Test Jig #4 on it's way

[Ben Fraser]

My threaded rod finally arrived a few minutes ago. I put a quick rig
together to test out the gear train (without pallets). It runs pretty
smoothly with moderate force, I'd say 5lbs or so applied to the drum
on the hour gear. I can stop the escapement wheel repeatedly and it
starts back up no problem. It feels like there is plenty of force
there to drive the pendulum.

I'm going to build a quick wood frame for this and try to get it
running. Will keep you posted!

[ruste...@prototribe.net]

Sweet!

[Mathieu Glachant]

I'm on 3D printed tenterhooks here. :-)

Can't wait!

[bre pettis]

Wow wow wow!

If you need any hardware from MakerBot to make it work, let me know!

Bre

MakerBot Industries
http://makerbot.com
Video - Time Magazine Game Changer: http://ti.me/nGtIrJ
Print - NYTimes: http://nyti.ms/oKITfc

NOTICE: This email may contain information that is confidential or
attorney-client privileged and may constitute inside information. The
contents of this email are intended only for the recipient(s) listed above.
If you are not the intended recipient, you are directed not to read,
disclose, distribute or otherwise use this transmission. If you have
received this email in error, please notify the sender immediately and
delete the transmission. Delivery of this message is not intended to waive
any applicable privileges.

[Ben Fraser]

Thanks for the offer Bre, I think I've got everything I need.

I built a quick wood frame and did some testing (still without pallets
in place). I was having some intermittent stalling. After some poking
around I noticed that it was the fat first layer on gear #6 impinging
on the escapement pinion. I decided to break it down and trim the
first layer on all the gears. What a difference. It now runs VERY
smoothly. I have to take a break now but will be back at it after my
kid's bed time.

Some interesting observations, here is the weight necessary to spin
the escapement wheel from different gears:

From the minute gear: 15g
From gear #2 (between minute and hour): 100g
From the hour gear: ~400g

That's all for now.

[Ben Fraser]

So freakin close.

Not sure why but when reassembled with the pendulum in place the
characteristics have changed a bit (not for the better). As a result I
am NOT able to run it off the hour gear. I still think it should be
possible but need to dig further. One possible issue is the amount of
leverage that the drum places on the hour gear twisting it out of
alignment. I think that leverage causes a lot of unpredictible forces
which are made worse by the loose slip fit around the minute pin/
barrel. I will continue to work on this.

That's the bad news, now here is the good news. I hung the weight
1400g on gear #2. Gear #2 is the intermediate gear between the minute
and hour gear. It ticks like a dream :)

http://youtu.be/UemE-tajAqc

[Mathieu Glachant]

That video is just beautiful :-)

[Ben Fraser]

:) Thanks, Pretty excited about it.

I was just thinking about what it would take to mount the drum on #2.
It might be over my head SCADwise, not sure. I'm about to poke around
in the library.

Correct me if I'm wrong:

#2 rotates 4 times per 12 hour period, drum circumference is ~ 9
inches. So the total drop of the weight would only be 3 feet over 12
hours. Maybe that's not so bad.

[Ben Fraser]

Also, if the drum were mounted on #2 it could ride on it's own
bearings. That would be a great improvement given the amount of weight
that it takes. If we added a pulley to the weight we could have a 24
hour clock (with 36in drop). That starts to sound pretty good.

[Mathieu Glachant]

Actually, what's keeping the drum from running on bearings now? The shaft linking the gear to the front of the clock? We can remove it in a handless version of the clock... Then figure out either concentric brass tubes or a separate shaft for the hands later.

On Tuesday, January 31, 2012, Ben Fraser <ma...@benfraser.com> wrote:

[Mathieu Glachant]

If you want, you can zip up your current version of the library, clock builder script and parameter file, email it to me, and I can make the mod. You'd only need to reprint the two parts of the drum gear.

[Benjamin Fraser]

Thanks for the offer, I've attached the library.

Here's how it's working right now:

1) Just a little pressure on the minute gear will cold start the pendulum no problem.
2) 5 ounces (or modest pressure with my fingers on the gears) will drive the pendulum, but not cold start.
3) Applying as much force as possible with my fingers to the hour gear (without stripping the gears) will just barely drive the pendulum, but not even reliably.

I'm all for trying to get the drum on the hour gear. That would be great, but right now it feels like 2 steps forward, one back.

[Mathieu Glachant]

I'll take a look at both options tomorrow: an hour drum that rests directly on the shaft via bearings, and a minute drum that can sit in the middle of the gear train.

I well remember the highs and lows of working on the clock. It always feels just out of reach... Addictive!

[bre pettis]

The wooden clock had a separate drum that was below the other gears
and had a smaller diameter. There may be something to that.

Bre

[Mathieu Glachant]

True... By getting the drum off the main shaft, it may be easier to control warping or binding and the like.

It sounds like Ben is having trouble driving from the hour gear even by hand, so the drum may not be the issue at all...

[Benjamin Fraser]

There's another possible issue here as well. The amount of force necessary to drive it from the our gear causes quite a bit of deflection in the 3mm threaded rod.

But, all that said. I just got it to run on the hour gear! Really not sure why it started to behave all of a sudden.

Here it is: http://youtu.be/wUIsHWfpWgw

5lbs of water hanging off the drum.

[ruste...@prototribe.net]

First of all I've got to say seeing that thing in action is just
awesome! I especially like how with each tick each successive gear moves
a little less than the prior! Heck, with the wood case, it even looks
like a clock!

I got distracted by the bar but am hoping to upload some video of the
prototype i've been tinkering with very soon.

I think at least part of the binding problem is related to gear wobble
and subsequently force vectors not being in-line the ideal plane of the
gear which subsequently leads to increased friction against the axle and
gears rubbing against each other and other fixed parts of the clock.

[ruste...@prototribe.net]

Here is a demo video of the prototype of the drive train from escape
wheel to hour drum. The bottle weighs about 750g. The gear at the very
top is minutes, the one at the very bottom will contain an escape wheel.
Drum is attached to the hour gear. Sorry the video is so dark. I need
more light!

http://www.prototribe.net/vidplay/prototypeclock_demo1.html

[Mathieu Glachant]

First off, yay!

Second, if the 3mm rod is deflecting under the weight hanging from the drum, Bre is right and we should relocate it to a separate, beefier shaft, even at the cost of an extra gear.

Alternatively, we move up to fatter shafts and a larger grade of bearings... There's room in the gear hubs from what I can see.

On Tuesday, January 31, 2012, Benjamin Fraser <ma...@benfraser.com> wrote:

[Mathieu Glachant]

Ack, can't see the video on the iPad... Will need to wait until I get to the laptop later this morning. :-)

Congrats anyway! Do you want the bot farm to print one out for this weekend for us to play with?

[Mathieu Glachant]

Wow, smooth as butter! :-)

[Ben Fraser]

Last night, it was running OK off the hour. I could get it to tick for
about 45 sec. or so before it would stop.

I think I figured out how to print the hour gear and drum without the
concentric shaft, and have it run on bearings. I'm going to give that
a try this morning.

Hopefully I'll make a little more progress today (while also getting
some work done) so that I can post a decent version for the Bot Farm
tomorrow morning.

Rustedrobot: That's super nice. I love how compact it is. Seems really
smooth. Integrated bearings/bushings? It would be very sweet to have a
clock that was completely printed.

[Mathieu Glachant]

It may just need a heavier pendulum, so that there is more inertia to carry it thru from tick to tock... Or a lighter one, so that it takes less torque to keep it running. :-)

[Mathieu Glachant]

Ok, made the mods, generating the stil files for the modified parts, which is going to take a while on this machine.

I took out the concentric shafts, including in the ratcheting drum. We'll need another way to work the hands, but hopefully this will make the gear train even smoother.

The modified parts are the hour gear, the drum that sits over it, and the minute gear. Files to follow shortly-ish.

[Benjamin Fraser]

Don't worry about the stl files, I can generate them.

[Mathieu Glachant]

Best way I know to make sure I didn't screw something up in the code... Won't take that long for three parts, one of which doesn't invoke involute_gear... :-)

I had to make small changes to all three files library,script,parameters, to get it to work right. Let's call that a side branch because with this version of the code, you can't just flip a parameter and get the concentric shafts back on the drum. I'll need to clean this up with some true/false flag so you can decide which type you want, and then merge it back into the main dev branch. I'll send the code with the stil files.

[ruste...@prototribe.net]

Yep the prototype includes the integrated bearings/bushings. While it would be cool to not need extra bits and pieces, i'll be happy with a working clock of any kind (it doesn't hurt that I already have ~20 of the tiny ball bearings lying around). =) Once we have that we can work on the refinements.

As far as the concentric shafts, what about a much larger bearing where the inner ring mounts on the inner shaft and the outer ring mounts into the gear of the outer shaft? Something like a skate bearing ought to have a large enough inner diameter for something like that.

Here's a pic of a non-working version of the prototype (next to an old test jig from the original effort).

[Mathieu Glachant]

Here are the files, 3 STLs and one zip containing the modified
OpenSCAD library, script and parameters.

It looks like it should all fit together and work on the assembled
view, but I guess we won't know for sure until it gets printed!

[Mathieu Glachant]

Nested bearings could work, but it will add up quick if we want seconds/minutes/hours/ratchetingdrum... Following Bre's suggestion of off-loading the drum on a separate shaft would help with a good chunk of that, tho.

I have more hope that the nested brass tubes will help. We could use your integrated bushings for the non-handed shaft(s) and design something similar that grips brass tubing very tightly without deforming it, and adapts to small variations of tubing diameter, for the handed shaft(s).

[ruste...@prototribe.net]

Sure. I'm hoping to have it refined enough to release files for it today so people can try it out even before hand.