Does Dell U2312HM not use PWM for brightness control?

[Crake8]

I have a Dell U2312HM. Doing the pencil waving test, I see no strobe effect at any brightness - even 0% shows none. On my old Viewsonic VP201S IPS monitor, I see the strobe effect under 90% when I wave a pencil in front of it.

I took some pictures with my Ipad. I do not know the shutter speed so I cannot calculate the frequency but I was just doing it to see if it confirmed the pencil waving strobe test - and it did. I took 3 pictures of the white line test, at http://dl.dropbox.com/u/2564500/Pics/flicker_test.jpg , with each monitor at 0%, 50% and 90%.

In all three (see below) the Dell U2312HM showed no signs of on and off with it being a consistent white across the whole line blur. At first I was not sure if the Ipad was capturing it properly since I did not know the shutter speed. So I thought, maybe it does have strobe effect but the Ipad camera is not sensensitive enough or long enough shutter speed to capture it. That is why I did it on the Viewsonic, so I could see if it also showed a solid line. Result: the Viewsonic showed on-off breaks in the pictures taken. At all tested brightness settings of the Viewsonic, 0% 50%, and 90%, you can see on-off breaks in the streak . But using the same Ipad camera, the Dell showed no such on-off phases at any brightness setting I tested - even 0%.

Pictures

Viewsonic at 0% and 90%

Dell at 0%, 50% and 90%:

Does this mean the U2312HM does not use PWM to control brightness or uses a very high frequency?

 

[CrabJuice]

CCFL are usually around 180hz. Dell could be 360hz or higher and it would be hard to tell with those pictures. I guess you need a camera with more detail and/or faster shutter speed to make sure.

 

[Crake8]

CCFL are usually around 180hz. Dell could be 360hz or higher and it would be hard to tell with those pictures. I guess you need a camera with more detail and/or faster shutter speed to make sure.

Thanks. So either flicker frree or pretty high frequency I guess.

What about the pencil waving test showing nothing, even at 0?

 

[CrabJuice]

Thanks. So either flicker frree or pretty high frequency I guess.

What about the pencil waving test showing nothing, even at 0?

Yes. Fairly high frequency. With PWM dimming, duty cycle on 0 should be relatively short. Maybe 30%, so it has to be high frequency or flickerless or you should be able to see it on those pictures. You could try to put the Ipad cam closer or move it faster but then it gets difficult to capture the line properly.

I managed to see the strobe effect from a pencil on a 500hz PWM Acer monitor. Thus, its possible. But you have to wave that pencil very fast. I suppose you also need to have a vision system in your skull that is capable of seeing it.

 

[CrabJuice]

Version: Rev 05 Date: May 2012 Location:I saw no location of manufacture anywhere on the back label.

Thanks! Its indeed a newer revision. Dell might actually have changed something up.

 

[k9wazere]

CCFL are usually around 180hz. Dell could be 360hz or higher and it would be hard to tell with those pictures. I guess you need a camera with more detail and/or faster shutter speed to make sure.

CCFL is an entirely different kettle of fish tho, isn't it. As I understand it, CCFL does not switch off instantly. It remains glowing for several ms after the power is stopped. Making direct frequency comparisons not really useful?

 

[CrabJuice]

CCFL is an entirely different kettle of fish tho, isn't it. As I understand it, CCFL does not switch off instantly. It remains glowing for several ms after the power is stopped. Making direct frequency comparisons not really useful?

True, but you can still measure and make comparisons even if the pulse is different. LED PWM is usually square while the CCFL is like a repeated shark-fin, from what I've seen. CCFL dimming also cause variation between color components. That is, RGB respectively are slightly out of sync. CCFL also have additional inverter flicker but that is very high frequency and should not be confused with dimming flicker.

 

[k9wazere]

Quick update, the u2312hm that's now sitting on my desk does not have the "strobing" effect at all. I'm running at 15% brightness.

 

[Crake8]

Can anyone confirm this technically?

 

[wyqtor]

I bought mine back in december (don't know the revision but probably an older one) and yes, PWM is unnoticeable.

 

[Pylon]

It's also possible they're using a circuit (low pass filter?) to smooth out the PWM.

 

[Crake8]

It's also possible they're using a circuit (low pass filter?) to smooth out the PWM.

Does that mean voltage is reduced and LEDs shine less bright versus being turned on and off?

 

[CrabJuice]

Does that mean voltage is reduced and LEDs shine less bright versus being turned on and off?

It would mean that the light level would have a smooth variance and not simply on-off, yes. Not very likely that Dell would use this, however.
There was someone over at OCN in the Yamasaki (or maybe Shimian) thread that measured that one to have sort of a sinus-wave for dimming control. Thats the only place I've seen some sort of proof of such a scheme.
A more clever solution, which I think, Samsung uses in their 850-series is to have different PWM-phase on different groups of LEDs. Dell could be using this.

The most likely answer is that they boosted PWM frequency, however. Doing signal smoothing or direct level control have some unwanted side-effects. PWM is the usual panel-manufacturers 'recommendations' anyhow and those recommendations are based on the capabilites of the LEDs arrays.

They could, however, quite easily offer both options in a monitor. Like having two 'brightness' controls. One that does PWM and one that does direct level control. But that will never happen. There is almost no demand and useless for marketing as very few understands what it means. On the other hand monitor manufacturers arent shy when it comes to useless features that are basically free.

 

[Mr.Pixel]

It's also possible they're using a circuit (low pass filter?) to smooth out the PWM.

Possibly, but that might defeat part of the purpose of using PWM. It should be about as easy to implement a non-PWM control.

Does that mean voltage is reduced and LEDs shine less bright versus being turned on and off?

Sort of. The voltage and current passing through the LEDs are interdependent. Adjusting the current (either linearly or using PWM) to alter brightness is usually better because situations where runaway current use might occur can be avoided.

 

[Mr.Pixel]

There was someone over at OCN in the Yamasaki (or maybe Shimian) thread that measured that one to have sort of a sinus-wave for dimming control.

That was probably me. I still don't have a definite answer about why it's like this, though I have several guesses.

A more clever solution, which I think, Samsung uses in their 850-series is to have different PWM-phase on different groups of LEDs. Dell could be using this.

Interesting, but I don't know why anyone would do this. It should be simpler to just increase the PWM frequency.

They could, however, quite easily offer both options in a monitor.

I believe there are some standard controllers out there that offer both functions, but don't know if any have ever been used for a backlight.

 

[CrabJuice]

That was probably me. I still don't have a definite answer about why it's like this, though I have several guesses.

It was most certainly you. What might those guesses be?

Interesting, but I don't know why anyone would do this. It should be simpler to just increase the PWM frequency.

Its not the same. Imagine summing 4 different square pulses that are have 1/4 phase shift. The amplitude sum will be lower with a minimum of 1/4th of using regular PWM dimming. It will give you better control and the ability to 'smooth the pulse' somewhat. This without causing color-shift or other LED issues.

I believe there are some standard controllers out there that offer both functions, but don't know if any have ever been used for a backlight.

They could actually be using it already for late production-tuning of maximum brightness.

 

[gauden44]

This is a great thread, and thanks to the OP for posting their results. I'm curious about this matter, as my U2412M does flicker, and I was curious whether the 23" variant did as well.

 

[Crake8]

Something else of note: I tried a new HP 2311XI (the new HP IPS monitor) and it had no flicker, as far as the pencil waving test showed - even on lowest settings. I did not do the camera test on it, but it showed no signs of strobe effect from waving a pencil/etc. in front of it - same as my U2312Hm showed no signs but my old Viewsonic did.

Could this be new backlight controllers used in many new IPS models of various brands? These monitors, from all the main brands, are probably all made by the same Asian OEMs and all along the same basic design, so if one uses it, all probably do too.

 

[CrabJuice]

Proof that the U2312HM does flicker in earlier revisions: http://www.youtube.com/watch?v=hoQEM5nu3rE

Something else of note: I tried a new HP 2311XI (the new HP IPS monitor) and it had no flicker, as far as the pencil waving test showed - even on lowest settings. I did not do the camera test on it, but it showed no signs of strobe effect from waving a pencil/etc. in front of it - same as my U2312Hm showed no signs but my old Viewsonic did.

HP is a different deal than Dell. They actually have a flicker free model already in zr2740w as shown in the Prad.de review of this model. Also zr2440w is quite high frequency PWM at 410hz.

Could this be new backlight controllers used in many new IPS models of various brands? These monitors, from all the main brands, are probably all made by the same Asian OEMs and all along the same basic design, so if one uses it, all probably do too.

There is probably a certain hivemind aspect to this, yes. I think its more about failsafing and convention, however. They do what the panel maker recommends and that is usually PWM dimming with 'adequate' frequency. But for this panel revision, LG might be recommending PWM free dimming. Due to improved WLED tech, perhaps.

 

[M4_ELITE]

Went to Best Buy today and I can also confirm the HP 2311xi passes the waving pen test at 0% brightness. The HP 2011xi did not pass even at relatively high brightness and showed multiple pens while waving.

 

[CrabJuice]

They could, however, quite easily offer both options in a monitor. Like having two 'brightness' controls. One that does PWM and one that does direct level control. But that will never happen. There is almost no demand and useless for marketing as very few understands what it means. On the other hand monitor manufacturers arent shy when it comes to useless features that are basically free.

Good news everyone! Looks like Eizo will step up to my challenge:
http://translate.google.com/transla...izo-ev2336w-hybrid-backlight-dimming/&act=url

Whats left is for me to do:
1. Wait for one of these models to come out and buy it.
2. Figure out how I could have called this before knowing anything about Eizo's plans.
3. ???
4. Profit.

 

[Aeluwas]

Good news everyone! Looks like Eizo will step up to my challenge:
http://translate.google.com/transla...izo-ev2336w-hybrid-backlight-dimming/&act=url

Whats left is for me to do:
1. Wait for one of these models to come out and buy it.
2. Figure out how I could have called this before knowing anything about Eizo's plans.
3. ???
4. Profit.

Sounds like they use both at the same time, though, not as a choice. I'd prefer it if there was *NO* flickering, but I suppose less flickering (perhaps between 50% and 65% brightness or something) is better than the PWM nightmare.

 

[CrabJuice]

Sounds like they use both at the same time, though, not as a choice. I'd prefer it if there was *NO* flickering, but I suppose less flickering (perhaps between 50% and 65% brightness or something) is better than the PWM nightmare.

Hehe. Yeh. I was quick in my interpretation. With two translation steps there, lots of information confusion. We'll see what it actually means later. But the fact that a monitor company are even publicly discussing PWM dimming, DC and flicker is a big step.

50% would help significantly. But a user option would be preferred, like I suggested earlier. Though, an on/off switch for 'eyecare' or something would suffice. Let us worry about minor gamut-shift or whatever!

 

[Mr.Pixel]

Just for reference, NEC has used a hybrid approach in the past where PWM would be used to dim the backlight most of the way, then switch to remapping the maximum digital value for the remainder. That's why the contrast ratio remains constant until reaching a certain point, after which it decreases rapidly. You can see it in the PA271w review here. This is, of course, still bad flicker-wise.

 

[CrabJuice]

Just for reference, NEC has used a hybrid approach in the past where PWM would be used to dim the backlight most of the way, then switch to remapping the maximum digital value for the remainder. That's why the contrast ratio remains constant until reaching a certain point, after which it decreases rapidly. You can see it in the PA271w review here. This is, of course, still bad flicker-wise.

Aha. That explains the contrast ratio graph of some of the Eizo monitors reviews. They must be doing something similar:
http://www.prad.de/new/monitore/test/2012/test-eizo-ev2335w-gb-teil7.html#Helligkeit

 

[webdude]

Did the hp2311xi pass the test on Windows 7? I have read that some monitors that passed the test previously did not pass on windows 7.