Elimination of PWM noise from TLC5947 LED

hey all, i need some help

I got a project, that i use a TLC5947 to drive some mostfets that drive 12v LED Strips
the thing works perfectly, well...execpt... that its bing used in a limo
(it controls cabin lighting)

but its putting noise on the AUDIO system.... when i look at the 12v supply to the LEDs, i can see a square wave on the power feed from the battery
that has some relation to the PWM...putting a 470uf cap across the input, does cut it down, but not enough

im think its a car, and there is other noise on the power, ie alternator... so... its the audio system not doing its job
but if i can kill audio at the source here... id be way ahead

this is my output stage
3 channels worth
thanks!
mitch

What current does each chan supply max?

mitchjs:
putting a 470uf cap across the input, does cut it down, but not enough

"470uf" makes me think you used an electrolytic capacitor.

If so, add a ceramic capacitor in parallel to the electrolytic (two capacitors across the same spot). Electrolytics don't have a very fast response time so they're not very good at decoupling.

each chain can supply up to 3 amps...

yes, 470uf is electrolytic cap

yes i would add a ceramic if it turned out that this was the solution...
prob adding a ceramic now wouldnt change the noise
ill try tonight

im putting the cap across the 12v input to the rgb strips, and the common anode
(so pins 1 and 2) on the output connector JLED0

yes i would add a ceramic if it turned out that this was the solution.

Even if it is not the solution you should fit it and keep it fitted. You might even have to use an inductor ahead of those capacitors.

Probably the noise is coming from the PWM in the LEDs and radiating off the connecting wires. Start by disconnecting all the LEDs and adding them one at a time. Until you can just here the noise. Then what ever you change will be more noticeable. If something improves things but does not affect a complete cure, keep it in and look for the next small improvement. You can also try dropping the LED current, adding screening, getting a proper ground scheme and so on.

It could be that you can't fix this at all, it would not surprise me if that were the case.

to test this i have < 4" of wire...
yes, less leds = less noise, im sure due to less current being drawn
maybe an inductor would help, shame i got nothing really laying around
that could handle the current, ill have to look... just a few leds connected
to see if it help

ill let ya know, soon... me just got home... play time in a few

thanks!!!

im sure due to less current being drawn

It is probably more to do with the radiated interference from the wires than current from the supply.

An inductor is a good suggestion. Without an inductor, the capacitor is basically trying to filter the entire electrical system.

A diode can sometimes help too... A diode prevents the capacitor from discharging into the entire electrical system, and it blocks positive noise spikes. But you will loose a little voltage across the diode, and your LEDs will be dimmer, assume they are running off the 12V).

well...

i made a small board that put inbetween my 12v source and the 12v input to the led chain
i feed 12v though a 330uH coil (only thing i could find), then i got 470uf electrolytic and a .1uf ceramic in parallel
does make a big difference, i dont think i hear any difference with out the coil, too hight uH?

any one ever mess with a Murata BNX012-01?
http://www.mouser.com/catalog/specsheets/Murata%20BNX%20Series.pdf

i need to get this now actually in the vehical, to see(err hear) if it makes enough difference

scope still shows the noise, but much less amplitude, and less spiky

mitch

i dont think i hear any difference with out the coil, too hight uH?

No too low if it is doing nothing for you. You can't have it too high.

I've built a 8x8 rbg matrix using the same kind of chip. It has a lot of ribbon cable connecting the led's in groups of 4. It's about 800mm square. Anyway When i switch it on it interferes with my FMradio big time, really loud an all over the house. You can listen to the radio with it on. So i had thought i would just shield the whole thing when it's finished. They produce a lot of high frequencies them chips i guess and with lots of wire to the led's that's lots of ariels. I had/have no idea how to fix it that's why i though shielding would work..i hope it does

If shielding worked that well I would never have had any trouble meeting EMC requirements for any of the set top boxes I designed.

Sadly it is very difficult to effectively shield a circuit. The problem is that wires must go into and out of the box and in the case of things like this light must come out. RF leaks out of the incomplete joint between two metal parts of the enclosure, along the wires and through holes and slots.

Sometimes shielding make matters worse by setting up large resonant standing waves inside the enclosure so that it is sometimes nosier inside a box than it is without.

The point is that reducing EMC emissions is not just a case of doing X, it is a matter of looking at every source of emissions, trying to minimise them and trying to stop them from radiating. It is a step by step process with no magic bullet, requiring test equipment so you know when things are getting better or worse, experience, patience and luck. Many times it also needs a different PCB layout.