2
\$\begingroup\$

I read that buck and boost converter due its switching mode produces too many noises compared with linear regulators but it is more efficient.

Can those those noises cause interference on boards with wifi chips on it?

How would be a good way to measure how big would be such interference at the point to determine that it is worth to use a linear regulator even without power efficiency?

\$\endgroup\$
  • 2
    \$\begingroup\$ Of course. Though interference from harmonics of switching may be weaker at 2.4 GHz than at lower frequencies. \$\endgroup\$ – Chris Stratton Jul 26 '16 at 13:59
4
\$\begingroup\$

Quick answer, as from Chris Stratton, is YES!

Longer answer: noise from electrical circuits is caused by currents, which turn any and every conductor into an antenna some some sort. The amount of noise they generate out of the unit will depend on the efficiency of the antenna (the length of the antenna vs the frequency of the signal) and the amount of current being switched. Buck and boost regulators work by switching currents through inductors, hence produce noise. A well designed switched converter will produce very little noise by the use of filters, shielding, well routed tracks etc). Meanwhile a linear regulator does not switch currents, it basically "burns off" the extra voltage, and so does not produce noise. However, if you want to boost the voltage, you pretty much have to switch currents in some form or another depending on the typology you are going for.

You mentioned the efficiency of a linear regulator is worse than a switcher, which is very much the case. If you are using a linear to get 3.3V from a 5V rail at 100mA, 100mA goes into the LDO (plus a little bit of overhead to power the LDO's internal circuitry). The device then dumps the excess power as head, so you have 100mA, 3.3V out (0.33 Watts), and you loose 100mA, 1.7V (0.17 Watts) as heat. Power out over power in gives you about 50% efficiency. Meanwhile a good, synchronous, switcher will give you about 70-80% efficiency.

As for interference with WiFi, again, the answer is that it is possible, but only if very poorly designed. A switcher tends to work in the 100kHz range, a lot lower than WiFi's 2.4 or 5 GHz, however you also have to worry about the harmonics of the aggressor's signal. It can get very complicated, it is worth being aware of it, but there will be bigger issues than your WiFi dropping out if your DCDC converters are causing that much noise...

\$\endgroup\$
  • \$\begingroup\$ Great answer! Added a bit of legal and signal perspective in my answer, so to extend, not replace yours. \$\endgroup\$ – Marcus Müller Jul 26 '16 at 15:07
  • \$\begingroup\$ Thanks Marcus, didn't want to go too far into legal aspects, as that can get complicated, and the question didn't explicitly ask \$\endgroup\$ – Puffafish Jul 26 '16 at 15:16
  • \$\begingroup\$ possible, but only if very poorly designed. Indeed. I'm certain that EVERY wireless AP on my network is powered by a switcher, as are all of the computers. It's a solved problem... \$\endgroup\$ – Ecnerwal Jul 26 '16 at 15:32
2
\$\begingroup\$

Can X cause interference with operation of Y ?

Yes. I was about to write a shorter version of Puffafish's answer, but I'd like to add:

All electrical devices typically have to comply with laws, governing the amount of energy they might spread around the spectrum. This means that legally, it's pretty safe to say that a compliant switch-mode power supply (SMPS) will not interfere with operation of a 2.4 GHz or 5 GHz system, as it's strictly forbidden to emit significant energy in that spectrum.

A problem that was far more wide-spread in the olden days was that wireless devices were actually sensitive to energy transmitted in bands that were actually far away their own band – that's why some FM radios make funky noises when you operate a cell phone close to them. Nowadays, devices need to be robust against legal operation of nearby devices in other bands, so that's not going to be a concern for any WiFi card.

About measuring: Get something that can work as a spectrum analyzer (calibrated spectrum analyzers can be very costly, but sometimes there's options to rent them – and some radio amateur clubs even measure devices with/for you as a service). As a starter, you could just get any cheap Software Defined Radio peripheral (e.g. a HackRF, or an Ettus USRP) and just compare the power of WiFi bursts with what you can receive in a sensible distance from your SMPS.

Generally, 2.4 GHz is very high compared to the kHz – single-digit MHz that SMPSes operate in, and WiFi devices are pretty robust, so I wouldn't worry. Really, a SMPS that spreads wideband noise at Gigahertzes powerful enough to negatively influence WiFi communication should probably not be used at all – it's not only illegal to operate or actually import with the intent of commercially selling it to end customers into most countries, it's also got serious design problems that cause that amount of interference, and might simply be dangerous.

\$\endgroup\$
  • \$\begingroup\$ I'm always dubious of saying "...a compliant unit... will not interfere..." without knowing more about the system. It is possible that compliant unit A paired with compliant unit B makes a NOT compliant unit C. Depending on which set of standards you need to comply to, this may be your issue (it would be under CE for instance) or not. But I agree with what you've said, especially about getting a spectrum analyser and having a look. \$\endgroup\$ – Puffafish Jul 26 '16 at 15:14

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.