Let's suppose I have a circuit which contains a 3.3V MCU and a buck converter (let's think that its frequency is higher than 400 kHz). I'm wondering if shielding the entire circuit would not a good idea because I think that the shield may reflect the noise generated from buck converter to the MCU. Is this a valid concern?
\$\begingroup\$
\$\endgroup\$
2
-
3\$\begingroup\$ "buck converter (let's think that its frequency is higher than 400 MHz" is very unlikely; did you mean 400 kHz? \$\endgroup\$– CurdSep 13, 2019 at 12:33
-
1\$\begingroup\$ Yeah off course i am sorry for this typo \$\endgroup\$– Faruk CEBECİSep 13, 2019 at 12:44
Add a comment
|
2 Answers
\$\begingroup\$
\$\endgroup\$
3
The short answer is that you're right. It is possible (and i mean possible separately from likely) for the shielding of a circuit to cause a problem because a circuit is not self-immune to the quantity of EMI it is generating when not allowed to radiate to free space.
The longer answer is that it's why you often see only select parts of circuits shielded. You might only shield the sensitive parts (like the MCU) to protect it from outside radiated EMI sources. You might equally shield things like switching power supplies to protect the rest of your circuit from them or to protect other circuits from the EMI generated by your supply. If you get problems from conducted EMI then shielding won't help you and conducted EMI may even get worse if you can no longer radiate the energy you're now containing.
-
\$\begingroup\$ Does a shielding reflect an em wave or absurb it and then converts it to heat or it depends on the wavelength? \$\endgroup\$ Sep 13, 2019 at 15:18
-
1\$\begingroup\$ It'll do both but it'll reflect far, far less if the shielding is grounded and can't change potential as it'll shunt current induced in it by incident EMI to ground instead. It'd be like a very poor capacitor referenced to the local ground potential if it was left floating which would mean it wasn't really doing it's job as very high frequency spectral content could be coupled in via it. \$\endgroup\$ Sep 13, 2019 at 15:25
-
\$\begingroup\$ solder a 1cm by 1cm loop, at end of coax cable. Use a scope to monitor the magnetic fields around the switching regulator. And be aware that any 1cm by 1cm loop in your MCU may experience similar voltages of interference. The XTAL oscillator of your MCU will be vulnerable. \$\endgroup\$ Sep 13, 2019 at 17:13
\$\begingroup\$
\$\endgroup\$
The MCU is not sensitive to "noise" as long as you don't have any analog circuit (e.g. ADC, analog comparator) and as long as the noise is not extreme (e.g. next to very strong and fast changing E/M fields). So you problably don't need to shied it from your small DC/DC converter. You just need to take care that the power supply is clean from spikes (some ripple doesn't matter), but that's not accomplished by shielding.
If you want to protect the outside wold from your circuit shielding may make sense.
