Cons of using main XLR/Balanced outs for subwoofer?

I am now way beyond my knowledge base and hearing the voice of common sense confusion- which could be just confusion, nothing more. You indicate that transformer roll-off will cause phase-shift up to about 100Hz and this is important to avoid. The frequency response of the Rowland's outputs is 5Hz-70kHz -3dB @ 8 ohms, so seems appropriate to output into the Jensen transformer (the SUB-1RR) with flatter F-R response. But immediately downstream from that transformer could be either a Dayton SA1000 amp which doesn't appear to have flat rating below 18 Hz ( its specs are hard for me to understand) or the first sub in a DBA chain (unlikely to have flat rating to less than 20 Hz for my system). Furthermore, setting up a "swarm" DBA usually involves intentionally producing a mix of different arrival times from the subs by adjusting position, phase, polarity, closing ports, etc. Duke always says the ear is very forgiving of small timing errors in the low frequencies such that the speakers + room = a minimum phase system. So I am unclear how a phase shift in the low (< 20Hz) bass is consequential. OTOH he also notes that the ear has exaggerated sensitivity to frequency response errors (non-flat curve below 20Hz) in the bass region, so maybe that is what you are trying to tell me. FWIW, Rowland recommended the 2 channel Jensen transformer for running subs off the balanced outputs in my Concerto (but that was 14 years ago). I again want to say I will be running my Salk mains (+/-3dB at 33 Hz) full range, no high pass crossover. I appreciate your efforts to understand my options.

Duke always says the ear is very forgiving of small timing errors in the low frequencies such that the speakers + room = a minimum phase system. So I am unclear how a phase shift in the low (< 20Hz) bass is consequential. OTOH he also notes that the ear has exaggerated sensitivity to frequency response errors (non-flat curve below 20Hz) in the bass region, so maybe that is what you are trying to tell me.

Duke is right; phase shift is different from delay, and 'yes' in that order :)

Apologies for not asking clear questions or understanding your brief answers, which prompted me to learn what a phase shift in a transformer means. I found a Floyd Toole article (Audibility of Phase Distortion) which reports studies showing that "Within very generous tolerances, humans are insensitive to phase shifts. He also points out that in a normal room sound is reflected many times from objects at multiple distances resulting in enormous amounts of phase shifts that dwarf the far smaller phase shifts in amplifiers (a lot of this read like Duke's take on this). Rowland tells me transformer bandwidth in this very low bass is not important for their solid state amp's line outputs. I did read that a possible exception to this is in valve amps that use a high level of negative feedback, which may be relevant to your fine Atma-sphere products. I thank you for spurring me to learn more. Decided to use my older Jensen 2 ch transformer.

Within very generous tolerances, humans are insensitive to phase shifts. He also points out that in a normal room sound is reflected many times from objects at multiple distances resulting in enormous amounts of phase shifts that dwarf the far smaller phase shifts in amplifiers (a lot of this read like Duke's take on this)

This is entirely correct; you can't sense phase at all if you're talking about a single frequency like a sine wave.


But leaving sound stage information aside (which might be messed with due to reflections) phase shift over a spectrum is perceived as a **coloration**, for example a rolloff above the audio band can be heard as a darkness in the highs if the rolloff is close enough to the audio band. In the bass its a similar phenomena, if rolling off prematurely (above 2Hz) even though your speakers are flat to 20 Hz it will be perceived as a lack of impact.


Because speakers are mechanical devices, a rolloff in them does not cause this problem.


I first became aware of how profoundly this was true a long time ago when trying to find out why an MFA Magus preamp sounded bright in the phono section. A dealer of mine wanted to sell the preamp instead of a competitor's but the brightness made it as bad as poor digitial of the day. I investigated the circuit and found that at 50KHz the circuit went from the RIAA equalization curve to flat. I removed the components responsible and the brightness was eliminated, with no sense of darkness- LPs sounded just fine after that. Now you'd think that at 50KHz there wouldn't be an issue, but phase shift can extend down to about 10th the frequency where this was happening (5KHz) so its easy to see how this was heard as a brightness. This principle is easy to demonstrate. 


I can put this another way: there is a **reason** that Jensen feels they needed to go so low (1Hz) with their subwoofer transformers. Jensen transformers are extremely well designed, are high quality, and the designer is held in high regard in home and professional audio. Anecdotally, I've tried using regular full range audio transformers made by Jensen for subwoofers and while they work OK, you can easily hear that something isn't right about the bottom octave- which is fixed using the Jensen subwoofer transformer. Preventing phase shift is is also why Stuart Hegeman of Harmon Kardon Citation fame insisted on wide bandwidth in his designs.


Since you already have the transformers I would go ahead and use them. But I would also get the subwoofer transformers and then audition them side by side- see what you think.

I run my two subs out of the single-ended (rca) outputs from my buffer, and feed my main amps from the balanced outputs. To the subs, I run longer rca cables into Jensen PI-RX transformers and then very short XLR cables from the transformers to the subs. Looking at the Jensen graphs and spec sheets, it appears the SUB-RX transformers would have been a slightly more optimum solution to use with the subs....but not by much, and not by enough to worry about. I actually purchased the PI-RX transformers for a different situation involving a single-ended tubed preamp into SS mono amplifiers that had only balanced (XLR) inputs. You do need to be careful with impedances in those full-range situations since having too high an output impedance into the transformers can roll off the high frequencies.