Balanced Speaker Outs for SII

[Deke609]

I just want to be make sure that this mod can be made to the SII.

As I understand it, all you need to do is: (1) remove the ground wires from a right/left pair of black (-ve) speaker posts; and (2) add to each of the 4 associated speaker posts, (red R/black R, red L/black L) a resistor, min. 1W each, wired to safety ground.

Assuming this works for the SII, what resistor value is recommended? I read in an older Stereomour thread that anything between 50-100 Ohms would work, with 70 Ohms suggested by PJ. 

Many thanks,

Derek

[Tom-s]

You're on the right path:
https://bottlehead.com/smf/index.php?topic=11053.0
https://bottlehead.com/smf/index.php?topic=2288.0
https://bottlehead.com/smf/index.php?topic=3750.0
https://bottlehead.com/smf/index.php?topic=10424.0

1W is fine. As long as you use 4 matched resistors in the mentioned range. Just try with whatever you have laying around first. Then change to non-inductive / audio grade variants and report back with your results !

[Deke609]

Many thanks Tom.

Yup, those are the threads I came across.

A few follow-up questions: does the R-value affect the sound? In other words, would there be an audible difference switching among matched quads of 40, 60, 80, 100 Ohms? 

If yes, I'll need to A/B/C/D test them and to do that I figure I'll need 2 dual-deck rotary switches - one for each channel -- so that I can quickly switch from one R-value to another in the hopes of hearing the difference.  Does anyone see a problem with switching while the amp is powered and and playing?  I don't want to toast the amp or my headphones.

Many thanks,

Derek

[Paul Birkeland]

If yes, I'll need to A/B/C/D test them and to do that I figure I'll need 2 dual-deck rotary switches - one for each channel -- so that I can quickly switch from one R-value to another in the hopes of hearing the difference. 

I think you're way, way overboard on this. 

[Deke609]

Thanks PB. But just so I understand: there's no magic in what R-value I pick between 50 and 100 Ohms so long as the 4 resisters are the same value and at least 1W. Is that right?

As you've no doubt already picked up on, I'm completely ignorant of the arcane art of conjuring great sound from electricity and wire    So I try to cover off as many possible bases as I can imagine might exist.

Many thanks,

Derek

[Tom-s]

Remember that the safety ground is solidly connected to the chassis, which will absorb a lot of heat from the soldering iron, so there is danger of a cold solder joint unless you apply a very hot iron or hold it there for a while. If the resistor body is closer to the speaker terminal, the the resistor won't get as hot while soldering. Fortunately, resistors can take very high temperatures.

[Deke609]

Got it.  Thanks Tom (and PJ!).  Resistor bodies will be well away from the ground solder joint, and I have my Weller 100/150W gun as backup if my WES51 doesn't have enough juice - but I suspect it will. Or even easier would be to detach the ground lug from the chassis before soldering and then carefully reattach after.

After sleeping on it, I've decided to go with a switch for A/B testing after all b/c at the very least I want to see if I can hear a difference between unbalanced and balanced - and for that I need quick switching from one setup to the other.  So, since I'll be switching, I thought I might as well use the switching approach for different R-values too. Am I expecting to hear any differences? Nope.  But I won't know until I put it to the test.

I found a 4 deck 6 position non-shorting switch on Digikey that should do nicely.  A bit pricey, but I can reuse it for other projects. I'll set up the first position as a traditional non-balanced configuration with only an insulated wire running to ground from each black speaker out, and then 4 successive positions will have wire/resistor running to ground from each of the 4 speaker outs, using quads of 40, 56, 82 and 100 Ohms (1W). I'll report back later next week on what I find.

cheers,

Derek

[Paul Birkeland]

After sleeping on it, I've decided to go with a switch for A/B testing after all b/c at the very least I want to see if I can hear a difference between unbalanced and balanced - and for that I need quick switching from one setup to the other.  So, since I'll be switching, I thought I might as well use the switching approach for different R-values too. Am I expecting to hear any differences? Nope.  But I won't know until I put it to the test.

Install your resistors, then you have a balanced amp.  Buy 2 clip leads.  Put each one on one black binding post.  When you want to "switch" to unbalanced, just connect the other end of each black alligator clip to a metal screw on the chassis or power transformer.

Maximum voltage out of the Stereomour would be on the 16 ohm tap.  4W into 16 ohms would be 8V.  We'll round that up to 10V.  If you put a pair of 50 ohm resistors across the posts, this will be 10V across 100 ohms.  This will induce a current of 100mA through the stack of resistors and dissipate about 1/2W through each resistor (at the very highest peak power levels).  This is where the 1W recommendation comes from.

[Deke609]

PB - Many thanks for explaining the 1W recommendation.  The lowest value I'll be using is 40 Ohms, so using your 10V estimation of max voltage at 16 Ohms configuration, I think that means I'd be looking at 125 mA and dissipation of 0.4W per resistor if the SII were configured for 16 Ohms (it's set for 8 Ohms). The highest value will be 100 Ohms, so 50 mA and 1W per resistor if the SII were configured for 16 Ohms.  I assume the voltage at 8 Ohms configuration puts me well under the 1W threshold? I don't know the math for calculating voltage from the impedance setting of the OT.

Re switching from balanced to unbalanced: I'm going to go the more expensive and elaborate route of using an actual switch b/c to get close-in-time A/B's. I don't want to have to power down the amp, flip it over and connect/disconnect the lead and power up again to do the comparison -- and I would not trust myself to try to do a hot switch: I don't want my butter fingers poking around a hot amp with clips. 

cheers and thanks,

Derek

[Paul Birkeland]

Re switching from balanced to unbalanced: I'm going to go the more expensive and elaborate route of using an actual switch b/c to get close-in-time A/B's. I don't want to have to power down the amp, flip it over and connect/disconnect the lead and power up again to do the comparison -- and I would not trust myself to try to do a hot switch: I don't want my butter fingers poking around a hot amp with clips. 

You don't have to power down the amp.  The clip leads can be connected on the top side. 

Adding the switch is going to give you something that I'm 99.9% sure is completely inaudible but will be an eye sore on an otherwise very nice build.

[Deke609]

@Pb - Ahh! Top-side - hadn't thought of that! 

RE eyesore: yeah, but this won't be permanent - it's just for testing (but will be in a nice little enclosure just for S and G's). If I decide balanced sounds better than unbalanced, I'll do a permanent install under the base, and re-purpose the switch and enclosure for something else in the future.

For the permanent install, 2 question about type of resistor:

(1) do I need non-inductive resistors? I would have thought that a 1W 50-100 Ohms resistor, no matter how inductive, would be incapable of releasing a big enough charge to damage anything.  But that's more guess than understanding on my part.

(2) What type/make of resistor would you recommend for best sound?  E.g., naked, foil, wirewound, etc.  I have no idea if it matters b/c I don't understand how this balanced speaker out mod works as an audio circuit - I'm just taking a paint by numbers approach here. 

Many thanks again,

Derek

[Paul Birkeland]

The amount of inductance in a 1W/100R is inconsequential, especially considering that very little signal current will flow through this part.  There is absolutely no concern about energy release from the magnetic field of a resistor in this situation.  You're going to need some very expensive equipment to even attempt to measure that the potential for this even to happen will be at frequencies that you can't hear.

The type of resistor is extremely unimportant here.  You are providing a ground reference, that is all.  It's not the safest practice, but the amp can be run as a balanced amp with no connection from the speaker taps to ground.  For the purposes of what you're doing, you could disconnect the grounds to each set of binding post,  put a 100 ohm resistor in each binding post on the top side, then you could listen to no resistors, unbalanced, and balanced with resistors without messing around with a switch box.  I would not leave the amp without some kind of safety connection for the transformer secondary to ground though for very long.

[Deke609]

Great!  Thanks PB. That means I can use the resistors I have on hand.

I'll ground the resistors/wires from the speaker posts just to be on the safe side.

cheers and thanks,

Derek

[Deke609]

I built and installed the switch this evening.  It has four positions: unbalanced, and three balanced using 56, 82 and 100 2W resistors, respectively.

Listening impressions: in balanced configuration, everything, from highs to lows, is tighter and clearer.  The added clarity makes previously unnoticed details stand out. Music sounds a bit more spacious, but that may just be an effect of the added clarity which creates greater separation between instruments.  A very worthwhile mod and one that can cost almost nothing -- just 4 resistors and some wire.  Pretty cool!

I can't hear a difference between the different resistor values.  But I will keep experimenting with the switch to make sure.

[Deke609]

After not hearing a difference between different R-values I arbitrarily chose to install the 80 Ohm resistors. But it just occurred to me that the 50-100 Ohm range for resistors originally suggested by PJ in another thread may have been based on the expectation that the amp would be used with speakers that typically present a 4-8 Ohm load.  I am using my SII amps exclusively with my LCD4 headphones. They present a flat 200 Ohm impedance.  Should I install much higher R resistors, e.g., 1 kOhm or higher?

My thinking is that with my amps as presently configured, the output signal has two possible paths: (1) the balanced mod path that places a combined 160 Ohm resistance between +ve and -ve speaker outs; and (2) the headphone path that places a 200 Ohm impedance between +ve and -ve speaker outs. Assuming that I can treat the impedance of the headphones as a resistive load (I may be completely wrong about this - if so please correct me), doesn't it follow that a good amount of available power is disappearing via my balanced speaker outs mod?

If I am completely wrong about this (specifically my hypothesized need for much higher r-values) and it will take too long to try to explain why, I'll be happy with a simple "Nope" response.

Thanks in advance,

Derek