Glo A pessimistic guess about Kobo Glo's brightness limit.

[xuyn2003]

Dear all,
Many people have asked KOBO for a less bright light at the minimal light level,
but till now KOBO never said anything about it.
I'm afraid the bright limit is denpend on hardware, not software.

I hope this is not real!

[stewacide]

I bet they could lower it, but I can think of two reasons they might not want to:

1. Lower brightness would make the lighting look uneven, which people will complain about and might cause returns of people with unreasonable expectations
2. If they're using PWM to control the LEDs, going too low might make 'strobing' apparent

p.s. perhaps an easier to implement solution is a white-on-black mode: if they did this at the display driver level would seem trivially easy to implement. It's how I use my iPhone and laptop in the dark comfortably (although on a laptop this is sub-optimal since it messes up the sub-pixel smoothing, which with eink is not a factor)

[murg]

Originally Posted by stewacide

2. If they're using PWM to control the LEDs, going too low might make 'strobing' apparent

I think the Glo is using PWM. I noticed that after reading in the dark for a while, when I turn off the Glo, I see a strobing in my eyes for a while, which is very similar to what I get when I stop watching TV.

[Uschiekid]

Originally Posted by stewacide

p.s. perhaps an easier to implement solution is a white-on-black mode: if they did this at the display driver level would seem trivially easy to implement. It's how I use my iPhone and laptop in the dark comfortably (although on a laptop this is sub-optimal since it messes up the sub-pixel smoothing, which with eink is not a factor)

I wonder about this though....the only time I notice ghosting (besides when using the browser) is when there is white "text" on black background. I wonder if having the black background would make the letters illegible if not on every page refresh......

[xuyn2003]

Originally Posted by stewacide

2. If they're using PWM to control the LEDs, going too low might make 'strobing' apparent

Oh my god！'strobing'！
Meaning that lower bright is achieved by slower LED flashing speed? not by lower electric voltage？ That's harmful to eyes!

[DNSB]

Originally Posted by xuyn2003

Oh my god！'strobing'！
Meaning that lower bright is achieved by slower LED flashing speed? not by lower electric voltage？ That's harmful to eyes!

Sheeshh,....

Their most likely method of controlling the LED brightness is using PWM. There will be no change in the flashing rate when using PWM as the brightness change will be done by modifying the amount of time the LED is on during each cycle.

If they are using PWM at 1KHz (a 1000 microsecond cycle width), I somehow doubt that anyone is going to notice any strobing effect or suffer any ill effects. Now if they were using between 5 and 30 pulses per second, you might have a case (and for the photosensitive epileptic, a seizure).

Say you are using a 1000 microsecond cycle and a resolution of 100 microseconds, you would get 11 steps from 0 microseconds on (dark) to 1000 microseconds on (full brightness) without having to change the cycle time.

Perhaps next time before getting angry, you should try getting educated?

Regards,
David

[DNSB]

Originally Posted by stewacide

I bet they could lower it, but I can think of two reasons they might not want to:

1. Lower brightness would make the lighting look uneven, which people will complain about and might cause returns of people with unreasonable expectations
2. If they're using PWM to control the LEDs, going too low might make 'strobing' apparent.

Out of curiosity, why would you think that using PWM and going too low might make 'strobing' apparent? Most of the PWM implementations I am aware of use a constant cycle time (cycle = on time + off time).

Regards,
David

[jusmee]

Originally Posted by DNSB

Out of curiosity, why would you think that using PWM and going too low might make 'strobing' apparent? Most of the PWM implementations I am aware of use a constant cycle time (cycle = on time + off time).

Regards,
David

And the frequency is way too fast for the eye to perceive. After effects from a TV were mentioned, but the frame rate for most TV systems is 50 or 60 Hz. PWM usually works in the thousands of Hz.

[EldRick]

Every adjustable-brightness LED flashlight in the world uses PWM to control the brightness.

[sysKin]

So has anyone managed to control the glow by directly interacting with hardware?
Looking at /sys/devices/platform/ -- which should have all Freescale hardware accessible - there exists /sys/devices/platform/pmic_light.1/, which has a virtual file "lit". It *looks* like writing to that file should control the light, just like writing to platform/pmic_leds.98/leds/pmic_ledsb/brightness makes the LED go blue.

Looking at constants from libnickel.so, it seems to do that in a few places:
echo ch %1 > /sys/devices/platform/pmic_light.1/lit
echo cur 1 > /sys/devices/platform/pmic_light.1/lit
echo dc %1 > /sys/devices/platform/pmic_light.1/lit
echo bp %1 > /sys/devices/platform/pmic_light.1/lit

However, it's anyone's guess what %1 should be at those points, and is there some sequence of "ch" and "dc" and "bp" and "cur" commands that need to be run together. My attempts to write numbers or letters there failed, and my attempts to read Freescale documentation on this device were unsuccessful. It seems to be designed for LCD backlight in mind.

It's also perfectly possible that the light is actually not controlled by pmic_light, but instead by one of the mxc-i2c devices or even mxc-pwm.1 device or even something else.

Ideas?

Maybe some linux guru knows a way to capture writes to that "lit" file, somehow, and dump them to a real file instead?
For example overriding "echo" command with our own?

[xuyn2003]

Originally Posted by DNSB

Sheeshh,....

Their most likely method of controlling the LED brightness is using PWM. There will be no change in the flashing rate when using PWM as the brightness change will be done by modifying the amount of time the LED is on during each cycle.

If they are using PWM at 1KHz (a 1000 microsecond cycle width), I somehow doubt that anyone is going to notice any strobing effect or suffer any ill effects. Now if they were using between 5 and 30 pulses per second, you might have a case (and for the photosensitive epileptic, a seizure).

Say you are using a 1000 microsecond cycle and a resolution of 100 microseconds, you would get 11 steps from 0 microseconds on (dark) to 1000 microseconds on (full brightness) without having to change the cycle time.

Perhaps next time before getting angry, you should try getting educated?

Regards,
David

OK, I understand, the flashing rate will not change at different level, but the on-off time was adjusted in one cycle.

Another question, will PWM ,have a minimal on time limit? For example, it had to be on for at least 0.1 microsecond, if the time is less than 0.1 microsecond, it will not be turn on at all？

[GeoffR]

Originally Posted by sysKin

So has anyone managed to control the glow by directly interacting with hardware?

Not in the way you mean, but I can reduce the (apparent) brightness by holding the hardware further away from my face :-)

When reading most paper books I hold the book quite close in order to see the small type clearly. When I first got my Glo I did the same thing, but then I realised it was much more comfortable to increase the font size and hold it further away. As a result I had to turn up the brightness setting, but that is no problem, the Glo can go brighter that I will ever need.

[DNSB]

Originally Posted by xuyn2003

OK, I understand, the flashing rate will not change at different level, but the on-off time was adjusted in one cycle.

Another question, will PWM ,have a minimal on time limit? For example, it had to be on for at least 0.1 microsecond, if the time is less than 0.1 microsecond, it will not be turn on at all？

The minimum on time limit depends on the resolution of the controller. A dedicated controller can easily have an on time of 0.1 microseconds, a controller being run by a microprocessor in software which is handling other tasks is likely to have a longer minimum pulse width. The other limit on the minimum duty cycle would be the LED itself as you generally need higher current at the shorter duty cycles to have the LED turn on and there is a minimum pulse width below which the LED will not turn on. Those numbers will depend on the LED manufacturing process and materials.

Using my 1KHz cycle time, an LED being turned on for 0.1 microseconds then off for 999.9 microseconds is not going to generate very much light. It's more likely to have a minimum pulse width (AKA duty cycle) determined by the minimum acceptable light level and then however many steps the designers felt would be useful between that duty cycle and the maximum duty cycle. If an external PWM controller is used, some of the parameters may be hard-coded into the PWM controller while others would be set by the microprocessor. If not PWM controller is present, the microprocessor would be doing all the control functions.

Regards,
David

[KevinShort]

Originally Posted by sysKin

So has anyone managed to control the glow by directly interacting with hardware?

I just figured it out today, see this thread.

[murg]

Originally Posted by xuyn2003

Dear all,
Many people have asked KOBO for a less bright light at the minimal light level,
but till now KOBO never said anything about it.
I'm afraid the bright limit is denpend on hardware, not software.

I hope this is not real!

You could always wear sunglasses while reading your Glo in the dark...