3
\$\begingroup\$

I have a power circuit that implements an AC-to-DC rectifier and a non-isolated buck converter to convert a 120 VAC 60Hz signal to 12V and then a linear voltage regulator to step the 12V down to 5V.

The 5V will be used to power a micro controller and there is a possibility that people will touch the pins of the micro controller. I realize that it is not ideal to use a non-isolated power converter, but due to size constraints I do not want to put an isolated power supply with a transformer in the design.

The question is:

Will a 500mA fuse after the 12-to-5V linear voltage regulator make the design safe against electric shock since the buck converter is non-isolated?

\$\endgroup\$
  • 4
    \$\begingroup\$ Seek professional help (i.e. from someone knowledgeable about electrical safety regulations) as getting this 'wrong' is not only dangerous but could expose you to significant liability. \$\endgroup\$ – Adam Lawrence Jul 25 '16 at 15:10
  • 2
    \$\begingroup\$ Considering how small and cheap professionally-designed 120VAC-to-5VDC "power bricks" are these days, and how disastrous getting it wrong could be, I don't see why you would even go there in the first place. \$\endgroup\$ – zwol Jul 25 '16 at 22:03
24
\$\begingroup\$

Will a 500mA fuse after the 12-to-5V linear voltage regulator make the design safe against electric shock since the buck converter is non-isolated?

No. A fuse does not protect against electric shock. Fuses are for limiting current in a fault condition to protect equipment, wiring and prevent fire. What you are proposing is potentially lethal. You must isolate the low voltage circuit from mains if there is any risk of contact with the circuit.

Just for reference, RCD / GFCI earth leakage protection devices generally trip at 30 mA as this is considered a safe limit for protection of humans against electric shock.

Put safety first. Then worry about size and cost. An isolated SMPS (switched mode power supply) with 5 V output won't be any bigger than what you are proposing.

\$\endgroup\$
  • \$\begingroup\$ Thanks transistor for your input. I am definitely switching to an isolated SMPS. \$\endgroup\$ – user41391 Jul 25 '16 at 22:15
  • 2
    \$\begingroup\$ ...generally trip at 30 mA - generally below that, even. Depending on whose code book or standards document you're reading, top class GFCIs trip at 4-6mA and standard class at 20mA. At 30mA and up you're talking more about ground fault protection equipment (GFPE). GFCIs also trip fast - faster than a 500mA fuse would, certainly. Even an FF (very fast) 500mA fuse would push 200% current (1A) for 10-20sec or so before blowing. That's more than enough time to cause harm - well beyond the AC-4 zone of 50%+ chance of fibrillation. \$\endgroup\$ – J... Jul 26 '16 at 0:41
  • \$\begingroup\$ "You must isolate the low voltage circuit from mains if there is any risk of contact with the circuit" - is there such a safety regulation, or is it just a good practice? \$\endgroup\$ – C K May 29 at 12:25
  • \$\begingroup\$ @CK: Safety, common sense, good design, ethics would be my drivers. Regulations will vary by region so I don't think I can answer that other than to look at the CE Low Voltage Directive for products for sale in Europe. \$\endgroup\$ – Transistor May 29 at 13:09
4
\$\begingroup\$

No. 500 mA is significantly greater than the current necessary to put a human heart into fibrillation.

The main potential failure mode here is if the device is plugged into mains wiring with the live and neutral conductors swapped (which is disturbingly common). Then both your "+5V" and "0V" outputs will be superimposed on top of line voltage, and will be an immediate hazard.

Or if you're using a full bridge rectifier on the front end then the output will be floating at \$\frac{1}{\sqrt{2}}\$ of mains voltage even if the outlet is correctly wired!

\$\endgroup\$
  • 3
    \$\begingroup\$ Your first paragraph seems to imply that if the fuse were small enough (smaller than the fibrillation current) it would protect users, but it would not! A fuse needs time to blow, and can conduct a lot more current than its rating before it does. I reiterate: no fuse value can make this approach safe. \$\endgroup\$ – marcelm Jul 25 '16 at 20:21
  • \$\begingroup\$ As well as L-N reversals you also have to consider broken neutrals. \$\endgroup\$ – Peter Green Jul 26 '16 at 4:33
4
\$\begingroup\$

There must always be a full isolation barrier between AC mains and anything a person might touch, and it must be arranged in such a fashion that no plausible single-point failure can breach it (certain components specified such that fail-shorted scenarios need not be considered plausible). That point should be considered non-negotiable.

If you don't want to put an isolation barrier between your electronics and the mains, you must put an isolation barrier between those electronics and all persons nearby, i.e. don't let anyone touch anything connected to the electronics. It's unclear why you are expecting people to touch the pins on your controller, but if you are expecting that, you must use an isolated supply.

A fuse will reduce the likelihood that a failure in your device will start a fire. If you're lucky, the fuse may reduce the harm caused by some other kinds of failures, but fuses aren't really good for much beyond fire prevention. Devices that don't burst into flames are generally safer than those that do burst into flames, so in that sense fuses do improve safety, but they in no way affect the isolation-barrier requirements.

\$\endgroup\$
  • \$\begingroup\$ Thank you. By this time I have realized that it is extremely important to isolate my low voltage electronics from the mains. The reason I expect people to touch the pins on the microcontroller is because it is a prototype that will be handled by individuals other than myself. I will also take great caution to enclose the AC part of the circuit before the isolation barrier (i.e. isolated SMPS) \$\endgroup\$ – user41391 Jul 25 '16 at 22:11
  • 1
    \$\begingroup\$ Thats good. The easiest, cheapest, and safest thing to do is to run off on off-the-shelf "wall wart" power supply. If your device can run off of 5V 500 mA-2A you can put a USB connector on and use a phone charger. If you need 12 V, you will have to go with a barrel connector. That should be plenty good for a prototype even if the final device will have an integrated PSU. \$\endgroup\$ – Evan Jul 26 '16 at 4:47
  • \$\begingroup\$ THIS! There are plenty of non-isolated electronics in daily use (eg remote-controlled sockets, socket-timers or kill-a-watts), but every 5V electronics inside must be treated as operating at mains voltage. Because it is operating at mains potential. \$\endgroup\$ – Agent_L Jul 26 '16 at 11:38

Your Answer

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