Transistor Switching Noise???

[eetrit]

Hello everyone,

I have a prototype board that has a microcontroller that sends out a status heatbeat signal. The signal is generated by 2 transistors switching from high to low. One transistor for each line. The transmission lines appear to be equal and opposite of each other which is the recommended waveforms.

My problem is that the heartbeat can be heard in the AM band on an AM Radio. I think I've tried almost every possibilty to fix this problem. Any suggestions will be greatly appreciated.

Thanks,
eetrit

 

[nbucska]

The hearthbeat is so slow, you don't need fast switching --
so slow it down. The radiated noise is proportional with the speed. <nbucska@pcperipherals.com>

 

[Lewish]

Without knowing your exact circuit topology, I suspect what you are getting is crossover noise. A transistor will turn-on faster than it will turn-off, so one transistor is trying to drive a large impedance (the not quite off other transistor), hence producing a large noise current.
Try using a differential driver IC instead.

 

[pebe]

If you have one transistor for each line, then it can't be crossover noise. I think your transistors are switching too fast. In the past I have used a 100pf capacitor between base and collector to get rid of RF input to an audio amp. If you are driving the base through a resistor, it would probably work in your case because the negative feedback will increase the collector rise time.

 

[nbucska]

The frequency of the hearthbeat is around 1 Hz, so a 1 KHz
bandwidth should be more than enough. <nbucska@pcperipherals.com>

 

[melone]

Try limiting the rise and fall times of the signal. Place a small cap across the base-emitter and another small cap across the collector-base junction. Vary the values of the caps to adequately round-off the corners.

Good luck and keep us posted!

 

[eetrit]

Thanks for all the good info. I'm back at work and i'll be trying all your ideas. Thanks again.

 

[eetrit]

Hello. I tried everything on my board. I succeeded in reducing the emissions off the board but the status heartbeat is still loud and clear. Oh well, we'll deal with that later. Thanks again for all the help. At least we pass the EMC tests now.

 

[circuitsage]

1. Put as much series resistance as you can at the input and output of the transistors.
2. Put the transistors as close as possible to the microcontroller and the transistor outputs as close as possible to the loads.
3. Keep the two lines as close to each other as possible.
(i.e. keep the entire circuit small so it doesn't act like a big loop antenna)
4. Use surface mounts instead of through holes, and try not to break the ground plane.
5. Bypass close to the microcontroller.