[sdiy] Noise source in Bryston 10B crossover

[cheater cheater]

Sorry, in the last email I used "headroom" incorrectly, I meant "SNR" there.

On Sun, May 9, 2021 at 5:16 PM cheater cheater
<cheater00social at gmail.com> wrote:
>
> Hi all,
> I am using a Linkwitz-Riley based three-way crossover from Bryston. It
> turns out to be the major source of noise in my setup and I am trying
> to see if anything can be improved on the circuit in order to lower
> its inherent noise, which is standard at -90dB. It uses linear voltage
> regulators (7824C, 7924C) as well as what I guess are discrete
> transistor based op amps. My main issue is with noise in the HF
> output. I think it is the first output from the subsequent filter
> stages, which should make it easier to analyze the schematic... but I
> can't really make much sense out of it.
>
> My crossover is a PMC 10B, which is a rebranded Bryston 10B hardwired
> in three-way mono mode, and is balanced.
>
> 1. First of all I am trying to understand where the noise comes from.
> Are the op amps adding this noise? Is it the voltage rail noise being
> compounded by several stages of op amps?
>
> 2. If I wanted to improve noise on the voltage rails, what would I do?
> The 7824C has noise specd at 10 uV/V_O which I assume means 240 uV at
> 24V, and the 7924C is specd at 170 uV at 24V. That gives us 410 uV /
> 48V = -101.36 dB - can this be improved upon? There isn't a huge
> selection of linear 24V 1A regulators on Mouser. The only other is
> from ROHM, and (I'll spare you translating the Japanese data sheet)
> their +24V regulator is at 170uV, which isn't that much better than
> 7824C.
>
> 3. What could I do to improve the filters themselves? Can I replace
> the transistors used with something that provides less noise? Could I
> replicate the boards with the same topology but with modern low noise
> op amps instead of the discrete transistor based ones used by Bryston?
> There are op amps that go as low as 1nV/sqrt(Hz) nowadays... I wonder
> how hard it would be to repurpose the schematic to a low noise
> integrated op amp? Note that this device runs at +/-24V rails - I'm
> not sure how one would use modern low level op amps like this - any
> suggestions here?
>
> 4. The noise is constant, white-ish, and every now and then it will
> fluctuate and become quieter for a second or so. What causes this
> fluctuation?
>
> BTW I know I can get rid of a bunch of the noise using gain staging,
> but I would like to eke out more dynamic range out of this set up, as
> the following Bryston electronics easily hit -113 dB noise, so it
> sucks to lose 24 dB headroom to noise.
>
> Schematics: https://archive.org/download/bryston_10B_SCHEMATICS/10B_SCHEMATICS.pdf
>
> Photos: https://imgur.com/gallery/CNcXtZn
>
> Discussion on DIY Audio:
> https://www.diyaudio.com/forums/analog-line-level/372026-noise-output-pmc-main-monitors-due-bryston-crossover.html