Bi-Wire Tuning


I'm wondering if anyone has experience tuning bi-wired loudspeakers with different sized wires in the high frequency and low frequency inputs. Does it work? If so, what sizes would you recommend switching to?

Basically, I'm trying to improve bass accuracy and overall soundstage/imaging on a recently acquired pair of towers that have four binding posts each for low frequency and high frequency inputs.

Here is the system:

* An older, high end Sony CD player
* Arcam A75 integrated solid state amp
* Nordost interconnects
* DIY speaker cables: Belden 5000UP 12 AWG with gold plated BFA and banana plugs
* Paradigm Reference Studio 100 V.2 towers

At first I used my old Nordost Flatline cables in single pair with the jumpers that came with the speakers. I liked the way they sounded but, partly owing to the room dynamics and partly to the two 8.5" woofers that handle the lows, found that the bass could overwhelm and subsume the mids and highs, especially in ensemble jazz and rock. To remedy the excessive bass, I experimented with speaker placement and decoupling the speaker feet from the floor with aluminum receiving cups, to which I later attached furniture floor sliders -- which to my surprise *dramatically* improved instrument separation and imaging. But the bass could still be overpowering at times, so I decided to try bi-wiring to see if I could improve accuracy and sound stage.

I made two 6' pairs of cables with Belden 5000UP 12 AWG wire. The Paradigms have separate crossovers for the high and low frequencies. The Arcam has two output pairs per channel that operate in parallel, so one pair in a channel goes to the high frequency binding posts on the speaker, while the other pair goes to the low frequency binding posts.

I do think the bi-wiring led to subtle improvements in sound stage, instrument separation, and bass accuracy, but I want even more. What if I went down to a 16 AWG (or something else) for the low frequency inputs, or even used 4-conductor wire from one of the amp outputs to the two speaker inputs? Is that likely to have a noticeable effect? Are there any rules of thumb for figuring out optimal wire thickness?

Thanks,

Jeff
jdrouin
I haven't tried different gauges but different brands. A long time ago I used Linn speaker cable for the highs and Naim for the lows and had pretty good success. If I switched them the bass diminished and the highs developed a harsher tone. I forget the models of the cables but they were the bulk cables available at most Linn and Naim dealers.
Jeff, Belden 5000UP is a stranded bare copper cable - the worse to use in my opinion. According to Audioquest thick cables produce skin effect that changes impedance of the cable with frequency. Skin effect starts at gauge 18 (@20kHz). In stranded cable current jumps from strand to strand to stay on the outside (skin effect) crossing surface of each strand where copper oxides are. Splitting cable into insulated strands will be the better choice but when strands are in magnetic field of each other skin effect is still the problem. For that reason cable companies use flat tape, flat wound wire or hellical twist on oversized hollow tube. Flat tape is thin defeating skin effect. Flat wound strands are only in magnetic field of neighboring strands. Twisting each wire with with GND (return) wire also reduces inductance. Hellical twist on oversized hollow tube is pretty much the same in different form factor.

I'm not sure if skin effect is audible, especially since tweeter's impedance goes up at high frequencies, but cable companies engineers know more on the subject while different manufacturers use exotic configuration to achieve the same.

Mixing thinner and thicker wires is often done in biwired cables (Audioquest), but thinner wires also have higher inductance. My Acoustic Zen Satori Shotgun have separate cables for Woofer and Midrange/Tweeter - both gauge 10. Hellical twist on hollow tube makes them almost an inch thick. Insulation of the cable is Teflon to reduce dielectric absorption. It uses 6N zero crystal copper.

Biwiring supposed to reduce effect of EMF from one speaker to another that still exists in spite of crossover filters. Some speakers show great improvement with biwiring while others show none.
Have you experimented thoroughly with speaker positioning? You'll get more noticeable results by moving the speakers a few inches than you will by varying cables. Especially for a speaker with the bass output of the Paradigm Studio 100.
Jdrouin, I'm using cables that have 2 different thickness on my Apogee Slant 6s in bi-wired form (as spec'd from my provider), and yes they made a difference and I have done a bit of what you've also said. The biggest way I have found to help is with room treatments as opposed to making wire changes. Right now I'm using 6 Room Tune floor standers, 4 corner Tunes and will be continuing to experiment with more (possibly), this has made a HUGE Difference! I've invested hundreds of hours moving and listening but the results have been more than worth the time and investment, so you may want to consider that. I continue to read of fourm members that have had great success in doing this, of course this is my subjective opinion and I hope that what I offer is of help.
Thanks for these replies. The points on room placement/treatment are well taken. I've spent about 15 hours so far experimenting with room placement and floor decoupling. Have hit a plateau due to room limitations, so I thought I'd work on the cables some more while continuing to work with placement.

The living room is pretty limited in terms of what can be done. It's 18'L x 12'W x 8'H, with suspended hardwood floors throughout. Normally I listen across the width, and the speakers need to be close to the wall in order to have space to move about in the room. Their normal position is about 1' from the wall, which is noticeably better than, say, 6" from the wall, while I sit on a couch against the opposite wall. I can move the speakers a couple feet closer when I want to do some more focused listening, and it's OK as long as the volume is not too much. It is certainly better than 1' from the wall. No matter how far the speakers are across the width, though, the wall tends to pick up sound and artificially amplify it as the volume increases.

For more critical listening, on a few occasions I've cleared the furniture out of the room and put the stereo at the end, centered on the long axis, with distances of 3-4' from the back wall and minimum 2' from the side walls, while the chair I sit in is a good 4-5' from the back wall to mitigate reflection around my ears. The speakers being near the corners, even at a distance of a few feet, is definitely not optimal, but soundstage/imaging and overall tonal accuracy is better than listening across the width of the room. I've found that the speakers sound best so far when 4' from the back wall and a bit closer together, say with about 5' separating them. Moving them in or out an inch or two at a time and adjusting the toe-in angle results in subtle improvements.

At first, in all of these placement configurations, the room tended to pick up mechanical vibrations and artificially reflect/amplify them to the point I could even feel individual waves in my feet as they moved along the floor. In order to reduce mechanical interference, since the speakers way 110lbs. each, I removed the brass feet and put the spikes in aluminum receiving cups (I'm on a tight budget). WOW -- what a difference! Accuracy improved across the board, and instruments sounded way more realistic right away, with a dramatic improvement in bass extension and accuracy. On a subsequent listening, I put Magic Sliders under the aluminum cups, and the improvement was very noticeable yet again. I'm actually thinking of replacing the Magic Sliders with felt pads, since on my dining chairs they have even less friction than the chairs with the sliders.

While I've been happy with these improvements, there is still more work to do. I'm intrigued with some of the decoupling products at Herbie's and might try to save up the $200 for a set of those.

I've been taking notes in a spreadsheet, using the same set of pieces to test various aspects of the music I enjoy:

* Amy Winehouse -- "Me & Mr. Jones," for that "big" 50s sound and female vocals
* Miles Davis -- "So What," for bass accuracy and realism, plus harmonic range of the ensemble, sound stage
* G. Love and Special Sauce -- "Stepping Stone," for guitar and bass realism, percussion musicality, and FUN!
* Sonic Youth -- "Schizophrenia," for percussion musicality and glorious guitar noise
* The Cars -- "Just What I Needed," for percussion musicality, precise guitar crunch, general tightness of the band and crispness of the recording
* Beethoven -- 9th symphony, for orchestration range, sound stage, instrument separation, male and female vocals.
* Radio Head's album In Rainbows and Beck's Morning Phase, which are just phenomenally well recorded and present a huge frequency range.

Granted, a lot of these recordings are bass heavy, but that's kind of the point. I don't find bass to be overwhelming in classical symphony, and in fact find these speakers to excel with that genre.

Didn't intend to write this much. Anyway, this all started when I traded in my old JM Lab Chorus 706 bookshelf speakers because I wanted something would not pull back with larger orchestration or could sound just a little bigger for older rock music.

The Paradigms certainly make up for the shortcomings of those tiny JM Labs, but are not quite as accurate in the mids and highs. In general I really like the Paradigms, but they might have been overkill given the size of the room. And due to our living habits, it's really not feasible to leave them set up across the length of the room, so I have to move everything all over again each time I want to try that way of testing, and inconsistency is a real problem.

The Paradigms are simply too big for the room, I think. But before I trade them in for something smaller/more appropriate, I want to try and learn as much as I can from them, since they have many more features than my old ones. Wiring changes are part of that process too.