the noise of pro tools HD converters. dirty dirty.

doing wiring the other day in a room at work that is complete pro tools hd with icon setup. With pro tools on input and feeding it a sine wave was amazed at how much extra crap the D/A stage of the 192 was adding in the signal. Had the o-scope set up to see the input and output waves. Coming from the signal generator had a nice clean sine wave on screen that was 1K @ 1.228V. Coming out of the 192 analog had a dirty sine wave that had a lot of artifacts and noise. It was not clean at all. This with a sample rate of 44.1khz. b24 bit. As I increased the sample rate, the noise was less noticeable on the scope. At a 192khz sample rate there was still some noise in the output sine wave but not nearly as bad at 44.1 Went digital out of the 192 into Tascam mx2424 and out of the 2424 analog. At 44.1 had much improvement with the sine wave looking almost identical to the one on the input. There was no noise in the wave at all.

so I'm trying to figure out where that noise is coming from. Obviously it's has to do with the d/a part of the 192. But is it the switching supply inside the 192 inducing noise? I have noticed this before a neve vr and pro tools but never actually put it on a scope.

This has been there in all versions of PT. 

I scoped the 888's and /24's years back.  Same deal except there wasn't a higher sample rate to smooth it out.  I just upgraded the clock which helped remove some of the hash I was hearing, but it was always there.  If someone has a better idea as to how to really improve this, I'm all ears.  I think it was also a little freq dependent.  The midrange always seemed to have more hash via my ear.  Dunno. . . .

Michael

anyone tried Prismsound ADA-8XR

But is it the switching supply inside the 192 inducing noise

Click to expand...

It's possible.  Broadband hash can introduce all kinds artifacts but you'd have to see if you can see if noise is on the powersupply outputs.  Other than that you could have clock problems too that could be from the noisy powersupply or simply a noisy XO/PLL that drives the converters.  Since you came out of an MX2424 and it was clean, I would say that the digital supply is likely clean and the logic clocks are clean which could also say that the supply is clean.  We all know that SMPS hash will get all over the place, not just a couple rails.

A full workup with a spectrum analyzer is warranted here.  An Oscope can only tell you part of the problem.

record 16- 24bit sine waves and we dissect if its applied  dither of some sort or not.

But is it the switching supply inside the 192 inducing noise?

Click to expand...

Very unlikely. Why should the PSU behaviour change with sample rate?

So I'm trying to figure out where that noise is coming from.

Click to expand...

We need at least a scope photo.

Samuel

Why should the PSU behaviour change with sample rate?

Click to expand...

nice catch!

the ole' 2424.... everyone has always liked them.

Samuel Groner said:

Why should the PSU behaviour change with sample rate?

Click to expand...

Noise at (a multiple of) the ADC's S/D modulator frequency is folded down (aliased) to the audio band. While unlikely, it's not impossible that EMI from the SMPS falls inside one of those bands for lower sampling frequencies. Still, this wouldn't be my first suspect.

JDB.
[and this is why you want to have your SMPS running synchronously with your sampling clock, so you have a measure of control over the noise spectrum]

Samuel Groner said:

We need at least a scope photo.

Samuel

Click to expand...

I will take one.

Noise at (a multiple of) the ADC's S/D modulator frequency is folded down (aliased) to the audio band.

Click to expand...

Aren't we talking about DA? I presume there can be similar issues though. In any case the PSU noise would need to be drastic if it is visible on a scope with a 1.2 V fundamental signal. I'd rather suspect some insufficient image rejection, even though with todays oversampling converters one ought to think this should be mostly independent of sampling rate too.

Samuel

the ole' 2424.... everyone has always liked them.

Click to expand...

Hated mine.  Everything sounded great while monitoring through it.  Record and playback = muddy poop.  Powersupply was an insufficient design and always had problems to be fixed.  The analog AD/DA boards were a monument to the engineers, sounded good but horribly over-engineered and expensive.

o.k. measurement test #2


the noise itself lies at  20 megahertz and appears to be random...

hmmmmmm.

pucho812 said:

o.k. measurement test #2


the noise itself lies at  20 megahertz and appears to be random...

hmmmmmm.

Click to expand...

Well...

Which is it?

(And how are you hearing 20MHz, by the way?)

well keith,

the noise is 20 megahertz and I got from the scope.

noisey isnt the word Id use to describe protools interfaces. But hey, Im no scope.

20mhz probably isn't powersupply hash.  Can we see the scope screenshot?