On Fri, 05 Mar 2010 13:18:32 -0800, Tim Wescott
wrote:



You wouldn't say that if you read the source code for a few products
with embedded processors. Embedded software is the ultimate Rube
Goldberg artifact, it's just well hidden.

You shred mechanical drawings so your competitors can't copy your product.

You shred electrical drawings so your competitors can't get an idea of
how you think.

You shred source code so your competitors won't laugh at you.

A corollalary is that you conceal test data to try to avert predation
by lawyers and politicans.

On Thu, 4 Mar 2010 14:31:33 -0500, "Buerste"
wrote:

Damn inertia!

I have an air cylinder, 1.5" x 4" stroke, that pushes a crank on a shaft to
rotate the shaft about 90 deg. On the shaft is a gear with a one-way
clutch. The 4" x 1/2" gear turns about 3" at the perimeter per stroke of
the cylinder. By moving the link to the air cylinder up or down on the
crank the gear will rotate more or less per stroke. A second gear is meshed
into the first gear with pressure from another air cylinder. Wire is fed
between the gears in a grove cut in the face of each gear. The idea is to
feed wire 3" per stroke.

Is works perfectly and fast! By varying the pressure in the cylinder that
presses the gears together, the wire is gripped very well at 60 lbs air
pressure and will slip through the gears easily at no pressure.

The only problem I have is that the gears over-run and feed too much wire.
I've considered lightening the gears by drilling a bunch of holes. Also, a
drag brake on one or both gears with screws pressing brass pucks against it.
Another idea is an air limit switch that is struck at the end of the stroke
by the crank. This switch would unload the pressure on the clamping
cylinder and let the wire slip between the gears when they over-run.

Would the air limit switch work fast enough to unclamp the wire in time?
Would I be better off with an electric switch and valve? I might have to
lighten the gears, brake them and switch off the air clamp The feed has to
take place in 300ms.

I would consider a low-inertia servo motor driving wire-drive
capstans. Recall the action on reel-to-reel tape digital data storage
years ago? Those fairly large reels started and stopped very quickly.

In your case, the diameter and mass of the drive capstans (gears if
you like) should be somewhat matched to the servomotors or vicey
versey. A servomotor is essentially a current-to-torque transducer
but you must factor in its own moment of inertia into torque
calculations. That torque can be braking torque as well as
accelerating torque.

If the wire comes from a large spool with significant moment of
inertia, then a secondary mechanism would be in order that feeds a
"wire buffer" like on a coil winding machine so that all the feed
mechanism need do is accelerate and decelerate a few feet of wire and
overcome the mild resistance of the springs or weights in the buffer.
The spool would have its own servo drive that would run at fairly
constant speed with the buffer handling the reciprocating difference
in slack. If you haven't seen a coil winding machine, visit a company
that makes transformers.

This is all moot if you're determined to Rube it with a completely
mechanical solution.

On Fri, 05 Mar 2010 23:40:26 -0600, the infamous Don Foreman
scrawled the following:

On Fri, 05 Mar 2010 13:18:32 -0800, Tim Wescott
wrote:



You wouldn't say that if you read the source code for a few products
with embedded processors. Embedded software is the ultimate Rube
Goldberg artifact, it's just well hidden.

You shred mechanical drawings so your competitors can't copy your product.

You shred electrical drawings so your competitors can't get an idea of
how you think.

You shred source code so your competitors won't laugh at you.

A corollalary is that you conceal test data to try to avert predation
by lawyers and politicans.

That brings a thought to mind: Has anyone sued the AGWK "pushers" for
pushing bad data on the world, costing it needless billions of
dollars? I haven't read of anyone trying to recoup their costs yet,
but stay tuned!

--
The blind are not good trailblazers.

-- federal judge Frank Easterbrook

On Fri, 5 Mar 2010 05:45:29 -0500, "Buerste"
wrote:

Hmmm, the wire needs to be in place and in control to be cut in the 300ms
window, that's all. This is the third wire feeder design in ten years.
We've had to use different wire manufacturers over the years as some plants
have closed and the qualities of the wire change...this or that feed doesn't
work on this or that wire too well...etc. This feeder is trying to address
the wire that has too much oil, too much scale, lead drag, ect. The wire
manufacturer can't fine-tune their quality any better and they are the only
domestic supplier. We also use a Taiwanese supplier but their QC and
lead-time is worse. Those are the only suppliers in the entire world! And,
there are only 4 customers in the world and we are by far the largest user.

If the problem is too much oil and crap on the wire.... Solve it!
Don't raise the drawbridge, lower the river.

Howzabout putting a cleaning station on the unspooling device?

Give the machine the despooled wire after you run it over a series
of counter-rotating cleaning rollers (short-nap paint roller covers,
cheap in bulk and easily replaceable) that are charged (and flushed)
with a small pump and a recycled mild solvent.

Think a wire straightener rig with the rollers staggered over-under,
and a gear drive to spin the rollers at maybe 5 to 10 RPM. With a
catch-pan under the whole thing.

The last few rollers can have a stripper wiper or a spring loaded
rubber brayer roller pressing on them to extract the solvent from the
roller cover, then the paint rollers absorb the excess solvent from
the wire as it goes over them.

You'll need an oscillator motor and eccentric to move the wire route
sideways through the rollers during the day to even out the wear.

(You're welcome. ;-P Run with it and see if you can get anywhere.
If it works, I'll give you an address for a little "Care Package".)

-- Bruce --

Don Foreman wrote:

A corollalary is that you conceal test data to try to avert predation
by lawyers and politicans.

And a right decent income is made by a fair number of folks doing
independent testing to develop unbiased data. Unfortunately this usually
happens after someone's died or been injured. A good friend worked for
Failure Analysis for several years. Much of his work is still covered by
NDA, but there's some cases he can talk about. He's got some interesting
stories...


Jon

Don Foreman wrote:

If the wire comes from a large spool with significant moment of
inertia, then a secondary mechanism would be in order that feeds a
"wire buffer" like on a coil winding machine so that all the feed
mechanism need do is accelerate and decelerate a few feet of wire and
overcome the mild resistance of the springs or weights in the buffer.
The spool would have its own servo drive that would run at fairly
constant speed with the buffer handling the reciprocating difference
in slack. If you haven't seen a coil winding machine, visit a company
that makes transformers.

Regardless of approach, this would be a good idea. I've seen very few
automated machines feeding any material from coil or roll that doesn't
utilize this principle in some fashion or other.

Jon

On Fri, 5 Mar 2010 14:11:32 -0500, "Buerste"
wrote:



Steppers don't have the speed or oomph and I'm a "MECHANICAL" engineer! Get
it...mechanical??? Anybody can just stick in electronics but it takes a
mechanical guy to do a "Rube Goldberg".


In my opinion, if this thing is always feeding the same length of
wire, a servo is the Rube Goldberg option. 300ms is an eternity to
feed and cut a short length of wire for a cam operated device, as long
as reasonable care is taken specifying the cam profiles. I've built
several wire feed, form and place machines over the past 25 years,
including one that operated at 1500 pieces/minute, though that one was
a continuous motion device and was servo driven to allow for quick
product changes.

--
Ned Simmons

On 2010-03-06, Don Foreman wrote:
On Thu, 4 Mar 2010 14:31:33 -0500, "Buerste"
wrote:

Damn inertia!

[ ... ]

I would consider a low-inertia servo motor driving wire-drive
capstans. Recall the action on reel-to-reel tape digital data storage
years ago? Those fairly large reels started and stopped very quickly.

But -- not quickly enough. The tape typically ran through a
vacuum column one each side to provide a very low inertia, buffer. The
The capstan can be pretty quick, but there is a lot of rotational
inertia in a full tape reel.

Sometimes, the capstan pinch roller would be engaged/disengaged
to start and top the tape, and the servos on the reels would be used to
keep the proper amount of tape in the buffers.

Of course, there were densities of up to 5600 BPI, so the tape
motion needed to be controlled to about 0.0002" to stop on a bit,
(normally, the stops were in an IRG (Inter Record Gap) whose length
would have to be determined by the worst case drive. :-)

For the slower ones -- the buffer would be a cats-cradle of tape
spooled between the fingers of a "dancing arm".

However, given the likely stiffness of the flat wire being fed,
I somehow doubt whether the air column would do anything useful, and
even the dancing arms would have to have pretty widely spaced rollers to
avoid too sharp a bend.


In your case, the diameter and mass of the drive capstans (gears if
you like) should be somewhat matched to the servomotors or vicey
versey. A servomotor is essentially a current-to-torque transducer
but you must factor in its own moment of inertia into torque
calculations. That torque can be braking torque as well as
accelerating torque.

Yes -- how much precision is needed in the feed again? Were we
told, or did I just forget it?

If the wire comes from a large spool with significant moment of
inertia, then a secondary mechanism would be in order that feeds a
"wire buffer" like on a coil winding machine so that all the feed
mechanism need do is accelerate and decelerate a few feet of wire and
overcome the mild resistance of the springs or weights in the buffer.

The "dancing arms" -- which with the wire in question, would
have to be rather large and at the same time rather low inertia
themselves.

The spool would have its own servo drive that would run at fairly
constant speed with the buffer handling the reciprocating difference
in slack. If you haven't seen a coil winding machine, visit a company
that makes transformers.

I'm sure that he has -- or other "dancing arm" mechanisms.

This is all moot if you're determined to Rube it with a completely
mechanical solution.

:-)

Enjoy,
DoN.

--
Email: | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---

On Thu, 4 Mar 2010 14:31:33 -0500, "Buerste"
wrote:
snip
The only problem I have is that the gears over-run and feed too much wire.
I've considered lightening the gears by drilling a bunch of holes.
snip
Lots of good observations and suggestions so far on what appears
to be a very interesting project.

Do you really need intermittent feed or is some sort of
continuous wire feed with a "flying cutoff" possible, with the
length adjustable when the extruded/extended wire trips a
proximity or optical switch.

The start/stop inertia of the "flying cutoff" shouldn't be a
problem as long as it is consistent and the cutoff trips/cuts at
the right time, which could also be controlled by an adjustable
switch/trigger.

Would an electric solenoid give faster response than an air
cylinder to operate the cut-off? How about a double solenoid to
give a push-pull action to avoid problems with spring return.
How about a linear motor or push-pull solenoids [no springs] to
cycle the flying cutoff back and forth?


Unka George (George McDuffee)
...............................
The past is a foreign country;
they do things differently there.
L. P. Hartley (1895-1972), British author.
The Go-Between, Prologue (1953).

On 2010-03-06, F George McDuffee wrote:
On Thu, 4 Mar 2010 14:31:33 -0500, "Buerste"
wrote:
snip
The only problem I have is that the gears over-run and feed too much wire.
I've considered lightening the gears by drilling a bunch of holes.
snip
Lots of good observations and suggestions so far on what appears
to be a very interesting project.

Do you really need intermittent feed or is some sort of
continuous wire feed with a "flying cutoff" possible, with the
length adjustable when the extruded/extended wire trips a
proximity or optical switch.

Hmm ... I've thought of another possible approach.

Are you familiar with "Geneva gears"? Input is a constant
rotational speed. Output is a precise 1/4 turn (or 1/5th turn or 1/6th
turn depending on the particular gear set involved). There should be no
overshoot on the output of this. Rotate the input gear with a constant
velocity motor, trigger cutoff just a little after the output rotational
position locks, and repeat until out of wire.

Here is an example of a Genevia Gear:

http://www.tangendrives.com/

Never dealt with them -- and there are other makers of Genevia
gears, they were just the first web search hit which looked fairly
reasonable.

The Genevia gears even have the advantage that the output
rotation accelerates for about half of its revolution, and decelerates
for the other half.

Enjoy,
DoN.

--
Email: | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---

On 7 Mar 2010 05:28:21 GMT, "DoN. Nichols"
wrote:
snip
Hmm ... I've thought of another possible approach.

Are you familiar with "Geneva gears"? Input is a constant
rotational speed. Output is a precise 1/4 turn (or 1/5th turn or 1/6th
turn depending on the particular gear set involved). There should be no
overshoot on the output of this. Rotate the input gear with a constant
velocity motor, trigger cutoff just a little after the output rotational
position locks, and repeat until out of wire.

Here is an example of a Genevia Gear:

http://www.tangendrives.com/

Never dealt with them -- and there are other makers of Genevia
gears, they were just the first web search hit which looked fairly
reasonable.

The Genevia gears even have the advantage that the output
rotation accelerates for about half of its revolution, and decelerates
for the other half.
=========
These work great at low speeds, but have high rotational inertia
due to lots of moving parts. Max rpm is about 300 or so. With a
4 slot cross this would be about 1200 cuts per minute max.

Another thought. Have you thought about buying the wires cut to
length?

For people that don't know what we are talking about see
http://en.wikipedia.org/wiki/Geneva_drive

Off the shelf Geneva units look like a spicy meatball...
http://www.thomasnet.com/products/un...4411308-1.html
http://www.wmberg.com/catalog/produc...pp/default.asp
http://www.thomasnet.com/heading.htm...9&cid=10056509

also see
http://www.accrainc.com/ctl.html


Unka George (George McDuffee)
...............................
The past is a foreign country;
they do things differently there.
L. P. Hartley (1895-1972), British author.
The Go-Between, Prologue (1953).

On 6 Mar 2010 22:03:04 GMT, "DoN. Nichols"
wrote:

On 2010-03-06, Don Foreman wrote:
On Thu, 4 Mar 2010 14:31:33 -0500, "Buerste"
wrote:

Damn inertia!

[ ... ]

I would consider a low-inertia servo motor driving wire-drive
capstans. Recall the action on reel-to-reel tape digital data storage
years ago? Those fairly large reels started and stopped very quickly.

But -- not quickly enough. The tape typically ran through a
vacuum column one each side to provide a very low inertia, buffer. The
The capstan can be pretty quick, but there is a lot of rotational
inertia in a full tape reel.

I used the example to keep my post simple, but I really did do the
math here. I guessed at 4" dia capstan with a mass of 1 lb,
accelerating and decelerating to have peripheral travel of 3" in 0.25
seconds starting and stopping at zero velocity. Peak torque was
trivial compared to the capability of even a small servomotor driving
direct -- no gears or anything. Torque required was 0.056
Newton-meters, trivial for even a very small coreless servomotor.
Accurate design calculation? Of course not, it's a quick sanity check
that clearly shows feasibility.

F. George McDuffee wrote:
These [Geneva gears] work great at low speeds, ... Max rpm is about 300 or so. With a
4 slot cross this would be about 1200 cuts per minute max.
....

1200/min = 20/sec, well beyond Buerste's 300ms speed. Bob

"DoN. Nichols" wrote in message
...
On 2010-03-06, F George McDuffee
wrote:
On Thu, 4 Mar 2010 14:31:33 -0500, "Buerste"
wrote:
snip
The only problem I have is that the gears over-run and feed too much
wire.
I've considered lightening the gears by drilling a bunch of holes.
snip
Lots of good observations and suggestions so far on what appears
to be a very interesting project.

Do you really need intermittent feed or is some sort of
continuous wire feed with a "flying cutoff" possible, with the
length adjustable when the extruded/extended wire trips a
proximity or optical switch.

Hmm ... I've thought of another possible approach.

Are you familiar with "Geneva gears"? Input is a constant
rotational speed. Output is a precise 1/4 turn (or 1/5th turn or 1/6th
turn depending on the particular gear set involved). There should be no
overshoot on the output of this. Rotate the input gear with a constant
velocity motor, trigger cutoff just a little after the output rotational
position locks, and repeat until out of wire.

Here is an example of a Genevia Gear:

http://www.tangendrives.com/

Never dealt with them -- and there are other makers of Genevia
gears, they were just the first web search hit which looked fairly
reasonable.

The Genevia gears even have the advantage that the output
rotation accelerates for about half of its revolution, and decelerates
for the other half.

Enjoy,
DoN.

--
Email: | Voice (all times): (703) 938-4564
(too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html
--- Black Holes are where God is dividing by zero ---

Was invented by a machinist at National Cash Register in Dayton, OH. Do not
know how it came to be called a Geneva gear. Trivia mode off.

Maybe retain these guys as design consultants?
http://www.youtube.com:80/watch?v=qybUFnY7Y8w

On 3/7/2010 6:17 PM, Bill McKee wrote:
(...)

Do not know how it came to be called a Geneva gear.

By convention, one assumes.

--Winston

On Mon, 08 Mar 2010 10:09:41 -0800, Winston
wrote:

On 3/7/2010 6:17 PM, Bill McKee wrote:
(...)

Do not know how it came to be called a Geneva gear.

By convention, one assumes.

--Winston
============
http://kmoddl.library.cornell.edu/model.php?m=481
snip
Geneva movement is so called because of its use in Geneva watches
as a stop wind.
snip
[Editor’s Note: This is a classic Geneva mechanism used in Swiss
watches in the late 19th century. It is also of a general class
of intermittent ratchet kinematic devices that change continuous
rotary motion into digital motion. Such devices are often used in
counting mechanisms. Similar Models in the Reuleaux-Voigt catalog
include N-8 and N-9. FCM]

Francis Moon 2005-00-00
Reference :

* Brown : Five Hundred and Seven Mechanical Movements (p. 54,
55, 1871)
* Reuleaux, Kennedy : Kinematics of Machinery (p. 564, 1876)
* Clark, Downward : Mechanical Models (1930)
snip


Unka George (George McDuffee)
...............................
The past is a foreign country;
they do things differently there.
L. P. Hartley (1895-1972), British author.
The Go-Between, Prologue (1953).

"F. George McDuffee" wrote in message
...
On Mon, 08 Mar 2010 10:09:41 -0800, Winston
wrote:

On 3/7/2010 6:17 PM, Bill McKee wrote:
(...)

Do not know how it came to be called a Geneva gear.

By convention, one assumes.

--Winston
============
http://kmoddl.library.cornell.edu/model.php?m=481
snip
Geneva movement is so called because of its use in Geneva watches
as a stop wind.
snip
[Editor's Note: This is a classic Geneva mechanism used in Swiss
watches in the late 19th century. It is also of a general class
of intermittent ratchet kinematic devices that change continuous
rotary motion into digital motion. Such devices are often used in
counting mechanisms. Similar Models in the Reuleaux-Voigt catalog
include N-8 and N-9. FCM]

Francis Moon 2005-00-00
Reference :

* Brown : Five Hundred and Seven Mechanical Movements (p. 54,
55, 1871)
* Reuleaux, Kennedy : Kinematics of Machinery (p. 564, 1876)
* Clark, Downward : Mechanical Models (1930)
snip


Unka George (George McDuffee)
..............................
The past is a foreign country;
they do things differently there.
L. P. Hartley (1895-1972), British author.
The Go-Between, Prologue (1953).

Here's a nice animation of a Geneva escapement:

http://www.brockeng.com/mechanism/Geneva.htm

It sounds like Tawwwm needs some kind of escapement for this job. There are
many types, including the Geneva mechanism, but I've never encountered one
that's adjustable. If you want to start and stop some fixed amount of
rotation, however, there are many of them.

If he'd Google around on "escapements," he'd probably find something
suitable.

--
Ed Huntress

Winston wrote:
On 3/7/2010 6:17 PM, Bill McKee wrote:
(...)

Do not know how it came to be called a Geneva gear.
A 4-step Geneva mechanism has a part that looks like a "Swiss cross",
the name must come from that. Of course, there are variants of the
mechanism that have a different number of steps per full rev of the
output shaft.

Jon

Don Foreman wrote:

I used the example to keep my post simple, but I really did do the
math here. I guessed at 4" dia capstan with a mass of 1 lb,
accelerating and decelerating to have peripheral travel of 3" in 0.25
seconds starting and stopping at zero velocity. Peak torque was
trivial compared to the capability of even a small servomotor driving
direct -- no gears or anything. Torque required was 0.056
Newton-meters, trivial for even a very small coreless servomotor.
Accurate design calculation? Of course not, it's a quick sanity check
that clearly shows feasibility.
Well, those vacuum-column tape drives had capstans weighing around 10
grams, and accelerating at insane rates, so the gaps in recorded info on
the tape could be kept short. To maintain a 0.6" gap, you had 0.3" to
start, and 0.3" to stop. The acceleration had to be performed in a
sub-millisecond interval. I've got some of the motors used. One is a
Yaskawa minertia motor with an ironless rotor and ceramic shaft, the
capstan was made of magnesium and fiber composite materials. Typical
capstans are about 1.5" diameter.

Jon

Bob Engelhardt wrote:
F. George McDuffee wrote:
These [Geneva gears] work great at low speeds, ... Max rpm is about
300 or so. With a
4 slot cross this would be about 1200 cuts per minute max.
...

1200/min = 20/sec, well beyond Buerste's 300ms speed. Bob
The Geneva mechanisms in 35 mm movie projectors run at 24 frames/second.
I think they use 4 steps/rev, so the output shaft is moving at 360
revs/minute. The mechanism is advancing 1440 times/minute.
They do use some VERY light parts in these, or they'd hammer themselves
to pieces.

Jon

On Mon, 08 Mar 2010 17:41:06 -0700, "Paul Hovnanian P.E."
wrote:

Don Foreman wrote:

[snip]

If the wire comes from a large spool with significant moment of
inertia, then a secondary mechanism would be in order that feeds a
"wire buffer" like on a coil winding machine so that all the feed
mechanism need do is accelerate and decelerate a few feet of wire and
overcome the mild resistance of the springs or weights in the buffer.
The spool would have its own servo drive that would run at fairly
constant speed with the buffer handling the reciprocating difference
in slack. If you haven't seen a coil winding machine, visit a company
that makes transformers.

Transformers are all made in China now. The coil winding machines you
refer to are 10 year olds.

I saw one in operation in an engineering prototype lab in Minnapolis
quite recently, though it very probably was more than 10 years old.
Yes, most production has gone offshore. Not all in China BTW.

On Mon, 08 Mar 2010 13:08:45 -0600, Jon Elson
wrote:

Don Foreman wrote:

I used the example to keep my post simple, but I really did do the
math here. I guessed at 4" dia capstan with a mass of 1 lb,
accelerating and decelerating to have peripheral travel of 3" in 0.25
seconds starting and stopping at zero velocity. Peak torque was
trivial compared to the capability of even a small servomotor driving
direct -- no gears or anything. Torque required was 0.056
Newton-meters, trivial for even a very small coreless servomotor.
Accurate design calculation? Of course not, it's a quick sanity check
that clearly shows feasibility.
Well, those vacuum-column tape drives had capstans weighing around 10
grams, and accelerating at insane rates, so the gaps in recorded info on
the tape could be kept short. To maintain a 0.6" gap, you had 0.3" to
start, and 0.3" to stop. The acceleration had to be performed in a
sub-millisecond interval. I've got some of the motors used. One is a
Yaskawa minertia motor with an ironless rotor and ceramic shaft, the
capstan was made of magnesium and fiber composite materials. Typical
capstans are about 1.5" diameter.

Jon

I'm starting to regret having mentioned tape.

I'm looking at a servomotor with stall (all day) torque of 0.734 N-m,
peak torque of 5.5 N-m. The whole damned motor weighs about 2.4 kg,
the rotor mass is surely less than half of that. Motor OD is about
80mm or 3.125". Do you think this motor could accomplish Tom's task?
I think it could do it easily and for a very long time.

Depends on the transformer. There is a company 30 miles from here
and they have a sister site in Mexico. These are power line xformers.

Martin

Paul Hovnanian P.E. wrote:
Don Foreman wrote:
[snip]

If the wire comes from a large spool with significant moment of
inertia, then a secondary mechanism would be in order that feeds a
"wire buffer" like on a coil winding machine so that all the feed
mechanism need do is accelerate and decelerate a few feet of wire and
overcome the mild resistance of the springs or weights in the buffer.
The spool would have its own servo drive that would run at fairly
constant speed with the buffer handling the reciprocating difference
in slack. If you haven't seen a coil winding machine, visit a company
that makes transformers.

Transformers are all made in China now. The coil winding machines you
refer to are 10 year olds.