CamBam
News: *** FORUM BACK *** please see 'Latest News' for updates
 
*
Welcome, Guest. Please login or register.
Did you miss your activation email?
May 22, 2018, 01:25:32 am


Login with username, password and session length


Pages: 1 2 [3] 4
  Print  
Author Topic: Automatic tool changer for the lathe  (Read 5765 times)
lloydsp
CNC Jedi
*****
Offline Offline

Posts: 7508



View Profile
« Reply #30 on: November 12, 2017, 18:54:32 pm »

Eddy,
If wires could be properly routed to the head and flexed as required, then there would be no need for batteries!  Cheesy

Lloyd
Logged

"Pyro for Fun and Profit for More Than Fifty Years"
EddyCurrent
CNC Jedi
*****
Offline Offline

Posts: 3430



View Profile
« Reply #31 on: November 12, 2017, 18:57:27 pm »

Good point  Grin  Grin  Grin

But wait a minute, what if the electronics required a totally ripple free DC supply  Cheesy
« Last Edit: November 12, 2017, 18:59:23 pm by EddyCurrent » Logged
Bob La Londe
CNC Jedi
*****
Offline Offline

Posts: 3230


^ 8.5 pounds on my own hand poured bait.


View Profile WWW
« Reply #32 on: November 12, 2017, 20:45:09 pm »

Bob,
You do know that redtube is a porn site?

Nope, didn't know that.  Just know Google (who owns YouTube) keeps trying to force me to buy an enhanced YouTube experience.  Must have gotten the name wrong. 

Logged

Getting started on CNC?  In or passing through my area?
If I have the time I'll be glad to show you a little in my shop. 

Some Stuff I Make with CamBam
http://www.CNCMOLDS.com
lloydsp
CNC Jedi
*****
Offline Offline

Posts: 7508



View Profile
« Reply #33 on: November 12, 2017, 21:46:27 pm »

"....what if the electronics required a totally ripple free DC supply?"
--------------
'Tain't no such thing, Eddy, MOST ESPECIALLY from batteries, which possess a specific series resistance, and ALWAYS cause 'ripple' (load variations) with varying loads.

A properly filtered power supply with a large post-filtering capacitor posessing a low series resistance will beat a battery every time!

Lloyd
Logged

"Pyro for Fun and Profit for More Than Fifty Years"
dave benson
CNC Jedi
*****
Offline Offline

Posts: 887


View Profile
« Reply #34 on: November 13, 2017, 05:03:59 am »

Hi Eddy

Yes it's been a interesting  project  'and' on the whole I'm satisfied with the result, But far from content.

When I started out, the goal was to make a cordless\Bluetooth tool changer that was Plug'n Play
for example if you had a job in the lathe which requires changing to  the manual tool post (perhaps you need to do some deep boring)  then you would unbolt the turret install the manual tool post
and continue on without changing tool scripts or fiddling with mach3 at all.

With the batteries, I was figuring on 500 tool changes at the outset, and so to get 864 just means that
I've got plenty of power left to do something else.

What I propose to do is make another two (based on the lessons learned from this build) and flog them off on Ebay and if  They sell I will use the proceeds to build a Beagle Bone Powered version
with a video camera (maybe a Xbox connect) to run Opencv and some python Ai to add →

1. Home switch functionality
2. True Tool identity
3. Tool Change complete confirmation
4. Tool damage and misalignment
5. Collision Detection\electronic fencing 

I've already drawn up the turret to standard Steel\Fasteners dimensions.
I have some cold rolled 12mm and 16 mm plate.
I've ordered  two more housings which have even heavier bearings but are shorter in the snout.

The new ones will have Bi directional control meaning that the turret will select the shortest path for the next tool change.
As part of doing this requires using two pawl's and a (RC servo or a solenoid ) .
The by product of this is that the turret will be rigidly locked in position for both directions.

I did look at 4 axis positioning capability and with the addition of a disk brake, it would be possible
But not very practical, what you would really need is a high torque compact motor and drivers ect.
that fit in the same space, as the present kit, I did look at one compact atmega2560 board and a slew of micro style stepper drivers, and so miniaturisation is possible for the controller but not the stepper drivers as they don't provide the mode I'm using.

One  thing I'll have a look at is to use one of those “One battery for many tools” and charger from the local big box stores, as the battery has visual indication of charge  as well as low voltage cut-out
in this way I could off board the battery and electronics and free up space inside the turret. And as you can see in the cad file image, there's not much room and that's without the LM298 driver.

The tricky bit would be to machine a fitting to suit the battery.
But I  will look into this as the benefits would clearly outweigh time spent making the CB file in the first place.

Dave








 
 


* steel1.png (1144.5 KB, 1000x750 - viewed 70 times.)

* freecad turret model Capture.PNG (41.53 KB, 479x387 - viewed 75 times.)
Logged
Bob La Londe
CNC Jedi
*****
Offline Offline

Posts: 3230


^ 8.5 pounds on my own hand poured bait.


View Profile WWW
« Reply #35 on: November 13, 2017, 15:35:19 pm »

Maybe this is over simplification, but why not setup so you can just add a tool post when you need to do that?  A lot of guys run a t-slot table on their cross slide for gang tooling. 
Logged

Getting started on CNC?  In or passing through my area?
If I have the time I'll be glad to show you a little in my shop. 

Some Stuff I Make with CamBam
http://www.CNCMOLDS.com
dave benson
CNC Jedi
*****
Offline Offline

Posts: 887


View Profile
« Reply #36 on: November 13, 2017, 23:15:46 pm »

Hi Bob

Here is a shot of the top-slide from the rear, when I was repairing the the top-slide during the lathe build, I added
some extra spigot location holes for the manual tool post, although in practicality I probably wont use it.(no space)
What I have thought about is a rear gang tooling block or a rear mounted parting off tool.

I do like the idea of a rear parting tool for small lathes like this, and I think I will make one.

If you were setting up the lathe for production work, for similar parts, you could fine tune the tool arrangements in the turret and the gang tooling ect. But for me the parts would be one offs so one day I might be doing long thin turning and the next chucking work. 




* space maybe for rear tool post .png (1248.43 KB, 1000x750 - viewed 44 times.)
Logged
dave benson
CNC Jedi
*****
Offline Offline

Posts: 887


View Profile
« Reply #37 on: December 04, 2017, 12:56:53 pm »

Another small update
I've been using the turret for a month now on a variety of jobs (including making more turrets rap rap style)
There are a couple of things that I've changed from the original turret specs.

1...Moved the mounting bolts to the outside of the unit this makes it faster to remove ( 1 minute), with the original it takes at least five minutes (you have to remove the top cover before removing the mounting screws).

2...Up-ratted the main bearing specs (there was no need to do this except the new heavier housings are cheaper to buy
and more readily available). I mainly did this to achieve less over hang from the cross slide, the snout is 40 mm shorter.

In the process I discovered the person that made the 4 jaw chuck that came with the lathe missed every dimmension
by some margin, the jaws  not only varied in height on each step,but were not ground flat (0.06 mm) difference between each side of the jaw  Shocked.

I've made two front bearing plates,Pawls, Ratchets and need to machine one more bearing housing.

I wouldn't characterise the The cast steel housings as hard, but "tough" as I eventually used three (6mm) cutters to do the job.

I haven't had to charge the batteries yet.

I've played around with the opencv libs and C++ for the beagle-bone  (industrial) and at this stage can't see any big issues in using the libs for what I want to do, but because you are dealing with the fixed set of methods in the library
and 900 Mhz, that if you  pre-process the images then tool detection and identification is good.

I'm also looking into the Raspberry Pi 3 but it's power consumption may be a little high for this purpose.
Of course the main thing I need is SOC that runs Opencv and can drive a stepper motor, that has a small foot print
and is easy on the power.

Dave
 




 


* One Turret Making Another (RapRap Style).png (1157.18 KB, 1000x750 - viewed 49 times.)

* Pawls done.png (797.46 KB, 1000x750 - viewed 50 times.)

* Work Done So far.png (914.07 KB, 1000x750 - viewed 51 times.)

* Lathe Tool Turret tool detection find tool tips.png (623.62 KB, 1000x789 - viewed 59 times.)
Logged
dave benson
CNC Jedi
*****
Offline Offline

Posts: 887


View Profile
« Reply #38 on: January 20, 2018, 05:47:45 am »

Another update:

The new goal of the project, was to make another couple of turrets from  standard  size stock materials and fasteners, and to set up the software (Freecad -->Cambam -->Camotics) work flow so as to make them in a quick and efficient manner as possible with the equipment that I have.

I'm using CamBam 0.98 and Camotics, both of which have worked well together.
I'm using the latest  Freecad binaries of 0.17 13050, like they say on the website these versions are under heavy development so save often.

There's practically no documentation for these versions so if you want a relatively painless (Import/Export) experience and want to use the Assembly workbench's, then these are the way to go however you have to be fairly proficient in 3d parametric modelling to get to grips with these versions.   Although a little persistence is well rewarded.
 
Now I have two turret bodies ready to except the tool holding platterns.

As you know I MacGyver'ed  the first turret together (Scrap Heap Challenge style) over a few weekends from material I found fishing around in the scrap bin. (16 mm plate was a little overkill)

I'd already started machining the second Tool Holding Plattern a couple of weeks after getting the
first turret running, after making some making some improvements to the design, for example
adding a longer toe to the tool rests to support the tools more (see picture), however I found out by using the first turret almost daily for the last 3 months, that I was fixing a non-existing problem.

I did move the mounting holes to the outside of the turret to make it quicker to dismount.
The other thing I did was to move the overload mechanism to the front of the turret so that you could reset it without having to pop the top and one side off the turret. I've only had to reset  the overload  three times crashing into the chuck, (Ezilathe was adding a G28 and I hadn't set mach3 home position as I use G53's.)

I also changed the shape, so that I could affix tools to the periphery of the tool holder and changed the dimensions of the tool rests to accept half inch tools rather than the 12mm ones that I originally
had planned for (this does mean that you have to shim the tools to get the correct centre height) just like the original tool post. see pic 2

The first design took 37 hours of machining time, even though I was using the Troc mops and had
set the code for 10 mm tooling and .5 mm DOC and 200mm Feedrate (on a normal profile mop I use 100mm).
This seemed to work for a few hours, but eventually the spindle motor overheated (it was a hot day 38 deg C  outside, don't know what it was in the shop but a lot more I guess).

The other thing using the Troc mops generates huge files and the mill computer took noticeably
longer to respond. 470,000 lines of code.

So instead of machining away the “Islands” between the tool rests (which was most of the work) I made the tool rests separately and did some hand preparation of the material.

And now the job takes 6.5 hours of machining time, plus the hand prep I'm doing. Mostly baby sitting the mechanical hacksaw. I could do this with the 9 inch grinder to make it a lot faster (but I like my neighbours.)

Dave






 


* New Tool Plattern Ver3 plus sensor wheel plus Disk Brake.PNG (76.19 KB, 626x567 - viewed 40 times.)

* New Lathe Plattern Turret ver 4.PNG (48.96 KB, 664x560 - viewed 47 times.)

* turret NMBR 2 and ver3 tool Plattern.png (1152.94 KB, 1000x750 - viewed 48 times.)
Logged
dave benson
CNC Jedi
*****
Offline Offline

Posts: 887


View Profile
« Reply #39 on: February 08, 2018, 23:38:30 pm »

Another small Update
Well I finally got the second turret Backing plate and tool rests made using this process:

1. mark out the bar stock with an endlmill and drill and counterbore the bolt hole recesses.
2. Cut out the shapes roughly to size.
3. Use a jig to do each tool rest separately. Full Depth cut shrinking toolpaths

This method worked reasonably well, although I made a mistake by machining a chamfer too early in the process and I will change this for the third turret. This was only problematic because I forgot to square one of the pieces of barstock before starting out.

This method is by far the fastest, but required  more setups and indicating back in so today I'll try a different method
where I'll cut the barstock a bit wider and cut three rests out at once, this will take longer than the grinding method
(to cut the pieces) but will reduce the handling time. and potential errors due to misalignment between setups.

Dave


* Various styls of tool tests.png (1123.25 KB, 1000x563 - viewed 37 times.)

* Cutting a tool rest.png (1026.5 KB, 1000x563 - viewed 42 times.)

* Tool Rest Fit Up to backing plate.png (1191.49 KB, 1000x563 - viewed 57 times.)
Logged
EddyCurrent
CNC Jedi
*****
Offline Offline

Posts: 3430



View Profile
« Reply #40 on: February 09, 2018, 08:30:04 am »

Dave,

You must have a nice range of machinery now.
Logged
dave benson
CNC Jedi
*****
Offline Offline

Posts: 887


View Profile
« Reply #41 on: February 09, 2018, 15:28:45 pm »

Hi Eddy

Don't know really how nice they are, the mill is home made from a Taylor and Hobs Dimensional
Co-ordinate Measuring machine, the Spindle is a all cast iron Lathe Head (Heavy), the ball screws are NSK and Lead shine drivers. Crappy non machinist vice, although I did modify it  with an Anti  lift Toe Plate on the moveable jaw it  still lifts almost consistently 0.2 mm
Because I was going to make interchangeable parts (the tool rests) I needed to know just how accurate the whole process was, so I measured everything as I went along.

With the heavier loads placed upon it, by my doing these turrets, I've come to grips with it's true capacity and find the Z Axis A little wanting in the rigidity Dept and need to do something about it like heavier capacity slides and a ClearPath servo for a spindle drive.

What I was really interested in, was  this problem -> The tool rests have a 3mm roll pin to position them on the backing plate and a M6 SHCS to lock them down. The M6 measures 5.85 and I drilled a clearance hole to 6 mm and so a clearance around the screw between the thread and the side wall of 0.075 mm.

I took some measurements “of holes drilled to full depth” on both top and bottom of  a test  plate but these were somewhat inconsistent, as if you saw the “Bull's Eye Plot” the holes were grouped mostly in the positive direction towards the operator,  we are  talking about  Drill positioning error not Milling error which was different.

The first thing I did  was checked the gib's and reset the spindle preload  and the tram was out a little too, I also changed the mops a little by adding a spot facing  milling op before the spot drilling op.

The other thing is that I'm using a 3 Jaw chuck which has it's own run-out and so I had to indicate each tool in (by undoing the key and rotating the tool 30 Deg  and doing up again and keying  the chuck in the right order) sometimes I 'd luck out and get it only after a few times, other times it took many attempts. 

After a lot of fiddling and quite a few dud parts, it turned out well in the end.
 
One other thing is that  3 mm HSS drill's aren’t very rigid, so   I used a new cobalt drill, as these
have a thicker core diameter compared to a standard HSS drill and are stiffer.

There's always going to be a bit of error in any machine,system or process, and if you know what that error “is” and then if you design your part “with that in mind” you can save yourself a lot heartache.

Dave
Logged
dave benson
CNC Jedi
*****
Offline Offline

Posts: 887


View Profile
« Reply #42 on: February 21, 2018, 23:05:00 pm »

Another progress update.

Well after what seems an eternity, I've managed to get both of the tool holding plates complete, as far as the cnc'ing goes, there is still some still some hand finishing and a chamfer on one plate which I'll do on the lathe.

The tool rests in the upper of the picture are the Failures, so to get 12 good ones I had to make 20 all up.

Three of  four on the far left were on size and would have been ok to use, but I thought I would add the recessed pocket to them, for style reasons only (should have given myself some Gansta Slaps at this point) as I didn't have enough room to put the M5 gib adjustment screws in afterwards.

The other four (the first one's I made) were  under size and had a poor finish, after this  I checked the mill over properly and trammed it in, and completely changed the method's used to to make them and opened up the tool recess enough to use ½ inch tooling which I do have some of.

NMBR 2 and NMBR 3 turrets differ slightly in the way tool plattern's bolt up as I realised that I could eliminate six extra M6's and the associated machining ops by using the tool rest securing bolts to fix the tool plattern to the  spindle shaft.

NMBR three's PCD is 84 mm the same as some chucks so that with an adapter plate you could use a chuck and use the turret as a (positioning only) 4 Axis.
Both new turrets will have an Disk Brake to enable you to do this.


What I have to do now is mostly lathe work, and because I decided that the overload mechanism securing bolt should be moved from inside the turret body to the outside (to make it easy to reset)
and so now it has to be left hand thread.

 I've ordered a M14 LH thread tap and die and intend to Get a couple of M16 x 1.5 RH Thread bolts and machine them down to suit the M14, that way  I can use still a standard 16 mm  socket with them.

On the programming side I've successfully set up cross compiling for both a Beagle Bone Black
and a Raspberry Pi with VS 20017, I've also setup a project that just uses an Arduino and uses the host computer to run OpenCV and will test out each option (they all have their good and bad points) when I've finally got the turrets operating.

I've got some time now so I'm of to the shed to bore the clearance hole for the M14 and drill the backing plate to accommodate tool plattern.

Dave


* Tooling Plate.png (1068.27 KB, 1000x563 - viewed 49 times.)
Logged
dave benson
CNC Jedi
*****
Offline Offline

Posts: 887


View Profile
« Reply #43 on: March 01, 2018, 04:30:29 am »

Face Plates shafts, and Ratchet wheels Finished.
Just waiting on tap and Die, just need to add the M6 Gib adjusters and some extra holes on the periphery for
the tooling that bolts on to the side of the tool holders (for boring bars and spotting drills).

Dave


* Complete Face plates shaft and ratchets.png (1041.73 KB, 1000x563 - viewed 41 times.)
Logged
Dragonfly
CNC Jedi
*****
Offline Offline

Posts: 1889



View Profile
« Reply #44 on: March 01, 2018, 07:26:20 am »

Look very solid.
Logged

Before asking a question do some effort and walk through all menus and options in CamBam.  Maybe the answer is there. Please.
Pages: 1 2 [3] 4
  Print  
 
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.21 | SMF © 2015, Simple Machines

Valid XHTML 1.0! Valid CSS! Dilber MC Theme by HarzeM
Page created in 1.374 seconds with 19 queries.

Copyright © 2018 HexRay Ltd. | Sitemap