Path: utzoo!utgpu!jarvis.csri.toronto.edu!cs.utexas.edu!uwm.edu!cbnewsc.ATT.COM
From: tjr@cbnewsc.ATT.COM (thomas.j.roberts)
Newsgroups: rec.audio.high-end
Subject: Re: Trimming MSB on DAC
Message-ID: <2653@uwm.edu>
Date: 28 Feb 90 18:15:34 GMT
Sender: news@uwm.edu
Lines: 45
Approved: tjk@csd4.csd.uwm.edu


> 
> In article <2589@uwm.edu> @sun.acs.udel.edu:hamilton@sun.acs.udel.edu (hamilton) writes:
>>
>>I read in the current issue of CD Review that it is possible to adjust the 
>>trim pots on the D-A convertors of a CD player to improve linearity.
>> ...  Linearity below -100dB is
>>something found only in the most costly CD players, and maybe in the new
>>1-bit machines.  Does anyone know what is needed to perform the adjustment?
>>Can it be done by the average joe?

This is a difficult adjustment to make, and one that only "purists" would
bother with (by "purist" I mean someone who is more concerned with talking
about a system than listening to it ... :-).

There is no hope of making this adjustment on normal music. You will need
a CD with special test tones (namely a very quiet sinewave, -60 to -80
dB below full scale). In virtually all cases, leaving the trimmer alone
(or at its center position) will suffice.

I have a 16-bit homebrew music synthesizer, and a 16-bit ADC, both
installed within a PC. The DAC has an MSB trimmer. I cannot hear the
effect of the trimmer on normal-volume music or sinewaves. The ADC
(used as an oscilloscope) cannot see it on normal-volume sinewaves.
When I run the DAC on a -72dB sinewave (wrt full-scale), and crank
up the gain in the speaker amp, I can barely hear the MSB adjustment when
it is at its extreme ends (pulling the MSB out of true by about 4
LSB counts) [note this "sinewave" has only 10 levels]. The ADC
oscilloscope (single sweep) has considerable noise at this level
(1 LSB = 0.08 millivolts), but I can infer the steps of the "sinewave";
I have not bothered to write the multiple-sweep averaging software
required to reduce the noise to the level needed to see the effect.
This is with the ADC Sample/Hold connected directly to the DAC output,
and differing sampling rates of the DAC (50 kHz) and ADC (33 kHz)
complicate the observation; looking after the DAC output filter washes
out the steps, and makes the trimmer effect invisible. 
The DAC is specified to have THD <0.1% WITHOUT TRIMMING;
I believe it (I installed the trimmer when I was having linearity
problems - they turned out to be completely unrelated to the DAC or
its trimming).

I simply leave the trimmer in the middle of its range, and am happy.

Tom Roberts
att!ihlpl!tjrob