PS Audio DLIII DAC Inverts Absolute Polarity


Just wanted to let everyone know, who has a Digital Link III from PS Audio, that they may not be realizing the full potential of this incredible DAC.

I was fortunate enough to come across a new old stock unit from a dealer in CA recently. As I was waiting for the unit to arrive, I decided to catch up on some reviews to get a feel on what others thought about this DAC. I then came across a press review from Audio Xpress magazine where they described a slight flaw with this DAC.

The article says that the DLIII actually inverts polarity. What this means is that if you connect your DAC normally with either the RCA or XLR outputs, the output will be 180 degrees out of phase (reverse polarity). As I read this, I started to panic and thought I just purchased a dud. But the article goes on to say that this can be easily corrected in one of two ways: by swapping pins 2 and 3 on the XLR outputs or if all the music sources are being routed through the DAC, then you can simply reverse your speaker connections.

So I then decided to enlist the help of Rob Fritz at Audio Art Cable to make a custom XLR-to-RCA cable that makes pin 3 from the XLR connector as the Hot signal on the RCA instead of Pin 2.

As I was waiting for the custom cable, I decided to break in the DAC for a few days with the normal connections.. After about 100 hours, I was not that thrilled with the sound. Though it was warm sounding, the soundstage was a bit odd and well in front of the speakers. The bass also was very loose and ill-defined and the highs seemed a little rolled off. When I started to compare my observations with other press reviews, others had similar things to say, like a very forward soundstage.

But once my custom XLR-to-RCA cable arrived, the sound went from good to phenomenal. The soundstage was incredibly open as well as wide and deep. The treble was smooth and had great extension and refinement. But the biggest improvement was the bass. It was more natural with great definition and punch.

The reason I am sharing this experience is that many people may not know about the phase issue with the DLIII and may not be realizing the full potential of this incredible DAC.
wkass
So I then decided to enlist the help of Rob Fritz at Audio Art Cable to make a custom XLR-to-RCA cable that makes pin 3 from the XLR connector as the Hot signal on the RCA instead of Pin 2.

As I was waiting for the custom cable, I decided to break in the DAC for a few days with the normal connections.. After about 100 hours, I was not that thrilled with the sound.
Can you describe more specifically the connections you were using prior to the arrival of the custom cable? The reason I ask is that many XLR-to-RCA adapters, while connecting XLR pin 2 to the RCA center pin, also short pin 3 to ground (pin 1). Which as I see it makes no sense when adapting XLR outputs to RCA inputs (although it is proper practice when adapting RCA outputs to XLR inputs). Although many designs will not have a problem with that, the output stages of some equipment will not be happy having the signal they provide on pin 3 shorted to ground, and their performance may be compromised. See this thread for example.

So I'm wondering if the difference you perceived might have been due mainly to eliminating an improperly configured adapter arrangement, rather than to correcting absolute polarity. The consensus of many discussions that have been held here about absolute polarity, as I see it, is that it matters mainly for recordings that have been engineered using purist techniques, i.e. a minimal number of microphones and minimal mixing and processing. On other recordings, which in most genres seem to comprise a considerable majority, a multitude of different mics are mixed together and processed with a hodgepodge of random polarities.

Nicely written post, btw. Regards,
-- Al
Al,
Thanks for your response. The way I connected it up orginally was just using the standard RCA outputs since my Simaudio integrated amp has only RCA inputs (so just used RCA cables).

If you do a google search on XLR-to-RCA adapter, every pin-out diagram I have seen shows Pin 2 to signal pin on RCA and Pins 1 and 3 to the GND shield on the RCA. All I did with my custom adapter was just swap Pins 2 and 3 so that Pin 3 is now the signal pin on the RCA and Pins 1 and 2 go to the GND shield on RCA.

This simple change made quite an audible difference for me, and for the better.
If you do a google search on XLR-to-RCA adapter, every pin-out diagram I have seen shows Pin 2 to signal pin on RCA and Pins 1 and 3 to the GND shield on the RCA.
Yes, most of them are (incorrectly) designed that way. My understanding is that Cardas adapters are one exception, which correctly leave pin 3 unconnected on adapters that have an XLR-female connector, and therefore would be used to adapt XLR outputs to RCA inputs. Shorting an output signal on pin 3 to ground, as I see it, accomplishes nothing whatsoever (apart from in certain situations involving transformer-coupled outputs; I presume that's why it's done), while creating the potential for adverse sonic effects on some output stages (as illustrated in the thread I linked to).

As I indicated previously, though, pin 3 SHOULD be shorted to ground on adapters that have an XLR-male connector, so that pin 3 on the XLR input they would presumably be connected to is not left floating.
The way I connected it up originally was just using the standard RCA outputs since my Simaudio integrated amp has only RCA inputs (so just used RCA cables).
I would point out, therefore, that the change you implemented actually changed three things at once:

1)You corrected the absolute polarity of the system (which per my earlier comment won't necessarily be better for many recordings).
2)You changed the output of the DLIII that is being used from its RCA output to its XLR output, which may have even called into play a different output stage.
3)You changed the basic design and internal physical configuration of the interconnect cable that is being used.

If the improvements you have noted can be heard consistently across a broad range of recordings, my suspicion is that most and perhaps nearly all of the improvement is attributable to some combination of nos. 2 and 3.

In any event, I'm glad it worked out well.

Regards,
-- Al
Thank you both for a very informative discussion, however some of these proposed changes are a bit beyond my comfort zone. However, given the feedback on the sound quality after the change I am very interested in investigating further. So my setupis as follows:
* PS Audio DL lll w/balanced XLR cables to integrated amp
* Ayre AX7e integrated amp
* Ayre C5xeMP CD
* Sonos connected via digital coax to PS Audio DAC
Overall sound is very good - clearly the Ayre CD is more detailed on most music but not all. So after reading WKASS' post I am very interested in a change that could impact the sound quality in such a big way.

With my setup is it as easy as swapping the XLR cable connections at either the DLlll or Ayre side - again using full XLR between both no RCA? Or could this do damage to any equipment?

Thanks for any feedback or direction.

S
Sbleam, yes, you could accomplish the polarity correction for the DLIII, without affecting the polarity of playback from the CDP, by having the wiring of pins 2 and 3 interchanged at one end (not both ends) of the XLR cables connecting the DLIII to the integrated amp.

That can't hurt your equipment, assuming proper workmanship on the cable.

Before having that done, however, you may want to assess how much difference it would make by temporarily interchanging the + and - (usually red and black) connections at the outputs of your amplifier. For playback from the Sonos/DLIII that would accomplish the same polarity correction as the XLR pin swap, although it would result in inverted polarity for playback from the CDP.

Regards,
-- Al
Sorry to be some 7 months past this discussion, but I just bought a Cullen-modded DL III and was wondering if anybody knows how these mods affect the phasing issue, or if they are addressed at all. I have the unit connected to a Hegel H200 via XLR. Thanks.