Irritating high-pitched noise with 555 PWM driving LEDs (Update: 12 Aug 2015)

[thejoggingmat]

An other module with constant current adjustment, should give a better control of the brightness http://www.ebay.nl/itm/Automatic-Converter-Boost-Buck-LX6009-4-35V-to-1-25-25V-CC-CV-Voltage-Regulator-/400873719224?pt=LH_DefaultDomain_3&hash=item5d55ef85b8

To avoid led flicker you could use a buck boost converter based on the XL6009 like this http://www.ebay.nl/itm/Hot-Selling-DC-DC-Boost-Buck-Step-Down-Up-Converter-XL6009-Solar-Voltage-Module-/301596262442?hash=item463889c02a

So I should use this module instead of the 555-based circuit?

Yes, considering the price you can't go wrong.

I'll order a few samples in a while to try out, but they're going to take a while to reach me.

In the meantime, are there any long-term implications if I were to use per the schematic I posted?
As 'dirty' as the circuit is, there's no denying that it makes the least audible high pitched noise. (Unless I can whip up a better one)

Running components at high temperature will decrease their life expectancy.
The audible noise could drive you really mad, since its a long wave transmitter the Pentagon may ask you to shutdown the device so the can communicate with their submarines.      

No worries, I'll just donate free tin foil hats to whoever they send knocking on my door. 

[Gyro]

Still a bit curious about what mechanism in the LED array is responsible for the noise production (the interconnects? thermal expansion of the LEDs themselves?)  - it is clearly related to the peak current or rise time which would have increased with the improved decoupling.

When the 1000uF cap was directly across the transistor it would have slowed the rise time significantly, hence the lower noise. Of course very bad for the health of the transistor and the cap so not a solution!

It implies that a low value series resistor between the transistor and LED array (or possibly in the emitter lead) to control / define the peak current would quieten it down, hopefully without too much loss.

I agree though, a module is probably the way to go.

[jlmoon]

Damping resistor (low value 5 watt or so) between supply rail (+) and anode of led. Then that 1000uF Cap and a .01 cap as bypass to ground at junction of resistor and Led to ground.   

[jlmoon]

I would also think the worst case Current gain of 12 for that TIP41 would stress the 555 timer output considerably with that kind of load and supply voltage.
 

[Gyro]

Damping resistor (low value 5 watt or so) between supply rail (+) and anode of led. Then that 1000uF Cap and a .01 cap as bypass to ground at junction of resistor and Led to ground.   

I'd try the damping resistor after the decoupling first - from the OP's comments the noise really is coming from the LED array itself. The 1000uF directly decoupling (discharging into) the array and transistor makes it louder.  It's remote from the rest of the circuit so the interconnects are already providing some resistance.

[IanB]

Still a bit curious about what mechanism in the LED array is responsible for the noise production (the interconnects? thermal expansion of the LEDs themselves?)  - it is clearly related to the peak current or rise time which would have increased with the improved decoupling.

I think feeding something (anything) with PWM is going to make it "sing".

The solution is to use a PWM frequency out of the audible range, or put an LC tank circuit on the output of the PWM (but then the tank circuit will sing).

[jlmoon]

Still a bit curious about what mechanism in the LED array is responsible for the noise production (the interconnects? thermal expansion of the LEDs themselves?)  - it is clearly related to the peak current or rise time which would have increased with the improved decoupling.

I think feeding something (anything) with PWM is going to make it "sing".

The solution is to use a PWM frequency out of the audible range, or put an LC tank circuit on the output of the PWM (but then the tank circuit will sing).

The third orders will get you as well.

[leblanc]

If you're gonna have a potentiometer, why not just use a common collector and avoid PWM altogether?



The potentiometer varies the base voltage from 0V to 11.4V, which makes the Ve vary from 0.6V to 12V. If you know the drop across the LEDs (the voltage at which they start conducting), you can add a resistor in series with D1 in order for R1 to give you a full range of dimming (no dead zone). A variation in Ve gives a variation in VLED, which will control the current (and therefore brightness) from 0mA to (almost) 800mA.

Q1 will need to be sized appropriately (I'd say 3W minimum, 5W is better).

You won't be as efficient, but you won't need PWM (no sound, lower EMI, less circuitry, no flicker).

There are a number of improvements that can be made to the circuit I attached, but it's a starting point.

[Zero999]

Or you could use constant current drive.

[leblanc]

I was going to suggest a constant current drive, but he has resistors in series with the LEDs already. It would work though, but you'd have a smaller range of brightness.

[SL4P]

Late arrival here...
The LEDs are salvaged from a Visualiser...
What's a visualiser?
Perhaps there is more to the load than 'just' LEDs ?

[poot36]

This might sound like a dumb question but how close to a object is the light output side of the leds?  I have had a pwmed light bar which drew around 800ma and had high quality red cree smd leds on it and when it got close to a object you could sometimes hear it buzzing.  The pwm frequency was 400Hz with a relatively low duty cycle.  It would even happen on carpet!  If you want a more extreme example of this effect lookup youtube user tesla500 and find the dlp tv take apart videos and you will find a section where he is focusing the light output of the dlp lamp with a magnifying lens on a piece of wood and the pulsed nature of the lamp can be heard as sound.

[Gyro]

Audible noise from LEDs / arrays seems to be a very common problem (haven't turned up anything describing any LED device specific noise mechanism though)

A search for "audible noise LED" brings up a fair number of hits, including a PDF app note from Linear Tech:

www.linear.com/docs/27822

And an article from LEDs Magazine (A Technological Lads' Mag clearly  )

http://www.ledsmagazine.com/articles/print/volume-10/issue-6/features/proper-driver-design-optimizes-led-lighting-magazine.html

The LinTech note looks as if it could be quite helpful.

[thejoggingmat]

Late arrival here...
The LEDs are salvaged from a Visualiser...
What's a visualiser?
Perhaps there is more to the load than 'just' LEDs ?

It's one of those gadgets used in classes to project hardcopies onscreen. The successor to an OHP.

[thejoggingmat]

I was going to suggest a constant current drive, but he has resistors in series with the LEDs already. It would work though, but you'd have a smaller range of brightness.

Or you could use constant current drive.

If you're gonna have a potentiometer, why not just use a common collector and avoid PWM altogether?



The potentiometer varies the base voltage from 0V to 11.4V, which makes the Ve vary from 0.6V to 12V. If you know the drop across the LEDs (the voltage at which they start conducting), you can add a resistor in series with D1 in order for R1 to give you a full range of dimming (no dead zone). A variation in Ve gives a variation in VLED, which will control the current (and therefore brightness) from 0mA to (almost) 800mA.

Q1 will need to be sized appropriately (I'd say 3W minimum, 5W is better).

You won't be as efficient, but you won't need PWM (no sound, lower EMI, less circuitry, no flicker).

There are a number of improvements that can be made to the circuit I attached, but it's a starting point.

I'd have used constant current drive, however whatever I'm doing needs to be as efficient as I possibly can. Fully charged 3x18650s already take around 2++hr to drain to ~10V under load in PWM adjusted to its max brightness as is. I'll be using whatever I build out of this to use on the field in my videography.

[thejoggingmat]

This might sound like a dumb question but how close to a object is the light output side of the leds?  I have had a pwmed light bar which drew around 800ma and had high quality red cree smd leds on it and when it got close to a object you could sometimes hear it buzzing.  The pwm frequency was 400Hz with a relatively low duty cycle.  It would even happen on carpet!  If you want a more extreme example of this effect lookup youtube user tesla500 and find the dlp tv take apart videos and you will find a section where he is focusing the light output of the dlp lamp with a magnifying lens on a piece of wood and the pulsed nature of the lamp can be heard as sound.

It's quite far from a carpet or any other electrical objects, so I doubt it comes into play here. I used a sheet of aluminium foil glued to a paper sandwich as a shield to test, just to make sure. Same result.

[thejoggingmat]

Audible noise from LEDs / arrays seems to be a very common problem (haven't turned up anything describing any LED device specific noise mechanism though)

A search for "audible noise LED" brings up a fair number of hits, including a PDF app note from Linear Tech:

www.linear.com/docs/27822

And an article from LEDs Magazine (A Technological Lads' Mag clearly  )

http://www.ledsmagazine.com/articles/print/volume-10/issue-6/features/proper-driver-design-optimizes-led-lighting-magazine.html

The LinTech note looks as if it could be quite helpful.

Hmm... Although most of them I googled are referring to the in-circuit caps singing. But that LT IC looks interesting.

[poot36]

Sorry I meant to say that the noise would happen if the light output of the LEDs was near any object soft or hard.  Ie. LEDs pointed at a table or in my other interesting case the carpet.  The carpet was interesting because you would think it would absorb the sound as well as diffuse the light enough to stop the sound altogether.

[ignator]

Maybe I'm missing something. I didn't see any comments about C3 the 1000uF cap. Every time Q1 is turned on, it has to discharge this. This is a large current spike. And this can produce audio sound. The current spike would be of interest, depending on how fast Q1 switches.

[Gyro]

C3 was drawn in the wrong place and got moved to the supply rails where it should have been (where it made the noise worse due to the higher peak current drive capability).

Edit: The OP did actually briefly try the cap as drawn, where it quietened things down a bit due to increased rise / fall times, but as you say, the transistor was having to discharge it every cycle!

[thejoggingmat]

I've finally went on to doing the veroboard version. Strange things are happening. I think I have narrowed down to:
1) The transistor
2) The 555 (I used a DIP socket the 2nd round and tried with a 2nd 555, still same result. Will try with more from my stash)
3) The rubbish veroboard connections (I am still a complete noob at veroboarding admittedly)

Unless parasitic capacitance or some third order is flirting with my circuit. Any ideas?