Building balanced headphone cables?

[BoilermakerFan]

Originally Posted by Steve Eddy

I think what can happen is, while removing the plug, the ring contact on the jack can contact the tip contact of the plug while the tip contact on the jack is still contacting the tip contact of the plug. The result being that the two channels' outputs get shorted to each other.

se

That is correct SE. The short only occurs when unplugging the headphones when the amp is powered up. There is a small short when plugging in too, but unplugging is worse and that is usually when the problems occur. Amps like the Beta22 and a modded F5 would be prone to damage during this unplugging without resistors on the jack, IIRC over 100R too, but with a powerful amp like the B22 and F5, the loss at the jack is really minimal, but the amp stays protected.

I'm building in Caddock or Rikens on my TRS jacks for my adapter box, to protect my amp, but also to knock down the power a bit to protect the cans since it will be putting out about 7W at 60R or over 15W at 6R.

[Steve Eddy]

Originally Posted by BoilermakerFan

That is correct SE. The short only occurs when unplugging the headphones when the amp is powered up. There is a small short when plugging in too, but unplugging is worse and that is usually when the problems occur.

Mmm. Don't see why plugging or unplugging would make any difference. A short's a short.

Amps like the Beta22 and a modded F5 would be prone to damage during this unplugging without resistors on the jack, IIRC over 100R too...

Don't see why they'd need to be any higher than the nominal impedance of the lowest impedance headphones the amp can handle without going up in smoke.

If the amp can drive say, 32 ohm 'phones, then it shouldn't require any more than 32 ohm resistors.

I'm building in Caddock or Rikens on my TRS jacks for my adapter box, to protect my amp, but also to knock down the power a bit to protect the cans since it will be putting out about 7W at 60R or over 15W at 6R.

If you're also wanting to attenuate the output instead of just protecting it, then what you might want to consider is a voltage divider instead of a single resistor. That'd let you to keep the output impedance lower than it would be with a single series resistor.

Use something along the lines of 30 ohms for the series resistor, then select the shunt resistor to give you the highest listening level you'll need plus a little extra. It will be only a fraction of the 30 ohm resistor.

Good luck!

se

[BoilermakerFan]

Originally Posted by Steve Eddy

Mmm. Don't see why plugging or unplugging would make any difference. A short's a short.



Don't see why they'd need to be any higher than the nominal impedance of the lowest impedance headphones the amp can handle without going up in smoke.

If the amp can drive say, 32 ohm 'phones, then it shouldn't require any more than 32 ohm resistors.



If you're also wanting to attenuate the output instead of just protecting it, then what you might want to consider is a voltage divider instead of a single resistor. That'd let you to keep the output impedance lower than it would be with a single series resistor.

Use something along the lines of 30 ohms for the series resistor, then select the shunt resistor to give you the highest listening level you'll need plus a little extra. It will be only a fraction of the 30 ohm resistor.

Good luck!

se

Yes, I should have edited my first post SE. The short occurs both ways, but it seems to always blow the B22 when the cans are unplugged. But if the amp was off when they were plugged in, then that would always be true. Most owners of the B22s know they have to shut down to change headphones. The damage tends to occur at meets or when alcohol is involved.

I think the higher values were recommended to protect the amps on the small surge with the short. There is certainly enough power there to drive right through a 100-200R network. My 35W/ch vintage Yamaha CR-620 has a 121R resistor network for each headphone jack.

Thanks for the value suggestions on the resistor network. There is a web page that has the formula for calculating the values based on the wattage output of the amp and the impedance of the headphones, but I don't have it handy.

[Pano]

Just stumbled across this thread. (oops!)

I've run balanced for years with a single connector. 4 pin DIN! The locking, screw type with metal body. Works great for me. And I have a short DIN to 1/4" cable that I use when I need to go back to unbalanced. Not tiny, but smaller and lighter than XLR, for sure.

Got the metal DIN at Mouser or Digikey, I think.

[Wolfsin]

Has anyone attempted using 3-wire shielded microphone cable? My ratoinale is that cross talk from the twisted pair to the third wire should be minimal because the equal and opposite currents should create canceling fields. Any 'crosstalk' from currents in the single wire would cancel in the can connected to the twisted pair.

There are 3 wire microphone cables, for example from Gotham, that are ultra flexible though not necessarily small diameter.

[nereis]

On top of the actual cable change, I'd be really tempted to add a small RC zobel right on the transducer to "terminate" the cable.

[Wolfsin]

Is the zobel to prevent reflection or compensate for transducer impedance? Targeted use is with HD6x0 cans.

[Wolfsin]

As the following thread shows, I am not the first to consider this.

Tying in shield for headphone cables (esp. balanced situations)?

The history of bal<->unbal is littered with hums, screeches, and worse between components, but allow me to point out that cans are not a component in the traditional sense. They are more like an MC cartridge, i.e. inherently balanced, while being capable of rejecting common noise.

Please help me discard this idea before I go off and build one if I am "thinking inside the cable" and missing something important.

[sgrossklass]

Any kind of zobel network would be for ensuring amplifier stability, as it keeps load impedance low up to the MHz range. This should only be of concern to amplifier designers (in an ideal world where they can be relied on to know what they're doing).

Indeed, headphones are much like a phono cartridge, inherently balanced and floating.

One of the more important points in wiring them up unbalanced is keeping the returns separate for as long as possible, as return resistance common to both channels degrades channel separation. Stock cables for the usual suspects (HD5x8, HD6x0, various AKGs etc.) implement that already.

Even a common ground connector does not necessarily spell disaster though, as long as its resistance is less than about 1% of minimum driver impedance. (Example: HD590, ~1 ohms vs. 100 ohms.) 40 dB of channel separation should be plenty on cans if even 12 dB of imbalance already is painful and vinyl gets along fine with <30 dB.

Ideally, the cable for a balanced headphone connection should have either 2 conductors and a shield per channel, or 4 conductors plus shield for both. Shield would be connected at the amp only.

Anyway, people tend to put a lot of emphasis on electrical characteristics. IMO, they're not even all that important (assuming capacitance doesn't skyrocket or resistance exceeds ~10% of minimum impedance). You do want the cable to be light and mostly devoid of microphonics though, and it would be nice if it didn't break after a week of use. Sennheiser had massive cable reliability problems with the thin copper cables they originally supplied with HD414s way back when, so they eventually switched to somewhat microphonic but tough steel-conductor cables (still to be had with HD25-1s these days). It would take them about 15 years to come up with the kevlar fiber reinforced copper cables they usually supply nowadays (low microphonics and decent durability).

[Wolfsin]

Thanks for the rules of thumb for cable resistance. My first attempt was with Mogami 3106 stereo microphone cable. Twin pairs of twisted conductors, separately shielded so it fully meets your ideal. It leaves something to be desired in terms of useability. I recently held a length of Gotham GAC-3 and thought how wonderful it would be to have that flex in a headphone cable. That started my 'inside the cable' thoughts.

Now I am questioning what parts of the ideal might be relaxed at little (or better, no) loss in quality. Is shielding really needed? Might it instead be used as the fourth conductor? It was at that point that I posted my thoughts above and then found RaneNotes on shielding, grounding, and audio interconnect. My thoughts fall under "Floating, Pseudo, and Quasi-Balancing" but I am undeterred :-(

Gotham has graciously provided three and four conductor cabling at what seem to me quite low prices. Surprisingly they sell cut lengths to one and all. I will add connectors and test for crosstalk differences between GAC-3 and GAC-4 (four conductors in a single shield). I will also test against Mogami dual, separately shielded, twisted pair.