110pFd is embarrassingly small; device and wiring C is super significant.
Specs work out to 1,000V and 9mA peak (at 10KHz). (I note that the companion amps mostly aim for 350Vrms 500V peak.)
I was under the impression that even shrill recordings did not have FS levels past 6KHz; if they do now, I would not want to hear them. 20KHz power response is not needed. We do not wish to strain at 6Khz; 10KHz seems a good pencil target for a very-good ear-burner.
Then the "ideal" amplifier is as shown in the attached?
NDFPD1N150CG is hasty-search candidate for an all-N-type output: high voltage, low C, ample dissipation.
Specs work out to 1,000V and 9mA peak (at 10KHz). (I note that the companion amps mostly aim for 350Vrms 500V peak.)
I was under the impression that even shrill recordings did not have FS levels past 6KHz; if they do now, I would not want to hear them. 20KHz power response is not needed. We do not wish to strain at 6Khz; 10KHz seems a good pencil target for a very-good ear-burner.
Then the "ideal" amplifier is as shown in the attached?
NDFPD1N150CG is hasty-search candidate for an all-N-type output: high voltage, low C, ample dissipation.
Attachments
Last edited:
yes the headphone cable C is very significant part of the load for ES headphones
and Coss, Crss are 1st order important for Q choice
NDFPD1N150CG seems to have good specs
I looked into IXTP01N100D - depletion mode is nice, a little high Coss, may need to cascade at higher V
and I believe the Cree SiC C2M1000170D is used by Gilmore's "Carbon"
and Coss, Crss are 1st order important for Q choice
NDFPD1N150CG seems to have good specs
I looked into IXTP01N100D - depletion mode is nice, a little high Coss, may need to cascade at higher V
and I believe the Cree SiC C2M1000170D is used by Gilmore's "Carbon"
Last edited:
Making a big fuss about specs sounds more like work than like hobby to me. The point of designing another hobby amp is that you like doing so.
Anyway, my direct-drive amplifier for an electrostatic loudspeaker had a distortion of about 0.2 % at 10 kHz, measured single-ended. The real (balanced) value was probably lower, but my distortion meter has no differential input.
Great idea, I 'll try to do this!
And I just thought how: It should sense the existence of a current drawn towards the ground (instead of the current drawn to the output terminals). When that current reaches a threshold, it will cut the power until the user resets it (like a home safety relay) in a millisecond or so.
The challenge is to make it 1) reliable, 2) with minimum part count and 3) not load or affect the outputs -it should monitor all four outputs BTW. Or six if you include the bias (which is rather not dangerous due to the high resistance in series).
Last edited:
yes, at a head-fi meet I quickly handed back a Bamboo shelled ES headphone when it made my hand tingle
I don't know if any regulatory agency has said anything about commercial ES Headphones or Amps - EN-60065 only costs US$514 from UL
Do It Yourself - www.Head-Case.org threads do show a specs race - PS stored energy higher than a defibrillator, bias currents to 10s of mA (direct contact shock protection RCD trip current is 5 mA)
Last edited:
The challenge to achieve a difficult goal that you desire (like awesome specs), makes design more fun and worthwhile.
The same with the challenge to make something novel and better -at least in some ways, and owing something I can't currently buy. If it's novel, that's even better as no-one else can buy something similar either
If you enjoy making circuits with awesome specs, go ahead and have fun! I generally like building somewhat unconventional circuits that I like conceptually. Of course they also should work well.
Congratulations! You are the first in the world who publishes his amp's specs at a frequency other than 1Khz!
That said, my goal is to beat that number by 20 -at least, and at 20Khz, not 10.
The measurements have been on the Internet since early 2004, in the zip file at the bottom of this page: Elektrostatic Loudspeakers
Frankly, my main goal when I made my direct-drive amplifier for electrostatic loudspeakers was to survive; its 2.66 kV supply that could easily deliver lethal currents scared the hell out of me. That's why I settled for a rather simple and inefficient topology that needed as little debugging as possible. It sounded fine, though, although the maximum volume was insufficient.
You missed the joy, brain tickle at high frequencies in sync with your favorite song -by design
What costs $514?
I downloaded the safety standards but there is no mention about electrostatics and high voltages. Obviously the maximum db are violated by Stax and others.
http://www.itu.int/en/ITU-T/Worksho...stening_devices_situation_analysis_report.pdf
I'll take a look at that site, thanks.
http://www.itu.int/en/ITU-T/Worksho...stening_devices_situation_analysis_report.pdf
In the post that's been deleted, you wrote about the current being dependent on whether someone touches the cable. Wouldn't capacitive currents from the cable to the body of the listener trip your protection?
Actually if they would, that would probably be desired behaviour. There should not be any significant signal current flowing through the listener's body, no matter whether it's due to faulty insulation or due to capacitive coupling.
You deleted a post it took me an hour to write? You could ask me to correct it instead -if it was wrong.MrMagic, i deleted your post with wrong quotation attributions, you can post again and please make sure you get it right...
Last edited:
Of course they would, but there is no comparison between a very low current due to the wire capacitance -which is harmless and a direct contact one -hence the threshold.
the cable itself is 45 pf. So a significant amount of the load. And yes if you hold the cable, the capacitance goes up. (measured stator to stator)
the 2.5 meter extension cable is 95 pf. (sre725)
you don't want the stator voltage swings to exceed the bias voltage, 580v otherwise the diaphragm attracts particulates.
the cable is a 6 wire flat cable, and has different capacitances depending on where you measure
pin 1 L+
pin 2 bias
pin 3 L-
pin 4 R-
pin 5 bias
pin 6 R+
the 2.5 meter extension cable is 95 pf. (sre725)
you don't want the stator voltage swings to exceed the bias voltage, 580v otherwise the diaphragm attracts particulates.
the cable is a 6 wire flat cable, and has different capacitances depending on where you measure
pin 1 L+
pin 2 bias
pin 3 L-
pin 4 R-
pin 5 bias
pin 6 R+
Last edited:
Thanks, Kevin.
All-- (available) P-type MOSFETs still only go to 600V?
Yes, that also looks promising. Depletion means we can have active pull-up without floating battery, or the significant "efficiency" hit of resistor loaded. The "efficiency" gain with active pull-up just about cancels the lower breakdown voltage for +/-500V swing. While we could brain-pain the different capacitances, active pull-up just seems sweeter.
All-- (available) P-type MOSFETs still only go to 600V?
Yes, that also looks promising. Depletion means we can have active pull-up without floating battery, or the significant "efficiency" hit of resistor loaded. The "efficiency" gain with active pull-up just about cancels the lower breakdown voltage for +/-500V swing. While we could brain-pain the different capacitances, active pull-up just seems sweeter.
commenting on moderation is against the rules. so is ignoring moderator directive, you are warned...ixys 10m90s cascade with dn2540 works well at up to 900v as current source.
there are other ixys depletion mode parts that work past 1kv but are slower due to input capacitance.
anything else that is better is far more complicated requiring a floating battery
and npn transitors >1kv are now hard to come by
there are other ixys depletion mode parts that work past 1kv but are slower due to input capacitance.
anything else that is better is far more complicated requiring a floating battery
and npn transitors >1kv are now hard to come by
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.