Maker Faire 2010 - IN MICHIGAN (7/31-8/1)

[Keith Mc]

It hasn't been officially announced yet...
The **FULL** Maker Faire 2010 is coming to Michigan! <GRIN>
... Date - July 31 & August 1, 2010
... Place - Greenfield Village and Henry Ford Museum grounds
They are CURRENTLY estimating about 20,000 attendees, but who knows...

Can we add this to our calendar? If everyone wants to shoot for it,
I would love to see us demonstrating RepRap there. It would give us
focus and a hard target of a bit over 8 months, to get all of our machines
fully up, and debugged. (IMHO, that should be plenty of time.)

If we wish to go for it, I will be see about becoming a contact for RepRap demos.

Personally, I like it. It is what we have needed. We can talk about
some CONCRETE presentation GOALS at the next meeting.

IMO it would help us focus, and spur development of a detailed action plan
and activity schedule for the next year to get this project DONE.
Regardless of what we show, I'd just be thrilled to shoot for it, just to see
everyone's machines running by next summer.

Thoughts?

- Keith Mc.

Dan <dandi...@yahoo.com> wrote:
> Dale Dougherty and Sherry Huss of O'Reilly Media informally passed the word that
> Maker Faire 2010 is coming to the Henry Ford Museum and grounds on July 31
> and August 1st, 2010. The public announcement will likely come in December
> with Maker applications being called for in March. Estimated attendance is
> 20,000 and depending on future developments attendance could possibly
> exceed this. Exhibit space will be in both the Henry Ford Museum itself and in
> outdoor tenting which will be erected in the IMAX parking lot.
>
> There is no attendance fee for Makers to exhibit although there is an simple
> application process for safety and facility purposes. The theme for the local
> Maker Faire will be "From Motor City to Maker City" and a web site has
> recently been set up at reflecting this:
>
> Maker City Web Site
> http://www.makercity.com
>
> The first Maker Faire was held in 2006 in California at the San Mateo Fairgrounds
> and had 20,000 attendees. The most recent Maker Faire at the San Mateo
> Fairgrounds in 2009 drew over 80,000 attendees.

[Keith Mc]

I was researching backlash nuts. Commercial ones are EXPENSIVE.
What backlash solution has everyone else come up with?

Here are some options I've found:

1) Here is a basic, machinable nut pair, using a compression spring between them:
... http://3dreplicators.com/cgi-bin/cblog/index.php?/archives/500-Anti-backlash-nut.html

2) A great explanation of the nuts, including drawings that SUGGEST a
PRINTABLE pair of Nut Things:
... http://www.cncroutersource.com/precision-acme-leadscrew-nut.html
... http://www.cncroutersource.com/image-files/acme-tension-nut-2.jpg
- RR two dual/tri-finger interlocking cylindrical Things. Optionally, Thing1
has a mounting flange or flat printed, to tie it to the slide. (Or, tie it with a clamp.)
- Both have EITHER hex holes on their far ends to nest-hold stock nuts for
the threaded rod, OR if we can RR strong enough, simply print it solid
then drill/tap each Thing to match the rod thread.
- The interlocking finger sets keep the two ends from rotating relative to
each other as they slide along the screw axis. (They only need to allow
for a few threads worth of sliding play.)
- Insert standard nuts into the ends of each Thing.
- Expand the pair with a compression spring, that slips over the cylindrical
finger sections during assembly.
- Screw Thing1 onto the rod, with pressure compressing the spring
while engaging the 2nd nut of Thing2.
- Tie Thing1 to the slide. Let Thing2 float and maintain tension.

Note that interlocking finger slop AND hex nut slop are BOTH irrelevant.
As long as the compression force stays above the slide's drag force,
there is NO loss of accuracy.

This could be made easily with a mill and some HDPE block stock, but I
was searching for printable solutions, and this seems the best one to me.

3) Here's a VERY interesting item from another of my groups, on a SIMPLE
solution using a SINGLE piece of brass rod, and NO springs:

In multim...@yahoogroups.com, Pat wrote:
] I have an interesting [backlash nut] on a $300 Grizzly cross feed table.
] The nut is brass and a little longer than normal. It has a saw cut that
] goes about 3/4 way through and probably was hammered closed a little
] before the Acme screw was screwed in. Not a ball screw but it works very well.
] My kind of solution

I was thinking we could simply take a piece of square or hex brass rod
and make this with hand tools. Drill and tap it, and do the same (partially
hacksaw through it, hammer it together gently a tiny bit to offset the threads,
then fit it onto the rod). Fasten it on one end to the slide frame with a clamp
(and maybe a spacer), or a couple of screws, leaving the other end free.

One length of brass square or hex rod could quickly make an awful lot of
anti-backlash nuts for everyone in the group.

4) Here's a similar system, with a lengthwise cut:
... http://cr4.globalspec.com/PostImages/200901/SlidingNut01_008B14C1-B0F2-2683-594E244977520BF0.JPG
... http://cr4.globalspec.com/thread/32226/New-CNC-Membership-Type-Under-Hobbies
Not sure of how well this one works. I would think you still need longitudinal
tension, not latitudinal "squeeze", to eliminate backlash.

5) Here's a way to measure backlash:
... http://www.britishideas.com/2008/06/19/measuring-cnc-backlash-and-software-compensation/

6) BTW... Here's an interesting DIY mill I ran across while researching this:
... http://www.buildyouridea.com/cnc/pcbmill/pictures.html
Reminds me a bit of Devin's McWire...

- Keith Mc.

[Nicholas McCoy]

The Darwin design uses a spring trapped between two nuts.  The two nuts are held by a single printed part that prevents them from rotating relative to each other.  The printed part has a nice 90 degree platform for the bed to bolt onto.

I've reproduced it in my machine.  The springs are there, but you can't see them.  I don't have any problems with backlash on the thread-drive axis, but the only thread-drive on my system is the Z-axis, which does not change direction during a build.

[Nicholas McCoy]

I meant to include this link:

http://reprap.org/bin/view/Main/AssemblingDarwinMachinery#Top_end_assembly

[Keith Mc]

I came up with another one, too. All it takes is one piece of wood.

- Take a solid block of wood, or thick plywood. Cut an "H" frame into it.
The gaps between the arms are nut-flat width. The block thickness is
a touch larger than a nut's point-point dimension.
- Drill a rod pass thru hole through the center of the H crossbar,
parallel to the arms.
- Bolt the long arms of the H to the bottom of your slide.
- One side of the H holds a nut.
- The other side of the H has a preloaded spring, THEN the second nut.
(If needed, include some flat washers to keep things from binding up.)

This can be done ENTIRELY with just ONE piece of wood, drill & bits,
and a hacksaw (or a Dremel). NO glue necessary.

In fact, searching Google Images, I just found a pic of a similar rig:
http://www.instructables.com/files/deriv/FY6/IE7F/FPKF9T72/FY6IE7FFPKF9T72.MEDIUM.jpg
(So much for originality! ;-)

Nicholas McCoy <mccoyn@[...]> wrote:
> The Darwin design uses a spring trapped between two nuts.
> The two nuts are held by a single printed part that prevents
> them from rotating relative to each other. The printed part has
> a nice 90 degree platform for the bed to bolt onto.

Thanks.

Well heck... I thought no one was addressing this one.
All I've been seeing are single nuts held by hose clamps at the builds,
and everyone talking about software backlash compensation.

(OOC, How come most of the guys aren't using this solution, or at LEAST
an assembled wooden or machined variant, until they can print one)?

> I meant to include this link:
> http://reprap.org/bin/view/Main/AssemblingDarwinMachinery#Top_end_assembly

Here's the image directly:
... http://reprap.org/pub/Main/AssemblingDarwinMachinery/28.JPG
Yea, that's a good solution for it. You could make a RepStrap version
of that easily from a block of drilled square wooden dowel, a couple
of slices with a hacksaw, etc... (Use a blind hole drill bit, though...)

I seem to keep running into the "there is too much unindexed data
scattered all over reprap.org to easily find any one answer" problem.
Even Googling "<keyword> site:reprap.org" doesn't always get me the
right answer right away, because most of the prose doesn't always
include industry standard terminology for what they are doing.

Is anyone else having trouble finding specific answers to things there?

- Keith Mc.

[Tim Schmidt]

There is an even simpler solution... let gravity manage your
backlash. Simply tilt your build surface ~15 degrees off level in
both the X and Y axes. As demonstrated by the upside-down mendel, it
should not affect printing quality.

--tim

[Tim Schmidt]

On Thu, Nov 12, 2009 at 10:43 AM, Keith Mc <ac...@provide.net> wrote:
> Can we add this to our calendar?

I am 210% behind this. We need to be there.

--tim

[Keith Mc]

That IS a pretty cool idea, Tim... <grin> (Some DIY mills do something similar.)
That's FUNNY. I always keep telling my FIRST robotics students: "Remember:
Gravity Is Your FRIEND, for its proper use can eliminate many actuators".

I'm not sure if I want a V shaped Machine though. It is easier to diagnose
sags and motion errors in parts if the gravity vector isn't skewed.

Now if something momentarily binds up (especially from slide contamination,
rust, etc), without at least a LITTLE bit of anti-backlash force on the screw
every so often you could get intermittent motion glitches. So -personally-
I still want to add SOME kind of "significant" anti-backlash mechanics to
my RR's axes.

BTW... Another method (vs anti-backlash nuts) would be to simply attach a
STRONG Constant Force Spring (like a big tape measure, a BIG key chain retriever,
or a power cord retriever... ;-) to each axis slide. If the axis friction is low
enough (and XY is level, so gravity isn't a factor), they would keep "pulling"
the X and Y axes ever so slightly toward one end. <grin>

But yea, IF the slides won't bind, tilting the entire rig in both X and Y IS a
simpler solution than adding CFSs !

- Keith Mc.

[stmaus]

I would have to agree with and earlier statement that since the only
theaded axis is going in one direction during operation the backlash is
a moot issue.

On my McWire the X,Y which move horizontally would be more susceptible.
By using a coupler bolt which has a very long thread I push that problem
further down the road since it will take a long time for the threads to
wear enough to have backlash. By then I can cheaply replace the parts
instead of adding more parts to my machine.

By the time backlash is an issue I hope to have a working Mendel.

-=[Devin]=-

--
Talent is hitting a target no one else can hit.
Genius is hitting a target no one else can see.

[Phil Frost]

On Thu, Nov 12, 2009 at 10:43:12PM +0000, Keith Mc wrote:
>
> Tim Schmidt <timschmidt@[...]> wrote:
> > There is an even simpler solution... let gravity manage your backlash.
> > Simply tilt your build surface ~15 degrees off level in both the X and Y axes.
> > As demonstrated by the upside-down mendel, it should not affect printing quality.
>
> That IS a pretty cool idea, Tim... <grin> (Some DIY mills do something similar.)
> That's FUNNY. I always keep telling my FIRST robotics students: "Remember:
> Gravity Is Your FRIEND, for its proper use can eliminate many actuators".
>
> I'm not sure if I want a V shaped Machine though. It is easier to diagnose
> sags and motion errors in parts if the gravity vector isn't skewed.

There's a simple solution to that, too. Don't align the motion axes with
the build platform. Just about any three degrees of freedom will do.
They don't even have to be linear.