High-output, low-duty cycle LED strobe circuit

[whit3rd]

[About using a flyback coil to store energy for an LED
pulse generator for stroboscope use... }

Could you expand a little on this?  I have some very vague things in
mind, but I'm falling short and could use a few pointers to consider.

 I'm assuming for a moment that the Ron is about 3
ohms or so with a Vfwd of 3.3V.  That suggests a dI=(3V-2V)/3 ohms or
(1/3)A.  The dt is 200us.  So the dI/dt, or V/L, is a little more than
1500.  With a mean V of about 2.5V, this is on the order of 1.5mH.
Once the 200us has expired, it would be desirable to ramp up the lost
(1/3)A of dI.  But over the much longer period of something on the
order of as long as 50ms (the OP mentioned 20Hz.)  This suggests a V
across the L of about 10mV

Yep, that'd be awkward, allright. With higher frequency input, you
still
want full charging of the inductor, so lets only allow 10ms (this
would
let the constructor reproduce 60 Hz and below, which is a useful
range).
To get a 1.5 mH inductor up to 0.3A with 5V (less a bit for a switch,
and for sense resistor and coil resistance), takes

t = L (0.3A)/(4.5V) = 0.1 ms

meaning that 1.5 mH is about one hundred times too small. It also
means that the turnoff transient at the regulated +5V supply is 300
mA.

Neither is good, so instead, use ten times as many turns on the
primary winding (the 5V one) as on the secondary winding (the 1.5 mH
one that drives the LED). Just as a car spark coil goes from 12V
on the primary (with -300V spike at points opening), while the HV
winding
generates 30 kV, so the doubly-wound core will have the
low-slew 5V side that ramps up, and the high-slew LED side that
quickly dumps the energy into that LED. The 'blocking diode' in
series
with the LED will have to hold off about 50V during charging.

Ten times the turns means instead of 1.5 mH, the primary winding has
N**2 * 1.5 mH,
or 150 mH inductance, and takes 0.030A instead of 0.300A at full
charge,
so the transient when it turns off is only 30 mA. I'd still prefer to
run this
off a filtered unregulated supply if that's available.

 

[Jon Kirwan]

[About using a flyback coil to store energy for an LED
pulse generator for stroboscope use... }

Could you expand a little on this?  I have some very vague things in
mind, but I'm falling short and could use a few pointers to consider.

 I'm assuming for a moment that the Ron is about 3
ohms or so with a Vfwd of 3.3V.  That suggests a dI=(3V-2V)/3 ohms or
(1/3)A.  The dt is 200us.  So the dI/dt, or V/L, is a little more than
1500.  With a mean V of about 2.5V, this is on the order of 1.5mH.
Once the 200us has expired, it would be desirable to ramp up the lost
(1/3)A of dI.  But over the much longer period of something on the
order of as long as 50ms (the OP mentioned 20Hz.)  This suggests a V
across the L of about 10mV

Yep, that'd be awkward, allright. With higher frequency input, you
still
want full charging of the inductor, so lets only allow 10ms (this
would
let the constructor reproduce 60 Hz and below, which is a useful
range).
To get a 1.5 mH inductor up to 0.3A with 5V (less a bit for a switch,
and for sense resistor and coil resistance), takes

t = L (0.3A)/(4.5V) = 0.1 ms

meaning that 1.5 mH is about one hundred times too small. It also
means that the turnoff transient at the regulated +5V supply is 300
mA.

Neither is good, so instead, use ten times as many turns on the
primary winding (the 5V one) as on the secondary winding (the 1.5 mH
one that drives the LED). Just as a car spark coil goes from 12V
on the primary (with -300V spike at points opening), while the HV
winding
generates 30 kV, so the doubly-wound core will have the
low-slew 5V side that ramps up, and the high-slew LED side that
quickly dumps the energy into that LED. The 'blocking diode' in
series
with the LED will have to hold off about 50V during charging.

Ten times the turns means instead of 1.5 mH, the primary winding has
N**2 * 1.5 mH,
or 150 mH inductance, and takes 0.030A instead of 0.300A at full
charge,
so the transient when it turns off is only 30 mA. I'd still prefer to
run this
off a filtered unregulated supply if that's available.

Thanks. This ties in the with the "impedance matching" comment
nicely. I think I see the picture, but I need to sit down a bit and
drill it into my brain some more so that it stays put. But the basic
intuition feels 'right', for now.

Appreciated,
Jon

 

[Jon Kirwan]

That's enough to dimly light up a really high efficiency LED.

Actually, I have some here that look pretty decent at half an mA. At
least, indoors.

Jon

 

[Bob Larter]

Hey, I didn't say everyone implements the standard* perfectly.

The PIC series has a long-standing hardware bug in slave mode.

Oh great. I had been thinking that PICs would make good I2C slave chips.
What sort of problems are there?

 

[Bob Larter]

Actually, I have some here that look pretty decent at half an mA. At
least, indoors.

Exactly.

If you're like me, & grew up with red LEDs that were reasonably visible
at 20mA, you'd have trouble picking them from modern LEDs at 500uA.

 

[Nobody]

Oh great. I had been thinking that PICs would make good I2C slave chips.
What sort of problems are there?

http://ww1.microchip.com/downloads/en/DeviceDoc/80132f.pdf

 

[RHRRC]

I'm looking for ideas on how to make an LED flash so brightly at a low
duty cycle that it's reasonably bright--maybe even close to what it
would be if it were on DC.

I'm building a project where I need to flash white LEDs very brightly
30-50 times a second at about a 0.4% duty cycle.....



With a 0.4% duty cycle you will need to produce 250 times as much output
over that time to equal the same as it would appear on d.c. (unrealistic!!)

However if low output devices like the ones you are using are ok on dc
(giving 0.13Cd) try some high output devices (Lumileds, Cree, Seoul Semi)
with optics if necessary (the optics are readily available for many of these
devices.

Powering leds @ 1amp (or more) is quite straightforward and is
commonly/cheaply done in some Infra-Red remote controls.

Use a decent high gain at current device such as the ZTX689B (amongst plenty
of others)

put led as collector load to +ve, ~0.7ohm emitter to ground, small npn:-
emitter to gnd, base to ZTX emitter, collector to ZTX base . (gives curent
limiting @~1amp)

Resistor from ztx base to uP (the ztx will need only a few mA from your uP
to switch 1 amp)

The duty cycle is so low that unless your supply volts are too large the ztx
will stand the power dissipation. Any 6V 1amp supply ok- just ensure there
is a low impedance (*not* a standard electrolytic) decent sized electrolytic
across the supply.



hope this helps

 

[whit3rd]

On May 5, 3:05 pm, Jon Kirwan <[email protected]> wrote:

[About using a flyback coil to store energy for an LED
pulse generator for stroboscope use...
with a primary side having ten times the turns of the secondary
side]

  The 'blocking diode' in
series
with the LED will have to hold off about 50V during charging.

Ten times the turns means instead of 1.5 mH, the primary winding has
N**2 * 1.5 mH,
or 150 mH inductance, and takes 0.030A instead of 0.300A at full
charge,

Oops; got that backward, the switch on the primary (high-turns)
side will have to hold off -50V during discharging.

 

[Bob Larter]

http://ww1.microchip.com/downloads/en/DeviceDoc/80132f.pdf

Lovely. *mutter*

 

[Jon Kirwan]

On May 5, 3:05 pm, Jon Kirwan <[email protected]> wrote:

[About using a flyback coil to store energy for an LED
pulse generator for stroboscope use...
with a primary side having ten times the turns of the secondary
side]

  The 'blocking diode' in
series
with the LED will have to hold off about 50V during charging.

Ten times the turns means instead of 1.5 mH, the primary winding has
N**2 * 1.5 mH,
or 150 mH inductance, and takes 0.030A instead of 0.300A at full
charge,

Oops; got that backward, the switch on the primary (high-turns)
side will have to hold off -50V during discharging.

Yeah, that's where I was. I was still trying to figure out what you
meant about the diode. Now it is back to making sense, again.

Jon

 

[Jon Kirwan]

Exactly.

If you're like me, & grew up with red LEDs that were reasonably visible
at 20mA, you'd have trouble picking them from modern LEDs at 500uA.

Much of that seems to be the shift in wavelength towards our eyes'
optimal sensitivies, while still being 'red', though I don't mean to
take away the idea that there have been some improvements elsewhere.

Jon

 

[Bob Larter]

Much of that seems to be the shift in wavelength towards our eyes'
optimal sensitivies, while still being 'red', though I don't mean to
take away the idea that there have been some improvements elsewhere.

Modern high brightness red LEDs are orders of magnitude brighter than
the ancient red LEDs that were available in the early days.

 

[Jon Kirwan]

<snip of question about driving more than 150mA through an LED>

Read the whole thread. 5 things stand out:
1) Put the LED in the collector side (you need the voltage advantage).
2) Bypass the LED circuit right damn close and maybe isolate it a bit
with a ferrite bead.
3) Even still you may have to jack up the collector/LED supply.
4) The LED output increases far less than linearly with overcurrent.
5) White LEDs are phosphor coated near UV devices, the phosphor decay
rates will have major impact in fast pulsed modes.

There are two basic categories of decay rates for wavelength shifting
effects -- fluorescence and phosphorescence. These are driven by two
different quantum mechanisms, the latter requiring an improbable state
transition in order to gain a longer decay time.

Some phosphorescent decays are long at room temperatures, with taus on
the order of a millisecond or two for some cheap, commonly used
materials. Fluorescence, on the other hand, tends to be very fast --
with nanosecond taus -- because the usual de-energizing state
transitions are quite fast. They are on the same order as for what
one observes when light passes through glass in an optical lenses, for
example. (At rates of roughly 10^8 absorption/ emission transitions
per second, broadly speaking.)

I've not tested white LEDs at all, but I do recall that many use a P46
YAG (a combustion-synthesized form of Y_3 Al_5 O_12:Ce+3(2) phosphor),
which has a tau on the order of some 70ns, memory serving. (I seem to
recall two emission peaks, one around 530nm and another around 550nm,
and I don't have my books handy to look this up to see which is which
here.) In any case, this is far, far faster than the OP's driving
period.

For an interesting page on lifetime measurement, see:

http://www.iss.com/resources/tech1/index.html

The term 'phosphor' is often applied to anything exhibiting a Stokes
shift with any useful efficiency -- conflating both fluorescent and
phosphorescent mechanisms into the term. So it isn't always possible
to tell what mechanism is operating, merely because a material is
called a phosphor.

There are also ZnSe substrate white LEDs that completely avoid the use
of a phosphor of any kind.

Jon

 

[mj]

I'm replying to myself here because I have a solution that works for
the moment.

It seems that it was enough to just have no current-limiting resistor
in the collector. We'll see how that goes with 12 LEDs; I may yet have
to get a low output impedance cap to avoid flicker.

Here's a short video demonstrating what I have so far on my breadboard
with this project:
http://sn.im/datastrobe1

Thanks to all who responded.

--Mark

 

[Jon Kirwan]

I'm replying to myself here because I have a solution that works for
the moment.

It seems that it was enough to just have no current-limiting resistor
in the collector. We'll see how that goes with 12 LEDs; I may yet have
to get a low output impedance cap to avoid flicker.

Here's a short video demonstrating what I have so far on my breadboard
with this project:
http://sn.im/datastrobe1

Thanks to all who responded.

--Mark

Thanks, Mark. That was really great in terms of letting me see
exactly what you are doing. That disk almost looks like those ink jet
printable dvd disks I've seen.

My wife didn't understand the video, at first. I had to add more
explanations to get her to accept that it wasn't just shining a
flashlight on a wheel with letters on it. Part of her distraction
about that was due to the time you spent talking about the IR LED,
which made her think that you thought it was the main attraction when
it was just part of the mechanism you wanted to highlight but wasn't
the feature attraction. I explained it more, then she got it. I
particularly just added the idea of a car's timing light. Then it was
much clearer to her.

Her immediate jump was then towards using this NOT for displaying
letters but instead for doing animation -- she draws cartoons and is a
writer. As a toy device, she thinks that this would be far more
interesting than just some character output device, which really
doesn't get her attention nearly as much.

You are probably already familiar with drawing stick figure cartoons
on the margins of a book and then flipping the pages to see the
action. A huge difference that this kind of device contributes to
something so low-tech is that the series of images are laid out on the
perimeter do NOT have to be seen in the same sequence. In other
words, a set of ... say, 50 images ... can be played in a wide variety
of ways, allowing not just one "story" or two, but a great many
different ways of being played out. Not all the same length of time,
either. In the end, a lot could be done. Even a game, perhaps, that
plays out differently with different people making different choices
in order to impact the way it proceeds.

It __may__ even be possible, with additional thinking, to consider
allowing end-users to do their own images -- even hand-drawn on those
dvds I mentioned -- with feedback coming from timing marks they also
draw on the perimeter edge with a pen. Some means of adjusting those
marks (making them wider and using the center, perhaps, allowing them
to go either way in adjusting the timing?) might allow them to "fix
up" what they are doing until it looks pretty good. Okay, practice
will help, too. But it might be fun. How to specify transition
orders might be a problem. Just thinking right now, that's another
issue perhaps.

Anyway, some neo-gaming possibilities are suggested.

Jon

 

[Jon Kirwan]

<snip>
Mechanical color video display as done CA 1900. disk with holes and color
filters.

I thought there might be some possibilities in varying the choice path
of a fixed set of frames. And I hope he gets some fun out of the work
he's doing.

Add some helpful ideas.

Jon