Led interference

[William_H]

When buying LED lamps look out for the specified supply voltage. If they say 12 ot 24 v then for sure they are switch mode. If they just say 12v likely resistive current limiting. Likewise with 240v LED lamps if they say 90 to 260v they are more sophisticated than the ones specifying 220v. Watch out for those as we get 250v and they are short lived. Especially the "corn cob" type. olewill

 

[whiteoaks7]

I think lw395 has the right take on this. LEDs are diodes they are dc current devices - to drive them you need to supply a dc current in one direction only. This is often done by connecting across a dc voltage supply and limiting the current with a resistor - the resistor is sized to drop the voltage supply down to the forward voltage of the LED (about 1.2V) at the operating current of the LED. LEDs come in all sorts of guises these days and may have widely varying current requirements so beware of trying too low a resistor value. Resistor power has to be taken into account too - calculate this as V*i across the resistor. So then came pwm supplies although I have to say I personally have never used one for an LED. It would seem though that the typical switching frequency of a pwm would be much much too low to interfere with radios at any of their amplifier stages, but that's a guess. DC convertors are ideal because they are so efficient (they amount to 'DC transformers' and have very little power loss) but they do operate at frequencies that are known to interfere with radios. Even the ones MoD use - which you can assume are the best - have given problems and designers are careful where they are mounted and how they are shielded.

 

[GHA]

Making your own RF free LED driver isn't actually that difficult ... =>

https://www.youtube.com/watch?v=Opm-3-hkDEM&index=1&list=WL

Might need a beefier transistor or FET though.

 

[JohnGC]

Making your own RF free LED driver isn't actually that difficult ... =>

https://www.youtube.com/watch?v=Opm-3-hkDEM&index=1&list=WL

Might need a beefier transistor or FET though.

That's got potential to oscillate!

 

[GHA]

That's got potential to oscillate!

It doesn't. Checked on a PC scope, rock solid, been used most nights for year and half.

Not sure how it could oscillate, but not being too up on circuit design like this, the TLV431 is supposedly really quick, why would it want to oscillate?

Glad it works though, cheap reliable RF free LED driver.

Ta.

 

[JohnGC]

It doesn't. Checked on a PC scope, rock solid, been used most nights for year and half.

Not sure how it could oscillate, but not being too up on circuit design like this, the TLV431 is supposedly really quick, why would it want to oscillate?

Glad it works though, cheap reliable RF free LED driver.

Ta.

Note I said "potential" not "will".

The TLV431 is a programmable zener (voltage reference). The centre pin is a feedback input. With no capacitors in the circuit, it becomes possible for the stray capacitance and stray inductance to create an unintentional oscillator. If and how it oscillates will depend on layout, temperature, supply voltage and a whole host of other possibilities. The addition of suitable suppression components will make oscillation very unlikely.

This is a general problem with just copying basic applications from the datasheet and is probably why the OPs switch mode regulator is so bad. It is perfectly possible to design a switch mode supply which does not cause problems. It's what EMC design and CE marking is for. But testing is very expensive, so there is always a risk that a cheap device hasn't been. It's also sometimes obvious from a PCB photograph that the designer knew nothing about the subject at all.

It's possible that by attaching your scope to such a circuit you stop it oscillating; scope probes have capacitance. I'm not saying it will, just that it might. Although to be fair, scope probes usually show more high frequency oscillations than are really there due to the large 0V loops created by the probe ground clips.

What I should be doing at the moment is getting on with an ECU design for a fork-lift truck manufacturer. (Too easy to get distracted by the forums isn't it?) I have just finished the CAN BUS interface section. It took me twice as long to design the EMC and spike protection as it did to design the actual interface which doubles the component count; that's normal.

 

[GHA]

Note I said "potential" not "will".

The TLV431 is a programmable zener (voltage reference). The centre pin is a feedback input. With no capacitors in the circuit, it becomes possible for the stray capacitance and stray inductance to create an unintentional oscillator. If and how it oscillates will depend on layout, temperature, supply voltage and a whole host of other possibilities. The addition of suitable suppression components will make oscillation very unlikely.

This is a general problem with just copying basic applications from the datasheet and is probably why the OPs switch mode regulator is so bad. It is perfectly possible to design a switch mode supply which does not cause problems. It's what EMC design and CE marking is for. But testing is very expensive, so there is always a risk that a cheap device hasn't been. It's also sometimes obvious from a PCB photograph that the designer knew nothing about the subject at all.

It's possible that by attaching your scope to such a circuit you stop it oscillating; scope probes have capacitance. I'm not saying it will, just that it might. Although to be fair, scope probes usually show more high frequency oscillations than are really there due to the large 0V loops created by the probe ground clips.

What I should be doing at the moment is getting on with an ECU design for a fork-lift truck manufacturer. (Too easy to get distracted by the forums isn't it?) I have just finished the CAN BUS interface section. It took me twice as long to design the EMC and spike protection as it did to design the actual interface which doubles the component count; that's normal.

Thanks, wasn't meant to come across as having a dig at all, just interested to learn more.
Last circuit was looked at with one of those little nano scopes, might be an interesting evening project to use a PC scope and do some more testing to maybe see a bit deeper if anything is going on.

 

[Graham_Wright]

Apart from the troublesome triads, I have various strings of leds under cupboards and the like.
During the last three years, suffering voltages generated by an unregulated mains charger, a regulated solar panel and the engine alternator, only one led out of several hundred has failed. They are all connected as groups of three with one series resistor.

With that abuse, why bother with electronic control?

 

[GHA]

Apart from the troublesome triads, I have various strings of leds under cupboards and the like.
During the last three years, suffering voltages generated by an unregulated mains charger, a regulated solar panel and the engine alternator, only one led out of several hundred has failed. They are all connected as groups of three with one series resistor.

With that abuse, why bother with electronic control?

Pretty much the same here, though had more failures than that, some start to flash when the voltage gets up high 14v's.
Buck converter would cost less than a beer and take a few minutes to fit on a lighting circuit.... one of these years....

 

[JohnGC]

Thanks, wasn't meant to come across as having a dig at all, just interested to learn more.
Last circuit was looked at with one of those little nano scopes, might be an interesting evening project to use a PC scope and do some more testing to maybe see a bit deeper if anything is going on.

I didn't read it that way and I thought you might be. The devil really is in the detail and it can be hard to grasp all of the possibilities. Semiconductor datasheets are vital, but always look for the manufacturer's application notes as well; if they exist it is where you'll find the details.

 

[Jungle Jim]

Whaat about wrapping the circuitry in a bit of aluminium foil?

But ensure that the foil doesn't actually touch anything with volts in it! Many years ago I dropped my foil-wrapped lunch right on top of a prototype controller board on my work bench. It didn't end well :ambivalence: