Direct-drive amplifiers can indeed be lethal, but for any ESL that has kilovolt signals on its stators you have to make very sure that no-one even comes close to the stators, no matter whether the signal comes from a direct-drive amplifier or from a transformer.
Is that true? How many milliamps would pass through you via a transformer? Especially if we are talking of getting shocked by the audio signal.
With DD, you have the whole bank of smoothing capacitors in the B+ supply to zap you and no series resistance.
B.
If you want to commit suicide, connecting yourself to the power supply of a direct-drive amplifier will probably give you more chance of success than touching the secondary connections of a step-up transformer while playing loud music, although I'm not even sure about that, because AC tends to be more dangerous than DC.
If you want to survive, you have to regard both as potentially lethal. The current levels I calculated in post #18 are high enough to cause heart fibrillation. They are also in the order of magnitude you would expect to get out of a step-up transformer. For example, a power amplifier with 5 A current limit connected to a 1:100 transformer would ideally result in a 50 mA short-circuit current on the secondary side.
I think it is relatively easy to put an insulating enclosure around a direct-drive amplifier. Putting an insulating enclosure around a loudspeaker is more difficult, because it needs to be acoustically transparent. The QUAD ESL-63 has dust covers (4 um thick mylar) around its panels, then an earthed metal grid, and then loudspeaker cloth. I think the main safety precaution is the earted metal grid.
I think the main safety disadvantage of direct-drive amplifiers is that they are likely to need debugging. That means taking measurements with the amplifier not enclosed. You have to be bloody careful while doing this. It is also the reason why I used a very straightforward topology for my own direct-drive amplifier design, I wanted to limit the debugging phase as much as I could.
If you want to survive, you have to regard both as potentially lethal. The current levels I calculated in post #18 are high enough to cause heart fibrillation. They are also in the order of magnitude you would expect to get out of a step-up transformer. For example, a power amplifier with 5 A current limit connected to a 1:100 transformer would ideally result in a 50 mA short-circuit current on the secondary side.
I think it is relatively easy to put an insulating enclosure around a direct-drive amplifier. Putting an insulating enclosure around a loudspeaker is more difficult, because it needs to be acoustically transparent. The QUAD ESL-63 has dust covers (4 um thick mylar) around its panels, then an earthed metal grid, and then loudspeaker cloth. I think the main safety precaution is the earted metal grid.
I think the main safety disadvantage of direct-drive amplifiers is that they are likely to need debugging. That means taking measurements with the amplifier not enclosed. You have to be bloody careful while doing this. It is also the reason why I used a very straightforward topology for my own direct-drive amplifier design, I wanted to limit the debugging phase as much as I could.
Helpful post. Thank you.
On my DD amp, picture posted somewhere, I had a draining resistor in the power supply (of course) as well as another just sort of attached on the outside for misc safety purposes.
In long R&D and several decades of use, I never received a lethal (or any) shock, that I can recall. Or maybe that's why I can't recall. But VERY dangerous unless you are lucky and wear rubber boots and work with one arm behind your back.
But about the AC/DC question, perhaps you are thinking of the infamous experiment Edison did where he publicly electrocuted a dangerous elephant using AC to prove Tesla wrong about AC power distribution.
B.
Last edited:
No, I vaguely remembered from university that 50 Hz to 60 Hz is the most dangerous frequency range because of heart fibrillation and that anything over 30 mA could be enough. As university is several decades ago, I also looked up the current levels required for electrocution on Wikipedia.
"Low" by elctrical distribution standards:
Electrical Safety of Low-Voltage Systems: Massimo Mitolo: 9780071508186: Amazon.com: Books
pretty good chapter with human safety curves, basic modeling
Electrical Safety of Low-Voltage Systems: Massimo Mitolo: 9780071508186: Amazon.com: Books
pretty good chapter with human safety curves, basic modeling
Years ago, the US Navy had more deaths from 110/120VAC than all other voltages.
2 reasons: more access to that voltage; and disrespect for that low of a voltage.
The early TV sets CRT anode high voltage did not have enough current to kill.
But there was a documented case where that low current caused a muscle reaction, and the person was thrown across the room and broke his neck.
As always, all electrical systems need to be treated as if they could kill.
2 reasons: more access to that voltage; and disrespect for that low of a voltage.
The early TV sets CRT anode high voltage did not have enough current to kill.
But there was a documented case where that low current caused a muscle reaction, and the person was thrown across the room and broke his neck.
As always, all electrical systems need to be treated as if they could kill.
... thinking about the safety challenge ...
... would it be possible to implement some kind of "HFI relay" (so it is called in Danish - don't know what the English word is) arrangement in an electrostat so that when an "error" current was detected - as if someone touched the stators and conducted current outside of the electrostat circuit - then immediately the voltage to the stators was reduced to virtually zero so that the person touching the stators was not harmed? ... An additional challenge could be to do this in a way that did not affect the sound-quality ...
But might it be an option?
Cheers,
Jesper
... would it be possible to implement some kind of "HFI relay" (so it is called in Danish - don't know what the English word is) arrangement in an electrostat so that when an "error" current was detected - as if someone touched the stators and conducted current outside of the electrostat circuit - then immediately the voltage to the stators was reduced to virtually zero so that the person touching the stators was not harmed? ... An additional challenge could be to do this in a way that did not affect the sound-quality ...
But might it be an option?
Cheers,
Jesper
Not possible, because no current is flowing in normal case, and couple of (lethal) milliampers are flowing through the body at failure. A residual current device working on the principle of balanced currents in the forward and return wires can't be applied here.
But if someone touched one of the stators, the current that this one stator draws would be different from the current at the other stator. In principle, that can be detected and the amp could be shut off. Or using a HV relay just shut off that one stator.
But that does impact the sound 😉
Jan
But that does impact the sound 😉
Jan
Hi icsaszar & Jan ... thank you both for considering my safety issue thoughts ...
Wouldn't it just be great if sometimes good solutions were simple - and also good sounding 🙄
Cheers,
Jesper
Wouldn't it just be great if sometimes good solutions were simple - and also good sounding 🙄
Cheers,
Jesper
- Status
- Not open for further replies.