Hey Ray L,
I know I just got my machine, but I can't wait to get this upgrade. All of the work you have done looks awesome.
Hey Ray L,
I know I just got my machine, but I can't wait to get this upgrade. All of the work you have done looks awesome.
Today I got the new broach, so I was finally able to put the new transfer arm drive on the machine. It fits perfectly, and seems to operate almost perfectly. I managed to screw up machining the leadscrew (managed to get the bearing shoulder off-center...), so I'll need to make another one. The arm works perfectly, and the new fork is a huge improvement. The lift is being a bit cranky. It's smooth as can be, but for some reason is lacking in power, even though it's the same motor, encoder, and leadscrew as the previous version. I'll dig into that tomorrow.
One thing I'm really pleased with is how easy it was to mount and adjust. There are now only two mechanical adjustments for the transfer arm. The first is sliding the whole assembly front-to-back, using the adjustment slots in its bracket, so the fork engages "cleanly" with a toolholder in the spindle, and the arm is level. This took less than a minute. Then, the carousel has to be adjusted on its brackets so the fork engages cleanly with a toolholder in the carousel. Assuming the carousel bracket was properly leveled when it was mounted to the column, this means just loosening three screws, and rotating the carousel, then re-tightening the bolts. This also took less than a minute. Then there are two final adjustments that are done in software. First is to calibrate the lift to the height of the spindle. This is done by jogging the lift up or down, using the keyboard, until the fork is perfectly aligned to the ATC groove in a toolholder held in the spindle. Once it's right, press a key, and that offset is stored away in EEPROM. The second adjustment is exactly the same, but sets the height for the carousel, rather than the spindle. Each of these also took less than a minute. Everything else will calibrate itself each time the system is powered up.
Regards,
Ray L.
Here's a new video showing current status. I think you'll be able to see how much smoother and more precise the motion on all axes now is, especially the lift.
Everything is now working with the new drive, except the lift still does not have quite enough "Oomph", in part because it's heavier than it was, and, no doubt, a bit more friction due to the larger size and closer fit on the guides. Still working on that, but suspect I'll likely end up adding a 2:1 reducer to make life easier for the motor. Otherwise, it's working extremely well. There was a loud screeching noise, which, oddly, turned out to be coming from the PTFE leadnut! A single shot of WD-40 completely eliminated it. I had put a little grease on the screw, thinking that would help. It seems to have done exactly the opposite. In any case, that's a minor issue that will get sorted shortly.
You'll notice the arm now parks itself up high, behind the head. In fact, when parked, the bottom of the transfer arm drive is actually above the bottom of the head casting, so it will be virtually immune from crashes due to the head being lowered over tall objects. This will also protect the arm from chips and coolant. There will be a shield installed on the back of the head, and the arm will actually be parked above and behind the shield, so it should stay nice and clean. The shield will also provide a convenient mounting point for a small air-blast nozzle to blow chips off the spindle nose, to prevent them from fouling the transfer arm fork when it engages the tool.
SAM 0419 - YouTube
Regards,
Ray L.
That thing is nice, I can't wait for the finished product.
Thanks for sharing!
Will
Here it is actually doing toolchanges. The lift problem was resolved in firmware - there was some kludge code in there to deal with some problems with the old lift. Once I removed that, it started working. It still doesn't have quite as much muscle as I'd like, so I will almost certainly add a 2:1 reducer to the lift drive for production. That will also speed up lift moves, and let me reduce motor current, which is always a good thing. But, for now, it's working just fine.
SAM 0422 - YouTube
Regards,
Ray L.
Very cool! I really like the robot sounds it makes, it's like a Terminator or something.
It is looking really good Ray. I noticed the carousel lifting a bit when putting a tool back in. Is that just an adjustment or is that some compliance you built in to make sure the tool holder seats properly?
Hirudin is right. It sounds way cool.
bob
Bob,
There is some compliance there by design. On the prototype it's more than intended, but harmless. It's necessary with TTS to push the tool up firmly before locking, to make sure it's really seated, so that's why it does that on both the ATC and the spindle. The prototype carousel is supported by kinda flexy aluminum brackets on the column (production brackets will be steel), and a single diagonal "hanger" brace, which has a turn-buckle in the middle, so it keeps the carousel from dropping down, but also allows it to be pushed up a bit. As I said, harmless, but the production version will be much stiffer.
I've now got logical-to-physical mapping of tools working, so I can put any tool in any position, and the ATC will know where to find it. It will also (eventually) be smart enough that if a tool is requested that is NOT loaded in the carousel, it will unload the current tool, and prompt the user to do a manual change. This will be nice for handling tools like large face mills, tapping heads, or R8 tools that are not ATC-friendly.
I also now have the ATC being controlled directly by the controller app on the PC. Up to this point, the controller app has still been treating toolchanges as manual affairs, and I would manually command the ATC, through its console port, to do the swap, so I could observe and watch for any hiccups or sequencing issues that could be improved. It's been working so well for several weeks now, I finally got off my lazt butt and wrote a "driver" so the controller app can now do the toolchanges autonomously, and it's working just fine.
I've just now finished designing a reducer for the lift drive, to give it the extra "Oomph" it needs. The reducer should speed up the lift as well, as the motor will no longer be struggling. With luck, the parts will show up later this week. In the meantime, other than one pesky firmware bug, it's all been working really nicely. I can leave it running random changes for hours at a time, and it just keeps on chuggin', until the firmware bugs hits, and it loses its mind. I think I know what's happening, and just need to make some time to get to the bottom of it.
Finally, I now have the first real PCBs in-process. I hope to have real encoder PCBs in the next 10 days or so, and the large main PCB a few weeks later, all in time for Cabin Fever.
Regards,
Ray L.
I guess I'd better get some of these projects cleared up and get back on my Pro! At my pace, I may be making my first part with your tool changer already installed! lol It looks awesome Ray!!!
Don't have to be too bright to be me
Ray,
Freaking awesome and great job. I have a question though, on my machine...I have "way" covers for the Z-AXIS. How much clearance is there from the arm to the Z-axis "strut" ?
Also, the last few days make me realize how much I hate manually changing tools!
Regards,
-Jason
I doubt your way covers will be a problem. During its homing sequence, and the first tool change, the ATC arm "feels" its way around to figure out where things are - the column, spindle, carousel, etc. So it will find your way cover instead of the strut, and accommodate it in its motions.
Regards,
Ray L.
A bit more progress. I spent yesterday putting in a detailed "trace" capability, for debugging. I can now get a detailed report, typically about 350 lines of data, showing exactly what happened during an ATC operation. This shows me in great detail what all the state machines are doing, encoder readings and PWM outputs for each axis, and more, so I can see exactly what path is taken through the code, and exactly where it went wrong, if it did. This will be a huge aid in debugging problems in the field. It's already helped me fix a couple of very minor bugs I didn't even know were there.
I've also got the PCB design for the motor encoders done, and will be sending them out for fab today. I should have boards back within two weeks. Price is almost ridiculously low - barely 50 cents each! Next up is the main PCB.
I expect today to receive the pulleys for adding the belt reducer to the lift drive, and the belts are only a few days behind. I've received the first sample of a new motor for the carousel drive. It's very similar to the one I've been using, but has a more convenient output shaft configuration, and mounting configuration, and appears to be a bit more robustly made. Finally, I received new solenoids that will be used for both the tool receiver lock, and the lock I am considering for the tool fork. This one is very nicely made, and much more robust than the one on the prototype. Things are really falling into place nicely....
Regards,
Ray L.
Ray, really an amazing job with this entire project/product.
Who are you using for PCB prototypes if you don't mind me asking? I have used PCB123 for many years but now it's been many years since that time. Will need to explore for other PCB fabrication.
Sounds like the "trace" capability is going to be a big help, good luck and keep us posted.
Regards,
-Jason
There are several good low-cost PCB places out there. At the moment, I'm using MyroPCB.com. Like most, they are out of China, but pricing, quality, and service are first-rate. Despite coming from half-way around the world, I can have boards in-hand in a week, at prices almost too good to believe.
Regards,
Ray L.
Lordy! This stuff can be frustrating sometimes! Just spent about 8 hours tracking down a problem I through was in firmware, but actually turned out to be in hardware. I'm getting rare, spurious triggers on the lift limit sensors. For reasons I can't understand, this became a major problem yesterday. Fortunately, it caused a very consistent failure, as the lift drive would be shut down mid-move by one of the many safety interlock systems. I was able to get around the problem, for now, by just disabling the limit safety during the one operation that seems to be triggering the noise. I'm sure when I get the real PCBs made, this will no longer be an issue at all.
Overall, I can't feel too bad, given the "explosion in a spaghetti factory" wiring. By rights, I should have had far more such problem than I have. And, along the way, I've found a few more small improvements to make. I am also cautiously optimistic that I did find and fix the threading bug that was causing the firmware to weird out at random times.
I also got the new, beefier lock solenoid mounted, and it fits, and works, wonderfully. A big improvement. The only problem I'm seeing now is one which should be completely eliminated by the addition of the lift drive reducer, and, in the new carousel I'll be building shortly, a minor dimension change in the tool receivers. I've seen no other errors in quite some time.
Regards,
Ray L.
Forgive me if this has already been addressed. I don't have one of these machines so am curious. It looks like the arm moves against the column after and between tool changes. Is there a gap between the column and head that allows the z to move down as it looks like the head would hit the arm moving down to cut. Would it be possible to see a video that shows where the arm is in relation to the head and column during cutting? Thank you.
Ben
Hi Ray
Welcome to the real world. It now works? That means you are half-way there - no further...
But I do like the safety interlocks bit.
Cheers
Roger
Cool. Thanks for the explanation. It makes perfect since now.
Ben