High voltage DAC
Dec 7, 2003 at 7:45 AM Thread Starter Post #1 of 5

Prune

Banned
Joined
Sep 16, 2003
Posts
349
Likes
0
There are direct digital amplifiers, which are essentially powerful DACs. So what's wrong with making a high voltage DAC that will be able to drive electrostatic headphones directly? And please don't tell me to just use a regular DAC and a Gilmore amp (as someone did on another board), since that doesn't answer my question.
 
Dec 7, 2003 at 10:12 PM Post #2 of 5
All dacs are current output devices with a maximum voltage
compliance of +/-15 volts. Many of the newer and faster ones
are +/-5 volts. So one way or the other you need a current
to voltage converter with a +/-400 volt swing.

Building a 24 bit dac with a voltage compliance of +/-400 volts
would be an item filled with lots and lots of discrete output
devices and would be just about impossible to trim.
 
Dec 7, 2003 at 11:03 PM Post #3 of 5
I think the output filter design (really the heart of the sound for a so-called digital amplifier) would be extremely difficult to design around those voltages. I may be offbase here, but I'd have to look into the response characteristics of power MOSFETs that can handle 400V. It may require using HEMTs or something like that to get fast enough switching?
 
Dec 8, 2003 at 12:14 AM Post #4 of 5
I tried to hook up a smART DI/O to my Headroom Little More Power. And it's output was way too strong. I was getting an excessive amount of staticy distortion. That same smART DI/O sounded fine with a Gilmore, though.

-Ed
 
Dec 8, 2003 at 12:31 AM Post #5 of 5
Quote:

Originally posted by kevin gilmore
All dacs are current output devices


The CS43122 (which is what's in my regular DAC) for example has voltage output.

In any case only ASDFer seems to have understood what I meant. I was talking about building a discrete DAC, not hooking up an HV output stage to the end of a DAC chip -- that would be stupid as it's no different from using an amp.

For a discrete R-2R stereo 24 bit DAC with differential output there would be 96 HV output devices driving the four R-2R networks, followed by some kind of sampler for deglitching, and the filter. It may be a lot of work, especially in matching the resistors in the newtorks, but that doesn't mean it couldn't work.
 

Users who are viewing this thread

Back
Top