for aircraft carriers.

The tool collet in the spindle is muuuch more efficient at grabbing the tool (polished collet, polished collet seat, polished tool shank) BUT the tube collet has muuuch more contact area. Even if the contact interface isn't very efficient, there's a whole bunch of it.

If you wanted the collet to slip on the tube, about the worst combination you could come up with would be relatively soft plastic being squeezed under relatively high pressure on a relatively soft aluminum tube.

I'm printing a worst case example: low resolution PLA print, thin walls and light infill. (a quick print to make sure everything fits together and the threads mesh). After putting a little hand pressure on the collet (waiting for the other parts to print) I have no doubt the "collet grab" is going to be more than adequate, ...but time will tell.
I'm not going to fix it unless it doesn't work

17. Oct 23, 2019

### addicted2climbing

#### Well-Known Member

Joined:
Jan 27, 2012
Messages:
957
Likes Received:
270
Location:
Glendale, CA
Collet looks awesome... So if your going down this route, have you sorted out the software to take a Solidworks file and figure out the path? Also since your using this on a large table, can you have it cut both sides so they are properly clocked. For example it cuts the side near the cut end, then moves say 20" up the tube and does a plunge cut and cuts 110 degrees of the profile lifts up and starts a new plunge cut of the profile. Ie leaving 3 small tabs to still translate the rotation that can be trimmed later. Is this the direction your going with this?

Heck doing it this way you could cut multiple tube profiles at once and separate them from the main tube after the cutting is done.

Marc

GeeZee likes this.
18. Oct 24, 2019

### FritzW

#### Well-Known Member

Joined:
Jan 31, 2011
Messages:
3,624
Likes Received:
3,251
Location:
Las Cruces, NM
Posts #19 and #26 explain (in my long winded way) how I see it working. The idea is you can cut as many finished tubes as you can fit on a blank tube that's as long as your longest axis.

CAM software lets you put hold down tabs anywhere you want but *you wouldn't need any (like you describe), you'd just need to put a **steady rest on each side of the cut(s). The first pass would drill all the holes, the second pass would start cutting profiles at the far end of the tube and work it's way back to the front.

*When It finishes cutting a profile the rotation will stop and the tool will retract. It might leave a little booger on the end of the free tube. I'd set the DOC to about 1/3 the wall thickness (a deep engraving pass) so the little booger would just snap off anyway.

To make life easy (and repeatable) I plan to screw a 1' x 8' piece of 3/4" MDF to the table, surface it and drill it like a cribbage board. That way I can move the steady rests wherever I need them without have to realign anything. The CNC machine drills the cribbage board holes so I know it's aligned with the machine and the steady rests are going to be as accurate as printer (which is way more accurate than it needs to be).

EDIT: since I found a profile wrapper plugin for my CAM software I won't be handling the gcode like I suggested in post #19.

19. Oct 25, 2019

### FritzW

#### Well-Known Member

Joined:
Jan 31, 2011
Messages:
3,624
Likes Received:
3,251
Location:
Las Cruces, NM
I finally got all three parts printed and gave them a test fit. So far it's working like a champ. Even just *hand tight I don't think there's a chance that the tube (at least a 2" tube) will slip in the collet.

*I could only go hand tight tonight, I sized the collet holder grip "lugs" to fit my water softener filter wrench but the collet nut is too big. I'll have to come up some sort of wrench'ish thing for that end.

Now I need to come up with a 3D printed stepper mount and an H series belt and pulley (I might try a 3D printed drive pulley for the motor, if one end is plastic the other end might as well be also. ...I'm still trying to keep this as simple and idiot proof as possible.

Last edited: Oct 25, 2019
Pops, petroelb, Aerowerx and 3 others like this.
20. Oct 25, 2019

### ScaleBirdsScott

#### Well-Known Member

Joined:
Feb 10, 2015
Messages:
1,033
Likes Received:
656
Location:
Uncasville, CT
I am curious to make one of these at some point. At a minimum being able to drill on the 4th axis is very cool.