PopularFX
Home Help Search Login Register
Welcome,Guest. Please login or register.
2026-08-16, 18:23:56
News: The text input boxes (where you write your messages) are resizeable.  Just drag the bottom border of the text box to size it appropriately to your device.  The changes are persistent across your devices.

Pages: 1 2 [3] 4 5 6
Author Topic: LED driver-may have hit the jackpot.  (Read 1175 times)
Topic Options img
img  poynt99
Re: LED driver-may have hit the jackpot. « Reply #50 on: 2014-09-18, 14:24:24 »

Group: Administrator
Hero Member
*****

Posts: 3538
It's not as complicated as it may seem...
Thanks Brad.

I created a model for the LED as PSpice doesn't have one. It is a diode configuration (12 x 1N4007) that yields about 2mA with 2.8V across it.

I will try the new coil inductance and resistance values, and add the CVR. ;)


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  Slider2732
Re: LED driver-may have hit the jackpot. « Reply #51 on: 2014-09-18, 16:04:26 »
Group: Elite Experimentalist
Hero Member
*****

Posts: 1602
... .-.. .. -.. . .-.
Ah! a question answered, the coil resistance being 9.2Ohms.
Can build to that spec now :)


---------------------------
ʎɐqǝ from pɹɐoqʎǝʞ a ʎnq ɹǝʌǝu
   
img  ion
Re: LED driver-may have hit the jackpot. « Reply #52 on: 2014-09-18, 16:37:15 »
Group: Elite
Hero Member
******

Posts: 4127
It's turtles all the way down
Be aware  that there is a color shift associated with LED's that changes with applied current. In pulse mode the LED is operating at one end of it's spectrum, at direct current it may be at the other.

The light box should be characterized for frequency of light output also narrow band, and wide band spectral response vs output to account for some minor differences in measurement.

I highly recommend getting familiar with the LED manufacturers curves from the data sheet, not just the box specs.

I have highlighted a few important points in the attached document on page 2.

Poynt: I'd be interested in taking a peek at your LED model and rationale.


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   
img  Itsu
Re: LED driver-may have hit the jackpot. « Reply #53 on: 2014-09-18, 22:12:58 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4848
Quick and dirty replication of the latest circuit presented by Tinman

CSR = 219.6 Ohm,  the inductor = 7.56mH / 9.4 Ohm (extra resistors added)
3 x 1n4148 diodes,  2x 10mm bright leds
input SG is 4.5V ALL positive square wave @ 130KHz @ 40% duty cycle
The led2 is NOT on, the led1 not full.

Screenshot shows:

yellow: input from SG
Blue:   signal across ground and led1 anode
purple: signal across ground and led2 cathode
green:  signal across the 219.6 Ohm csr

The Fluke meter set to current shows 0.281mA rms, but its not designed to be used at 130KHz and the current calculated via the CSR is 90mV rms / 219.6 Ohm = 0.4mA rms (confirmed lateron (not on video) sort off via my AC/DC current probe working at its lower limits, 1mA/div.)

Video here: https://www.youtube.com/watch?v=bJM7qvI6hBY&feature=youtu.be  

Regards Itsu

  
« Last Edit: 2014-09-18, 23:59:07 by Itsu »
   
img  TinMan
Re: LED driver-may have hit the jackpot. « Reply #54 on: 2014-09-18, 22:22:09 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 5677


Buy me some coffee
Quote from: ION on 2014-09-18, 16:37:15
Be aware  that there is a color shift associated with LED's that changes with applied current. In pulse mode the LED is operating at one end of it's spectrum, at direct current it may be at the other.

Thanks for your input ION.

This is why i have the LED's running at there rated voltage in the circuit. LED 1 on 100% at the 2.8v and LED 2 on for 60% of the cycle at 2.8 volt's. This is the same voltage applied using DC. Once the new circuit is made with the 10mm LED's,then we can go through a wide voltage range,both with the circuit and a DC supply. From there we can make a graph showing the difference in light output.

Brad


---------------------------
Never let your schooling get in the way of your education.
   
img  TinMan
Re: LED driver-may have hit the jackpot. « Reply #55 on: 2014-09-18, 22:28:39 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 5677


Buy me some coffee
Quote from: poynt99 on 2014-09-18, 14:24:24
Thanks Brad.

I created a model for the LED as PSpice doesn't have one. It is a diode configuration (12 x 1N4007) that yields about 2mA with 2.8V across it.

I will try the new coil inductance and resistance values, and add the CVR. ;)
Darren

Thanks for your time on this,but i must ask-why are you using a sim?. All that i asked was that a real modle be used to do the testing on. The circuit is very simple,and very few part's. Here you have a situation with the sim where as you have to simulate part's to simulate the circuit. This also brings the question-how are you going to use a light box to see the difference in light output from the circuit to that of a DC supply with a string of diodes?.


---------------------------
Never let your schooling get in the way of your education.
   
img  Chet K
Re: LED driver-may have hit the jackpot. « Reply #56 on: 2014-09-18, 23:29:58 »
Group: Ambassador
Hero Member
*****

Posts: 4750
Brad
It would seem from the link which ION posted that getting more Lumens per watt with a Pulse is a design feature of these LED's
Strait DC will never throw as much LUMEN per Watt as a Properly  Pulsed LED [pulsed at its design frequency]

Link below
snip

Power efficiency is another advantage of the pulse width
modulation approach. Because pulse width modulation uses
a constant current level for each pulse, it’s possible to select
a pulse level where the LED operates most efficiently, that is,
where the lumen output per watt is the greatest.

to Summarize ,these LED's are very inefficient when Driven with strait DC and a properly pulsed LED will thro more Lumen per watt  then an improperly run unpulsed LED

a Manufacturers spec sheet for the particular LED will give its design parameters and its most efficient watts per Lumen spec [pulse freq]

thx
Chet
   
img  TinMan
Re: LED driver-may have hit the jackpot. « Reply #57 on: 2014-09-19, 00:49:45 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 5677


Buy me some coffee
Quote from: Chet K on 2014-09-18, 23:29:58
Brad
It would seem from the link which ION posted that getting more Lumens per watt with a Pulse is a design feature of these LED's
Strait DC will never throw as much LUMEN per Watt as a Properly  Pulsed LED [pulsed at its design frequency]

Link below
snip

Power efficiency is another advantage of the pulse width
modulation approach. Because pulse width modulation uses
a constant current level for each pulse, it’s possible to select
a pulse level where the LED operates most efficiently, that is,
where the lumen output per watt is the greatest.

to Summarize ,these LED's are very inefficient when Driven with strait DC and a properly pulsed LED will thro more Lumen per watt  then an improperly run unpulsed LED

a Manufacturers spec sheet for the particular LED will give its design parameters and its most efficient watts per Lumen spec [pulse freq]

thx
Chet
Thanks for the info Chet.

In regards to the constant current pulse-as you will see from the last scope shot I posted, the current is anything but constant. The reason is that the current is already flowing in the right direction when the 40% on time starts. The only thing we change at that instant is the voltage polarity. The current starts to build up in a steddy ramp, and peaks at the switch off of the cycle.


---------------------------
Never let your schooling get in the way of your education.
   
img  poynt99
Re: LED driver-may have hit the jackpot. « Reply #58 on: 2014-09-19, 03:19:06 »

Group: Administrator
Hero Member
*****

Posts: 3538
It's not as complicated as it may seem...
ION,

Here is my model in schematic form. This is what is inside the LED symbol on my diagram (next post).

Also shown is the I-V characteristic of the LED model.

Apologies to anyone who may have thought I was trying to model the optical characteristics of the LED. No, simply the current vs. voltage (I-V) curve. Why? As a first step so I can try to match the scope shots.


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  poynt99
Re: LED driver-may have hit the jackpot. « Reply #59 on: 2014-09-19, 03:21:47 »

Group: Administrator
Hero Member
*****

Posts: 3538
It's not as complicated as it may seem...
Here is the schematic and resulting voltage plots across D1 and D2 LED1 and LED2.

Fairly close I think to your scope shots Brad.

Correction for voltage plot: I really mean voltage across LED1 and LED2, not D1 and D2.


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  TinMan
Re: LED driver-may have hit the jackpot. « Reply #60 on: 2014-09-19, 06:21:32 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 5677


Buy me some coffee
Quote from: poynt99 on 2014-09-19, 03:21:47
Here is the schematic and resulting voltage plots across D1 and D2 LED1 and LED2.

Fairly close I think to your scope shots Brad.

Correction for voltage plot: I really mean voltage across LED1 and LED2, not D1 and D2.
Yes it is close Darren. The peak voltage slope is incorrect, but not bad at all. Could you put a current trace in there as well, like I showed?  .
Cheers


---------------------------
Never let your schooling get in the way of your education.
   
img  TinMan
Re: LED driver-may have hit the jackpot. « Reply #61 on: 2014-09-19, 06:42:38 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 5677


Buy me some coffee
Quote from: Itsu on 2014-09-18, 22:12:58
Quick and dirty replication of the latest circuit presented by Tinman

CSR = 219.6 Ohm,  the inductor = 7.56mH / 9.4 Ohm (extra resistors added)
3 x 1n4148 diodes,  2x 10mm bright leds
input SG is 4.5V ALL positive square wave @ 130KHz @ 40% duty cycle
The led2 is NOT on, the led1 not full.

Screenshot shows:

yellow: input from SG
Blue:   signal across ground and led1 anode
purple: signal across ground and led2 cathode
green:  signal across the 219.6 Ohm csr

The Fluke meter set to current shows 0.281mA rms, but its not designed to be used at 130KHz and the current calculated via the CSR is 90mV rms / 219.6 Ohm = 0.4mA rms (confirmed lateron (not on video) sort off via my AC/DC current probe working at its lower limits, 1mA/div.)

Video here: https://www.youtube.com/watch?v=bJM7qvI6hBY&feature=youtu.be  

Regards Itsu

  
Itsu
The voltage should be 9vpp , with a 4.5v offset to get a 0 volt 60% off time


---------------------------
Never let your schooling get in the way of your education.
   
img  poynt99
Re: LED driver-may have hit the jackpot. « Reply #62 on: 2014-09-19, 13:51:23 »

Group: Administrator
Hero Member
*****

Posts: 3538
It's not as complicated as it may seem...
Brad,

I'll not only add the current trace, but power dissipation as well. Do you have a recent scope shot of the SG voltage as well?


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  poynt99
Re: LED driver-may have hit the jackpot. « Reply #63 on: 2014-09-19, 13:56:20 »

Group: Administrator
Hero Member
*****

Posts: 3538
It's not as complicated as it may seem...
Quote from: TinMan on 2014-09-19, 06:21:32
Yes it is close Darren. The peak voltage slope is incorrect, but not bad at all.

I've seen this in other of my sims as well where one slope is opposite. I suspect it might be because I am using an air-core (ideal) inductor, whereas I'm sure yours has a ferrite core? I've also not included some small capacitance parallel to the inductor.


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  ion
Re: LED driver-may have hit the jackpot. « Reply #64 on: 2014-09-19, 14:53:27 »
Group: Elite
Hero Member
******

Posts: 4127
It's turtles all the way down
This article might be of interest, not exactly the same but raises some interesting application ideas.

http://www.edn.com/design/led/4368392/An-LED-s-intrinsic-capacitance-works-in-a-650-mV-LRC-circuit


---------------------------
"Secrecy, secret societies and secret groups have always been repugnant to a free and open society"......John F Kennedy
   
img  TinMan
Re: LED driver-may have hit the jackpot. « Reply #65 on: 2014-09-19, 17:34:24 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 5677


Buy me some coffee
Quote from: poynt99 on 2014-09-19, 13:56:20
I've seen this in other of my sims as well where one slope is opposite. I suspect it might be because I am using an air-core (ideal) inductor, whereas I'm sure yours has a ferrite core? I've also not included some small capacitance parallel to the inductor.
I'm using one of those small inductors that look a bit like a 1/2 watt resistor. I bought a heap of different ones to try.
I'm not sure why,but the best i can get with these 10mm LED's is only 125.4% better than with a DC power input. I might have to have a dig around my old salvaged TV board's,and see if i can find a larger inductor.

I found one strange thing while trying different inductors-with one inductor (not sure of the value's yet)i can turn the SG off,and by touching one of the LED legs,i can get it to light up??? -not very bright,but definitely visible.

Oh,by the way Darren,these 10mm LED's draw 25mA @ 3.2 volt's. I dont have a screen shot of the new setup yet,but will get one tomorrow for you,as it's bed time for me after the long week i just had.


---------------------------
Never let your schooling get in the way of your education.
   
img  poynt99
Re: LED driver-may have hit the jackpot. « Reply #66 on: 2014-09-19, 21:45:17 »

Group: Administrator
Hero Member
*****

Posts: 3538
It's not as complicated as it may seem...
Quote from: TinMan on 2014-09-19, 17:34:24
I found one strange thing while trying different inductors-with one inductor (not sure of the value's yet)i can turn the SG off,and by touching one of the LED legs,i can get it to light up??? -not very bright,but definitely visible.
Now you understand why many of us have been saying for quite some time that superbrights are not a good choice for barometers of power output. It takes next to no power to illuminate them.


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  Itsu
Re: LED driver-may have hit the jackpot. « Reply #67 on: 2014-09-19, 22:14:04 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4848
Quote from: TinMan on 2014-09-19, 06:42:38
Itsu
The voltage should be 9vpp , with a 4.5v offset to get a 0 volt 60% off time



Ok,  done, now both leds are on.

With an "isolated from ground" SG i measured directly across the both leds, like seen in point99 his drawing only i had to reverse
the ground leads so that both ground leads are connected on the junction between the leds.
I inverted the both channels, so the screenshot 1 below shows the correct signals across led1 (blue) and led2 (purple).


I disconnected both probes and now measured the input voltage from the SG (yellow) and the input current with my current probe (controller set to 10mA/div. so
similar as the scope setting) (green), see screenshot 2.
The math function (red) shows the ch1 * ch4 = input power = 18.9mW

Regards Itsu
   
img  poynt99
Re: LED driver-may have hit the jackpot. « Reply #68 on: 2014-09-19, 22:52:34 »

Group: Administrator
Hero Member
*****

Posts: 3538
It's not as complicated as it may seem...
Itsu,

Can you also do a power dissipation measurement for each LED?

I've changed to a new LED model (much more realistic), and my Pin is about 15mW. Power dissipation in each LED is about 6mW. I will post my results in a day or so.


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  TinMan
Re: LED driver-may have hit the jackpot. « Reply #69 on: 2014-09-20, 07:58:08 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 5677


Buy me some coffee
i will be going back to the 5mm LED's ,as these 10mm ones seem to behave in a very different way-- just cant get the efficiency i was with the 5mm LED's for some reason???. Maybe i need a larger inductor(more robust) for these 10mm ones. The best i can achieve is 126.$% efficiency over the DC supply.

Before i change back to the 5mm LED's,i will plot a graph between the circuit and DC output from the 10mm LED's. Hopefully i will have this done tomorrow-along with that screen shot Darren.

Cheers
Brad


---------------------------
Never let your schooling get in the way of your education.
   
img  Itsu
Re: LED driver-may have hit the jackpot. « Reply #70 on: 2014-09-20, 11:25:35 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4848
Quote from: poynt99 on 2014-09-19, 22:52:34
Itsu,

Can you also do a power dissipation measurement for each LED?

I've changed to a new LED model (much more realistic), and my Pin is about 15mW. Power dissipation in each LED is about 6mW. I will post my results in a day or so.

Sure,

screenshot 1 is from led1,   screenshot 2 from led2:

Regards Itsu
   
img  Itsu
Re: LED driver-may have hit the jackpot. « Reply #71 on: 2014-09-20, 11:35:51 »

Group: Elite Experimentalist
Hero Member
*****

Posts: 4848
One observation is that with the SF "isolated from ground", as been used in the above measurement/screenshots, the LED1 (only) is dimly on without
a signal from the SG, infact even with this SG powered off it still is dimly on!!

So there is a strong 50Hz signal inserted somewhere, not sure this will skew the above measurements.

See screenshot taken with no signal from the SG across the dimly lid led1, but with current controller set to 1mA/div, so the Ch4 rms value and the power math value need to be divided by 10, so we have ch4 0.281mA rms and 0.244mW power.

Regards Itsu
   
img  poynt99
Re: LED driver-may have hit the jackpot. « Reply #72 on: 2014-09-20, 16:33:54 »

Group: Administrator
Hero Member
*****

Posts: 3538
It's not as complicated as it may seem...
Thanks Itsu.

My wave forms now look as per the attached.

Brad, I think your inductor is as you originally had thought; 1.5mH. That is the only value that will match the current wave form and peak level of 1.5V across the 200R CVR resistor (7.5mA peak). Note the value in my new schematic here.


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  poynt99
Re: LED driver-may have hit the jackpot. « Reply #73 on: 2014-09-20, 16:46:49 »

Group: Administrator
Hero Member
*****

Posts: 3538
It's not as complicated as it may seem...
Brad,

You were wondering about your input current and Pin calculation. This is how I would eyeball my wave forms to do a rough manual calculation.

So I have a peak current of 7.5mA. I know my duty cycle is 40%. I estimate the current wave form to be approximately an equilateral right triangle, plus a bit more (we are trying to determine area under the curve there). I would estimate a factor of about 0.6 then. Now to multiply it all out:

7.5mA x 0.4 x 0.6 x 9V = 16.2mW Pin

This is close to my estimate (15mW) from looking at the averaged wave form in my plot. If I let the sim run with more cycles on the screen, the actual level would be more accurate, but you get a pretty good sense on where it is going to settle, even with as few as 4 or 5 cycles on the screen, as I have.
« Last Edit: 2014-09-21, 15:01:53 by poynt99 »


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
img  poynt99
Re: LED driver-may have hit the jackpot. « Reply #74 on: 2014-09-20, 16:53:38 »

Group: Administrator
Hero Member
*****

Posts: 3538
It's not as complicated as it may seem...
Quote from: TinMan on 2014-09-16, 14:56:33
The scope shot below is over a 220 ohm 5 watt resistor with an exact value of 216.1 ohm's.
This is as clear as i could get the wave form over a resistor to get our I/in. If some one could tell me the current value from this shot,then i can compair that to what my meter says-which is 2.59mA

Cheers
Brad


So Brad,

Using the same method I used above, I would estimate the following:

Ipeak is 6.94mA (1.5V / 216.1)
Area factor about 0.6
duty cycle 40%
VSG = 9V

So to calculate Pin we have: 6.94mA x 0.6 x 0.4 x 9V = 15mW Pin

From this, is the way you are computing your Pin resulting in the same value?


---------------------------
"Some scientists claim that hydrogen, because it is so plentiful, is the basic building block of the universe. I dispute that. I say there is more stupidity than hydrogen, and that is the basic building block of the universe." Frank Zappa
   
Pages: 1 2 [3] 4 5 6
« previous next »


 

Home Help Search Login Register
Theme © PopularFX | Based on PFX Ideas! | Scripts from iScript4u 2026-08-16, 18:23:56