PWM - What is it, How does it work and how to detect it.

[jon_slider]

The Tikka is a PWM light. You call it unregulated. I find that confusing only when you call the LOD regulated and PWM in the same breath.

anyway, when Fenix says digitally regulated now, they mean no PWM, I think we can agree.

Note the use of pwm as opposed to current controlled regulation in this self built review:
http://www.candlepowerforums.com/vb/showthread.php?321587
"No PWM/Strobe
Here's a nice find – there is no sign of PWM on the Lo mode of the Worm.

It thus seems to use current-control for its low mode.
…
Current-controlled circuit performance is excellent – very flat regulation, and top-of-class runtimes at both levels."

 

[Cataract]

The Tikka is a PWM light. You call it unregulated. I find that confusing only when you call the LOD regulated and PWM in the same breath.

anyway, when Fenix says digitally regulated now, they mean no PWM, I think we can agree.

Let me clarify things a bit for you. I believe your confusion is coming from the fact hat 2 concepts are intermingled here, something I did not quite realize when I started this thread.

There are 2 concepts that should be distinguished from each other, namely "output regulation" and "mode regulation".

"Output regulation", short for "Constant output regulation", means the number of lumens coming out from the LED should be constant over a period of time until the battery(ies) can no longer give out enough energy to maintain the selected output.

"Mode regulation" or "mode control" do not necessarily mean that the lumens output will trace an exact flat line over time. Mode regulation can be achieved through current control (sucking a proportional number of amps from the battery to feed the LED the appropriate number of amps for a specific output) or through PWM (it is simpler to send you back to my first post graphs than to elaborate again on this one.) In theory, current control should give a very constant result like a flat line (if we ignore varying qualities in circuitry).

PWM alone does not mean the output will be constant in intensity; PWM switches the LED on and off at maximum achievable power with various on-to-off ratios for each mode (on for less time and off for more time to achieve lower modes; just refer to the graphs in my original post). If the manufacturer did not include circuitry to compensate the fact that batteries lose power as they discharge, you can have a PWM light that does not have a constant output since the maximum output power will diminish over time. AFAIK, PWM controlled lights pretty much all include some form of output regulation - else I doubt they would still sell their lights. However, since the entire goal of PWM is simplicity and cheapness of the circuits, most PWM controlled lights have less-than-ideal output regulation circuits.


Let me know if this explanation clears things up for you and which part is not clear enough. I'll be glad to help and this would also help me update my original post to make things as clear as possible for Mr. everybody. I'm also open to hear from other experts to correct me if I made a mistake. I'm not looking into making a Wikipedia-worthy entry, just a good description that can be understood by the Mr. Everybodies out there.

 

[masterP]

the foursevens mini 123 is PWM

http://www.laymanslights.venturous.org/blog/2010/01/4sevens-q-mini-123/

 

[masterP]

my email response from Fenix.....



Dear customer,

Thank you for supporting Fenix.

Fenix products are Current Control reglated.

Best regards.

Linda Yao
Fenixlight Limited
8/F, 2nd Building,
DongFangMing Industrial Center
33rd District, Bao'an,ShenZhen 518133 China
0086-400 886 6093
www.fenixlight.com
www.facebook.com/fenixproducts

 

[Hondo]

The Tikka is a PWM light. You call it unregulated. I find that confusing only when you call the LOD regulated and PWM in the same breath.

Thanks Cataract for the detailed explanation. I was just going to say that use of PWM to control brightness levels does not ensure regulation of brightness throughout the discharge of the battery. It does not preclude it either, the above two lights are an example of both conditions. And while I don't have an example I can think of, I am sure it is possible to control brightness with a current controlled circuit and fail at achieving flat regulation as well.

 

[jon_slider]

the foursevens mini 123 is PWM

http://www.laymanslights.venturous.org/blog/2010/01/4sevens-q-mini-123/

Total respect for your research and for providing a link as supporting evidence.

Fenix products are Current Control reglated.

Well researched! Clearly the word Regulated as used by Fenix means Current Controlled, not Pulse Width Modulated.

 

[reppans]

I agree with Cataract and Hondo... PWM vs Current Control and Regulated vs Unregulated are independent of each other, you can have all four combinations. Two examples of current controlled, yet unregulated, are the ET D25A and 47 Atom on moonlight mode - both outputs will vary ~5x depending upon voltage (ie, new 1.7v L91 to dead 1.2v NiMh). My Malkoff MDC AA is PWM light, and it's moonlight mode also displays the same voltage sensitive, although it's other modes seem well regulated.

Some may actually claim all CC, but still have some PWM, like this ET D25A on 3lms on the left. On the right is a 47s Mini (PWM is spec'd) - these are the two fastest PWM lights I've seen - this sweep is as fast as I can move my arm.

Btw, 47s uses PWM across the Mini line and Preon P1 and P2, but otherwise is CC AFAIK (although I know Hondo has, and certain other early samples I've seen, Quarks with PWM warble effect). I have quite a few (I suppose later model) Quarks and they are all nicely CC and well regulated. My D25A also has that warble effect on moonlight at low V, pics posted earlier in this thread.

Don't own any Fenix lights, but the definitive way to tell is a fast sweep across a DSLR on time exposure, side-by-side with a known fast PWM light, and a true CC light (as above). My own eye-ball detection method is catch the spill in a reflection/mirror and move my eyeballs left to right quickly - I can quite easily see the PWM on the 47s Mini, but it is very hard to detect it on the D25A.

 

[hiuintahs]

............I don't have an example I can think of, I am sure it is possible to control brightness with a current controlled circuit and fail at achieving flat regulation as well.

Well, not entirely true if designed properly. The nature of LED's is that their output is constant per current at a given junction temperature

.

Current controlled is sensing the current through the LED and translating that into a voltage feedback signal back into the driver that then regulates or holds the Vf across the LED such that it holds a constant current. Light output will be pretty flat with a slightly down slope as the junction temperature of the LED heats up. It all depends on how hard the LED is being driven. Low and medium modes will be very flat over time if done right.

As I have seen some reiterate, the confusion that I see is that the term "regulated" means constant output. That can be constant voltage or constant current as in ideal power supplies or constant brightness as in the output of an LED. If the circuit is regulated, then there is a feedback signal to the driver. The problem with PWM is that unless, the circuit compensates by increasing the duty cycle of the pulse width as the battery voltage drops, you will see the output drop with depletion of the battery. Constant light output in a PWM system would require feedback of some type and more complexity thereby defeating the simplicity of PWM........then might as well go with current controlled.

To me I don't care how they derive the driver, the winner is the one with the most efficient and flat line output over the time the battery discharges. Once I got a data logging light meter, built a light box patterned after Selfbuilts, this hobby took on new meaning as not all lights are created equal.

 

[Cataract]

Thanks Cataract for the detailed explanation. I was just going to say that use of PWM to control brightness levels does not ensure regulation of brightness throughout the discharge of the battery. It does not preclude it either, the above two lights are an example of both conditions. And while I don't have an example I can think of, I am sure it is possible to control brightness with a current controlled circuit and fail at achieving flat regulation as well.

1- As far as I can tell, good manufacturers that use PWM for mode control, also include some form of output regulation. How they perform on output regulation does depend on the quality of the circuitry. Some are so bad the output curve pretty much parallels a battery output curve and it would only be normal to question if the even try to keep the output constant.

2- Most fail to achieve absolute flat output regulation with current control to varying degrees; it is a very tricky thing to come up with an absolute flat line like Fenix does (just check out Selfbuilt's collection of output graphs). That said, I don't really mind if a light can give me at least ~75-80% flat output regulation since I will probably not notice it even if I try hard. Sometimes other factors are more important that a brick solid output intensity.

I agree with Cataract and Hondo... PWM vs Current Control and Regulated vs Unregulated are independent of each other, you can have all four combinations. Two examples of current controlled, yet unregulated, are the ET D25A and 47 Atom on moonlight mode - both outputs will vary ~5x depending upon voltage (ie, new 1.7v L91 to dead 1.2v NiMh). My Malkoff MDC AA is PWM light, and it's moonlight mode also displays the same voltage sensitive, although it's other modes seem well regulated.

Some may actually claim all CC, but still have some PWM, like this ET D25A on 3lms on the left. On the right is a 47s Mini (PWM is spec'd) - these are the two fastest PWM lights I've seen - this sweep is as fast as I can move my arm.

Btw, 47s uses PWM across the Mini line and Preon P1 and P2, but otherwise is CC AFAIK (although I know Hondo has, and certain other early samples I've seen, Quarks with PWM warble effect). I have quite a few (I suppose later model) Quarks and they are all nicely CC and well regulated. My D25A also has that warble effect on moonlight at low V, pics posted earlier in this thread.

Don't own any Fenix lights, but the definitive way to tell is a fast sweep across a DSLR on time exposure, side-by-side with a known fast PWM light, and a true CC light (as above). My own eye-ball detection method is catch the spill in a reflection/mirror and move my eyeballs left to right quickly - I can quite easily see the PWM on the 47s Mini, but it is very hard to detect it on the D25A.

I found a very old post from 47's (before the mini line) that stated they did not use PWM in their lights, despite the light showing signs of it when you move it as fast as you can. They said that this is actually an artefact from the circuits used. When viewing a graph of the current output at the LED, it became obvious that the time on - time off proportions never changes, but there was a small dip at regular intervals on the graph. In a sense, it is not actual PWM, but does behave somewhat the same (it should be much, MUCH easier to detect on lower modes, but isn't really that much easier to detect). The frequency of that dip is very high though, and it is hard to find as you must have experienced yourself. I don't know if the same really applies to the quark mini line, but I can say that none of my quarks bother me with signs of PWM and I have not heard from anyone saying it does. It does look the same and I thought it was actual PWM until I read their post. Now I give them the benefit of the doubt.

 

[Cataract]

Well, not entirely true if designed properly. The nature of LED's is that their output is constant per current at a given junction temperature .

Current controlled is sensing the current through the LED and translating that into a voltage feedback signal back into the driver that then regulates or holds the Vf across the LED such that it holds a constant current. Light output will be pretty flat with a slightly down slope as the junction temperature of the LED heats up. It all depends on how hard the LED is being driven. Low and medium modes will be very flat over time if done right.

As I have seen some reiterate, the confusion that I see is that the term "regulated" means constant output. That can be constant voltage or constant current as in ideal power supplies or constant brightness as in the output of an LED. If the circuit is regulated, then there is a feedback signal to the driver. The problem with PWM is that unless, the circuit compensates by increasing the duty cycle of the pulse width as the battery voltage drops, you will see the output drop with depletion of the battery. Constant light output in a PWM system would require feedback of some type and more complexity thereby defeating the simplicity of PWM........then might as well go with current controlled.

To me I don't care how they derive the driver, the winner is the one with the most efficient and flat line output over the time the battery discharges. Once I got a data logging light meter, built a light box patterned after Selfbuilts, this hobby took on new meaning as not all lights are created equal.

A "slightly" more advanced detailed explanation than I meant to see here when starting the thread, but very well explained (at least, for those who can follow). The feedback loop is what creates that small high frequency dip on the Quarks output IIRC. I might have learned a thing or two from your post...

Your post started with the answer to the riddle: circuit quality differs and some (most, really) fail to achieve absolute flat output and my belief is that has more to do with not being able to completely compensate for the battery discharge state since that is the core of the challenge (but I could be wrong). Different reasons could help explain this: not enough research, trying to keep the cost down, trying to minimize the size of the driver, etc. I think others might also just go with the flow so they can state longer runtimes on the box.

I think "output regulation" or "regulated" are, in general, used by manufacturers to mean that the average output intensity is better than just connecting the battery(ies) straight to the LED and I wouldn't say it is a deception in itself to claim so; they did do something. I'm not arguing that flat output isn't ideal, but the word "regulation" as used in the industry IMO simply means that circuitry was used to give a more constant output than doing nothing. Some are just so much better than others at it and can simply stand on their reputations without needing to flip backwards to explain it in more details than it should be needed to. I still do love my less-than-ideal current controlled lights so long as I don't notice the difference as I'm using them

 

[jon_slider]

I was unable to detect the PWM of my Maratac Rev 3 by waving the light, video, or still photo. However I can see the PWM strobe effect very slightly on a fan, or on running water.

some discussion that includes the Constant Color benefits of PWM
http://www.candlepowerforums.com/vb/showthread.php?331102-Please-comment-Quark-Mini-AA-quot-PWM-quot

The Quark's use a very high frequency PWM - it's extraordinarily unlikely that your eyes will notice that it's there.

On a HCRI light like the one you purchased, you should be grateful that the levels are PWM-controlled. Constant current drivers cause noticeable tint shift as you decrease the output, meaning you might get an ugly orange instead of a creamy white. Any PWM-generated levels will have the same tint as full brightness.

Probably, though the amount of shift varies by each emitter so you might wind up with a model where it's not even noticeable. It's pretty random, but I don't like rolling the dice on whether it turns green at a low setting.

I'm personally of the opinion that if I'm buying a light for it's premium tint (like a HCRI) then I want that tint at all levels, even if it means losing some runtime. McGizmo's Sundrop or HCRI Haiku is a good example of a light where I'm glad he went with a PWM driver.

The Prometheus Beta Copper, which has a N219 and a low of 1 lumen, also uses PWM.

Just playing devils advocate so we can be aware that not all PWM is slow, bad, or limited to cheap lights.. PWM has benefits in terms of color fidelity on low levels.

 

[reppans]

To avoid confusion, 47s separated the Quark line from the Mini line - earlier on the Mini was called "Quark Mini." 47s does specify "PMW" for the Mini, and "Current Regulation" for the Quarks on its website.

I found a very old post from 47's (before the mini line) that stated they did not use PWM in their lights, despite the light showing signs of it when you move it as fast as you can. They said that this is actually an artefact from the circuits used. When viewing a graph of the current output at the LED, it became obvious that the time on - time off proportions never changes, but there was a small dip at regular intervals on the graph. In a sense, it is not actual PWM, but does behave somewhat the same (it should be much, MUCH easier to detect on lower modes, but isn't really that much easier to detect). The frequency of that dip is very high though, and it is hard to find as you must have experienced yourself. I don't know if the same really applies to the quark mini line, but I can say that none of my quarks bother me with signs of PWM and I have not heard from anyone saying it does. It does look the same and I thought it was actual PWM until I read their post. Now I give them the benefit of the doubt.

All my later model Quarks are very smooth (and have very flat output "regulation," aside from Max), and I love them, but I bought a used Titanium 123^2 (circa 2010 or 11?) that has very noticeable PWM-like behavior (or artifacts). Haven't take a picture of it yet, but I suspect it's the same as what Hondo posted here (post #115) and what Ti-force posted here (I posted smooth samples):

http://www.candlepowerforums.com/vb...Quark-Tactical-QT2L-X-Burst-Mode-review/page3

 

[reppans]

.... PWM has benefits in terms of color fidelity on low levels.

I know this is conventional wisdom but don't see the current regulation tint shift thing. This is an old photo of some lights on moonlight (top) and something around 100 lms (bottom) - camera is on auto white balance and auto exposure (-2? exposure comp) for both - can anyone tell which are current regulated and which are PWM?

I also wonder about current regulation efficiency thing too.. I test my lights for output and runtime on the sub-/low-lumen modes I use most (with a lightbox). My PWM Malkoff MDC AA tested about double the lumen-hours on moonlight mode vs a bunch of current regulated lights in a side-by-side test (note though, the MDC is not particularly well "temperature regulated" ).

 

[Cataract]

Thanks for the info. I can understand they wanted to keep it simple early on and I definitely am not bothered by PWM on my mini's. I will admit that I'm not sure at which stage of the line's evolution, but I personally don't mind if a light does not bother me and I don't plan on putting myself in a situation where it could make a difference, like trying to stop the fast moving blades of a huge propeller

 

[reppans]

Yeah the Minis are generally fast enough to be hard to visually notice... but that line's efficiency, unfortunately, takes quite a hit .

 

[Cataract]

@ reppans: tint shift VS current input (or difference in tint on lower modes) seems to be much more obvious on modern high power LED's from what I read around here, but I admit that I barely can see it myself.

By "temperature regulated", do you mean the light gets hot or the tint shifts a lot?

 

[reppans]

The bulk of my collection are EDC AA/14500 lights so I'm usually below ~400 lms @ Max - so I guess I'm not really in the "high powered" class .

Temperature regulation is an weird MDC thing - its moonlight output will vary 4x from 0.15 to 0.6 lms depending upon the temperature of the head (from >80F to <60F, respectively). So while it doubled the lm-hrs efficiency of my CR lights @60F (in a chilly attic), it might have also come in half as efficient if tested during the summer heat.

 

[Cataract]

Interesting...

400 lumens is still a lot more than you need unless you're in deep, deep doo-doo IMO.

That temperature related regulation issue is kinda weird and not so weird for someone who understands electronics to a deep level. Temperature related resistivity is definitely a subject for another thread, although it is worth mentioning in here just to get people to understand how many factors can enter the equation.

 

[SemiMan]

I also wonder about current regulation efficiency thing too..

LEDs are more efficiency current dimmed down to about 5% of maximum. Below that, you are better to drive them at 5% and PWM them. As well, your circuitry is likely to be more efficient this way too.

Semiman

 

[Cataract]

{disclaimer for all readers}:


I realize how the latest posts seem like they might not be directly related to the original intent of this thread/sticky, but, as this thread is due to mature, I welcome them as an introduction to how wide the entire world of flashlight "regulation" really is.

I also want to mention that there are no reasons (other than subjective preference ) to reject a potential light based simply on any single spec alone; If it's in your hands and allows you to see and do what you need to do without annoyance or hindrance, this light is your best friend in the situation you are in. That is my pure and honest opinion.

When I started this thread, I simply wanted to help people understand the element of PWM (Pulse Width Modulation) since I saw many member struggling to understand what it is at the time I wrote the original post. I have been made to feel I achieved this purpose with all the positive feedback and I thank you all for letting me know I helped in understanding the phenomenon. All "thank you" notes are definitely warmly welcomed. These let me know I have achieved my goal and definitely encourage me in making an effort in keeping this thread up to a certain standard. Such remarks also inspired me in writing an article at work on basic lithium-ion battery care, which I will likely share here in the near future. It is my turn to thank you for letting me know I was able to help and for the strength you gave me to put the extra effort needed. I did get good words from my boss and I feel I owe it to you all here and to those who gave me feedback in my "basic flashlight troubleshooting guide" thread.

I now hope the extra information contained in the later posts may help the more curious people and followers to understand how much more is really involved in making a well made "regulated" light and that PWM is not necessarily evil. I do my very best to keep things simple or, at least, interpret what is being shared with as much simplicity as I can as people share more directly or indirectly relevant and expansive information on the core of the subject of LED "regulation". This is not easy to do as a thread is fluid and is built on as many points of view as there are participants. To top that off, I also do tend to many other hobbies at the same time (and try to have a life too )


A BIG thank you to all that have shared from post 2 to this post and I feel this is all going on very well.