DAC left channel dropping out. A mystery.

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Administrator
Joined 2004
Paid Member
OK, here is one for the "If I didn't hear it myself, I would not believe it" chronicles.

Last week I was having trouble with my system, the left channel kept dropping out. I thought it was a loose speaker cable or interconnect on the amp, but nothing seemed to be wrong there. Then I noticed that the left channel wasn't completely gone, just so low I had to put an ear to the speaker to hear it. Strange.

I found by accident that the cause was the Toslink optical cable coming into the DAC! It was stretched tight and maybe not seated well. Wiggling it would cause the left channel to come and go. Moving the DAC a bit closer and reseting the cable fixed it.

OK, how could this happen? The S/PDIF signal carried over the fiber is serial. It's not like analog connections with a wire for left and a wire for right. The left and right are combined in the S/PDIF stream. How can a bad connection cause just the left to drop out? Why not everything dropping, or random noise?
 
Hi,

OK, how could this happen? The S/PDIF signal carried over the fiber is serial. It's not like analog connections with a wire for left and a wire for right. The left and right are combined in the S/PDIF stream. How can a bad connection cause just the left to drop out? Why not everything dropping, or random noise?

I have no explanations but I have had the same experience, except NO SPDIF INVOLVED! Right channel 100%, left distorts or drops out.

It is not really relevant how we fixed it, or what the cause was, but it also affected left!

Ciao T
 
Administrator
Joined 2004
Paid Member
Come on. No more takers? A weak S/PDIF stream that causes the attenuation of the left channel only? Certainly that is bizarre enough to warrant further comment and speculation.
How can corruption of an interleaved stream affect only one channel?
 
Perhaps it was the bandwidth of that part of the stream effected ? I am talking out my pooper mind you.. I definately understand digital broadcast but not really spdif. . . . I use to think I had a failing amp, then preamp, or perhaps a DA converter in that order...and I bought new gear up to the point of the AD/DA interface.....Turned out my volume pots on my interface ( a line6 UX2 ) were oxidized, not only causing a channel imbalance but noise. I cleaned all the pots in my interface and never had channel imbalance or un-natural noise again...
Perhaps you inadverntantly cahnged your volume settings during your experiments and had not realized it ....especially since this was/is or began as an intermittent problem as did mine.. . but I got a spare amp and wicked preamp out of the ordeal . . . .if you want or need some good recommendations on how to clean pots properly let me know.
~erick~
 
I don't think the stream was corrupt...was my point...I thought I had many problems only to find out my volume knob on my interface was not tracking right. . . .giving me 10-20db channel imbalance.....digital tens to work or not work. . . .check your analog gear original poster. . . .mainly the volume and trim pots...if they are oxidized they not only will introduce noise and distortion they can mistrack with varying levels.....the intermitency of the original post points to thise very problem I had....
 
digital streams...are nothing but on/off signals...binary...they work or don't...they can get corrupt...but basically it works or it does'nt and toslink/spdif is a singal glass conductor of a light stream of on/offs.... you have analog problems for it to be so isolated...if it was global i would suggest a receiver or transmitter problem....
peace out. . . .
"MG- Chemical super contact cleaner/lube" has brought back to life ever pot I serviced with it. . . but I wash the board afterwards as there is lube and grease everywhere...
be sure to use anti-static measure and give your caps time to discharge as the cleaner is conductive....
 
Administrator
Joined 2004
Paid Member
Eric. Believe me, I checked and triple checked all the analog connections and the volume pot. That was my first thought, of course (see original post) That was not it. It really and truly was the Toslink connection. That's what makes it so interesting. Of course there is a small chance I missed something, maybe even something bad on the PCB, but I tried very hard not to.

Thorsten, I agree that it probably is a word clock issue, but I don't understand "how". That's what I'd like to know. This DAC chip has the S/PDIF receiver built in, as it's an ESS 9006. The signal does pass thru a SN74HC151D that acts as in input selector switch. But this DAC chip is hit with S/PIDF, not IIS like most DACs.

So what I find interesting about this will be trying to learn how the stream could be corrupted enough to drop the Left channel only, but continue to work. It may teach something about the S/PDIF signal. Or maybe someone with a strong knowledge of the S/PDIF signal will say "Impossible" and I'll have to go back and look elsewhere. For the moment I've put that DAC aside and am using another.
 
Word clocks only do one thing...they time the sampling. I know clocks. I use a $2200 Lucid word clock in our recording studio...clocks tell the ad/da when to trigger a sample...clock jitter is noise because your sound wave is not accurate because your timing is not accurate. Nothing more nothing less is done by a Clock ( unless you are talking about synchronization for video production)..
I am not going to look up your convert but is it dual DA? (some are intervleaved 1 bit delta sigma designs with one da converter- does anybody make this crap anymore ? doubt it ) perhaps one converter is failing...
do you use recorded material (not commercial recordings) if you were passing clipped signals through your DA converter excessively I could see an output amp failing as a clipped digital signal DA will tell the amp to stay on for the suration of the clipped samples. I could see this causing cheap gear to fail over time... I use a digidesign 192 I/O with a Lucid clock so we do not have these problems ...we even have soft clip limiters on the buffers to smooth out and soften clips protecting our recording and gear. . . just ideas for you...I am just a recording engineer not an electronic engineer. . .
 
Administrator
Joined 2004
Paid Member
My guess... you have a dry/intermitent somewhere and touching the SPDIF cable "appears" to fix it as it flexes the PCB :)

Yeah Mooly, it could be as simple as that. I'll have to check into it. If that's the case I've been chasing a phantom and wasting everyone's time. :eek:

Eric. This is a DAC board I designed. The output stage is transformer only, no active devices outside the chip.
 
I also think it is a poor connection elsewhere that get's fixed with pcb flexing.
But grasping at straws here, the pulse-width or rise/fall time coming out of the TOSLINK receiver might be marginal which gets aggravated with a weaker (optical) signal. Why the right channel is favored does seem odd.
I can hold an optical cable over an inch away from a TOSLINK rx and get clear audio.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.