whats with the sudden dislike of pwm?

when the jetbeam mk2 came out... nobody cared. not a single mention on all those pages of the mk2 thread.

makes me think of the time when the inova XO2 first came out...back in early 2006

not one thread...

I don't care much for PWM, but I'm willing to tolerate it. I see it as a convenient way to acheive a wide range of brightness levels. My HDS light uses PWM, and it goes from very dim to very bright. My Surefire U2 does not, but does not go anywhere near as dim. I assume if it was practical and reasonable to use a constant current driver and still get low levels like the HDS, Surefire would have done it.

I know that PWM dimming exhibits less tint shifting than current dimming. I also know that one of the two techniques is typically more efficient, although I cannot recall at the moment which one it is.

People dislike PWM because they can see/sense the flickering, even if only on a subconcious level. I notice the flickering of my HDS out of the corner of my eye, and it bothers me slightly. I also notice computer monitors with a 60 Hz refresh rate, and feel a strong sense of discomfort and a need to "fix" it to a better refresh rate.

I've also noticed that PWM lights are useless for certain tasks. I've worked with a piece of equipment called a Sweco screener. It's got a large seive mounted in a tub. You pour product in the top, and the Sweco's motor shakes the heck out of it. The fine particles get shaken through the sieve, and the crud stays on top. The side of the Sweco has an optical vibration gauge. It's basically a picture of the letter V with little cross hatches across it. You look at the gauge, and where the lines intersect gives an indication of how much vibration is present. When the vibration level is too high, the machine requires service. Guess what - you can't read it if you shine a PWM light on it.

I think people will complain about non-PWM lights given the chance. They'll say they turn too green on the lower settings. Guess what - they're right! PWM flickers on low settings, and constant current exhibits tint shift. You can't have your cake and eat it too.

Given the choice, I'll take the tint shift. But, if the light is good, I'll learn to live with the PWM too.

It depends on the frequency of the PWM, if its low it is irritating also it's not as efficient as current regulation to the LED

Well, if you are LOOKING for reasons to dislike PWM i have a new one for you. I
have one of the Eddie Bauer lanterns on my nightstand next to a am/fm radio
(Radio Shack ac/4 C cell [ $ 5.00 at the thrift store ]. The more i dim it the louder it is on the radio. This, of course, just annoyed the h**l out of me until, after many arduous hours of reviewing electromagnetic theory and the inverse square law, i discovered that if i move the lantern 3-4 inches away from the radio the evil noise goes away. Be warned, PWM is an evil alien device designed
to disrupt your ability to hear Art and George on Coast to Coast.
Sorry gang! Just my deeply warped sense of (hu)mor seeking balance after a long day at work. I will now step onto my transporter and go home:devil:

cheapo said:

when the jetbeam mk2 came out... nobody cared. not a single mention on all those pages of the mk2 thread.

Click to expand...

PWM has long been known to be sub-optimal and the various resistor limited 2-stage lights (McLux PD, etc) and the regulated Li14330 all avoided it. The dimmable light par excellence is the Rigel Systems Nightstar(?), I mean the ultra-dim astronomy flashlight, and it's all analog I'm pretty sure. Surefire likely didn't bother with an ultra-dim level for the simple reason that they weren't trying to make a do-everything light with the U2.

That said, PWM is better than nothing, as a battery stretcher or if you don't want overpowering brightness.

I must be blind, because I just checked my Photon Freedom and I did not see any flickering at all, even on its lowest setting. Are y'all using some type of special equipment to see it?

I recognize most lights that use PWM but I didn't know or forgot that the HDS uses it. I had no idea in use that it used PWM as it isn't at all noticeable like some other lights I've got that use it (I just checked the HDS on a very low setting and it's PWM is pulsing a lot faster than another light I have that uses PWM thus making it less noticeable). But to answer the question - I seem to recall that it's not a sudden dislike but one that has been around for a long time. Maybe just more people suddenly became aware of it. When I think of PWM I think of my Eternalight which is great in how long it can run on a set of batteries but it is quite noticeable at lower levels - and yet it still produces good useable light and if things aren't moving in the light you'd hardly know it.

Both my Photon Freedom and X Light show it very noticeably BUT, only whenu look directly into the LED itself.

Wave it moderately fast n u will see "dots" of light instead of a continueous "line".

I think there is a "sudden" dislike of PWM because of the new Fenix lights that boast that they are current regulated.

If you look at Fenix as a sort of benchmark "cheap" or rather mass-produced, consumer-centric flashlight company, then if they offer a certain set of features it is assumed that said features have reached a norm in all similarly priced and more expensive brands.

Of course, this isn't really the case, and Fenix is certainly ahead of the curve in many respects. If current regulation can be had in a smaller package (notice the added girth and length of the new AA Fenixes- the L0P still uses PWM because it is too small to fit anything else), it will likely replace PWM, and just like luxeons, PWM will be SO 2006 . As paulr said, PWM is certainly better than no low mode at all, or a very inefficient resistor mod. I do tend to find it annoying for extended periods reading, but it's not so bad, and it is quite rare when such a situation arises. That said, I do look forward to seeing some cool new current regulated lights coming down the pike.

I've always hated PWM. I see the flashes when I'm walking and it drives me nuts. It can give me a headache, too.

Solstice said:

As paulr said, PWM is certainly better than no low mode at all, or a very inefficient resistor mod.

Click to expand...

No I didn't say that. Resistor limiters (hard to call them "mods" when they're designed into the light as with the McLux PD, Firefly 2, etc) work quite well. Even if there's some efficiency cost, at low power there's generally plenty of runtime regardless. It might even be comparably efficient in lumens/watt to PWM, even though the power conversion is less efficient, because the LED is more efficient in the lower power range.

I do notice that my PT Aurora headlamp has much faster PWM than the L0D CE and maybe for that reason, it never bothered me much. The L0D CE PWM is fairly noticable whenever anything near the light is moving.

paulr said:

No I didn't say that. Resistor limiters (hard to call them "mods" when they're designed into the light as with the McLux PD, Firefly 2, etc) work quite well.

Click to expand...

McLux PD is not resistor limited, it is current controled in low mode as well.

Many LED's are less efficient when driven PWM....


OTOH, many boost regulators are more efficient when driven PWM...




The higher the PWM rate, the less of an issue flicker is. If it's fast enough >100-200Hz, it's not much of a real deal issue except for the "connoisseurs". But it is a reality...

It's all about electronics design choices and compromises. In broad general terms:

- PMW is less somewhat less efficient than current regulation at levels below 100%. I remember a Newbie post on this subject which presented the evidence clearly in graphical form. Basically, it is a physical property of all LED's that efficacy increases when underdriven. Current regulation is underdriving the LED by definition; whereas PWM drives at full rated current during the on portion of the duty cycle.

- as PWM frequency goes up, the ability to notice it goes down (which is obviously a good thing); however, there is a tradeoff: as PWM frequency goes up, converter efficiency goes down. If the designer is trying to make as much light and as little heat as possible, then he wants a relatively low PWM frequency. The trick is to strike a happy balance between a higher PWM frequency and reduced converter efficiency.

- Current regulation is typically controlled by a feedback resistor which sets the output current of the regulator IC. Varying this resistor value will provide varying current. But to maximize regulator efficiency at high light output, the resistor values need to be very small, fractions of an ohm, which makes it difficult or impossible to accomplish varying resistor value with a potentiometer. (The vast majority of small, cost-effective potentiometers do not adjust down as low as fractions of an ohm.) Therefore, you sometimes see current regulation accomplished by using a bank of several discrete feedback resistors, which are selected via firmware or various other means. This gives a light with a selection of discrete output levels rather than a continuous output adjustment.

- PWM very simply and readily allows use of a potentiometer in the timer circuit, therefore it is the design choice when fully variable dimming is preferred over a few discrete levels.

- I seem to recall that Newbie's data also showed that tint-shift was only slightly better with PWM.

I have found that my personal preferences for PWM driven lights is a drive frequency of 1KHz or higher. That seems to be where I no longer can notice the flicker. I have been playing around with PWM for my home lighting project and currently have 2 designs that I like a lot. One using a 555 timer and which has a frequency that varies between 1.1-1.5 KHz. I also have a PSoC micro consoler design that runs a 5 KHz and is great. I was concerned about EMI with that high of a frequency switching 4 amps through hundreds of feet of wire but I have yet to notice any ill effects in other electronic devices.

Brlux said:

I have found that my personal preferences for PWM driven lights is a drive frequency of 1KHz or higher. That seems to be where I no longer can notice the flicker. I have been playing around with PWM for my home lighting project and currently have 2 designs that I like a lot. One using a 555 timer and which has a frequency that varies between 1.1-1.5 KHz. I also have a PSoC micro consoler design that runs a 5 KHz and is great. I was concerned about EMI with that high of a frequency switching 4 amps through hundreds of feet of wire but I have yet to notice any ill effects in other electronic devices.

Click to expand...

Hi Brlux:

Frequencies over 1KHz can be very difficult to achieve with the typical IC's used for LED regulator circuits in battery-powered flashlights. In the case of our FlashFlood we're forced to keep the frequency well below 1KHz due to the IC constraint.

- Current regulation is typically controlled by a feedback resistor which sets the output current of the regulator IC. Varying this resistor value will provide varying current. But to maximize regulator efficiency at high light output, the resistor values need to be very small, fractions of an ohm, which makes it difficult or impossible to accomplish varying resistor value with a potentiometer. (The vast majority of small, cost-effective potentiometers do not adjust down as low as fractions of an ohm.) Therefore, you sometimes see current regulation accomplished by using a bank of several discrete feedback resistors, which are selected via firmware or various other means. This gives a light with a selection of discrete output levels rather than a continuous output adjustment.

Click to expand...

You can very easily do continuously variable constant current regulation with a potentiometer. You don't need to modify the sense resistor, you use a potentiometer to modify the reference voltage. I've built many constant current regulators with continuously variable output with this method - it's a very basic electronic circuit - a reference voltage that's changed with a pot, a small value sense resistor, and an op-amp controlling a pass transistor. I can vary the output from 25mA to 1A with this method.

Hi evan9162:

PM sent.

soffiler said:

- PMW is less somewhat less efficient than current regulation at levels below 100%. I remember a Newbie post on this subject which presented the evidence clearly in graphical form. Basically, it is a physical property of all LED's that efficacy increases when underdriven. Current regulation is underdriving the LED by definition; whereas PWM drives at full rated current during the on portion of the duty cycle.

Click to expand...

Agree. The LED's are least efficient when driven at the high end of their capability.

soffiler said:

- as PWM frequency goes up, the ability to notice it goes down (which is obviously a good thing); however, there is a tradeoff: as PWM frequency goes up, converter efficiency goes down. If the designer is trying to make as much light and as little heat as possible, then he wants a relatively low PWM frequency. The trick is to strike a happy balance between a higher PWM frequency and reduced converter efficiency.

Click to expand...

The current regulators are nearly always a boost topology switch-mode regulator (a PWM circuit) operating at 100Khz on up to 2Mhz. As long as the on/off PWM brightness control is well below the switch frequency, and it is well designed (*important*), I have seen the PWM frequency make very little (almost no) difference to boost converter efficiency (I'm just finishing a project that drives about twenty LED's that are PWM'd) and don't see why it should. A less well designed one can have nasty transients, overshoot and ringing during the switch that might cause some loss, but if it is well designed for this purpose, this shouldn't be [much] of an issue.

soffiler said:

- Current regulation is typically controlled by a feedback resistor which sets the output current of the regulator IC. Varying this resistor value will provide varying current. But to maximize regulator efficiency at high light output, the resistor values need to be very small, fractions of an ohm, which makes it difficult or impossible to accomplish varying resistor value with a potentiometer. (The vast majority of small, cost-effective potentiometers do not adjust down as low as fractions of an ohm.) Therefore, you sometimes see current regulation accomplished by using a bank of several discrete feedback resistors, which are selected via firmware or various other means. This gives a light with a selection of discrete output levels rather than a continuous output adjustment.

- PWM very simply and readily allows use of a potentiometer in the timer circuit, therefore it is the design choice when fully variable dimming is preferred over a few discrete levels.

Click to expand...

Pretty much. You can get just a bit exotic and get around this, but it's not particularly easy and it will give up at least some efficiency.

evan9162 said:

You can very easily do continuously variable constant current regulation with a potentiometer. You don't need to modify the sense resistor, you use a potentiometer to modify the reference voltage. I've built many constant current regulators with continuously variable output with this method - it's a very basic electronic circuit - a reference voltage that's changed with a pot, a small value sense resistor, and an op-amp controlling a pass transistor. I can vary the output from 25mA to 1A with this method.

Click to expand...

You can do this, but as soffiler mentioned, it's not very efficient. For high efficiencies, these sense voltages are low, sub-bandgap. You can't go lower with a particular device. You can go higher at the expense of efficiency.