If its not PWM, nor Ripple, is it Constant Current Circuit Noise Flicker?

Recently I have learned that Constant Current lights can have flicker. Not all flicker is visible to the naked eye. The following images are links to the original source, but the image itself is hosted on my photo site, so as not to violate any hot linking rules.


http://www.hdssystems.com/Articles/P...lashlights.pdf
There are two different types of resolution errors that are important to us. The first is the absolute output error. As long as the absolute output changes slowly over time an absolute output error of 10% has very little visual affect - i.e., the user will never see the error. The second type of error is the inter-voltage error. This is the suddenchange in output voltage for a very small change in input voltage. The inter-voltage error results in a sudden change in light output which is much easier for the eye to detect. The inter-voltage errors must be kept small and should result in less than a 5% change in light output to keep the user from noticing.


HDS



Muyshondt Beagle



Manker E02

Re: Its not PWM its Constant Current Circuit Noise Ripple

Those runtime graphs have nothing to do with ripple. Ripple doesn't mean "wavy line in the output," it has a technical meaning: the remnants of incomplete suppression of an AC waveform when converting AC to DC. Ripple is observable on the time scale of microseconds, not minutes. The output level of a light can fluctuate for any number of reasons but temperature is the big one. Not just the light either, uncompensated light meters can be extremely sensitive to temperature. I've got a few (unpublished) runtime plots that were not usable because the normal temperature changes in the room showed up as 10% swings in the output. (A better luxmeter fixed that.)

The Beagle might be ripple. I'd need to know the full settings of the camera. Vertical lines generally mean extremely high speeds but it is rare to have them be perfectly aligned like that. There is also the technique used by some AMC7135 linear lights where it alternates between two brightnesses. Technically this is still PWM but with a DC offset. It is much easier on the eyes than normal on-off PWM.

What does it take to produce detectable ripple? Heavier loads will exagerate ripple but practically speaking it is easier to observe at low loads. First is the frequency of the SMPS. The ripple will (under normal circumstances) be that frequency. A 1MHz SMPS will have invisible ripple.

Under low loads, many SMPS start to cheat a little in order to dramatically improve efficiency. For example, they might start pulse skipping or reduce the operating frequency. This might reduce the frequency enough to make the ripple observable. If the duty cycle is decreased enough the SMPS can enter discontinuous mode. This can offer dramatic power savings but the "discontinuity" refers to the fact that no current is flowing through the regulator. In an LED, this will appear as the LED completely turning off. A large enough capacitor can smooth this out but space as at a premium and it might smooth out the strobe mode too. Superficially this will look similar to PWM except that the LED output will follow a triangle wave instead of a square wave.

Some Zebralights have detectable ripple on the lower modes. It is very hard to design a SMPS that works perfectly across a 1:100000 output range! They have the best electrical design of any flashlight, but ZL has to work with an extremely small PCB.

The only feasible way to measure ripple is to connect an oscilloscope directly to the LED. Or use an extremely fast light sensor (with sub-microsecond response time) connected to an oscilloscope. Runtime plots will not show anything useful. A photo might indicate that there is something interesting going on, but you still need the 'scope before drawing any conclusions.

All due respect to his genius, but Henry is making up words. "Absolute output error" is more commonly called stability. Thermal stability is the major one, a typical spec might be 2% variation across a 0C-100C temperature range. "Inter-voltage error" is more commonly called line regulation. (The other major spec for regulated PSUs is the load regulation, but that doesn't really apply to flashlights since the load is a fixed LED.)

If you want to get really technical, you can view a regulated power supply as a special class of op-amp. Some regulators will then specify their unity gain bandwidth, open loop voltage gain, DC transconductance, harmonic suppression and phase response.

Re: Its not PWM its Constant Current Circuit Noise Ripple

Here's some more background: https://en.wikipedia.org/wiki/Ripple_(electrical)

Re: Its not PWM its Constant Current Circuit Noise Ripple

I have found my recently purchased Duracell brand flashlight to be making high-pitched noises. I'm pretty sure it's from the voltage regulator, however I have not taken it apart to find out yet. I found the light on sale at Home Depot for $19.88. It also came with a smaller light emitting 250 Lumens. What do you all think is making noise inside the circuitry?

Re: Its not PWM its Constant Current Circuit Noise Ripple

M4GL173M0DD3R said:

noise inside the circuitry?

Click to expand...

Welcome to the forum.
it is probably normal for that model.
none of my flashlights make whining noises... not all brands do it.
Its your choice to live with it or return it.

here are some opinions from other people that own lights that whine

Re: Its not PWM its Constant Current Circuit Noise Ripple

M4GL173M0DD3R said:

I have found my recently purchased Duracell brand flashlight to be making high-pitched noises.... What do you all think is making noise inside the circuitry?

Click to expand...

My guess would be ...

https://en.m.wikipedia.org/wiki/Coil_noise

Re: Its not PWM its Constant Current Circuit Noise Ripple

According to 4sevens, constant current always has ripples, its normal

4sevens said:

There are small ripples from the regulation which actually exist in all constant current drivers.

Click to expand...

Here is another Constant Current light that shows ripples (I think in this case due to thermal management, possibly similar to HDS)
pic, hosted on my photo site, is a link to the original source

Re: Its not PWM its Constant Current Circuit Noise Ripple

M4GL173M0DD3R said:

I have found my recently purchased Duracell brand flashlight to be making high-pitched noises..........

Click to expand...

Inductor whine caused by instability from poorly designed feedback of switching type of regulator circuit. It doesn't mean that a failure is immanent but is indicative of cheap design..........or I'll give the benefit of the doubt........perhaps something simply went out of whack with one of the components and its a one time problem. As an electronics guy, I expect better and would get rid of it if it was me.

jon_slider said:

According to 4sevens, constant current always has ripples, its normal.........

Click to expand...

Yes and no. Yes, if you put the output on an oscilloscope, you can always see a certain amount of ripple on every switching power supply type of regulator. The amount depends on a number of factors such as output capacitance and switching regulator oscillator frequency. However the ripple should always be low enough that you cannot detect it with your eyes or even with your data logging light meter...........thus it would look like a perfect current regulated output. I tend to lose interest in flashlights that cannot achieve that.

To me the art in a flashlight is not simply the mechanical look to it but also the electrical driver inside it. On a side note, Fenix does a pretty good job with their regulator circuits..........except once I believe they had a manufacturing assembly mistake with the LD12 XP-G2 2nd gen light. They of course denied it but there was definitely circuit noise on two separate units that I had. I returned them and didn't get back into an LD12 until the 2016, NW model came out. It's a totally different driver circuit and thus the problem gone.
http://www.candlepowerforums.com/vb...or-runtime-chart-(no-PWM-just-noisy-circuit-)

Re: Its not PWM its Constant Current Circuit Noise Ripple

(my bolding added)

hiuintahs said:

Yes and no. Yes, if you put the output on an oscilloscope, you can always see a certain amount of ripple on every switching power supply type of regulator. The amount depends on a number of factors such as output capacitance and switching regulator oscillator frequency. However the ripple should always be low enough that you cannot detect it with your eyes or even with your data logging light meter...........thus it would look like a perfect current regulated output. I tend to lose interest in flashlights that cannot achieve that.

Click to expand...

thanks for the food for thought

selfbuilt said:

it is not PWM, but it is potentially detectable as a perceptible visual flicker ... I recommend you scroll down to post #2 for more info on what PWM actually is.
...
a current-controlled light – but one with an oscillating signal that may potentially be visible to some people

Click to expand...

I agree, things that the oscilloscope can show, may or may not matter to the user of the flashlight.
For many people, the decision is based on whether they notice any flicker, in their actual use.

Re: Its not PWM its Constant Current Circuit Noise Ripple

hiuintahs said:

Inductor whine caused by instability from poorly designed feedback of switching type of regulator circuit.

Click to expand...

This would be about the most unlikely reason for the whining. Likely reasons would include:

- PWM slamming inductors/ceramics at audible frequencies
- Dimming mode that reduces the switching frequency into the audible range
- Switching topology that when dimming goes into burst mode

A cheap design? ... not necessarily, but poor for the usage and likely fixed with some potting or silicon on the inductor.

Re: Its not PWM its Constant Current Circuit Noise Ripple

This is how I understand inductor whine within the confines of a switching power supply. The audible is coming from inductor vibration which is caused by erratic current in the inductor. Usually a constant PWM and uniform switching will not be heard.............because those frequencies should be higher than the audible range. It's when you get jitter where the PWM frequency is changing and isn't stable and thus can cause harmonics within the audible range. When inductor current changes erratically you get a magnetic field that changes and its this magnetic field that is causing the inductor to vibrate. Of course this vibration is very small. If the vibration is big enough and within the audible range you hear it. I believe this to be caused by a term called jitter where the frequency of the PWM control signal to the switching mosfet is erratic and not maintaining a fixed frequency. The jitter is caused by improper board layout or improper feedback stability values. At least that is my opinion. I've had designs where I've had to work on getting the jitter out. Swithcing power supply design and associated board layout requires a bit of expertise.
http://www.ti.com/lit/an/slua747a/slua747a.pdf

I test every one of the USB charging adapters I get. You can get some of the very same brand and model (ie: Samsung) and some you can hear and others you cannot. I'm not sure how to determine if one with whine is more or less reliable than one that isn't.........just that I know that if you cannot hear it, then that at least eliminates a possibility of any jitter being present...........

Re: Its not PWM its Constant Current Circuit Noise Ripple

With the exception of "If the vibration is big enough and within the audible range you hear it", I would consider this pretty much all wrong. All that is needed in order hear noise in a power supply is for some switching element to be switching in the audible range. Again, that could happen in several ways:

1) Overall PWM (i.e. for dimming). This could be a switching supply that is PWMed for brightness control, essentially turning the whole supply on/off in the audible range, but could even be a linear supply that is PWMed. Some capacitors can generate audible noise.

2) The primary switching frequency for the switching power supply drops into the audible range. This could be caused by a topology that is using discontinuous conduction at a low frequency when dimming, or using a burst mode at low current (essentially internal overall PWM), etc.

None of these situations is an indication of bad design, a manufacturing failure, etc. Variations in winding, vacuum impregnation, etc. can cause the level of noise generated by inductors/transformers across a manufacturing batch to be different, and a conscious design decision can be made to make the switching frequency drop slightly into the audible band, and how far will vary with component tolerance which again could be amplified by variances in inductor/transformer construction.

In your case of USB chargers, especially today with the very low standby requirements, burst mode operation is almost a given as it is the only way they can meet the very low standby requirements. Depending on the mechanical construction, a little bit of noise, probably worse when lightly loaded, is possible and for the reasons stated above, some may be audible to you, and some may not, though they both are likely making noise if the same design.

In general, inductor noise is not a concern for reliability.

hiuintahs said:

This is how I understand inductor whine within the confines of a switching power supply. The audible is coming from inductor vibration which is caused by erratic current in the inductor. Usually a constant PWM and uniform switching will not be heard.............because those frequencies should be higher than the audible range. It's when you get jitter where the PWM frequency is changing and isn't stable and thus can cause harmonics within the audible range. When inductor current changes erratically you get a magnetic field that changes and its this magnetic field that is causing the inductor to vibrate. Of course this vibration is very small. If the vibration is big enough and within the audible range you hear it. I believe this to be caused by a term called jitter where the frequency of the PWM control signal to the switching mosfet is erratic and not maintaining a fixed frequency. The jitter is caused by improper board layout or improper feedback stability values. At least that is my opinion. I've had designs where I've had to work on getting the jitter out. Swithcing power supply design and associated board layout requires a bit of expertise.
http://www.ti.com/lit/an/slua747a/slua747a.pdf

I test every one of the USB charging adapters I get. You can get some of the very same brand and model (ie: Samsung) and some you can hear and others you cannot. I'm not sure how to determine if one with whine is more or less reliable than one that isn't.........just that I know that if you cannot hear it, then that at least eliminates a possibility of any jitter being present...........

Click to expand...

Re: Its not PWM its Constant Current Circuit Noise Ripple

I like your other reasons. Sounds like you are pretty knowledgeable in this area. For me, I've only encountered inductor whine associated with excessive jitter during the development phase of a switching power supply which had instability and jitter. So I equate it to poor design. I haven't used all forms of topology. So I agree that it can happen under normal designed circumstances and will not be a concern for reliability.

Re: Its not PWM its Constant Current Circuit Noise Ripple

You are likely misusing the term jitter. Jitter is by its definition random. Random perturbations do not create non-random noises. What you mean is unintended oscillations caused by stability issues either from poor circuit design and/or poor layout. In a completed and shipping product, that is likely to be rare or you would likely have stability issues across the manufacturing distribution likely leading to flicker and/or it just not working.

Re: Its not PWM its Constant Current Circuit Noise Ripple

HDS is known for whining, but does not use PWM.

Circuit "noise" from Constant Current, is not necessarily audio noise, it shows as waves on a scope, per the selfbuilt post I quoted in post #9, and can also create visible flicker (per selfbuilts use of terms).

The interference in the photo of the Muyshondt in post #1, is caused by Constant Current, not PWM.

None of the circuit noise, flicker, and runtime ripples I am reporting in this thread, are from PWM.

Re: Its not PWM its Constant Current Circuit Noise Ripple

This is not what 4sevens meant by ripple. The lack of smoothness in the graph is as likely to be measurement issue as an actual change in light level.

jon_slider said:

According to 4sevens, constant current always has ripples, its normal



Here is another Constant Current light that shows ripples (I think in this case due to thermal management, possibly similar to HDS)
pic, hosted on my photo site, is a link to the original source

Click to expand...

Re: Its not PWM its Constant Current Circuit Noise Ripple

ssanasisredna said:

This is not what 4sevens meant by ripple. The lack of smoothness in the graph is as likely to be measurement issue as an actual change in light level.

Click to expand...

In the use of your own words...."I'd pretty much consider this all wrong" . It's not likely a measurement issue, it's a graph of the actual light level that the data logging light meter is picking up. The sample level is fairly course such as every 2 seconds or thereabouts and thus data sampling just so happens to correspond to a high spot or low spot based on the higher speed ripple voltage across the LED which Jon is referring to as circuit noise ripple. Your eyes can't see the difference in output, but the meter can pick it up. I've tested some lights that were a lot worse than this and I still could not detect it with my eyes. But from a purist standpoint, I've seen some outputs that were very constant and solid too.

Re: Its not PWM its Constant Current Circuit Noise Ripple

hiuintahs said:

Your eyes can't see the difference in output, but the meter can pick it up. I've tested some lights that were a lot worse than this and I still could not detect it with my eyes. But from a purist standpoint, I've seen some outputs that were very constant and solid too.

Click to expand...

thank you
I feel you understand what Im trying to talk about
fwiw, both HDS and Boss use Thermal management, and Imo the ripples in the chart are caused by changes in brightness of the LED due to Thermal Management. I agree the human eye can not detect these changes, because of the small size of the changes.

otoh, the wavy lines on the E02 are just circuit noise, non audible. The brightness is changing because as 47s pointed out, all Constant Current circuits have
small ripples from the regulation which actually exist in all constant current drivers.

another example of Constant Current Ripple that produces Visible Flicker that some people confuse with PWM, is this Fenix Selfbuilt posted,

Originally Posted by selfbuilt


another Constant Current light that actually also produces Visible Flicker is this Zebralight
#199

jon_slider said:

to the best of my present understanding, Zebralight uses constant current with a feedback loop that creates a type of "circuit noise".

see #80
it's more than just a very low amplitude flicker especially on the SC5Fc.

#101
Here is a photo, thanks to maukka, that overlays a theoretical PWM square wave over the Zebralight SC5fc pulsed power management system


#74
Here is a photo of Low1 on an SC52 (photo from Samsung Galaxy S7)

#59
Constant-current regulators... use feedback-monitoring circuitry to enhance efficiency, and incorporate current balancing protection, sometimes manifesting as pulsing that can resemble PWM.

#61
Zebralight et.al. can rightfully claim not to use PWM (pulse wave modulation), but they still use PWM (pulse whatever modulation)

Click to expand...

I completely agree that there are lights that do not have such obvious wavy lines on a scope output. And there are Constant Current lights that that do not produce Visible Flicker, or at least, none that I have been able to photograph, detect visually, nor have I seen them produce obvious wavy lines on a scope image. For example Maratac AAA 2016, Tool 2016, PT-16 and PT-18.

There are also lights that have flat output lines on some levels:

maukka said:

Click to expand...

and pulsed output on other levels

maukka said:

Click to expand...

Re: Its not PWM its Constant Current Circuit Noise Ripple

All you have done is verify exactly what i said. It's a measurement issue. It's a side effect caused by sampling below the nyquist frequency. A good measurement would either sample faster and average or use an analog filter so that a graph representative of what the eye sees would result.

hiuintahs said:

In the use of your own
words...."I'd pretty much consider this all wrong" . It's not likely a measurement issue, it's a graph of the actual light level that the data logging light meter is picking up. The sample level is fairly course such as every 2 seconds or thereabouts and thus data sampling just so happens to correspond to a high spot or low spot based on the higher speed ripple voltage across the LED which Jon is referring to as circuit noise ripple. Your eyes can't see the difference in output, but the meter can pick it up. I've tested some lights that were a lot worse than this and I still could not detect it with my eyes. But from a purist standpoint, I've seen some outputs that were very constant and solid too.

Click to expand...

Re: Its not PWM its Constant Current Circuit Noise Ripple

No, all constant current circuits do not have ripple. Only ones with switching technology always do, and even some of those have post linear regulation to eliminate flicker.