Voltage regulators and intrinsic safety

[BenYL]

Hi Everyone,

What PCB-mount voltage regulators are usable in an intrinsic safety design? By "usable" I mean they would not make the system near impossible to get certified. My impressions from what little I know:
-DC-DC converter modules (made up of discretes, either re-enclosed or not) are not good because you need at details about the discretes.
-Linear regulators are fine as long as the temperature is kept down.
-Change pump regulators are generally fine (buck, boost, or buck/boost) because they don't have inductors. Perhaps ones that boost would need a clamp on the output.
-Buck converters with an inductor might work if the size of the inductor were not an issue, but analysis is difficult.
-Boost or Buck/Boost converters are difficult because they boost, they have an inductor, and analysis is difficult.

 

[Electric-Light]

Hi Everyone,

What PCB-mount voltage regulators are usable in an intrinsic safety design? By "usable" I mean they would not make the system near impossible to get certified. My impressions from what little I know:
-DC-DC converter modules (made up of discretes, either re-enclosed or not) are not good because you need at details about the discretes.
-Linear regulators are fine as long as the temperature is kept down.
-Change pump regulators are generally fine (buck, boost, or buck/boost) because they don't have inductors. Perhaps ones that boost would need a clamp on the output.
-Buck converters with an inductor might work if the size of the inductor were not an issue, but analysis is difficult.
-Boost or Buck/Boost converters are difficult because they boost, they have an inductor, and analysis is difficult.

If you're asking for a part # for a SoC based solution, there is no answer for you your request is too vague. What exactly are you trying to do?

You should read this. It gives you a damn good idea of limits and what-ifs you need to consider.

http://www.omega.com/temperature/Z/pdf/z131-148.pdf

Just as an example, many LED ballasts have an electrolytic capacitor directly in parallel with the LED stack (an arrangement of LED elements). This would fail, because the entire stored energy in the capacitor is available at the point of load side short.

 

[IntrinsicSafety]

My experience, for what it's worth.

-Commercial off-the-shelf DC-DC converter modules: These are very bad because you must have details about the internal construction of the device (bills of material, schematics, PCB layouts, encapsulation details, etc.) so worst-case parameters can be established (capacitance, inductance, thermal, etc.). They are fault-able and must be limited by infallible zener diodes, which are themselves difficult to make work. Thermal parameters and/or adequacy of potting may be impossible to establish. And they contain inductance, which is pure poison to IS designs.

-Linear regulators: No worries. (They are fault-able for thermal, but that's true for any device in an IS design.)

-Change pump regulators: They are fine but will be considered to boost to the limit of their capacitor-ladder architecture (i.e. doubler, tripler, etc.) unless otherwise limited by infallible zener clamps.

-Buck converters: Contain inductance, which is poison. Most agencies consider them fault-able to boost operation, so infallible zener clamps are still required, which are themselves a headache.

-Boost or Buck/Boost: As for buck, above, except everyone considers them boost-capable (obviously).

Moral: Don't use switching power supplies unless you have no other option.