maemo.org - Talk

maemo.org - Talk (https://talk.maemo.org/index.php)
-   Nokia N900 (https://talk.maemo.org/forumdisplay.php?f=44)
-   -   Infrared LED brightness (https://talk.maemo.org/showthread.php?t=82567)

Copernicus 2012-02-24 17:06

Infrared LED brightness
 
Question for N900 hardware experts: is there any way to control the brightness of the infrared LED? I'm working on a program that uses infrared (Pierogi), and it appears that the brightness of the LED can vary somewhat between N900s. I don't see any brightness controls in the lirc_rx51 device driver, but maybe there is some other mechanism to adjust it? Thanks!

Mike Fila 2012-02-27 04:56

Re: Infrared LED brightness
 
iirc the ir port was tied in some way to the kernel, maybe the use of a different kernel like kp could affect it's output.

ajack 2012-02-27 07:54

Re: Infrared LED brightness
 
@Copernicus: May I suggest you ask Pali!

sulu 2012-02-27 08:17

Re: Infrared LED brightness
 
Just a wild guess (I don't have my N900 here atm):
Have a look at:
Code:

/sys/class/leds/
IF there are any devices listed, one of them MIGHT be the IR LED.
If so you should find a file called brightness which accepts values of 0-255.
Assuming the LED supports different brightness levels (most of them only switch on and off) you can check the result with any digital camera (photo, video, webcam) pointing at the N900's IR emitter since their sensors partially cover the IR spectrum.

reinob 2012-02-27 12:40

Re: Infrared LED brightness
 
@sulu,

Unfortunately I only see the R/G/B, the keyboard leds and the vibrator.. so no IR there :(

Copernicus 2012-02-27 18:22

Re: Infrared LED brightness
 
Hey, poking around in the rest of the /sys directory turned up a /sys/devices/platform/lirc_rx51 directory. I have no clue what the contents signify, but it even has a subdirectory called "power" (although it contains just one file called "wakeup", and that looks empty).

I guess I should ask, might there be any documentation around for any of this? :)

blue_led 2012-02-27 23:02

Re: Infrared LED brightness
 
CIR output is PWM type so there is no brightness control, just only on ( maximum ) and off
but between "always on" and off state can be a lot of average brightness values.

from datasheet :
• Transmit mode only (receive mode is not supported)
• Free data format (supports any remote-control private standards)
• Selectable bit rate
• Configurable carrier frequency
• 1/2, 5/12, 1/3, or 1/4 carrier duty cycle

dani846 2012-02-28 19:48

Re: Infrared LED brightness
 
please come up with something soon guys.people like me,who don't wanna mess around with their lovely N900s so much,would all be very grateful.Pierogi is a great programme,but the fault lies in my mobile phone.The Infra red LED brightness is very low.There HAS to be a way to adjust this somehow.its an open challenge to all you good people who're good at crackin' stuff.cheers people.

davdav 2012-02-29 09:09

Re: Infrared LED brightness
 
just an idea
why we dont connect an infrared usb chip to the n900 via hen
and make pierogi transmit via it ?

Copernicus 2012-02-29 13:37

Re: Infrared LED brightness
 
Quote:

Originally Posted by davdav (Post 1171886)
just an idea
why we dont connect an infrared usb chip to the n900 via hen
and make pierogi transmit via it ?

Actually, make it a transmitter and a receiver, and I could turn the N900 into a "learning" remote. :) (And it might be useful for the N9 too, if it has USB host mode.)

I'm not a hardware guy, so I don't really do this sort of thing, but I think it'd be easy to rework the Pierogi back-end to tie it to other IR devices...

vi_ 2012-02-29 13:58

Re: Infrared LED brightness
 
http://i44.tinypic.com/k1tc02.png

According to schematics, the current through the LEDs is controlled by the the 220R resistors. Replace them all with 110R resistors for maximum IR lols.

...Although, the fact they have used two transistors in parallel implies that the forward current is already quite high for them (the transistors). So, will doubling it cause any problems? If not, why use 2 transistors?? Can the LED even take any MOAR juice anyway?

Who will try to find out?

sixwheeledbeast 2012-02-29 19:04

Re: Infrared LED brightness
 
Quote:

Originally Posted by vi_ (Post 1172064)
According to schematics, the current through the LEDs is controlled by the the 220R resistors. Replace them all with 110R resistors for maximum IR lols.

...Although, the fact they have used two transistors in parallel implies that the forward current is already quite high for them (the transistors). So, will doubling it cause any problems? If not, why use 2 transistors?? Can the LED even take any MOAR juice anyway?

Who will try to find out?

AFAIK messing with transistors linked like this can cause problems.
Two transistors in parallel like that have to have the same resistance (with not much tolerance).
If there not one transistor will end up taking all of the load. This will destroy both transistors quickly.

Two transistors allow double the current of one.

Most 5mm 860nm infra-red LED's I have seen have a max forward voltage of 1.8V. 940nm LED's can have even lower voltage (1.2-1.6)
Increasing voltage above this will reduce light output due to increased temperature and eventually fry the LED.

I am certainly not going to try.

Mike Fila 2012-02-29 21:15

Re: Infrared LED brightness
 
Quote:

Originally Posted by Copernicus (Post 1172049)
Actually, make it a transmitter and a receiver, and I could turn the N900 into a "learning" remote. :) (And it might be useful for the N9 too, if it has USB host mode.)

I'm not a hardware guy, so I don't really do this sort of thing, but I think it'd be easy to rework the Pierogi back-end to tie it to other IR devices...


It fairly easy to make a receiver and cheap

http://www.networkedmediatank.com/sh....php?tid=29013

that is a receiver using a1/8 phono jack, Im sure you could use a mini usb wire instead. For a sender there are plenty that are pre made for iphones using the headphone jack or that could easily be made as well

blue_led 2012-02-29 23:31

Re: Infrared LED brightness
 
Quote:

Originally Posted by sixwheeledbeast (Post 1172226)
AFAIK messing with transistors linked like this can cause problems.
Two transistors in parallel like that have to have the same resistance (with not much tolerance).
If there not one transistor will end up taking all of the load. This will destroy both transistors quickly.


Two transistors allow double the current of one.

Most 5mm 860nm infra-red LED's I have seen have a max forward voltage of 1.8V. 940nm LED's can have even lower voltage (1.2-1.6)
Increasing voltage above this will reduce light output due to increased temperature and eventually fry the LED.

I am certainly not going to try.

you are wrong
case ONLY one transistor is on the current flow thru this transistor ( one , singleton ) will be the same as both transistors are on state. only the led current will double wen both transistors are on.
Ic( max) = ( Vbat - Vled - Vcesat ) / ( R / 2 ) -----> ~ 15 mA
all values can be considered constants !
i am sure, transistor can sink more than that

the maximum brightness without mods is achieved on 1/2 duty cycle. ( register settings )

average led current will be ( 50% duty , both on ) : 1/2 * 15 mA * 2 = 15 mA and this value can be supported by any led continuous mode
!
AFIK peak led current on traditional ir remote is 1A !!! much higher than n900 ir led current
across all n900 the ir power is low compared with an genuine ir remote control

sixwheeledbeast 2012-03-01 17:41

Re: Infrared LED brightness
 
Quote:

Originally Posted by blue_led (Post 1172331)
you are wrong

This confirms what I thought.
Quote:

On to paralleling the transistors - yes, you can parallel them, but it isn't as simple as just attaching the leads together. Transistors experience a positive temperature coefficient in gain. So higher temperatures mean less voltage drop. With two in parallel, they won't be identical to start with so one will have a little less voltage drop, which means more current flows through it (current follows the path of least resistance). So more heat, which raises the temperature of that transistor, causing a greater imbalance, until that transistor is hogging enough current to overheat it and burn it up. Then the second transistor has all the load and also burns up.

blue_led 2012-03-01 20:37

Re: Infrared LED brightness
 
@ sixwheeledbeast
you have the right to believe what you want but who wants to be always right lose the chance to know the truth

yahoo answers in not a reference, at least for me, over my personal achievement when i designed animated led panels for public advertising . ( from processor assembly code to schematics and mechanical design )

finally, who is right ? you or nokia engineers ?
over and out

sixwheeledbeast 2012-03-01 21:19

Re: Infrared LED brightness
 
@blue_led
It wasn't the only reference I found, but the best explained one.
This is what I was taught, not what I believe.

It sounds like you have a lot more experience in this, so naturally I will take what you say as the truth.

Finally, why does this have anything to do with comparing myself and Nokia or there engineers?

davdav 2012-03-02 08:58

Re: Infrared LED brightness
 
so now we must ask H-E-N developers to attach an external ir device ?

dani846 2012-03-08 17:30

Re: Infrared LED brightness
 
still nothing for me?very sad.:(:(:(

woody14619 2012-03-08 19:04

Re: Infrared LED brightness
 
Just a few things:

The brightness of the IR transmitter in the N900 is what it is. The best you can do is select the 1/2 duty cycle. You can't make it brighter with software, nor would you want to if you could.

As for the bits about transistors: The case sixwheeledbeast is making is valid for some applications, but not this one. Part of the reason being, the system is controlling the duty cycle on the led, so the chances of either heating up enough to cause this effect is minimal.

As for why it's using two, it is probably to control current. When one is on, there are only 2 resistors (equaling 110 miliohm), when both are on, there are 4 (equaling 55 miliohm). And yes, the calculations involved are simple enough to determine that most leds would have no problem handling that.

blue_led 2012-03-09 00:49

Re: Infrared LED brightness
 
1)
the only situation when sixwheeledbeast's statement is true,is bipolar transistors PIN to PIN direct connection due to transistors electrical characteristics misfit
is not the case here because base pin have a large resistor ( 4.7 k ) and colector resistors 110 ohm and this make the equilibrium between transistors
2)
@ woody14619
resistor notation is 220R ----> 220 Ohm, 2R2---> 2,2 Ohm, etc
220 R is noT 220 miliohm

55 miliohm is very close to short circuit or high power FET transistor on resistance, with this value is like direct apply battery voltage on led. for sure is not real case.

without specs i say the ir led on n900 can handle 100 mA pulsed current and the resistors can be replaced with a lower value. the problem is low voltage in transistor's "base" combined with low hFE ( 80 typical for digital transistors ) ----> low base current -> low colector current ( resulting 2 transistor needed ) .
a transistor with hfe = 80 & 1,8 V Vout of CIR output ( not sure ) can handle ~ 20 mA in cold weather
far lower for high peak current needed for "mall tv fun shutdowner"


All times are GMT. The time now is 11:04.

vBulletin® Version 3.8.8