Little side project. I want to build a vertical stabilizer arm for my
Glidecam camcorder stabilizer. The theory is similar to the Steadicam.
Picture a 10"x3" parallelogram pinned with bearings at the corners. Long
dimension is horizontal. One 3" side is fixed and the other floats and
holds the Glidecam. A spring runs diagonally between the top end of the
fixed side to the bottom of the floating side. The moving parts of the arm
weigh 2 lbs and when level the COG is 7" from the fixed end. The Camcorder
and Glidecam weigh 4.2 lb and the COG is 12.75" from the fixed end.

How do I figure out which spring to use to just maintain the arm slightly
above level? If the spring rate is to high the stabilization effect will be
diminished and if to low it won't recover fast enough. The spring can't be
longer than 6" with no tension and will need about 3.5" of extension to
allow the arm to move from 30 degrees above to 30 degrees below level.
There will also be an adjustment screw to pretension the spring.

--
Glenn Ashmore

I'm building a 45' cutter in strip/composite. Watch my progress (or lack
there of) at: http://www.rutuonline.com
Shameless Commercial Division: http://www.spade-anchor-us.com

Hi Glenn,

I'm sure I could figure out the answer to this problem if it wasn't 2
am! If no one has solved it for you by tomorrow I'll have a go (let me
know if I forget).

Best wishes,

Chris

Glenn Ashmore wrote:
Little side project. I want to build a vertical stabilizer arm for my
Glidecam camcorder stabilizer.
How do I figure out which spring to use to just maintain the arm slightly
above level?

It's not the answer to your question, but it might be interesting for you:

http://www.tiffen.com/dynamic%20primer.pdf

Please put a sketch in the dropbox after you make it. I would like to
add one to my (never completed) project list.

Kevin Gallimore

----== Posted via Newsfeeds.Com - Unlimited-Unrestricted-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =----

On Sun, 2 Apr 2006 20:50:50 -0400, "Glenn Ashmore"
wrote:

Little side project. I want to build a vertical stabilizer arm for my
Glidecam camcorder stabilizer. The theory is similar to the Steadicam.
Picture a 10"x3" parallelogram pinned with bearings at the corners. Long
dimension is horizontal. One 3" side is fixed and the other floats and
holds the Glidecam. A spring runs diagonally between the top end of the
fixed side to the bottom of the floating side. The moving parts of the arm
weigh 2 lbs and when level the COG is 7" from the fixed end. The Camcorder
and Glidecam weigh 4.2 lb and the COG is 12.75" from the fixed end.

How do I figure out which spring to use to just maintain the arm slightly
above level? If the spring rate is to high the stabilization effect will be
diminished and if to low it won't recover fast enough. The spring can't be
longer than 6" with no tension and will need about 3.5" of extension to
allow the arm to move from 30 degrees above to 30 degrees below level.
There will also be an adjustment screw to pretension the spring.

I'll assume that the 6" rest length and 3.5" max extension are due to
unmentioned constraints because that's too short for the geometry. The
spring will need a "dead" extension -- a link or piece of wire 2.29"
long.

The load moment at level is 67.55 lbf*in. With the 10" member
horizontal the diagonal spring is at a 16.7 deg angle from
horizontal. It will therefore need to exert 23.5 lbf to balance at
level. Extension at horizontal is 2.15", so spring constant k is
10.93 lbf/inch. To balance "slightly above" level the spring constant
will need to be "slightly greater".

"axolotl" wrote

It's not the answer to your question, but it might be interesting for you:

http://www.tiffen.com/dynamic%20primer.pdf

Please put a sketch in the dropbox after you make it. I would like to add
one to my (never completed) project list.

I have already found that and used it to rework the Glidecam. The Glidecam
is basically the front end of a Steadicam. With a gimbal and
counterweights it uses balance spreading to give the camcorder some
stability. I addeded an aluminum camera sled and threaded and knurled
counterweights to make it less bulky and aid in balance adjustment. Works
great for roll and yaw but I want to mount it on the rail of a sailboat so I
need to deal with bouncing and vibration. That is what the arm is for.
Just can't swing $20K for a mount for a $1K camcorder. :-)

--
Glenn Ashmore

I'm building a 45' cutter in strip/composite. Watch my progress (or lack
there of) at: http://www.rutuonline.com
Shameless Commercial Division: http://www.spade-anchor-us.com

Thanks! That may give me enough to navigate McMaster's spring selection.

I actually have about 9" of room but I wanted the no load length 6" or less
to allow for extension and the adjustment mechanism.

--
Glenn Ashmore

I'm building a 45' cutter in strip/composite. Watch my progress (or lack
there of) at: http://www.rutuonline.com
Shameless Commercial Division: http://www.spade-anchor-us.com

"Don Foreman" wrote
I'll assume that the 6" rest length and 3.5" max extension are due to
unmentioned constraints because that's too short for the geometry. The
spring will need a "dead" extension -- a link or piece of wire 2.29"
long.

The load moment at level is 67.55 lbf*in. With the 10" member
horizontal the diagonal spring is at a 16.7 deg angle from
horizontal. It will therefore need to exert 23.5 lbf to balance at
level. Extension at horizontal is 2.15", so spring constant k is
10.93 lbf/inch. To balance "slightly above" level the spring constant
will need to be "slightly greater".

Glenn Ashmore wrote:
...
How do I figure out which spring to use to just maintain the arm slightly
above level? ...

There are a couple of other things to consider here, if you haven't
already. One is damping. With bearings at the corners, your mount will
be very free and in constant motion. One good bump and your camera will
be bobbing like one of those silly car rear window statues. You'll need
some form of (adjustable) friction.

Another is accelerations. Given the size of your boat, they will be
pretty small, I'd guess. So you won't have a problem with your mount
bottoming out. But the mount might be too stiff for the small
accelerations that you will have. You could calculate the deflections
that you will have, given the geometry and Don's spring constant, but I
don't know what numbers you'd use for the accelerations.

Hmm - it just occurred to me that the way to think about your mount is
as a high frequency filter. Of the spectrum of frequencies that the
boat is moving at, you want to camera to only move at low frequencies.
The camera/mount frequency is mostly a matter of the spring constant:
high spring constant equals high frequency. A lower spring constant
means more extension, means a longer horizontal arm. The gotcha, of
course, is *how* low the frequency needs to be and how to calculate the
spring constant and geometry from that. I'm sure that there's an ME
reading this that can help.

Don't you just hate it when the problem becomes *much* more complicated
that you thought it was. Sorry about that.

Bob

Not answering your question, but I thought this might interest you and
others on this newsgroup. You might have already seen it. A hand-held
steadicam:

http://www.steadicam.com/handheldmerlin.html

There is a quicktime movie on this link that shows it in operation.

Sam


"Glenn Ashmore" wrote in message
news:Hj_Xf.69191$YX1.63070@dukeread06...
Little side project. I want to build a vertical stabilizer arm for my
Glidecam camcorder stabilizer. The theory is similar to the Steadicam.
Picture a 10"x3" parallelogram pinned with bearings at the corners. Long
dimension is horizontal. One 3" side is fixed and the other floats and
holds the Glidecam. A spring runs diagonally between the top end of the
fixed side to the bottom of the floating side. The moving parts of the
arm weigh 2 lbs and when level the COG is 7" from the fixed end. The
Camcorder and Glidecam weigh 4.2 lb and the COG is 12.75" from the fixed
end.

How do I figure out which spring to use to just maintain the arm slightly
above level? If the spring rate is to high the stabilization effect will
be diminished and if to low it won't recover fast enough. The spring
can't be longer than 6" with no tension and will need about 3.5" of
extension to allow the arm to move from 30 degrees above to 30 degrees
below level. There will also be an adjustment screw to pretension the
spring.

--
Glenn Ashmore

I'm building a 45' cutter in strip/composite. Watch my progress (or lack
there of) at: http://www.rutuonline.com
Shameless Commercial Division: http://www.spade-anchor-us.com

On Mon, 03 Apr 2006 10:26:21 -0400, Bob Engelhardt
wrote:

Glenn Ashmore wrote:
...
How do I figure out which spring to use to just maintain the arm slightly
above level? ...

There are a couple of other things to consider here, if you haven't
already. One is damping. With bearings at the corners, your mount will
be very free and in constant motion. One good bump and your camera will
be bobbing like one of those silly car rear window statues. You'll need
some form of (adjustable) friction.

Another is accelerations. Given the size of your boat, they will be
pretty small, I'd guess. So you won't have a problem with your mount
bottoming out. But the mount might be too stiff for the small
accelerations that you will have. You could calculate the deflections
that you will have, given the geometry and Don's spring constant, but I
don't know what numbers you'd use for the accelerations.

Hmm - it just occurred to me that the way to think about your mount is
as a high frequency filter. Of the spectrum of frequencies that the
boat is moving at, you want to camera to only move at low frequencies.
The camera/mount frequency is mostly a matter of the spring constant:
high spring constant equals high frequency. A lower spring constant
means more extension, means a longer horizontal arm. The gotcha, of
course, is *how* low the frequency needs to be and how to calculate the
spring constant and geometry from that. I'm sure that there's an ME
reading this that can help.

Don't you just hate it when the problem becomes *much* more complicated
that you thought it was. Sorry about that.

Bob

In order to calculate the resonant frequency we'd need to know the
mass moment of inertia, which depends on the distribution of mass as a
fn of radius from the pivot point. Since other constraints have
already determined the spring constant, the simplest way to determine
resonant freq will be just to see what it turns out to be. My hunch
is that with this spring constant and mass the resonant freq will be
several hertz, well above the roll, pitch or yaw rate of a boat.

Addition of some viscous friction, as a dashpot, would damp the
system as you say. If the damping is right, the system will then
become a second-order highpass filter with flat response above the
cutoff freq, 3 dB down (.707 of amplitude) at the resonant freq and
diminishing at 12 dB per octave of frequency below resonance. With
proper damping there will be no "ringing" in response to an impulse or
step disturbance. I say "highpass" because the system attenuates
boat motion conveyed to the camera at low frequencies.

I'm not an ME, but the differential equations describing such a system
are exactly the same in form as those describing an L-R-C electrical
lowpass filter with mass analogous to inductance, 1/k analogous to
capacitance and viscous friction anlogous to resistance. If these
quantities are expressed in MKS units (henrys, farads, ohms, meters,
kilograms, force in newtons, torque in newton-meters, etc) then the
units even turn out right with frequency in radian/sec.

The viscous friction would determine the type of low-pass response.
Butterworth response would probably be most suitable here, having
flat compliance up to cutoff freq with no ringing or overshoot.
Response would be 3dB down (.707 amplitude) at resonant frequency,
then diminishing by 12 dB/octave thereafter. A sutable damper might
be something like a pneumatic screendoor closer without the internal
spring. The damping (viscous friction) can then be adjusted by
adjusting the screw that determines the leak orifice.

A really steady steadycam could be made using a silicon accelerometer
(about $15 from DigiKey), a torquer (small DCPM motor) and an
electronic feedback control system. That camera would be steady
even with the boat doing the watusi. Might be kinda neat....

"Don Foreman" wrote

The load moment at level is 67.55 lbf*in. With the 10" member
horizontal the diagonal spring is at a 16.7 deg angle from
horizontal. It will therefore need to exert 23.5 lbf to balance at
level. Extension at horizontal is 2.15", so spring constant k is
10.93 lbf/inch. To balance "slightly above" level the spring constant
will need to be "slightly greater".

Trying to set this up in a spreadsheet. Is the spring force the
moment/(Sin(A)*arm length)?

--
Glenn Ashmore

I'm building a 45' cutter in strip/composite. Watch my progress (or lack
there of) at: http://www.rutuonline.com
Shameless Commercial Division: http://www.spade-anchor-us.com

On Mon, 3 Apr 2006 14:27:59 -0400, "Glenn Ashmore"
wrote:


"Don Foreman" wrote

The load moment at level is 67.55 lbf*in. With the 10" member
horizontal the diagonal spring is at a 16.7 deg angle from
horizontal. It will therefore need to exert 23.5 lbf to balance at
level. Extension at horizontal is 2.15", so spring constant k is
10.93 lbf/inch. To balance "slightly above" level the spring constant
will need to be "slightly greater".

Trying to set this up in a spreadsheet. Is the spring force the
moment/(Sin(A)*arm length)?

Yup.

On Mon, 03 Apr 2006 10:26:21 -0400, Bob Engelhardt
wrote:

Glenn Ashmore wrote:
...
How do I figure out which spring to use to just maintain the arm slightly
above level? ...

There are a couple of other things to consider here, if you haven't
already. One is damping. With bearings at the corners, your mount will
be very free and in constant motion. One good bump and your camera will
be bobbing like one of those silly car rear window statues. You'll need
some form of (adjustable) friction.

Another is accelerations. Given the size of your boat, they will be
pretty small, I'd guess. So you won't have a problem with your mount
bottoming out. But the mount might be too stiff for the small
accelerations that you will have. You could calculate the deflections
that you will have, given the geometry and Don's spring constant, but I
don't know what numbers you'd use for the accelerations.

Hmm - it just occurred to me that the way to think about your mount is
as a high frequency filter. Of the spectrum of frequencies that the
boat is moving at, you want to camera to only move at low frequencies.
The camera/mount frequency is mostly a matter of the spring constant:
high spring constant equals high frequency. A lower spring constant
means more extension, means a longer horizontal arm. The gotcha, of
course, is *how* low the frequency needs to be and how to calculate the
spring constant and geometry from that. I'm sure that there's an ME
reading this that can help.

Don't you just hate it when the problem becomes *much* more complicated
that you thought it was. Sorry about that.

Bob

I got curious. I figure the resonant frequency of this system with
the recommended spring will be about 2 Hz. This is just a rough
SWAG , not knowing how the mass is distributed as fn of radius.

Don Foreman wrote:
... My hunch
is that with this spring constant and mass the resonant freq will be
several hertz, ...

Oh, that sounds bad. Imagine trying to watch a video that jiggles up
and down at several hertz! Seasickness in your living room - talk about
capturing the experience on tape. I'd think an order of magnitude lower
would be better ("sway" rather than "jiggle"). Of course, the "badness"
would depend upon the amplitude and the distance to the taped object. A
1" amplitude while taping the horizon would be imperceptible, but if
taping a person 6' away, probably pretty bad.

... the system will then
become a second-order highpass filter with flat response above the
cutoff freq, ... I say "highpass" because the system attenuates
boat motion conveyed to the camera at low frequencies. ...

I don't think that you'd want to pass the high frequencies, e.g.,
vibration - that would give a fuzzy picture, wouldn't it?

A really steady steadycam could be made using a silicon accelerometer
(about $15 from DigiKey), a torquer (small DCPM motor) and an
electronic feedback control system. That camera would be steady
even with the boat doing the watusi. Might be kinda neat....

Oh oh - another project for Don 8-)

Bob

On Mon, 03 Apr 2006 19:14:53 -0400, Bob Engelhardt
wrote:

Don Foreman wrote:
... My hunch
is that with this spring constant and mass the resonant freq will be
several hertz, ...

Oh, that sounds bad. Imagine trying to watch a video that jiggles up
and down at several hertz! Seasickness in your living room - talk about
capturing the experience on tape. I'd think an order of magnitude lower
would be better ("sway" rather than "jiggle"). Of course, the "badness"
would depend upon the amplitude and the distance to the taped object. A
1" amplitude while taping the horizon would be imperceptible, but if
taping a person 6' away, probably pretty bad.

... the system will then
become a second-order highpass filter with flat response above the
cutoff freq, ... I say "highpass" because the system attenuates
boat motion conveyed to the camera at low frequencies. ...

I don't think that you'd want to pass the high frequencies, e.g.,
vibration - that would give a fuzzy picture, wouldn't it?

You're right. I don't know what I was thinking. It's a lowpass.

If it's damped properly, the resonance doesn't mean that it'll jiggle
at that frequency. The resonance just determines the corner
frequency. Response to stimulus would be flat below resonance, 3dB
down at resonance, and then go down from there.

It probably won't do much aside from couterbalance it. I doubt if
there is much boat motion component above 1 Hz, unless it's going fast
and hitting small waves.

A really steady steadycam could be made using a silicon accelerometer
(about $15 from DigiKey), a torquer (small DCPM motor) and an
electronic feedback control system. That camera would be steady
even with the boat doing the watusi. Might be kinda neat....

Oh oh - another project for Don 8-)

Uh huh. I happen to have a few sample 2-axis accels in my
goodiebox. Working on something else....don't need yet another
unfinished project.

I wonder if angular motion wouldn't be more objectionable than linear
motion. It sure is with binoculars. Wonder if a little gyro
stabilizer might work better than a spring-mass system.

Thanks to same day delivery from McMaster I cobbled together the first
prototype last night. I quickly figured out that the spring rate and
initial tension are critical. A rate of 10 lb/in is way to fast. It is not
sensitive enough to prevent the camera moving. Had to cut back to a 5.29
rate with a 6.41 initial tension. Preloaded the spring to 17 pounds and the
arm settled at about 5 degrees up. Rapidly moving the fixed side up and
down 7-8" the camera stayed at the same level but when I moved it up and
stopped the camera followed a bit to quickly and overshot. It needs to
approach the equilibrium point more slowly.

I think I need a rate of about 4.5 to 4.8 and ideally with a higher initial
tension but McMaster doesn't have one in that range that will fit in the
available space and I don't think you can increase the initial tension and
lower the rate at the same time.

Many thanks to Don for the formula. With it working things out on a spread
sheet greatly eased the design.

--
Glenn Ashmore

I'm building a 45' cutter in strip/composite. Watch my progress (or lack
there of) at: http://www.rutuonline.com
Shameless Commercial Division: http://www.spade-anchor-us.com

"Glenn Ashmore" wrote in message
news:Hj_Xf.69191$YX1.63070@dukeread06...
Little side project. I want to build a vertical stabilizer arm for my
Glidecam camcorder stabilizer. The theory is similar to the Steadicam.
Picture a 10"x3" parallelogram pinned with bearings at the corners. Long
dimension is horizontal. One 3" side is fixed and the other floats and
holds the Glidecam. A spring runs diagonally between the top end of the
fixed side to the bottom of the floating side. The moving parts of the
arm weigh 2 lbs and when level the COG is 7" from the fixed end. The
Camcorder and Glidecam weigh 4.2 lb and the COG is 12.75" from the fixed
end.

How do I figure out which spring to use to just maintain the arm slightly
above level? If the spring rate is to high the stabilization effect will
be diminished and if to low it won't recover fast enough. The spring
can't be longer than 6" with no tension and will need about 3.5" of
extension to allow the arm to move from 30 degrees above to 30 degrees
below level. There will also be an adjustment screw to pretension the
spring.

--
Glenn Ashmore

I'm building a 45' cutter in strip/composite. Watch my progress (or lack
there of) at: http://www.rutuonline.com
Shameless Commercial Division: http://www.spade-anchor-us.com

On Tue, 4 Apr 2006 10:03:12 -0400, "Glenn Ashmore"
wrote:

Thanks to same day delivery from McMaster I cobbled together the first
prototype last night. I quickly figured out that the spring rate and
initial tension are critical. A rate of 10 lb/in is way to fast. It is not
sensitive enough to prevent the camera moving. Had to cut back to a 5.29
rate with a 6.41 initial tension. Preloaded the spring to 17 pounds and the
arm settled at about 5 degrees up. Rapidly moving the fixed side up and
down 7-8" the camera stayed at the same level but when I moved it up and
stopped the camera followed a bit to quickly and overshot. It needs to
approach the equilibrium point more slowly.

I think I need a rate of about 4.5 to 4.8 and ideally with a higher initial
tension but McMaster doesn't have one in that range that will fit in the
available space and I don't think you can increase the initial tension and
lower the rate at the same time.

Many thanks to Don for the formula. With it working things out on a spread
sheet greatly eased the design.

Resonant frequency is determined only by springrate, regardless of
initial tension. One possibility might be to use a torsion spring,
like a clock spring or the spring from a recoil starter on a small
engine. You can wind in a lot of initial tension without needing a
lot of space.

Another possibility might be to add a "negator" spring. Those are
dished flat strips that wind onto a roller, provide an essentially
constant pull rate regardless of extension. It could provide some of
the bias tension, enabling use of a soft spring to make up the
difference. The springrate of a negator is nearly zero; it's more
like a counterweight but without the mass. That and a soft helper
spring would provide very low resonance frequency.

AxMan Surplus sometimes has those. I'll look when I next visit.
They're about 2 bux if they have 'em.

It sounds like you about have it, though. A little dashpot damper
would cure your overshoot. Just a plastic or metal cylinder with a
leaky piston -- like a screendoor closer without the spring. You can
also make a torsional viscous damper with a disc in a cavity filled
with grease. Somebody, perhaps Airpot, used to make little glass
cylinder dampers with graphite pistons. Very smooth, no stiction,
last forever. Yeah, it *is* Airpot!
http://www.airpot.com/

What a viscous damper does is offer resistance that is proportional to
velocity. That quells overshoot. The reason your car doesn't
continue to bounce after hitting a bump, even though it is a spring
mass system with a resonant frequency, is viscous dampers AKA "shock
absorbers" though they are exactly the opposite. They transmit abrupt
shock but offer very little resistance to slow motion. In the UK
they are called "dampers".