White led stroboscopic effect

M

mezzrow

Member
Dears, first of all excuse my english.
I'm an italian guy with a problem to solve before the 24 of december.

I need to reach a fast ( more then 30 Hz i need the light to appear still ) strobo effect to freeze some droplets of water as a decoration for a little christmas tree.
I have tried with xenon flash of disposable camera, but the frequency is to fast to reach, and i don't like to work with 220 Volts AC and water at the same time.
So after years of attending this forum just reading how you guys solve the problems of everyone i hope you can help me to.

How to use a picaxe chip to drive some bright white led to achive the effect?
More specifically:
i know how to get a cicle of:
a pulse of 1 or 2 milliseconds
then a pause of 30 millisecond
and back to start...
But how to have a enough bright flash from a white led ( well... 20 leds) in so little time?

Be kind please if the answer is obvious or the question already treated
and thanks to everyone for the help.
 
B

BeanieBots

Moderator
Welcome to the forum, glad you've posted at last!
Is your problem that you don't know how to drive the LED from the PICAXE?
If yes, we need to know more about the LEDs you are using. Ideally, a datasheet.
Most LEDs can tollerate a pulse current many times larger than their normal max forward current but only the datasheet will confirm if YOURS can.
Possibly the best way to drive it would be to use a FET and a low value current limit resistor. You will also need a power supply capable of supplying such a large current. This can be acheived with a moderate supply with plenty of capacitance on it's output.

Supply the LED data and we can get more specific.
 
M

moxhamj

New Member
main:high 1 ' led on
pause 2 ' pause 2 milliseconds
low 1 ' led off
pause 30 ' pause 30 milliseconds
goto main

Get it working with one led on pin 1 and a 220 ohm resistor.

When it works with 1 led, you can look at a mosfet or a transistor driver and lots of leds in parallel and series.

But I think one white led might be bright enough. I've seen some quite bright ones especially out of torches or garden lights.
 
Last edited:
M

mezzrow

Member
thanks Beanie and Dr. for the very fast replies.
Sorry I don't reply so fast, but i started the thread "very" early in the morning and then go to bed.
To Dr. : the problem is not the programming of the chip (very easy after years of frequentation) but what Beanie sed.
I don't have the datasheet of the leds so how can i try to achive the goal?
Why a fet?
How about a ULN2003?
How can I slowly increase the pulse current and find when to stop? (led becomes HOT ?)
I' m going to use a M8 to get the strobo effect an a m8 with sound to make regular droplets from some vibrating little tubes throwing water.
With the correct frequency of sound i can achive the effect of the drops apparing to came out of water, slowly go back to the tubes passing in front of the christmas tree. Well i hope so...

thanks to all for the patience.
 
M

Michael 2727

Senior Member
WTFGL -
Welcome to the Forum and Good Luck ! ;)

What He ^^^^^^ said and - Edit: Oops,,,, you squeezed a relpy in first :)

I have the Silicon Chip Strobe (single LED) from years back, then White LEDs
cost around $9.00 AUD each, how things change.
Edit: Correction, that Strobe was published by Electronics Australia, AUG 97, p88
(now dead burried and gone, archives owned by Silicon Chip magazine) and sold by
Dick Smith's as a kit (K-3160) now also dead and gone by the looks of things.
Although Sil Chip did have a Strobe/Tacho in the August 2008 edition.
Now may be the time to upgrade the LED to a superbright version. /Edit:

Anyway that used a 555 chip pulsed the LED quite hard (overvoltage) for a very
short period of time (to make it appear brighter than it actually was).

As Doc said you may not need as many LEDs as you think if you drive them hard.
A MOSFET (Logic Level switching) would be the way to for a lot of LEDs and run
them at 2 x or 3 x their rated current but only when the Duty (time ON)
is very short.
If you are lucky enough to get or find a data sheet on the particular devices you
are using it may state the MAX pulsed current/time that can be used.

Firstly get a single LED up and running, both electrically and code wise, then expand
from there.
You can then try overvoltage (say from a 9V or 12V source) pulses, via a MOSFET
or Transistor.
You could later experiment with PWM and Poking Registers and Underclocking the
Picaxe for very slow (30Hz) operation.

A simple ON / OFF code should get you going for starters :)


Edit:A quick way to find the MAX voltage/current a LED will take is use a variable supply and
ramp up the voltage till the LED starts to get Dim, they usually do this just before
the burn out voltage/current. You need to do this very quickly, don't sit and count or watch,
it will be too late by then.
Try a 10 Ohm resistor for this and note when the brightness starts to fall and at
what Voltage. From there you can work backwards using the 10R and ? Voltage as a reference.

This will find the MAX continuous current you can use, reduce by say 10% to 15% for longer
LED life.
A pulsed short Duty (ON time) will allow more current to be used, but the above will be your
starting point.
You may have to sacrafice a LED or two in the process, but unless you have the exact Data Sheet
for your LED, this is a quick way to find out.
The ancient Romans were familiar with a little sacrafice here and there ;)
 
Last edited:
M

mezzrow

Member
hi Michael,
(yes, said not sed... pronunciation different from writing... )
i'll try yours suggestions.

Do you think i can use the ULN2003 ( i have it on the desk ) to drive the led?
Using 4 outs from picaxe and 4 groups of leds i can use them cyclically reducing the stress maybe.

I agree a little sacrifice (or sacrafice as you write?) is more than tollerable.

Hope to come back with good news, time is passing and i can hear the
sleigh of Santa Claus approaching.

Merry Christmas to everybody.
 
M

Michael 2727

Senior Member
I have not used the ULN2003, but it looks a little neater than a bunch of transistors.

I'm not sure whether the 500mA rating (although rated per Driver) would actually be
for the total package as many IC packages are. I don't think you will get 3.5A out of
that little IC package without frying it. I could be wrong.
Couldn't find any information on the data sheet I had.

Edit: I should have mentioned, when testing any device to the point of destruction
you should wear, Glasses, Safety Glasses or Reading Glasses.
With enough voltage the top of an LED can pop like a Champagne Cork. And if you don't
already know about Capacitors they can do the same or *worserer.

I guess I need to update my spell checker ;), but then again I know there are a lot
of teachers who visit this forum, and they *can't correct my work, so it's my revenge
for all those years of homework. :)
 
Last edited:
M

manuka

Senior Member
Most off the shelf white LEDs are now so bright, even drawing 30mA at 4V,that using clusters "as is" should be sufficient.

But unlike pure LEDs, which certainly can be pulsed extremely rapidly, a significant white LED strobe limit may arise, as the phosphor on these white LEDs has significant afterglow. The human eye is most sensitive at 555nm - so consider using bright green LEDs, which also will blend in well with the Xmas tree!
 
M

moxhamj

New Member
Good point manuka. Many white leds are actually fakes in that they use phosphors and these keep glowing. Some use true red/green/blue leds. But I'd go for a string of very bright green leds like you say.

I'd start by keeping the leds within their rated current eg 20mA. ULN2003 is fine and I've used them a lot for driving relays.

If you have 12V you can string together 5 green leds and then a dropping resistor for that string of leds. Then put another string on another ULN2003 output. Parallel up all the inputs to a single picaxe pin.
 
D

Dippy

Moderator
The ULN2003 is probably good enough. It's pretty meaty. Don't forget to connect things correctly, so many people think 'OUT' means the current comes out (if you see what I mean).

Good points Stan re: phosphor.
With risk of sounding like someone who has nothing better to do but Google... have a look at this - on the section about pulsing white LEDs fast.
http://www.stockeryale.com/i/leds/lit/app001.htm

Edit:
'Many' Drac? Surely 'all' ?
 
M

moxhamj

New Member
Dippy - "'Many' Drac? Surely 'all'"

I was thinking of the leds that can change hue, ie the ones with several legs so you can control the RGB individually. Probably overkill for this situation.

Re bright leds - two sources I've found that are very convenient are the brake light clusters, and led clusters for garden lights. Both are designed for 12V and include the dropping resistor. All you would need are the ULN2003 and a 12V supply. The brake light clusters are available from car stores. The garden ones from hardware stores in the garden lights section.
 
M

mezzrow

Member
Good day.
Nice to see you all answer to me, this is the Christmas Spirit!

I've tried a bit yesterday with a simple code and a white led:
'b1, b2 time of pauses on off
b1=1
b2=20

Main:
high 1
pause b1
low 1
pause b2
goto main

b1 is the shortest pause i can have and 20 is the minimum to avoid the flicker
the led is very dim also at 12 v.
To Michael: why i have to use the 10 hom resistor?
To Manuka: The point of "phosphor" is surely true. I can't freeze the blades of a fan they seem always a bit blurry.

I will try with green and red strips on opposite sides to obtain more colored reflections on the drops.
To Dippy: I'm going to read your google search, seems very good.
Hope to hear more from all of you thanks a lot!!
 
Last edited:
D

Dippy

Moderator
How about strobing using Red+Gren+Blue LEDs? Fantastic response times and change the hue/colour of your strobe. That would be really flash (haha).

You can overdrive LEDs quite a lot if you have a low duty cycle and a brief pulse.
Its heat that kills.
I don't know which LEDs from Ebay you intend to use, but as an example some Infra-Red LEDs have a continuous rating of 100mA but can be pulsed at >4 Amps - but for very short durations (microseconds) with a low duty cycle.

You will have to read up, but it may be that some of the more juicy LEDs could be driven at a few hundred milliamps for a few mS at a low duty.
If it were me, I'd read up, then buy a few and give it a go. I'm not scared ;)

Andrew: I've only briefly popped back. The Martian winter is drawing in and I'm fed up with sitting in that Mars Rover doing expriments and pretending to be a robot. I just came back to place an order with Damart.
 
B

BeanieBots

Moderator
Or just keep increasing current & duty on a test sample until brightness reduces then back off as described by Michael 2727

Dippy, Damart, the hight of fashion:cool: Saved my life on many an occaision:D
 
D

Dippy

Moderator
... noting that the falling in brightness may be permanent!!

I'm sure a little investigation will help to save some pennies.

If I had a spare fiver I'd buy you a boxfull so you could test to destruction....

Have a go.
 
M

Michael 2727

Senior Member
Testing for the Continuous MAX voltage/current rating, if you don't dilly-dally when
testing and back off the power as soon as you see a decrease in light output, many
of the LEDs survive quite well. As Dippy said they don't like the heat, it melts them inside.

For the Pulsed Mode you are more likely to do the damage there as you are way
beyond the limits of the LEDs designed operation.

With todays High Brightness LEDs you can't look directly at them or you run the risk
of frying your retinas. Try directing the light at some white paper and watch from there.

I had a little sacrifice of my own :( I found my E.A. LED MiniStrobe and connected it to
a 12V DC supply incorrectly. I usually include a doide on most things I make, NOT this time.
One NE555 went to the Smoke God, the other to the Paperweight God, 2 in 1 shot,,grrrr.
Fortunately after some scrounging I found 2 replacements, the last 2 chips I had.

The 10 Ohm resistor was for a reference when testing, Supply ?? Volts, -minus the
LED Forward Voltage drop (e.g. 3.6V) via 10 Ohms, I=E/R and all that stuff ;)
 
D

Dippy

Moderator
And, of course, don't forget that LED Vf increases with forward current, so there is a touch of 'diminishing returns' too.
But you'll all know that as you have read the Data Sheets thoroughly already ..... ;)
 
M

mezzrow

Member
Michael
i'm sorry for the sacrifice you suffer.

i'm now trying with a yellow led and the drops to see if it works.

Dippy
Really for the data sheet i don't have them...but what is a fiver?

tomorrow (monday for me) i'll go and buy some bright coloured leds With the data .

Again thanks for the help.
 
M

manuka

Senior Member
Mezzrow: "What is a fiver" - you are apparently not of UK origins!

"A fiver" refers to a five pound note (£5)- & although now barely enough for a fast food snack, was in living memory almost a weeks wages for a UK manual worker. As a kid I recall even holding such a note was akin to experiencing the wealth of Bill Gates, but these days "a fiver" is considered modest payment to cover incidentals.
 
E

eclectic

Moderator
The site in Dippy's posting #10 mentioned

< 5kHz for White LED phosphors.
~10 MHz for single colours

LED datasheets talk about ON times of < 10%, for high current.

Might PWMOUT work, with Duty <10%,
using a ULN2003 and a low resistance?

Attached example, 5kHz, using the Pwm Wizard

e
 

Attachments

M

mezzrow

Member
to Eclectic:
Not sure to understand what you mean...
The effect is achieved if the drops can travel a bit between flashes
5000Hz means a flash every 1/5000 of a second... to fast
i think.

Now really i am not sure to get the goal.
My test during the day have not give any good result.
Light from leds really dim or to slow to freeze the drops.

Hope to get some better leds in the shop today or i have to
live Christmas without...

anyway all the best to you all.
 
M

Michael 2727

Senior Member
This project has been around for a couple of years-
http://cre.ations.net/creation/the-time-fountain

It uses UV LEDs and UV Dye in the water.
WARNING:
Not sure it would be the best idea to have childern/anyone staring at it for long periods.
UV will damage your eyesight.

It also uses an ADC input to detect and trigger an interrupt when a droplet passes the
output tube to start/sync the timing sequence.

Without a detector to sync the droplets the drips will have to be in very constant flow
for the illusion of freezing in space, otherwise the timing will be all over the place.
(BTW this project used a Raw PIC and not a Picaxe, thats not saying a Picaxe can't do it also)

The proof of concept is there.
Just how well it will work with White LEDs, or others and plain water remains to be seen.
Timing is the critical part for a good (Freeze Frame) effect.
 
M

mezzrow

Member
@ Eclectic

Nice, i give it a try...
do this work for the m8?
i don't want to crash the little chip with the wrong POKE ...
 
E

eclectic

Moderator
I've just tried this version of Dippy's program
Code: 
#picaxe 08M
looper:
pwmout 2 , 249, 500 &#8216; set pwm to 4000 Hz
peek $12,b1
sertxd ("Normal:",#b1,13,10)
pause 4000
b1 = b1 + 1 ' add 1 (bin 1) for TMR2 prescale
poke $12, b1
peek $12,b1 ' peek to confirm
sertxd ("Prescale / 4: ",#b1,13,10)
pause 4000 &#8216; pause 1 s
b1 = b1 + 1 ' add 1 (total 2 bin 10)for TMR2 prescale
poke $12, b1
peek $12,b1 ' peek to confirm
sertxd ("Prescale / 16: ",#b1,13,10)
pause 4000 &#8216; pause 1 s
b1=b1-2 ' Back to normal (i.e. TMR2 prescale = 1)
poke $12, b1
goto looper &#8216; loop back to start
Using a Piezo sounder, then an oscilloscope.

It works.
Frequencies measured at
3.987 kHz
996.8 Hz
249.2 Hz

The figures won't be very accurate, but do show the ratios.
e
 
Last edited:
M

mezzrow

Member
Great!

i'll try the code right now.
This morning i get 10 brand new white leds ( also yellow ones )
hope not to fried them all.

Thanks Eclectic and Dippy
 
D

Dippy

Moderator
An interesting project.
Some things will be obvious but I'll say them regardless as there nothing like stating (and repeating) the bloomin obvious.

1. You can play with the PWM duty so that flashes can be longer or shorter for a given frequency.
2. This will obviously affect the perceived brightness.
3. Never assume hardware PWM frequencies are bang on accurate. The frequency is 'generated' within the PIC using divide registers, or more precisely 'scaling'. This will limit the precision of frequency and duty. Though for this project that isn't really important I'm just saying for future reference.
4. To get adequate brightness you may need some big butch LEDs or a handful of 5mm style. Depends on the distances/LED angle etc.

If white LEDs are too 'slow' (i.e. fuzzy) then consider using single colour LEDs or, as I suggested before, red+green+blue.
It could be used to produce some impressive (and rather pretty) effects with a little imagination.
 
M

mezzrow

Member
to Eclectic:
unfortunately not, the shop owner look at me as i ask something quite unusual...

But i hope to have more light from ten of them.

If i can see the effect then i will improve the thing.

Sure guys if it will work i will post the result in a clip. ( timing will be a problem, better a photo )

To Dippy: nothing is really obvious to me...

Tanks again
 
Last edited:
D

Dippy

Moderator
I'm sure it'll work just fine.

Tip: When going into shops and asking for something 'unusual' tell the shopkeeper that a friend asked you to get whatever it is.
 
M

mezzrow

Member
Dear all,

I have had 2 days of hard work ( not only picaxe of course!)
i post a jpg. I also have an avi of much more effect,
but don't know how to post it.
It works, easily.

Picaxe m8 at 8 mhz only this code:
main:
high 1
pause 1
low 1
pause 20
goto main

The uln2003 give power to 8 white leds at 5,0 V ( 4 in the jpg)
and to a little loudspeaker near the sprinkler.

Problems are not in picaxe part.
The system is very sensible.
Pressure of water, vibrations, bubbles...
a stable condition is very difficult to reach.
In the jpg the high of the stream is not more then 7 or 8 cm.
and far from sprinkler it became worst.
I'm trying hard but only few hours more and Santa will came...

Merry Christmas especially to Dippy, Eclectic, Michael, Beanie and Dr_Acula
your help was much appreciated
 

Attachments

Last edited:
You must log in or register to reply here.
Top