Input/output filter capacitor values for PWM dimmer

[RichLD]

Hi all

I am currently using the classic eBay LED 12v dimmers as dew band controllers with great success. However, to be sure I'm not putting too much noise onto my main 12v supply line (leisure battery) I'd like to put some filter caps on the input and also on the output to smooth off the edges of the square wave output, hopefully reducing any interference.

I know I should be using a separate power source for the PWM dimmers but at the moment that's not really an option.

The supply cables to the dew bands are sheilded and the circuitry is in an ali box. The output sockets are isolated from the case as well. I can't say that I've noticed any visual artefacts in my subs from the use of these controllers but would rather go belt and braces on this to be sure 

I'm not sure if this makes much difference but the frequency of the dimmers is 680hz - I believe that this can be changed by fitting different caps that bridge pins 2 and 6 on the 555 part of the circuit, preferably to lower the frequency in this particular application. I'm wondering if this will make any appreciable difference to any potential interference?

So, if anyone has any advice concerning a simple filter circuit (component values would be great, a schematic would be awesome!) or to address my frequency change question I would be extremely grateful 

Many TIA

Rich

[symmetal]

Hi Rich,

I haven't noticed PWM dew controllers causing any artefacts on my images either when they all ran off the same 12V switch mode PSU but I run the camera off a separate 12V battery just to get the cleanest power supply.

The square wave output going to the dew bands shouldn't create any real interference as they are purely resistive and the frequency is not high enough to cause any significant rf radiation. The harmonics of the 680Hz square wave will be very low amplitude when you get up to 1 MHz and beyond.

Any noise on the 12V supply due to the dew controller switching on and off at 680Hz will depend on the length and thickness of the wire from your 12V battery. To filter this a simple RC network could be inserted between your 12V battery supply and the 12V to the dew controllers. I've attached a scan of a sketch I did as it was quicker than drawing on the PC. Hope you can read it OK. 

I suggested the pi filter configuration (the bottom one) as your interference is from the load (pwm controller) rather than the source.  I assumed your dew band had about 10 ohms resistance meaning about 1 amp current. The resistor (R) in the filter will take some power from your dew band but at 0.22 ohms it's only about 2%. You can increase the value if you wish to get better noise isolation but note its power rating. At 0.22 ohms and 1 amp it's 0.2W. A 0.47ohm will need to be 0.5W. The electrolytic capacitor value isn't critical. Anything near will do. The capacitors should be 25V or higher to give a safety margin.

Replacing the resistor with an inductor will give much better noise isolation but they are bigger and more expensive. Let me know if you want me to work out a value for one.

If you really wanted to filter the output the single RC circuit (top) will do to remove the higher harmonics. I'll work out some values if you wish.

I should leave the 555 frequency as it is. Lowering it will mean higher value capacitors while raising it will create higher harmonics though smaller capacitors can be used.

Hope that helps. 

[RichLD]

Wow, that is fantastic! Thank you so much for the info, it's very, very much appreciated!

I'll let you know how I get on 

All the best

Rich

[Dr_Ju_ju]

You could also try RFI filters (frequency dependent) i.e. http://uk.rs-online.com/web/p/power-line-filters/0209083/

[Stub Mandrel]

555s are notorious for putting big spikes onto the power rails. Rule of thumb, you need a capacitor ten times the size of your timing capacitor, in parallel with a smaller capacitor with different construction to catch high frequency glitches. So, if the timer capacitor is 1uF fit a 10uF electrolytic and a 0.1 uf ceramic or polyester cap across the power rails as near to the 555 power pins as reasonably possible while keeping their leads short.

I wouldn't worry about the output - caps there will make the surges worse, not better, as they will increase the switching current. An inductor in series with the output could work but might fry the 555 with back emf...

[RichLD]

Thank you all so much for taking the time to share your knowledge and experience!  I'll report back when I've done the deed...

[symmetal]

Glad it helped Rich.   I should point out a little error in my sketch when I drew the graph of the frequency response.  At the -3dB point on the graph (also called corner frequency or cut-off frequency) the voltage output is not 50% input but 70.7% input. This is due to the voltage phase change that occurs at that frequency. It doesn't affect you as you're not concerned about the signal (noise) phase. Also the graph axis are logarithmic which is why it doesn't look  like 70.7% input.  Just thought I should mention it. 

[robinst]

Another thing for easily reduce electrical noise is to add a clamp on ferrite to the cable close to the noise source, see link below. It wont probably make that much of a difference in this application but if belt and braces is wanted it is a common solution for EMI problems. 

Regarding capacitors, a 100 nF capacitor directly over the output (and input) from the controller might be a good idea to further reduce high frequency noise. As stated above, a big capacitor will only generate current spikes when it is to be charged each cycle but a small one only will take out some of the sharpness of the rising voltage and reduce its high frequency content.  

 http://www.mouser.se/Passive-Components/EMI-Filters-EMI-Suppression/EMI-RFI-Suppressors-Ferrites/Ferrite-Clamp-On-Cores/_/N-bw7t7/