abbey road d enfer said:
Maybe I got it wrong but I think the orientation of the zener is not correct, on account BSS110 being a P-channel.
opps... yes I grabbed the wrong schematic symbol... the mosfet I used is actually a BSS87 a N-channel 200V part. [edit] I went back and used a correct schematic symbol but still wrong PN [/edit]
The circuit would probably work with a P channel MOSFET but C1 might need to be larger to support the relay coil current draw. As drawn when the N channel MOSFET turns on, the current can be directly supplied from the mains. A minor point but not an accident.
The zener clamp is probably not even needed if my resistor divider can keep the gate voltage within the +/-20V range.
Now I don't understand at all what the circuitry around OK1 does, except heating R6 in case of ground potential difference
.
I want to be enlightened.
OK1 is actually an opto-triac (not opto-transistor). PN MOC3023.
The diode part of OK1 is in series with roughly 1/2 the ground path current so more than a few mA of ground current will cause the OK1 output triac to conduct and shunt current from the line output of the GFCI to it's neutral input. This will cause a current imbalance and trip the GFCI, which will de-energize the relay and lift the ground. This ground current sense works independently of the GFCI so will protect the muso from a console with faulty power that energized the mic.
The interesting thing about this (for me) was the need to add C3 to prevent the triac in OK1 from false triggering from the rise time at turn on. The data sheet specs 1000V/uSec for max risetime. Without the cap the triac inside OK1 fires at turn on and trips the GFCI. The C to output side neutral keeps that cap current inside the valid GFCI loop, until the triac in OK1 is properly turned on and trips the GFCI. The triac has enough body capacitance that I could get the circuit to work with no cap for a larger value of R6 but I didn't want to be soft when tripping the GFCI in case there were competing currents in the regular GFCI loop. With the cap it will be more repeatable.
Note: my cheap and dirty (KISS) solution to just put a 0.15uF*** in series with the GFCI ground lead would work too. While it doesn't open the ground it keeps the fault current from an external shock hazard below lethal levels (for 120VAC).
In terms of operation, I think you should add a "polarity" inverter ahead, because as it is, in case of line/neutral inversion, the muso's life is saved, but he can't earn it because the goddamn thing wouldn't turn on!
I thought about this and don't like any solution that perpetuates using bad power. I'd prefer to have it zap any miswired outlets with a lightning bolt so they have to repair it.
I could add an automatic polarity swap with a second relay but that makes this even more expensive so even less commercial .
Again the KISS GFCI with stinger cap in place of the ground path, can still work and be used with miswired outlets.
JR
*** note 0.15uF is a stinger cap value for 120VAC 60Hz... for 240VAC 50Hz that value is too large and should be adjusted for that market.. In fact this design is for 120V only