The rebirth of GUS #1

The rebirth of GUS #1

Postby owens-bill » Sun Jan 12, 2014 3:49 am

Tonight I finished the retrofit of GUS #1 (not the original - that would be Nicholas' GUS #0 of course) and managed to get a print of the official Simpson Test Squirrel:

Photo Jan 11, 22 15 12.jpg
Squirrel just completed
Photo Jan 11, 22 15 12.jpg (2 MiB) Viewed 7029 times


Photo Jan 11, 22 15 19.jpg
Squirrel closeup
Photo Jan 11, 22 15 19.jpg (1.55 MiB) Viewed 7029 times


I discovered a couple of things. First, I needed to measure and calculate my own calibration number. It turned out to be about 123 steps/mm, which is close to Nicholas' 120, but far enough off that I could not get a set of calibration points that would work with segmentize.py. Second, I have some kind of hysteresis in the printer, perhaps due to the spring not maintaining constant tension. I'm not entirely sure what's going on there, more testing will be needed. I'll probably fall back to 20 mm calibration cubes now that the official print has been made.

Incidentally, the combination of issues that we had with the original incarnation of GUS #1 meant that it never managed to make a complete squirrel, so we're already well ahead in that respect ;)
owens-bill
 
Posts: 35
Joined: Tue Nov 26, 2013 3:17 am

Re: The rebirth of GUS #1

Postby Nicholas Seward » Sun Jan 12, 2014 4:55 am

I definitely have hysteresis in my system. It is not super noticeable when I fully tension. I will probably get stiff springs and really crank up the tension if I want to be a speed demon. These springs can only do about 3.5lbf so I am going to bump up to 10lbf max and give it a go.

Super exciting that you got a whole squirrel. Easier from here on out.
Nicholas Seward
 
Posts: 738
Joined: Mon Nov 25, 2013 10:41 pm

Re: The rebirth of GUS #1

Postby brandonh » Sun Jan 12, 2014 6:00 pm

Nice! Motivating to those of us not quite at the print stage yet.
brandonh
 
Posts: 93
Joined: Tue Dec 03, 2013 8:18 am

Re: The rebirth of GUS #1

Postby owens-bill » Mon Jan 13, 2014 2:19 am

Considerable progress today; I was able to track down two causes of hysteresis in the arms. The first one that I resolved was contributing just a little, I think; the tension in the strings wasn't high enough. The springs would slightly extend and contract during the arm motion, and once I cranked up the tension to the point where they stayed extended, that error disappeared.

The other problem was harder to track down, and I finally figured it out by constructing some test gcode files that moved just one axis at a time. The Z axis, and to a lesser extent the X, had just barely loose pulleys, enough to let them turn by a few degrees. Of course, tightening the pulley screws with the arm fully assembled is very difficult, so I improvised a solution with a soldering iron:
Photo Jan 12, 15 07 47.jpg
Motor arm with access hole
Photo Jan 12, 15 07 47.jpg (1.63 MiB) Viewed 6966 times


Once the pulleys were tight I re-did the calibration (fifth time, I think) and tried a 20 mm cube. The result wasn't perfect; the first layer is still dicey and there's definitely some skew, which Nicholas suspects is incorrect calibration on one or more of the arms. Nevertheless I tried a somewhat larger print, and got a pretty decent result:
Photo Jan 12, 18 25 38.jpg
Simpson prints a moai
Photo Jan 12, 18 25 38.jpg (1.73 MiB) Viewed 6966 times


I intentionally chose something that would look okay even a little distorted, even though that's really cheating! Looks like one of those 12" calipers will be in order for the next stage of calibration. . .
owens-bill
 
Posts: 35
Joined: Tue Nov 26, 2013 3:17 am

Re: The rebirth of GUS #1

Postby brandonh » Mon Jan 13, 2014 6:37 am

FWIW, I tighten the pulley screws with a long-length ball driver, through the far gear end. An access hole would have the side benefit of giving a view into any string overlap. Nice first print!
brandonh
 
Posts: 93
Joined: Tue Dec 03, 2013 8:18 am

Re: The rebirth of GUS #1

Postby owens-bill » Mon Jan 13, 2014 1:39 pm

I don't have any hex wrenches long enough to reach all the way in, the best I could do was work in between the halves of the gear, and that wasn't good enough to get the screws tight (as I discovered the hard way). I also didn't use nyloc nuts on the pulleys, but I'll probably change them.

The hole is small enough that I can't see anything except the center of the screwhead. I was actually thinking that a larger window in the side of the arm would be nice for watching the tracking but I wouldn't want to do anything that would reduce stiffness. I suppose I could reprint the arms in T-glase and just look through them ;)
owens-bill
 
Posts: 35
Joined: Tue Nov 26, 2013 3:17 am


Return to GUS Simpson

Who is online

Users browsing this forum: No registered users and 1 guest

cron