Hello Guest it is July 22, 2019, 08:22:40 AM

Author Topic: BT30 spindle from scratch - with power drawbar and ATC of course  (Read 402338 times)

0 Members and 1 Guest are viewing this topic.

Re: BT30 spindle from scratch - with power drawbar and ATC of course
« Reply #370 on: December 16, 2012, 05:17:11 PM »
Articulated is just another way of saying 'jointed' so I would submit that a 4bar is a specific type of articulation . . although there are numerous types of 4 bar arrangements, so I would then use the term 'specific' only in a general way. How's that for an oxymoron?
I think the distinguishing characteristic would be the independence of motion of the various arms in the system.  In a 4-bar, there is a fixed relationship between them, regardless of position, as the arms are pinned together at the pivot points.  If you know the position of any one of the four arms, you know the position of all of them.  To me, at least, an articulated arm implies having two or more independently controllable arms connected together, so you need multiple data points to know the position of the business end of the arm.  Semantics.....

There is the answer. I have not the time nor talent to create a protocol, so I shall have to be satisfied to huddle with the rest of the masses.
 I mean VERY simple, and easily modified/extended.  If you'd like to see what it does, I'd be happy to fill you in.  The protocol can be used as-is, with the specific "command set" defined by just changing a few very simple files. It's partitioned so the command set can be changed without touching the underlying "routing" code, and the hardware-specific parts of it are isolated in a few files, so the hardware platform can be changed quite easily, to move it to another MCU.  I had to do this, since the same base code must be able to run on both the KFlop and the Arduinos.  The only downside is this also force me to write it in old-fashioned C, rather than C++ or some other modern OOP language, which would have made the code a lot "cleaner".  But, I did it in a pseudo OOP manner, so it's still quite easy to follow.  And, again, the only parts you'd ever need to mess with are really dead-simple, and just have to do with validating incoming messages, and actually executing the commands.  All the housekeeping of the actual "protocol" is handled by the low-level code, which neither knows nor cares what the platform is, nor what the messages actually contain, or what they actually do - it just makes sure they get where they're supposed to, and that valid responses find their way back to the initiator.

I'm looking to get rid of a USB smoothstepper and early reports on the new Ethernet version were all good, but now it is starting to look a bit like the earlier incarnation with problems going unfixed for extended periods. Kflop seems to be on top of things and responsive to questions.
If you'd like to try one, I have one (and several USB ones) I'd be happy to loan you.

On topic, of particular interest is the Kflop is extensible and may be a better portal for the ATC operational data than trying to communicate with MACH. I like the idea of having Kflop orchestrate separate CPUs which each have their own process to control. Kflops user code interface is where I would seek whatever advice you might be willing to provide, but the prerequisite of course is to have the Kflop.

To that end, initially all I really need to hear is (regarding the MACH plug-in) 'It works' or 'It works except for this bug, that anomaly, this workaround, etc' (in which case I would pass).

Your endorsement would be enough for me to acquire a Kflop and invest some time in it.
On that basis, I would recommend KFlop without hesitation.  Tom Kerekes, the designer, is readily available, and extremely responsive.   He's already made several changes based on my input, and has always helped come up with a solution to any problem I've thrown at him.  I usually get a response within hours.  My only hesitation would be that the learning curve on the KFlop can be a bit of a bugger.  But, I'd be happy to give you assistance, and get you pointed in the right direction, so you can largely avoid flailing around in the dark as I had to at times.

You could easily have the basic machine up and running within an hour or two after unpacking the KFlop.  Adding all the ancillary features can then be done incrementally.  

As far as stability, I'm aware of no problems with the KFlop plug-in, and certainly saw none when I did work with it, albeit briefly.  I've never seen any problems raised on the Dynomotion Yahoo Group, which it the primary support portal.  If you give me information on your machines configuration, I could probably provide sample code that would get you started with only minor modification and debug.  Significant parts of the code I have on my machine might well work on yours with only minor modification.  My homing and probing code (perhaps the most complex code, other than my pendant "driver"), for example, I suspect you could use as-is.

With the KFlop, you do have to write code for things that you get out of the box with Mach3.  For instance, you have to write code to control the spindle speed, do homing, probing, handling limits, etc.  Very simple controls like relay outputs for coolant, etc, can be handled without code, but things like a PWM-driven spindel require some simple C code to be written.  But, if you're not doing anything exotic, this code will generally be quite simple, in comes cases only a few lines.  If you're at all familiar with C, you'll have no problems whatsoever. But, if you want exotic, it can easily handle that as well.  I've added some unusual capabilities to mine, but the code is still trivial.  For example, I can re-program all axis parameters on the fly from the PC application.  This allowed me to store all configuration data in a single XML file, rather than having some in the C code, some in the application, and some hard-coded.  You can also do things like have the KFlop processor hand off lines of G-code to be executed by the CP application, which comes in handy at times.

The KFlop is VERY powerful, and almost infinitely flexible.  Once you understand it, it's a truly wonderful thing to work with.  As an example, the code to get my ATC working took, total, probably two hours to write, counting the several iterations as the hardware progressed.  It stands right now at perhaps two pages of dead-simple code with lots of white space.

Regards,
Ray L.
« Last Edit: December 16, 2012, 05:25:46 PM by HimyKabibble »
Regards,
Ray L.
Re: BT30 spindle from scratch - with power drawbar and ATC of course
« Reply #371 on: December 24, 2012, 06:15:39 PM »
Steve better be making progress!  My ATC officially went into service yesterday, and it has now run its first three jobs without a single problem.  Today I got the "skirt" on the carousel, so all I have left to do is to get to a bicycle shop for the cable to operate the carousel door, and then do a cosmetic clean up the wiring and plumbing.  Other than that, I'm done!

I did try to get a video of it in-action, but coolant and the table enclosure make it pretty near impossible to get anything you can actually see, without the camera taking a bath.  Since it's a brand-new camera, I'm reluctant to subject it to that...

Regards,
Ray L.
Regards,
Ray L.
Re: BT30 spindle from scratch - with power drawbar and ATC of course
« Reply #372 on: December 26, 2012, 03:22:47 PM »
And, we now have a door on the carousel.  Just need to get into town and pickup a spring (to close the door) and a bicycle brake cable (to open the door), and we're done!

I can taste Steve's beer already....  :-)

Regards,
Ray L.
Regards,
Ray L.

Offline simpson36

*
  •  1,374 1,374
    • View Profile
Re: BT30 spindle from scratch - with power drawbar and ATC of course
« Reply #373 on: December 29, 2012, 07:53:28 AM »
Ray, truth be told, I fully expect you to win. I just want the competition to be at the highest level, and I know you are capable of that. I have a selfish motivation, of course, but it also hopefully benefits the many readers of this thread to observe the process.

I think we agree that the mechanics, complicated as they may be, are not the lion's share of an ATC, but rather the control and sensors. I would speculate that those who can build the mechanism are many, those who can create an autonomous and safe control system are few. That's where I see the real competition. The finish line includes all sensors and 'release candidate' level fully functional control code. I also want to follow what Hood and other 'non-contestants' are doing with regard to sensors and control code.

To consolidate your questions from the last couple posts, let me say that yes, I am familiar with 'C'. There are several thousand lines of it in my InTurn™ motor control system. And less volume but far greated intensity code in my balancer. Although I am a lightweight with electronics, I get by with the 90% learn and 10% do methodology. 

I am going to acquire a Kflop board based on your endorsement. Initially, I just want to replace the smoothstepper, but I have similar needs to yours down the road, so the sizeable learning curve is an investment that will pay off, methinks. You mentioned loaning a board, but the way it is worded, I cannot tell if you mean loading an ethernet Smoothstepper or a Kflop board.   ???

As to machine complexity, I have 4 axis of Mitsubishi J3 series industrial drives (and motors) and the spindle is a Copley Controls Xenus drive pushing a big DC brush motor. I am starting to talk to these devices in real time to get at some of the pertinent staus info they provide and create a 'reactive' system that can act on the data from the drives to adjust the paramaters on a CNC machine in real time. Another machining center feature to bring down to we mere mortals. Hence my interest in your amazing work with the Kflop. Yeah, I did some homework on that.

Anyway, I threw this machine together quick and dirty to get enough Y travel to do a specific project. I have since boxed and welded up the column and now it has decent performance.  My 'permanent' mill will be similar, but larger, much heavier and of course will have the spindle and tool changer from this thread. Since the spindle is completed (except for the die spring) and the ATC is almost done, I am revisiting the pile of parts and the frame design. Today I purchased a 9" x 42" Bridgeport mill table to add to the pile.

There has been no progress on my ATC because I had some unanticipated work to accomplish, however, I expect to start making progress again on the ATC today. I am simultaneously making a new spindle from A6 tool steel, (which will be fully hardened and ground) and ditching the bellevilles. I expect there will be a lot of interest in my big fat die spring because it *probably* can be implemented on an existing spindle which has no way to get the typical disc spring arrangement installed. Next couple days I will post some photos of the ATC assembled on the side of the new head.

I see from the photos you posted that the swing arm has been eating its Wheaties. Very nice piece!  8) I also note the pivot has an upper and lower mounting. Double nice. Significant improvements from the prototype, as one would expect. Q: curiosity, how did you form the smooth curve in the skirt?




Re: BT30 spindle from scratch - with power drawbar and ATC of course
« Reply #374 on: December 29, 2012, 10:46:49 AM »
Ray, truth be told, I fully expect you to win. I just want the competition to be at the highest level, and I know you are capable of that. I have a selfish motivation, of course, but it also hopefully benefits the many readers of this thread to observe the process.

I think we agree that the mechanics, complicated as they may be, are not the lion's share of an ATC, but rather the control and sensors. I would speculate that those who can build the mechanism are many, those who can create an autonomous and safe control system are few. That's where I see the real competition. The finish line includes all sensors and 'release candidate' level fully functional control code. I also want to follow what Hood and other 'non-contestants' are doing with regard to sensors and control code.

Actually, the sensors and code on mine turned out to be quite trivial.  The code is barely two pages of dead-simple C that took only maybe two hours to write and debug.

To consolidate your questions from the last couple posts, let me say that yes, I am familiar with 'C'. There are several thousand lines of it in my InTurn™ motor control system. And less volume but far greated intensity code in my balancer. Although I am a lightweight with electronics, I get by with the 90% learn and 10% do methodology. 

I am going to acquire a Kflop board based on your endorsement. Initially, I just want to replace the smoothstepper, but I have similar needs to yours down the road, so the sizeable learning curve is an investment that will pay off, methinks. You mentioned loaning a board, but the way it is worded, I cannot tell if you mean loading an ethernet Smoothstepper or a Kflop board.   ???

Must've been a SmoothStepper I was referring to.  I only have the oneKFlop of my own, and a backup that actually belongs to a friend.

As to machine complexity, I have 4 axis of Mitsubishi J3 series industrial drives (and motors) and the spindle is a Copley Controls Xenus drive pushing a big DC brush motor. I am starting to talk to these devices in real time to get at some of the pertinent staus info they provide and create a 'reactive' system that can act on the data from the drives to adjust the paramaters on a CNC machine in real time. Another machining center feature to bring down to we mere mortals. Hence my interest in your amazing work with the Kflop. Yeah, I did some homework on that.

Anyway, I threw this machine together quick and dirty to get enough Y travel to do a specific project. I have since boxed and welded up the column and now it has decent performance.  My 'permanent' mill will be similar, but larger, much heavier and of course will have the spindle and tool changer from this thread. Since the spindle is completed (except for the die spring) and the ATC is almost done, I am revisiting the pile of parts and the frame design. Today I purchased a 9" x 42" Bridgeport mill table to add to the pile.

There has been no progress on my ATC because I had some unanticipated work to accomplish, however, I expect to start making progress again on the ATC today. I am simultaneously making a new spindle from A6 tool steel, (which will be fully hardened and ground) and ditching the bellevilles. I expect there will be a lot of interest in my big fat die spring because it *probably* can be implemented on an existing spindle which has no way to get the typical disc spring arrangement installed. Next couple days I will post some photos of the ATC assembled on the side of the new head.

I see from the photos you posted that the swing arm has been eating its Wheaties. Very nice piece!  8) I also note the pivot has an upper and lower mounting. Double nice. Significant improvements from the prototype, as one would expect. Q: curiosity, how did you form the smooth curve in the skirt?

The pivot arm bracket is actually the prototype.  The top support for the pivot was always planned, it just took a while to get it built, and made it much easier to remove the whole assembly from the machine during initial tinkering and debugging.  The "skirt" was just rough hand-formed over a piece of 4" diameter aluminum round, then fastened in place with 6-32 screws.  I'm very pleased with how nicely it came out.  I now have the door operating as well.  I used a simple tension spring to pull the door open, and a cable from the door to the mounting arm on the ram to close.  The geometry is such that with just a short (18") piece of bicycle brake cable inner wire fastened between those two points, it pulls tight when the arm parks, and goes slack when the arm pivots to the quill, allowing the spring to pull the door open.  Could'nt be any simpler or cheaper.

Now that my annoying spindle problems appear to be resolved, I can finally get back to work!  Looking forward to seeing what you come up with.  Your work always leaves me feeling totally inadequate!  :-)

Regards,
Ray L.

Regards,
Ray L.
Re: BT30 spindle from scratch - with power drawbar and ATC of course
« Reply #375 on: December 29, 2012, 06:31:33 PM »
Here is one final video, making cuts and doing toolchanges.

http://www.youtube.com/watch?v=zjGi1WBTMBE&feature=youtu.be

Regards,
Ray L.
Regards,
Ray L.

Offline derek

*
  •  201 201
    • View Profile
Re: BT30 spindle from scratch - with power drawbar and ATC of course
« Reply #376 on: December 29, 2012, 07:15:21 PM »
Nice Ray!
But with the cover on you cant see the magic!

Offline simpson36

*
  •  1,374 1,374
    • View Profile
Re: BT30 spindle from scratch - with power drawbar and ATC of course
« Reply #377 on: December 30, 2012, 01:34:11 AM »
Some progress to report.

Ding, dong, the wicked Bellevilles are dead and gone. Long live the spring!

Only after scrubbing BOTH sides of each belleville by hand to create flat mating surfaces would the springs stay aligned when compressed. However, that solution created another problem by reducing the travel to barely what was needed to release the gripper. It still worked, but it would not take much variation in a stud to have a tool hang up in the spindle due to insufficient travel to fully release. Worse, even after all of that effort, the disk springs still chafed on the shaft as seen in the following photo. Using disk springs specifically made for a machine spindle or 'nesting' much thinner springs to get up to the required force might have solved these issues. Nitriding the drawbar and tube would certainly help, but my patience ran out for the whole spring stack arrangement. There may be some compelling reason to use this method in certain cases, for example, in-line nesting of coil springs increases travel, but not spring force. However, for the BT30 spec, I see no reason not to replace a troublesome $60 stack of problems with a $19 die spring and not have to deal with the the spring stack issues at all.



Jumping ahead to the completed components:



Switching over to the coil spring was relatively simple matter of cutting a pocket for the spring and making retainers and guides. The angle in the next photo is intended to show the bottom of the pocket and it makes the snap ring grove look a bit odd, but it is a simple square bottom grove. One of the unsatisfying aspects of the disc spring stack was that with no tool in place, the drawbar would extend all the way up and press against the actuator making the spindle difficult to turn (for belt changes, etc). If the spindle was started in this condition, it would damage the actuator plunger and the top of the drawbar.

In the new arrangement, a snap ring at the top of the spindle makes a positive stop for the spring when there is no tool holder in the spindle. The snap ring is on order so it is not pictured.



And here is the spring installed on the drawbar and partially inserted into the spindle. Most likely, this arrangement could be added to an existing spindle with the spring on top of the spindle. This particular spring was selected for BT30 and the rate is 2550lbs. At coil bind you have approx 2,000lbs and for a gripper with a .21" release travel, you get a max hold of about 1,450lbs. Note that this in not adequate for R8, but larger springs are available that for that application. I have the arrangement set up so that any hold tension from 600lbs to the 1,450 max is available by simply adjusting the threaded retainer cap.
Re: BT30 spindle from scratch - with power drawbar and ATC of course
« Reply #378 on: December 30, 2012, 01:43:38 AM »
Steve,

Nice work as always.  The time I spent playing with Belleville drawbars a few years ago convinced me I never wanted to mess with them again, and is what led to my stepper-driven drawbar.  I wonder why so many VMCs use them?  And huge stacks of them at that.

BTW - I sent you a e-mail earlier today re: KFlop.  Did you get it?

Regards,
Ray L.
Regards,
Ray L.

Offline simpson36

*
  •  1,374 1,374
    • View Profile
Re: BT30 spindle from scratch - with power drawbar and ATC of course
« Reply #379 on: December 30, 2012, 01:48:17 AM »
This post is just some misc stuff that has been completed for the ATC.

Following photos are of the completed carousel mounting arm. In the next few days, I hope to have this and the swing arm mounted to the side plate of the new mill head.



A previously missing detail is the safety interlock for the drawbar actuator. This little guy sports a hardened pin that runs thru a hardened die pin guide bushing and interferes with the actuator movement unless it is retracted. i.e. the drawbar actuator cannot release the tool unless this pin is first retracted. Totalling all of the special parts and the time to design and build this component adds up to a fairly expensive little gadget.  However, safety features are immune from cost cutting and/or production considerations.