Yes, thats what i ment by putting the pads off center.
But i would use pins 13 and 23 for the ground side (instead of the amps hv-output).
That way A1, A2, 13, 23, all would be low ground potential, no need for "off center" pads on the 4-pin "crammed" side, only need for "off center pads on 2-pin side of the relay (more clearance between the amps hv-output and relay coil, and propably also a bit lower capacitance to the coil).
Originally, when i suggested 4 relays instead of only 2, i was thincking using 13 for ground and 23 for protective earth and the 2 pins on the far off side (connected together) to each of the amps bipolar hv-outputs. No offcenter pads would be needed then.
This would also shortcut possible amplifier ground and protective earth voltage differences without introducing any hum issues under normal operation.
Last edited:
I don't quite get how this would work. If the pins are connected together, and to the bipolar amp outputs, they would short the bipolar output. I don't think that's what you had in mind. Can you clear up the fog in my brain?
I see your point, but (i) the aim is to sense high-voltage faults at the amplifier output and protect against these and (ii) proper grounding of the amplifier must always be in place and is independent of high-voltage issues.
Last edited:
Sorry Mathias, i was was a bit unclear.
With 4 relays, 1 for each output of the 2 channel differential outputs:
From your picture in #99, DPST NO relay, i would connect pin 13 to GND and, when the relay activates, it would make contact to 14 ( the, what i called "far off side").
23 would be connected to the grids protective earth and, when the relay activates, make contact to 24 (again, what i called "far off side").
The "far off hv-side", pins 14 and 24 would be connected together to one of the outputs of the differential amplifiers output.
In your conceptional drawing #79, GND and protective earth are not connected together, and the relay shorts the output only to protective earth.
To protect, GND and protective earth must be interconnected by means that assure that both are at the same safe potential under all circumstances.
I would interconnect GND and protective earth either directly (or if possible groundloops are a problem by a pair of sturdy bidirectional paralelled diodes)
or at least with the additional contacts 4 DPST relays would provide during a fault.
I hope it easier for you now to understand what i ment
I
With 4 relays, 1 for each output of the 2 channel differential outputs:
From your picture in #99, DPST NO relay, i would connect pin 13 to GND and, when the relay activates, it would make contact to 14 ( the, what i called "far off side").
23 would be connected to the grids protective earth and, when the relay activates, make contact to 24 (again, what i called "far off side").
The "far off hv-side", pins 14 and 24 would be connected together to one of the outputs of the differential amplifiers output.
In your conceptional drawing #79, GND and protective earth are not connected together, and the relay shorts the output only to protective earth.
To protect, GND and protective earth must be interconnected by means that assure that both are at the same safe potential under all circumstances.
I would interconnect GND and protective earth either directly (or if possible groundloops are a problem by a pair of sturdy bidirectional paralelled diodes)
or at least with the additional contacts 4 DPST relays would provide during a fault.
I hope it easier for you now to understand what i ment
I
Last edited:
I would suggest a different solution to remove dangerious HV when needed. A couple of SCR's could remove the HV in microseconds once tripped, (crowbar) probably fast enough to be really safe. The sense/control circuit is then the issue. And you could borrow the circuit concept from a GFCI. Need a ferrite ring to sense imbalance in the currents going to the transducers. If it exceeds 2 mA? then crowbar and disconnect the supply. Its not trivial circuitry but not really difficult today. Maybe scavinge components from a GFCI ($10-15 here) to start with. If you can make it work maybe get a patent?
I see Stax is now making a planar Magnetic headphone. Nothing is sacred any more.
I see Stax is now making a planar Magnetic headphone. Nothing is sacred any more.
Ah, ok, I see what you mean. I didn't show any connection between GND and PE in my drawing because I assumed this connection is already in place in the amplifier (either a direct connection, or through a ground-loop breaker as you described). I'd like to stick to the goal of protecting for high-voltage failures that cause a current flow from the amplifier output through the user/listener. Proper grounding should be dealt with anyway, so I feel protecting GND just adds unnecessary complexity.
I don't quite see the advantage of SCRs over a good old relay. Most mechanical relays out there have much faster switching times than the time constant of the fault sensing needed here.
That would work only for faults that result in a current imbalance between the two rails of the amplifier output. It would not sense the fault if both output rails short to B+ or B- in the same way, and both rails would carry the same current.
Did you see my post #79 illustrating my idea?
Are you protecting the user or the circuit? Any current going through the user is returning on a different path than the cable.
In any case RE fuses there are small fuses I.e. 50 mA that act quickly ( I know from experience) and there are fuses specifically for protecting semiconductors but they start at 40A.
In your circuit do you have enough current to drive the Led's in the optocouplers? If you crowbar the supplies do they have foldback or fuses?
In any case RE fuses there are small fuses I.e. 50 mA that act quickly ( I know from experience) and there are fuses specifically for protecting semiconductors but they start at 40A.
In your circuit do you have enough current to drive the Led's in the optocouplers? If you crowbar the supplies do they have foldback or fuses?
The user. When the user is dead, he, she, they or it can't listen to the circuit anymore anyway.
A 50 mA fuse cannot (by itself) protect the user sufficiently. Anything above 30 mA can be quite lethal if it is AC with the wrong frequency, and the time for the fuse to blow will be quite long when the current is just above 50 mA.
A 50 mA fuse cannot (by itself) protect the user sufficiently. Anything above 30 mA can be quite lethal if it is AC with the wrong frequency, and the time for the fuse to blow will be quite long when the current is just above 50 mA.
Again is this protecting a user or the circuit? And even with a crowbar you will need a fuse or foldback regulated power supply.
Worse; that chart is only for electrocution though the hands/feet I believe - ear to ground would be another chart, and many electrical related injuries are indirect due to the victim reacting to the surprise of the shock (even mild shocks can make you have a full limbic system startle response, which can have significant risks if, say, driving, holding a hot cup of coffee, etc.). I wouldn't be surprized if an ear shock could directly activate jaw muscle motor-nerves and make you bite your tongue tip off either... ECT is best administered by professionals(!)
No but I assume its from a 5-digit ISO standard relating to industrial safety and not specifically electrostatic headphone safety - surely?
@Mark Tillotson See post 54 and 68 for a reference of the current/time diagrams. I was asking if you have anything to substantiate your claim that limits for current from head to feet would be different than for (left) hand to feet. As far as I can tell, the major factor in such assessments is if (and to which extent) the current flows through the heart.
Prisoners on the electric chair are electrocuted from head to leg, so there should be plenty of experience with the required current levels in the USA.
