1

Firstly, I used TIP120 for switching. There was a problem of leakage current may be, so I had one out of four panels glowing. So I decided to switch to MOSFET IRF510. I soldered my circuit from scratch, but the problem remains. Is there something wrong with my circuit?

VERY IMPORTANT NOTE: It works fine on a breadboard! enter image description here

enter image description here

Mat_python
  • 59
  • 1
  • 9
  • Maybe it's like this because different led colours turn on at different amounts of currents. – Bradman175 Oct 01 '16 at 14:35
  • @Bradman175 I have not seen something like this before too. I actually soldered three - four times with these transistors/MOSFET: BC547, 2N2222, TIP120, IRF510. The problem remains. I am not sure what is the problem. Though, answer by Tony Stewart is a bit acceptable but complex for me to understand. I should try that soon. – Mat_python Oct 01 '16 at 14:40
  • what colours LEDs are lighting up? (I'm kinda colourblind) – Bradman175 Oct 01 '16 at 14:44
  • @Bradman175 There are four panels: Red, Green, Blue, Purple. I got purple by soldering Red/Blue terminals together on RGB LED strip. In the picture, the brightest panel is purple and dim LED panel is red color. – Mat_python Oct 01 '16 at 14:46
  • Gotcha I think my answer will work. – Bradman175 Oct 01 '16 at 14:47
  • I cannot imagine a better place to put unprotected, open circuitry than this "low-esd" carpet... – Xaser Oct 01 '16 at 14:48
  • 1
    @Xaser Can I suggest a [rubber ESD mat](http://www.eevblog.com/forum/buysellwanted/esd-mat/) in case the OP didn't get that one? – Andrew Morton Oct 01 '16 at 14:53
  • @Mat_python It might have worked on the breadboard because that little PP3 battery wasn't worn down then. – Andrew Morton Oct 01 '16 at 14:54
  • 1
    @AndrewMorton Nope. I changed the new battery. I tried all the possible combinations of battery, transistors, circuit soldering, LEDs. – Mat_python Oct 01 '16 at 15:09
  • @Mat_python Well my answer didn't work. Have you tried disconnecting some of the boards where the LEDs are working. Have you tried using the boards that are working and see if you can the the other LEDs to work individually. – Bradman175 Oct 01 '16 at 15:20
  • Without a ground wire to balance the PWM switched power lines , you are broadcasting harmonics of the PWM all the way up to frequencies matching the rise time and then beyond with more harmonics. You are essentially radiating RF noise in a local area which can be picked up by diodes or even an AM radio – Tony Stewart EE75 Oct 01 '16 at 15:22

1 Answers1

1

This is an excellent example of diodes detecting stray common mode CM E-fields and conducting current by rectification in the LEDs.

The unbalanced switch resistance causes this common mode field.

The solution is to reduce CM EMI by twisted pars then raise the CM impedance above ~1MHz for the fast rise times with a ferrite CM choke or BALUN. These are needed on all high speed high current switched wires.

It can be a clam-shell ferrite like those used in all VGA cables which is moulded in plastic with no air gap or a large torroid with all wires wrapped thru the torroid as many times as possible.

Start by adding a wire from ground and wrap around each cable from source and terminate with a low ESR capacitor on LED boards. Since you have no ground on the LED boards, adding this cap and ground wire, reduces the loop area of the offending board circulating current but will easily couple within the same distance gap as the size of the board loop.

Shielding or connection from an earth ground will also help absorb stray switched E fields greatly by lowering the impedance of your floating battery ground or a sheet of alum. foil connected to earth ground also works.

Community
  • 1
Tony Stewart EE75
  • 1
  • 3
  • 54
  • 182
  • Thank you for your help. I didn't understand few things here. Should I add 'single' wire from ground and wrap around each cable and terminate with a capacitor? How can I terminate it with a capacitor with one wire? Which value of a capacitor is suitable ? – Mat_python Oct 01 '16 at 14:35
  • 1
    I haven't heard of this phenomenon before. It would be better with some links as I would like to learn as well. – Bradman175 Oct 01 '16 at 14:37
  • There are many books on this subject. Try touching earth ground (grounded monitor) and the battery ground or hand around cables – Tony Stewart EE75 Oct 01 '16 at 14:53