Searching \ for '[PIC]: propellor clock as planetarium ?' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/microchip/time.htm?key=clock
Search entire site for: 'propellor clock as planetarium ?'.

Exact match. Not showing close matches.
PICList Thread
'[PIC]: propellor clock as planetarium ?'
2005\09\04@181044 by Stef Mientki

flavicon
face
hello all,

We want (well better saying my son wants) to build a propellor clock
and make a projection of the actual star map,
or in other words, "build a desktop planetarium".

Although I like the idea,
I've the feeling that are quiet a lot large problems to overcome.
The first problem I see (or better said I don't expect to see anything ;-)
is how to get such a small spot like "a star at the sky" projected ?

If searched google and ixquick, but couldn't find any usefull links.

So if anyone has some good ideas or  links,
I would be much obliged.

thanks,
Stef Mientki

2005\09\04@190135 by Bob Blick

face picon face
On 5 Sep 2005 at 0:10, Stef Mientki wrote:
> We want (well better saying my son wants) to build a propellor clock
> and make a projection of the actual star map,
> or in other words, "build a desktop planetarium".
>
> Although I like the idea,
> I've the feeling that are quiet a lot large problems to overcome.
> The first problem I see (or better said I don't expect to see anything ;-)
> is how to get such a small spot like "a star at the sky" projected ?

Hi Stef,

You will use lots of LEDs. Also, have two strips of them so you have
an interlaced display. You can put the two strips 180 degrees apart
or next to each other as long as the timing is right.

Cheerful regards,

Bob

2005\09\04@213313 by Robert Rolf

picon face
Bob Blick wrote:

> On 5 Sep 2005 at 0:10, Stef Mientki wrote:
>
>>We want (well better saying my son wants) to build a propellor clock
>>and make a projection of the actual star map,
>>or in other words, "build a desktop planetarium".
>>
>>Although I like the idea,
>>I've the feeling that are quiet a lot large problems to overcome.
>>The first problem I see (or better said I don't expect to see anything ;-)
>>is how to get such a small spot like "a star at the sky" projected ?
>
>
> Hi Stef,
>
> You will use lots of LEDs. Also, have two strips of them so you have
> an interlaced display. You can put the two strips 180 degrees apart
> or next to each other as long as the timing is right.

Or 4 strips, at 90 degrees, to get better resolution yet.
Use the smallest LED's you can get with adequate brightness.
White or yellow to look more like stars. RGB LEDS would be cool but
much more costly.

Use more than one pulse per revolution in your
speed sensor to reduce jitter (assuming perfect construction
symmetry). Be sure to filter the trigger pulses to remove
mechanical jitter jumpyness in the display. Software PLL works
well for this, hence the desire for multiple pulses per rev.

If you are building a 'planetarium' type display, you
probably want a hemispherical arc instead of a straight one.
Radio Shaft (shack) sells a propeller style clock which
is hemispherical and would give you an idea of how fuzzy
T1 LEDs are. Around $40.

How big a dome were you thinking of building?

Robert

2005\09\04@224918 by Bob Blick

face picon face

> White or yellow to look more like stars. RGB LEDS would be cool but
> much more costly.

White LEDs are too slow to use in a propeller clock :(

-Bob


2005\09\05@034850 by Jason

flavicon
face
What are the speed restrictions with white LEDs?

I built a stroboscope a while ago with white LEDs that used 10 or 50 uS
pulses and operated as fast as 200Hz, and I noticed no issues with the LED
response time.  I was easily able to view crisp stop motion on my CPU fan
(around 4500 rpm).


----- Original Message -----
From: "Bob Blick" <spam_OUTbblickTakeThisOuTspamsonic.net>
To: "Microcontroller discussion list - Public." <.....piclistKILLspamspam@spam@mit.edu>
Sent: Sunday, September 04, 2005 7:49 PM
Subject: Re: [PIC]: propellor clock as planetarium ?


>
> > White or yellow to look more like stars. RGB LEDS would be cool but
> > much more costly.
>
> White LEDs are too slow to use in a propeller clock :(
>
> -Bob
>
>
> --

2005\09\05@101939 by olin piclist

face picon face
Robert Rolf wrote:
> Be sure to filter the trigger pulses to remove
> mechanical jitter jumpyness in the display.

That is exactly the wrong thing to do.

If there is jitter in the display rotation speed, the PIC needs to know
about it to compensate for it.  Another way of looking at what you are
saying is to low pass filter the feedback signal.  Basic feedback theory
says the overall result is increased high frequencies to the output, which
in this case means more jitter.

The best method is to make the rotation rate inherently stable in the first
place.  This is done by:

1 - A good motor.

2 - Sufficient rotational inertia compared to the air resistance and the
motor torque.

3 - General shape of the rotating assembly to make air flow as smooth as
possible.

4 - Smooth and tight housing around the rotating assembly so that air flow
is uniform at all points of the rotation.

5 - If you're not using all 360 degrees of display, put the rotation sensor
so that the tick happens just before the start of the desired visible part.
That will be the most stable part of the display.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\09\05@104218 by Bob Blick

face picon face
White LEDs are Blue LEDs with a phosphor layer. So although the
blue part will have a rise time <1usec, the rest of it will be slow
responding. How slow, I don't know. White LEDs tend to be quite
blue-looking in general.

Cheerful regards,

Bob


On 5 Sep 2005 at 0:51, Jason wrote:

> What are the speed restrictions with white LEDs?
>
> I built a stroboscope a while ago with white LEDs that used 10 or 50 uS
> pulses and operated as fast as 200Hz, and I noticed no issues with the LED
> response time.  I was easily able to view crisp stop motion on my CPU fan
> (around 4500 rpm).

2005\09\05@105244 by Spehro Pefhany

picon face
At 07:49 PM 9/4/2005 -0700, you wrote:

> > White or yellow to look more like stars. RGB LEDS would be cool but
> > much more costly.
>
>White LEDs are too slow to use in a propeller clock :(
>
>-Bob

Really?  Do you have any further information on this?

Maybe it varies by type.

Best regards,

Spehro Pefhany --"it's the network..."            "The Journey is the reward"
speffspamKILLspaminterlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com
->> Inexpensive test equipment & parts http://search.ebay.com/_W0QQsassZspeff


2005\09\05@115052 by Bob Blick

face picon face
Here's a wiki on white LEDs:
http://tinyurl.com/cy2jo

They suggest that the coating is actually a scintillator rather than a
phosphor.

To tell the truth, when white LEDs first came out, it was said that
they were slow. While I am sure this is true, it's not something I have
actually tested personally. But now I will, I'll pop one in a propeller
clock and see what it looks like :)

Cheers,

Bob


On 5 Sep 2005 at 11:00, Spehro Pefhany wrote:

> >White LEDs are too slow to use in a propeller clock :(
> >
> >-Bob
>
> Really?  Do you have any further information on this?
>
> Maybe it varies by type.

2005\09\05@130305 by Stef Mientki

flavicon
face


Bob Blick wrote:

{Quote hidden}

hi Bob,
first of all thanks for the great idea of the propellor clock,
a year ago we already build one from Lego, really magnificant.

>You will use lots of LEDs.
>
Yes, thinking of 100 .. 200 over the 90 degrees of the dome.

> Also, have two strips of them so you have
>an interlaced display. You can put the two strips 180 degrees apart
>or next to each other as long as the timing is right.
>  
>
and the alignment of both strips to each other !
But I want small spots, so I think interlacing only solves the problem
if not all LED fits on one strip.
Another good point of using 2 stripes is of course balancing.

cheers,
Stef

>Cheerful regards,
>
>Bob
>
>  
>

2005\09\05@131506 by Stef Mientki

flavicon
face


Robert Rolf wrote:

{Quote hidden}

hi Robert,
(what a coincidence, my son who wants to build this, is also Robbert ;-)

> Or 4 strips, at 90 degrees, to get better resolution yet.
> Use the smallest LED's you can get with adequate brightness.

Yes and whit small angle, maybe we even need a diafragma
(with a 100 or more LEDs, lenses are too large and expensive)

> White or yellow to look more like stars. RGB LEDS would be cool but
> much more costly.

Yes if you want to show the moon and planets too,
some coloring would be nice, but that's a second order problem for now.

>
> Use more than one pulse per revolution in your
> speed sensor to reduce jitter (assuming perfect construction
> symmetry). Be sure to filter the trigger pulses to remove
> mechanical jitter jumpyness in the display. Software PLL works
> well for this, hence the desire for multiple pulses per rev.

I saw Olin reacted, so I can think for a while about this ;-)

>
> If you are building a 'planetarium' type display, you
> probably want a hemispherical arc instead of a straight one.

Yes, we need a half sphere.

> Radio Shaft (shack) sells a propeller style clock which
> is hemispherical and would give you an idea of how fuzzy
> T1 LEDs are. Around $40.

Do you've an weblink ? (I couldn't found anything)
btw, what are T1 LEDs ?

>
>
> How big a dome were you thinking of building?

Let's say 20 .. 40cm.

Cheers,
Stef (and Robbert)

>
> Robert
>

2005\09\05@131837 by Stef Mientki

flavicon
face
hi Jason,

Jason wrote:

>What are the speed restrictions with white LEDs?
>
>I built a stroboscope a while ago with white LEDs that used 10 or 50 uS
>pulses and operated as fast as 200Hz,
>
What current did you apply (I assume you can use very high peak currents) ?

cheers,
Stef

>  
>

2005\09\05@132620 by Stef Mientki

flavicon
face


Olin Lathrop wrote:

{Quote hidden}

I dont think this completly right.
If you want dynamic responses you need to increase the high frequency gain.
But if you're goal is to achieve a static situation (which here is the
case), you must increase low frequency gain (integral part) and reduce
high frequency gain (differential part).

>The best method is to make the rotation rate inherently stable in the first
>place.  This is done by:
>
>1 - A good motor.
>
>  
>
fully agree

>2 - Sufficient rotational inertia compared to the air resistance and the
>motor torque.
>  
>
Large inertia is good (he Olin, but this is low pass filtering ;-).
If you motor torque is low, you can not control much, but this might be
good too (supposing a slow start is good enough).

But for the rest I can agree with you.

cheers,
Stef

2005\09\05@164846 by Jason

flavicon
face
From: "Stef Mientki" <.....s.mientkiKILLspamspam.....mailbox.kun.nl>
Sent: Monday, September 05, 2005 10:18 AM

> What current did you apply (I assume you can use very high peak currents)
?

I'm using 100mA per LED.  Every time I use it, I'm worried the PIC will
glitch with the LEDs on and fry them in a very small fraction of a second :)

Jason.


2005\09\05@165009 by olin piclist

face picon face
Stef Mientki wrote:
> I dont think this completly right.
> If you want dynamic responses you need to increase the high frequency
> gain. But if you're goal is to achieve a static situation (which here
> is the case), you must increase low frequency gain (integral part) and
> reduce high frequency gain (differential part).

But the discussion was about responding to rotation position jitter, which
is a high frequency problem and not static.  Assume the mechanical system
will add a little noise to the rotational position of the LEDs.  The
question was how the PIC uses the position pulses to reduce the apparent
jitter of "fixed" pixels to the extent possible.

>> 2 - Sufficient rotational inertia compared to the air resistance and
>> the motor torque.
>
> Large inertia is good (he Olin, but this is low pass filtering ;-).

Yes, but not in the feedback path.  You want to reduce the noise in the
first place, but then have the PIC respond as best as possible to what is
left.  Mechanical low pass filtering does the first, low pass filtering the
position signal does the opposite of the second.

> If you motor torque is low, you can not control much, but this might be
> good too (supposing a slow start is good enough).

There is no need for any active control over the motor.  This would probably
just add positional jitter anyway.  You don't care that much what the exact
rotation rate is as long as it's steady.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

2005\09\05@165706 by Jason
flavicon
face
From: "Bob Blick" <EraseMEbblickspam_OUTspamTakeThisOuTsonic.net>
Sent: Monday, September 05, 2005 7:41 AM

> > What are the speed restrictions with white LEDs?

> White LEDs are Blue LEDs with a phosphor layer. So although the
> blue part will have a rise time <1usec, the rest of it will be slow
> responding. How slow, I don't know. White LEDs tend to be quite
> blue-looking in general.

I knew that, but since my experience with white LEDs they are fast enough
for a stroboscope they should be fast enough for a propellor clock.  I
thought you might have been offering some more quantitative information.

I'll be interested to hear the results of the experiment you mentioned you'd
do elsewhere on this thread :)

Jason


2005\09\05@174450 by Jason

flavicon
face
From: "Robert Rolf" <Robert.Rolfspamspam_OUTualberta.ca>
Sent: Sunday, September 04, 2005 6:33 PM

> Or 4 strips, at 90 degrees, to get better resolution yet.
> Use the smallest LED's you can get with adequate brightness.
> White or yellow to look more like stars. RGB LEDS would be cool but
> much more costly.

I've been toying with the idea of building something similar with RGB LEDs.
I'd have around 16 of them 1" apart along one end of a meter stick.  The
system would build pictures that are visible as the meter stick is swung.

Where I ran into problems was controlling the LEDs.  16 RGB LEDs is 48 LEDs
to control, and to get full colour, I 'd need to PWM them individually  I
couldn't come up with a good way to drive them.  I know there are LED driver
chips that can do it, but they're only in tiny surface mount packages with
lots of pins for all those LED, so they look like too much work to use for
what I'll get out of the project.

This is a bit off the thread's topic, but what interesting things have
people done with RGB LEDs?  I've seen things like OGGz, the Flexiglow
mousepad, etc that just light things up with cycling colours.  I'm looking
for something more novel.

Jason


2005\09\05@175639 by Peter

picon face

On Mon, 5 Sep 2005, Spehro Pefhany wrote:

> At 07:49 PM 9/4/2005 -0700, you wrote:
>
>> > White or yellow to look more like stars. RGB LEDS would be cool but
>> > much more costly.
>>
>> White LEDs are too slow to use in a propeller clock :(
>>
>> -Bob
>
> Really?  Do you have any further information on this?
>
> Maybe it varies by type.

I think that the high efficiency ones have long (green) afterglow. The
cheap ones do not.

Peter

2005\09\05@181330 by Stef Mientki

flavicon
face

>Where I ran into problems was controlling the LEDs.  16 RGB LEDs is 48 LEDs
>to control, and to get full colour, I 'd need to PWM them individually  I
>couldn't come up with a good way to drive them.
>
yes, that's one of my next problems,
I want to drive 100 or more at 100 mA (or perhaps even more ??) each.
So for the final stage I think I'll use ULN2003 (500 mA) because they
are cheap.
But have no idea yet, how to multiplex all these LEDs

>  I know there are LED driver
>chips that can do it, but they're only in tiny surface mount packages with
>lots of pins for all those LED, so they look like too much work to use for
>what I'll get out of the project.
>
>  
>
Stef

2005\09\05@184713 by William Chops Westfield

face picon face
On Sep 5, 2005, at 2:56 PM, Peter wrote:

>>
>>> > White or yellow to look more like stars. RGB LEDS would be cool but
>>> > much more costly.
>>> White LEDs are too slow to use in a propeller clock :(
>>> -Bob
>>
>> Really?  Do you have any further information on this?
>>
>> Maybe it varies by type.

There are, as I understand it, three types of white LEDs currently
available:

1) Blue LED with yellow phosphor
2) Blue or UV LED with more complex (full spectrum) phosphors
3) multiple emitted (RGB) LEDs

Phosphors are black magic and subject to wildly varying characteristics
in such areas as "persistence", so the behavior of any particular white
LED is likely to be dependent on the particular manufacturer.  And the
usual datasheets don't seem to include phosphor decay info :-(

BillW

2005\09\05@230026 by Robert Rolf

picon face

Stef Mientki wrote:
> Olin Lathrop wrote:
>> Robert Rolf wrote:
>>> Be sure to filter the trigger pulses to remove
>>> mechanical jitter jumpyness in the display.
>>>  
>> That is exactly the wrong thing to do.

Not if you're using a mechanical contact like a reed switch
and moving magnet for your trigger.

>> If there is jitter in the display rotation speed, the PIC needs to know
>> about it to compensate for it.

Yes, of course. But if the sensor is noisy, you need some filtering.
My apologies for not being clearer about that.

T1 LED is 3mm diameter. Quite a bit smaller than the usual T1 3/4.
rocky.digikey.com/WebLib/Lite-on/Web%20Data/LTL-4296N.pdf
There are smaller LED's yet in surface mount.

As for the blue LED clock, it was actually a caller ID unit that
shows time when there is nothing else to display.
(The Source CC used to be called Radio Shack Canada).
http://www.thesourcecc.com/estore/Product.aspx?language=en-CA&catalog=Online&category=Caller+Id+Modules&product=4313350

This is what Radio Shaft US has.
http://www.radioshack.com/product.asp?catalog%5Fname=CTLG&product%5Fid=43-3913

Here is the USA Circuit City link.
http://www.circuitcity.com/ccd/productDetail.do?oid=131887&com.broadvision.session.new=Yes&BV_SessionID=@@@@1557928383.1125974671@@@@&BV_EngineID=ccdgaddfimeeilgcfngcfkmdffhdfkl.0


I am still not clear on whether you are trying to make this a
direct viewing application, or you want to project.
If the latter, you could use fiber optics to get you a very
high density line with very fine pixels into a small space.
You could then layer the LED's behind the swinging arm to get
high density while still giving good brightness.

Is the point to have a spherical display, or to play with
the LED's and motion. If the former, I'd just project a regular
LCD panel onto a frosted dome using a suitable wide angle lens
(a fisheye lens from an old 35mm camera for example).

Robert

2005\09\06@040559 by Alan B. Pearce

face picon face
>This is a bit off the thread's topic, but what interesting things
>have people done with RGB LEDs?  I've seen things like OGGz, the
>Flexiglow mousepad, etc that just light things up with cycling
>colours.  I'm looking for something more novel.

I have thought for a long time that one could do some neat tricks with an
ice sculpture, having LEDs shine up through it from the base. Not
necessarily RGB, but that could make it even more interesting.

2005\09\06@040903 by Alan B. Pearce

face picon face
>1) Blue LED with yellow phosphor

I presume this one is what is used for what is variously known as "warm
white" or "golden white" which makes a pretty good imitation tungsten lamp.
Used a fair bit in the model railroad world for headlamps, where the "blue
white" looks out of place.

2005\09\06@064616 by William Chops Westfield

face picon face

On Sep 6, 2005, at 1:08 AM, Alan B. Pearce wrote:

>> 1) Blue LED with yellow phosphor
>
> I presume this one is what is used for what is variously known as "warm
> white" or "golden white" which makes a pretty good imitation tungsten
> lamp.
>
Actually, I think the blue-led/yellow-phosphor LEDs end up having too
much blue in the output to appear white.  If you look at the spectrum
diagrams (page 8 of the DS25 lumiled datasheet, for instance), the
"normal" white LEDs have a very definite blue spike and yellow bump,
while the "warm whites" have a much broader spectrum with a lot of
power output well on the LONG side of "yellow."

I think the idea of the blue/yellow leds is to lose less energy in
the phosphors (which are generally not very efficient...)  (and they're
probably easier and cheaper to make, too.  But getting that blue/yellow
balance right and even seems to be a pretty tough trick, thus many
"color bins" with unfavorable comments from the masses, etc.)

BillW

2005\09\06@073238 by olin piclist

face picon face
Robert Rolf wrote:
>>>> Be sure to filter the trigger pulses to remove
>>>> mechanical jitter jumpyness in the display.
>>>>
>>> That is exactly the wrong thing to do.
>
> Not if you're using a mechanical contact like a reed switch
> and moving magnet for your trigger.

So don't do that.  The wrong tool will produce bad results.  The EyeBall
products we did used opto interruptors to sense the rotation position once
per revolution.  That worked nicely.


*****************************************************************
Embed Inc, embedded system specialists in Littleton Massachusetts
(978) 742-9014, http://www.embedinc.com

More... (looser matching)
- Last day of these posts
- In 2005 , 2006 only
- Today
- New search...