Distance EMI is radiated into traces

[Zak28]

Whats the necessary spacing required in which negligible amounts of EMI would be picked up by traces and components emanated from a flat heatsink raised to ~800VDC with a period of 8mS-12.5mS- 50% duty which would be right under a board with its traces?

Heatsink is isolated from all traces on board thus capacitve coupling is the mode in which EMI would be induced into components and traces. There are many different standoff studs available the proper distance would likely diminish any EMI into the board. The board is going to be mounted onto the flat heatsink.

 

[betwixt]

Not enough information by far!

1. define 'negligible'
2. what amplitude is the signal? (not the DC, that doesn't radiate EMI at all)
3. what are the rise and fall times of the signal?
4. what is the substrate Er and thickness?
5. how much capacitance is between the source and the traces?
6. how thick are the 'standoff studs'?
7. Are there current loops that would add magnetic coupling to the EMI?
8. What impedance are the traces at the frequency of the EMI?

Point 8 is important, if the trace is high impedance it will be far more vulnerable than one at low impedance. Even with all this information an accurate calculation would be almost impossible.

Brian.

 

[Zak28]

Rise is ~3ms fall is ~850nS high impedance nodes are very short each under 5mm there are only a few high impedance traces with their signals are 4-5v. Studs are 4mm diameter. Not sure how high the high impedance traces are they are only gate driver input and the very few found in a 555 oscillator.

Doesn't coupled EMI drop substantially with distance?

 

[betwixt]

Yes EMI does drop with distance, the exact amount depends on the geometry of it's surroundings but could be as much as inverse square of the distance.

The rise and fall times you quote are quite slow and it the rise and fall that primarily produces the EMI so the chances of you having interference problems are very small. None of the signals around a 555 and most gate drivers are high impedance, it needs to be more than about 50K to be prone in that kind of circuit. With traces as short as 5mm you shouldn't worry.

Brian.

 

[Zak28]

Yes EMI does drop with distance, the exact amount depends on the geometry of it's surroundings but could be as much as inverse square of the distance.

The rise and fall times you quote are quite slow and it the rise and fall that primarily produces the EMI so the chances of you having interference problems are very small. None of the signals around a 555 and most gate drivers are high impedance, it needs to be more than about 50K to be prone in that kind of circuit. With traces as short as 5mm you shouldn't worry.

Brian.

TC4422 says driver input is high impedance but does not say how high.

 

[betwixt]

The input to the TC4422 may be high impedance but that is so it provides minimal load to the circuit driving it. The impedance at the trace is a combination of the TC4422 and the output of the device driving it which probably has quite a low impedance.

Brian.