16 ohm speakers: any amp sounds better with more resolution. speaker cables less critical.

speakermaster, There issome truth to your claim that 16-ohm speakers were more often the norm in the early days of hi-fi audio than they are now.  And maybe the rationale did have something to do with the high-ish output impedance of the then prevalent tube amplifiers.  However, I look at it the other way around: The current and rather longstanding trend toward 4-ohm speakers is the product of speaker designers insisting upon multi-way, multi-driver speakers, which inevitably requires complex crossover networks that typically will present a net low impedance load, because one is paralleling several different drivers that individually are likely to be high-ish in impedance, and because the networks per se tend to reduce net impedance.  In other words, "speaker wars" (3 or many more than 3 cheap drivers in a box so one can claim a wide bandwidth of response) led to 4-ohm speakers.  I admit that I hate the sound of such concoctions and wouldn't have any of them in my house.

Ralph, I am not sure what you are saying, because the terms "multimeter" and "DVM" are so often conflated with one another.  VTVM is a clear acronym, because you are talking about a vacuum tube voltmeter.  But let's take my Fluke 87 meter.  It is a "multimeter" in the sense it can read DC and AC voltage, capacitance, DC and AC current, and it includes a diode checker.  But it is also a DVM in the sense it is digital and solid state.  On voltage settings, it offers 20Megohm resistance across the leads.  So, what is a "multimeter" and how do you distinguish it from a DVM?  Or, why isn't my Fluke a multimeter?  Folks like Elliot might benefit from this clarification even more than I. For sure, any respectable voltmeter can measure the -10V to -50V range of the grid bias voltage that is typically seen for most tube amplifier output tubes.

0.4V = 400mV, not 40mV
And 0.4V doesn’t make sense as a bias voltage for an amplifier output tube. It’s way too little. Also grid bias should read out as a negative voltage with respect to ground or with respect to cathode voltage.

Ralph, What I recommended to Elliot is that he first experiment with the settings of the L-pads in his speakers until he arrives at a solution that he wants to live with long term.  Doing this would automatically involve accounting for the driving ampifier and the length of speaker cable.  Once that steady state is arrived at, do you see anything wrong with swapping the L-pads for high quality high wattage discrete resistors that represent the resistance across each L-pad after their final adjustment?  Anyway, that is what I would do.  My approach is based only on listening to my Sound Labs speakers (first the M1s, then the 845PXs) with vs without the Brilliance control in circuit.  Big improvement in transparency and air without that L-pad.

Elliot, What you say to Cisco (Cisgo) may be true for an SS amplifier but for a tube amplifier, especially a transformer coupled one, the reverse is more likely, for a given tap on the output transformer:  An amplifier that makes 12W into 8 ohms might make only 6W into 4 ohms.  This is not always the case, and actual measurement is in order.  Likewise, an SS amplifier is likely to double its power into 4 ohms vs 8 ohms, provided it can deliver the needed additional current, which depends upon the power transformer.  But Cisgo apparently owns a First Watt amplifier, designed by Nelson Pass, one of the best and most innovative of all designers of SS gear.  In the First Watt series, he is experimenting with different topologies and different types of output devices.  So if Cisgo's information about his own amplifier comes from NP, I would accept it as gospel.  And Cisgo, if you want to know why the power rating changes as you say, ask NP.

For those who don't know, Sound Lab drive their full range ESLs using two audio step-up transformers wired in parallel, one for bass and one for treble.  Audio frequencies are divided by a first order crossover network (inductor in series with the bass transformer, RC network before the treble transformer) before being fed to the bass and treble transformers, respectively.  I have done impedance vs frequency curves for my 845PXs, measuring Z at 50, 100, 200, 500, 1K, 2K, 5K, and 10K Hz, with the original bass and treble transformers in place, with only the OEM SL bass transformer in place, with only the original treble transformer in place, and with the aftermarket full-range 1:90 ESL transformer alone and in combo with the OEM SL bass transformer.  I have the data in a notebook.  So I think I have a pretty good understanding of the impedance of an 845PX.   I use the speaker with the OEM bass transformer preceded by a small inductance, in parallel with my aftermarket audio step up transformer preceded by no filter at all.  Efficiency is hugely improved; I would wager it could easily be driven by less than a 50W amplifier.  My Atma-sphere amplifiers probably make 100W into a 16-ohm load, and they are coasting at high SPLs while driving the 845PXs, which appear in my system as about a 20-ohm load, using the generalization that Raul dislikes, but I can guarantee that Z never goes below 20 ohms below 5kHz. All ESLs are in effect giant capacitors, so it is inevitable that Z goes down as frequency goes up above 5kHz. As Ralph said, the big problem was the resistor in the RC network that was there to create a high pass filter for the (old) treble transformer.  That resistor, depending upon the vintage of the speaker, could be as low as 5 ohms, or in later speakers as high as 8 or 10 ohms.  The resistor creates an absolute upper limit of impedance at frequencies around the crossover point, because it is in parallel with the output of the amplifier.  Thus, if you were unlucky enough to have one of the older speakers with a 5 ohm resistor, you were dealing with 5 ohms or less impedance at around 500Hz to 2kHz.  That's also why you benefited from a very high wattage resistor in that RC network.  Raul has a point about the very high impedance in the bass region using the OEM bass transformer; it's way up around 100 ohms and maybe higher below 100Hz.  Yet, my OTLs drive it fine, as does Raul's friend's SS amplifier.

Elliot, For Sound Lab owners, the first thing I recommend is to bypass the "Brilliance" control, which is an L-pad in series with the treble transformer, much like the controls on your speaker.  Alternatively, if you feel you need some treble taming, you could use a discrete resistor of an effective value in place of the L-pad.  L-pad bad.  No L-pad good.

Dear Raul, Of course you are correct that no speaker has a flat impedance.  Yet it's common practice to refer to speaker impedance with a single value.  I am guilty of that as is almost everyone else in the audio world.  If it were of some importance to me personally in the course of evaluating a speaker for purchase, I would demand to see the full impedance curve over the entire frequency range.  As you know, an impedance dip in the bass or midrange can be much more consequential than a falling impedance at very high frequencies, for the interaction of any speaker with any amplifier.  I don't argue with any of the points you made about speaker impedance.  Also, all remarks about speaker impedance are subject to the choice of amplifier to drive that speaker, unless the speaker is just crazily variable in impedance across the audio frequency range, which will be challenging for any amplifier.

Ralph, The older Sound Labs sounded anything but "transparent" in the midrange, using your OTLs to drive the unmodified version of my 845PX. I remember listening to Frank Sinatra during the first week or two after I bought the 845PXs and thinking to myself, "Is that Frank Sinatra?" Yes, Dr West did respond to our findings and our fix by revising his circuit and substituting the old treble audio step-up transformer for a new one that works down to lower frequencies, at least an octave lower.  This allowed him to make changes in the crossover such that it sucks less power and provides for a higher impedance at midrange frequencies.  I am told it's a big improvement.  I continue to drive my own 845PXs using a treble transformer capable of full-range response that does not require any crossover components at all to differentiate the input to the bass vs treble transformers.  Now, getting back to the amplifier factor, it is quite likely that the old version of the 845PX could have been driven more satisfactorily at midrange frequencies using a typical SS amplifier, because of the lower output impedance exhibited by most SS designs, but then I wouldn't have the OTL-ness to which I am addicted. I was told that SL use SS amplifiers at their factory, which is probably why the problem went unnoticed for a while.

Reactance of the cable is also important. But now you’re starting a different topic.

I am not sure that generalization (16 ohms = lower amplifier distortion) would apply to every variety of solid state amplifier. Tube amplifiers as a class tend to have a much higher output impedance than do solid state amplifiers.  This does give an advantage for 16 ohm speakers. What I abhor about "4-ohm" speakers is that they tend to be multi-driver arrays with very complex crossover networks.  This is a set-up for low input impedance and for power-sucking crossovers.  You can get a Pass or other high quality solid state amplifier to drive these beasts because they generally have an output impedance well below 0.5 ohms, and there will be some listeners who like the results.  Speaking only for myself, I never have.  But it is wise to think about the speaker first, and then to choose an amplifier well suited to driving it.  Once that process is done, I have always tended to prefer at least modestly efficient speakers with high input impedance that can be driven by tube amplifiers (my preference has always been for OTLs) or low power Class A solid state amplifiers.  Not everyone will agree, and that's fine.

Don't forget to get rid of those "rheostats", which are probably L Pads.  L pads sound terrible and one is always better off without them. Replace them with discrete resistors or no resistance at all. See my earlier post.

The Zero (Speltz) Autoformers are not a free lunch. As with any transformer, they trade off current and voltage. So if you mate an amplifier with a high-ish output Z to a speaker with a low input Z, via the Zero, you lose voltage in direct proportion to the gain in impedance seen by the amplifier. This affects apparent amplifier power. It does work, but it’s not a perfect world. I happily used the Zeros for many years as an interface between my Atma-sphere amplifiers (OTL amplifiers with a high output Z) and my Sound Lab M1 speakers, which had a midrange impedance dip down to 2 ohms and was under 5 ohms impedance across a considerable portion of the midrange. The Zeros made it possible to drive them. Then I traded my M1s for a pair of Sound Lab 845PX speakers. The match was just barely acceptable, because the 845s + Zeros were sucking more amplifier power than the M1s, but with the intellectual input of a guy in Australia and by changing one of the two audio step-up transformers in the SLs and ripping out the crossover components, I was able to completely eliminate the impedance dip typical of SL ESLs. (Ralph knows this story very well.) In fact, my 845PXs now present an impedance that is never lower than 20 ohms from around 200Hz to 5kHz, and they are also much more efficient, for technical reasons too arcane to go into here. The load presented by the Sound Labs is just heaven for my Atma OTLs, and I don’t need autoformers. So, use ’em if you need ’em, but it’s still best to select either a speaker suitable for your amp or vice-versa, in terms of impedance.
To the OP. I advise you to set those controls the way you like them, then measure R across each rheostat, then substitute discrete resistors of that value, one for each rheostat (if I am correct that each speaker has 2). This will get you a big upgrade in clarity. When you do that, select very high quality resistors of very high wattage rating; that also pays off.