yet another mechanically scanned display

[ER]

I would like to solicit help in reverse engineering a mechanically
scanned LED display that I've seen in specialty stores such as the
Discovery Channel store.

As far as I can tell, the unit operates as follows:

1. A column of LEDs sits on top of an arm made out of a springy metal
(dimensions similar to a hacksaw blade). The base of the arm is fixed so
the top of the arm with the LEDs can sway back and forth.

2. Just below the column of LEDs on the arm is a bulk of metal which is
probably a magnet.

3. The arm is positioned between two parallel electromagnet bars which
are about 6-inches in length. The bars are at the same height as the
bulk of metal on the arm.

4. There is a photo-interruptor positioned so it gets triggered by the
moving arm, and the pulsing of the LEDs is controlled by a microcontroller.

I imagine the thing works by energizing the bar electromagnets in one
direction and then in the opposite direction to create an oscillating
magnetic field. The field pushes and pulls on the arm's magnet making it
swing back and forth. The springiness of the arm gives the arm a quick
stop at the limit of its swing as well as helps to propel it on its
return swing. I estimate the frequency of oscillation to be in the low
tens of Hertz (10 - 20 Hz).

One thing I eespcially like about this mechanically scanned display is
that there are no gears to jam up if you forcibly stop the arm. Indeed,
it is safe to stop the moving arm with your finger.

I'm just a computer guy so the microcontroller stuff isn't a problem.
The questions I have relate to the design of the electromagnetic
subsystem. Specifically,


1. Does my theory of operation have a chance of working? After all, the
magnetic field along the length of a bar magnet is much weaker than at
the ends. Is it possible that the two bars actually form a U-shaped core?


2. What kind of parameters should I be considering for the design of the
electromagnets? If possible, I'd like to use something like a 12V/1Amp
power supply, although my prototyping suggests that this won't be
powerful enough. What about using an AC-powered (120V/60Hz)
electromagnet? Has the potential of killing two birds with one stone
since the oscilliation would be taken care of by the power company.
(Yes, I know it also has the potential for killing myself, although
there are AC electromagnets in many household appliances, so I know they
can be safely made.)


3. Are there any other considerations I should be aware of when
generating these fluctuating magnetic fields (such as telling friends
who have watches with expensive Swiss movements to stand back.)

All help will be appreciated,

 

[Spehro Pefhany]

I would like to solicit help in reverse engineering a mechanically
scanned LED display that I've seen in specialty stores such as the
Discovery Channel store.

As far as I can tell, the unit operates as follows:

1. A column of LEDs sits on top of an arm made out of a springy metal
(dimensions similar to a hacksaw blade). The base of the arm is fixed so
the top of the arm with the LEDs can sway back and forth.

2. Just below the column of LEDs on the arm is a bulk of metal which is
probably a magnet.

3. The arm is positioned between two parallel electromagnet bars which
are about 6-inches in length. The bars are at the same height as the
bulk of metal on the arm.

4. There is a photo-interruptor positioned so it gets triggered by the
moving arm, and the pulsing of the LEDs is controlled by a microcontroller.

I imagine the thing works by energizing the bar electromagnets in one
direction and then in the opposite direction to create an oscillating
magnetic field. The field pushes and pulls on the arm's magnet making it
swing back and forth. The springiness of the arm gives the arm a quick
stop at the limit of its swing as well as helps to propel it on its
return swing. I estimate the frequency of oscillation to be in the low
tens of Hertz (10 - 20 Hz).

It would be operating at a (mechanically) resonant frequency. The
magnets are almost certainly not strong enough to yank that springy
bit over, but they *are* strong enough to feed in the little bit of
energy that's lost to air resistance and heating of the spring.

One thing I eespcially like about this mechanically scanned display is
that there are no gears to jam up if you forcibly stop the arm. Indeed,
it is safe to stop the moving arm with your finger.

You probably have to start it that way too.

I'm just a computer guy so the microcontroller stuff isn't a problem.
The questions I have relate to the design of the electromagnetic
subsystem. Specifically,


1. Does my theory of operation have a chance of working? After all, the
magnetic field along the length of a bar magnet is much weaker than at
the ends. Is it possible that the two bars actually form a U-shaped core?

Most likely. Maybe a permanent magnet at the end of the arm and a
U-shaped core, although it could be done with ferromagnetism. If it's
ferromagnetic, you would just have to energize the coil for a bit as
it approaches each pole of the core. Polarity could be the same for
each end of travel.

2. What kind of parameters should I be considering for the design of the
electromagnets? If possible, I'd like to use something like a 12V/1Amp
power supply, although my prototyping suggests that this won't be
powerful enough. What about using an AC-powered (120V/60Hz)
electromagnet? Has the potential of killing two birds with one stone
since the oscilliation would be taken care of by the power company.
(Yes, I know it also has the potential for killing myself, although
there are AC electromagnets in many household appliances, so I know they
can be safely made.)


3. Are there any other considerations I should be aware of when
generating these fluctuating magnetic fields (such as telling friends
who have watches with expensive Swiss movements to stand back.)

All help will be appreciated,

Best regards,
Spehro Pefhany

 

[Mathew Orman]

But the OP did not ask about anything like that did he?

Chuck
--
... The times have been,
That, when the brains were out,
the man would die. ... Macbeth
Chuck Simmons [email protected]

If I am inventing or designing something I would like to know
if someone did it before.
And I am shore that OP would like to know that too!

Sincerely,

Mathew Orman
www.ultra-faster-than-light.com
www.radio-faster-than-light.com

 

[Ben Bradley]

It would be operating at a (mechanically) resonant frequency. The
magnets are almost certainly not strong enough to yank that springy
bit over, but they *are* strong enough to feed in the little bit of
energy that's lost to air resistance and heating of the spring.


You probably have to start it that way too.

I wouldn't think so. Pulses into the electromagnet at the
approximate resonant frequency should be enough to get it oscillating
with enough amplitude to trigger the optointerruptor.

Since, as Spehro points out, this is surely a mechanically resonant
system, it really only needs a short pulse in one direction from the
electromagnet. This will pull the thing offcenter, and it will then
swing back the other way, and then back again, in a damped sinewave.
Just giving pulses at the right frequency will get it swinging back
and forth.

Most likely. Maybe a permanent magnet at the end of the arm and a
U-shaped core, although it could be done with ferromagnetism. If it's
ferromagnetic, you would just have to energize the coil for a bit as
it approaches each pole of the core. Polarity could be the same for
each end of travel.

With a resonant system, this should be more than powerful enough.
You're getting this thing to swing much like a pendulum in a pendulum
clock. The mechanical components (the thin piece of metal as the
spring, the mass of the arm with LED's) move the mechanism back and
forth, while the electromagnet only has to supply enough energy to
overcome frictional losses.
This mechanism works much like the electromechanical
battery-powered clocks from (I think) the '60's. Maybe someone else
remembers these and can give a description.

Don't use AC. Use 12VDC and switch it on and off with the
microcontroller. You don't need to wind it yourself either, you can
just use the coil out of a relay.

If you put it next to or ontop of a traditional CRT-based TV or
computer monitor, it may interfere with the picture on the screen, but
other than that it's no worse a source of fluctuating magnetic fields
than an operating CRT. It's probably a lot less.

 

[ER]

With a resonant system, this should be more than powerful enough.

Here's the issue I'm running into... 40 ft. of #28 gauge magnet wire has
a resistence of 20 ohms, so a 12V supply will give me .6 Amps. According
to my preliminary testing, this isn't powerful enough. I haven't yet
tried tweaking the mechanical side -- maybe I'll try that next. As for
making the electromagnets stronger, besides increasing the voltage, the
only other way I can think of is to wrap another coil around the first
so I am drawing double the current. Are there any pitfalls to this
approach? Or, would it be better to re-wrap the coil with th two strands
next to each other? In this case the turns/inch locally is halved, but
overall it is doubled since I'm going back and forth over the same
6-inch bar with two strands instead of one. Or maybe that's not the way
the math works...???

You're getting this thing to swing much like a pendulum in a pendulum
clock.

Exactly, and you don't need to start it by hand. When it's free to move
at rest, the arm quickly attains its maximum amplitude within a few seconds.

Don't use AC. Use 12VDC and switch it on and off with the
microcontroller. You don't need to wind it yourself either, you can
just use the coil out of a relay.

Okay, I'll hold off the AC.

 

[Frank Bemelman]

Here's the issue I'm running into... 40 ft. of #28 gauge magnet wire has
a resistence of 20 ohms, so a 12V supply will give me .6 Amps. According
to my preliminary testing, this isn't powerful enough. I haven't yet
tried tweaking the mechanical side -- maybe I'll try that next. As for
making the electromagnets stronger, besides increasing the voltage, the
only other way I can think of is to wrap another coil around the first
so I am drawing double the current. Are there any pitfalls to this
approach? Or, would it be better to re-wrap the coil with th two strands
next to each other? In this case the turns/inch locally is halved, but
overall it is doubled since I'm going back and forth over the same
6-inch bar with two strands instead of one. Or maybe that's not the way
the math works...???


Exactly, and you don't need to start it by hand. When it's free to move
at rest, the arm quickly attains its maximum amplitude within a few

seconds.

[snip]

FWIW, a while ago I bought such a clock, but wanted to change the fixed
messages. It runs on a PIC16F674, and when I had it apart, I did some
measurements as to find out what pins did what. I noticed that during
start, the coil is pulsed at a varying frequency, where it tries to
find out the resonance of the mechanism. The inital movement is rather
small, then increases until the arm passes the sensor, and from then
on the frequency seems fixed. The PIC code was protected, had no strong
desire to write it all from scratch, so I reasembled it and gave it to
my brother, as a present.

It is a bit marginal designed, for instance, if you place it on a
soft surface, like carpet, it's not able to 'tune' itself into
succesful swinging.

 

[Joseph.D.Warner]

I would like to solicit help in reverse engineering a mechanically
scanned LED display that I've seen in specialty stores such as the
Discovery Channel store.

As far as I can tell, the unit operates as follows:

1. A column of LEDs sits on top of an arm made out of a springy metal
(dimensions similar to a hacksaw blade). The base of the arm is fixed so
the top of the arm with the LEDs can sway back and forth.

2. Just below the column of LEDs on the arm is a bulk of metal which is
probably a magnet.

3. The arm is positioned between two parallel electromagnet bars which
are about 6-inches in length. The bars are at the same height as the
bulk of metal on the arm.

4. There is a photo-interruptor positioned so it gets triggered by the
moving arm, and the pulsing of the LEDs is controlled by a microcontroller.

Doesn't seem like you need to build an AC magnet operating at 110 V.
Seems like it would be best to go to 24 V or 12 V and build a
programmable pulse generator. That way you can control the stroke
length, acceleration, and speed by varying the shape of the pulse.
You'll need some magnetic field sensors like a Hall device along the
path of the piston for feedback to your central processor. You can
control the vibrations by using multiple coils and programming the
controller appropiately.

This is very similar to the controls for a pulse tube cooler.
Using FPGs and switching power supplies the controller shouldn't weight
more than a pound and might fit all on a medium size pc board.

 

[Wade Hassler]

You probably have to start it that way too.

Best regards,
Spehro Pefhany

Seems that way: the demo clock at Fry's would just sit there if you stopped it.
(Fry's! The time and money I've pissed away there!)
Wade Hassler

 

[Chuck Simmons]

If I am inventing or designing something I would like to know
if someone did it before.
And I am shore that OP would like to know that too!

The OP said in his first line or so he was reverse engineering an
interesting bit of hardware. Really a demonstration of persistence of
vision. Now be nice and go back to hunting for a sucker for your FTL
garbage or selling some stupid idea about particle accelerators. In any
event, it would be nice if you would go back to proving how incompetent
you are at physics and stop explaining things that everybody knows but
is not interested in because of context. You might give my signature a
read while you are about it.

Chuck