Valve amps and loudspeaker impedance curves

[ToTo Man]

c: What do you mean with testing method?

I'm not familiar with the method of measuring a valve amplifier's output, is it the same method as measuring a solid-state amplifier's output?

I don't have the proper electronic equipment to measure this anyway, I can only measure it acoustically using loudspeakers as a proxy. So, unless I can find speakers with a perfectly flat impedance curve, then the measurements I make are always going to be a combination of the valve amp's inherent 'tuning' plus the interaction between the amp's output impedance and the impedance curve of the load.

 

[Tony L]

The LS50s have always struck me as the sort of small speaker to throw a Krell-grade solid-state powerhouse at. They remind me of AE1s in a way in that they can sound decidedly average and one wonders what the fuss is about until that time you hear them on a real clean current monster. It is why I’m drawn to the other side of the mini-monitor market such as LS3/5As, JR149s etc as they can sound fabulous on a 10 Watt valve amp as whilst they are still obviously inefficient they are such a benign load.

 

[ToTo Man]

With regards to load impedance curve, is the Q of the impedance peak/dip as influential as its magnitude when it comes to determining how much frequency response variation will occur? Assuming for simplicity that a SET's output impedance is linear across all frequencies, based on the JR149's impedance curve for example, would you expect the SET to give a bigger boost in SPL at 60Hz or 1.5kHz? Or do we need to know other parameters such as phase angle? I find this stuff fascinating but I always end up with more questions that need answered!

 

[Tony L]

Assuming for simplicity that a SET's output impedance is linear across all frequencies, based on the JR149's impedance curve for example, would you expect the SET to give a bigger boost in SPL at 60Hz or 1.5kHz? Or do we need to know other parameters such as phase angle?

Phase angle/reactive load in combination with an impedance dip is I suspect the big one. To my understanding it is that which makes so many aggressively ported small stand-mounts so unsuited to valve amps, mandates a fainting couch for a Quad 303 etc. Without seeing the phase angle I’d expect a valve amp to have a slight lift at both 60Hz and 1.5kHz on the 149 along maybe with a slight dip at 10kHz. I prefer both the 149s and LS3/5As running off the 16 Ohm tap of the Leaks. 8 Ohm tap for the Tannoys downstairs. It is certainly noticeable and I fairly swiftly spot it if I’ve forgotten to switch it.

 

[ToTo Man]

Phase angle/reactive load in combination with an impedance dip is I suspect the big one. To my understanding it is that which makes so many aggressively ported small stand-mounts so unsuited to valve amps, mandates a fainting couch for a Quad 303 etc. Without seeing the phase angle I’d expect a valve amp to have a slight lift at both 60Hz and 1.5kHz on the 149 along maybe with a slight dip at 10kHz. I prefer both the 149s and LS3/5As running off the 16 Ohm tap of the Leaks. 8 Ohm tap for the Tannoys downstairs. It is certainly noticeable and I fairly swiftly spot it if I’ve forgotten to switch it.

OK so, AIUI, the 16Ω tap will exaggerate the frequency response variations more than the 8Ω tap, but this needs to be weighed up against ensuring there's enough voltage available to drive the speakers properly. Is your preference of the 16Ω tap for the JRs and the 8Ω tap for the Tannoys based on the tonal balance you're hearing from each tap or is it obvious that one tap is much better at driving the one pair of speakers than the other from a dynamic perspective?

EDIT - I'm forgetting your Leak Stereo 20 isn't an SET so it probably has a lower output impedance and therefore won't induce such obvious shifts in frequency response as an SET.

 

[S-Man]

The LS50s have always struck me as the sort of small speaker to throw a Krell-grade solid-state powerhouse at. They remind me of AE1s in a way in that they can sound decidedly average and one wonders what the fuss is about until that time you hear them on a real clean current monster.

The Meta is a significantly worse load for a valve amp (or other high o/p impedance amps) that the original LS50.
The Meta has similar characteristics using all of my SS power amps including with my Leach amp, which drives pretty much anything. It is noticeably more forward sounding with the valve amp.

 

[Tony L]

Is your preference of the 16Ω tap for the JRs and the 8Ω tap for the Tannoys based on the tonal balance you're hearing from each tap or is it obvious that one tap is much better at driving the one pair of speakers than the other from a dynamic perspective?

Both, they just sound better to my ears on those taps. Subjectively more heft, scale and apparent headroom. I do everything by ear. I’ll read and understand the theory to the best of my ability, but always trust what I hear more. If it sounds better it *is* better!

PS FWIW I view the 303 as a valve amp too, it really needs to be in that 8-16 Ohm non-reactive zone otherwise it starts sounding exactly the way people who hate them think they sound!

 

[ToTo Man]

Hmm... I don't think I'm a fan of JR149 mk1's and SETs, - it emphasises characteristics I'd prefer weren't emphasised, namely forward mids and rolled-off highs.

Note: Ignore the huge drop in bass output below 100Hz, this purely due to my laziness in speaker and mic placement. I didn't bother moving my Dittons out of the room so the JR149s are plonked in front of them more than two feet from the front wall and the mic is still 40% away from the back wall so the poor JRs are getting very little bass support in the measurements. When I did my listening I was sitting two feet closer to the rear wall so I was hearing more bass than the measurements suggest!

On the 8Ω tap the mids are so strong I find it difficult to focus on other aspects of the presentation. The tonal balance improves on the 4Ω tap and is much closer to what I hear through my Yamaha.

I can't say I notice a reduction in control or dynamics when I switch from the 8Ω tap to the 4Ω tap, but this may be because the mids are so prominent on the 8Ω tap that I wrongly perceive the 8Ω tap's presentation as less dynamic to begin with? I can perhaps hear a very slight fuzziness to the sound on the 4Ω tap, but I could just be imagining this. Nothing sounds obviously wrong or lifeless. I'd need to create EQ presets for both taps so that I have a level a playing field in terms of frequency response which would allow me to focus on assessing the dynamics.

The REW measurements show no statistically significant differences in THD between the 8Ω and 4Ω taps, but I'd need to run the tests at much higher SPLs in order to drown out the external noise pollution for more reliable results.

I should also qualify my findings by noting that the JR149s used for these tests are completely original and that the forward mids and rolled-off highs could at least be in part due to the drivers and/or crossovers drifting off-spec. I ought to refurbish them but I've always got too many other projects on the go!...

EDIT - A note about the effect of output impedance on the bass response. This JR149 impedance plot shows an impedance of 10Ω at 200Hz that gradually rises to 16Ω at 100Hz and shoots all the way up to 60Ω at 60Hz, yet the boost I'm measuring at 60Hz is only +1dB higher than the boost at 100Hz. If an increase in impedance from 10Ω to 16Ω causes an increase of +1dB, shouldn't an increase from 10Ω to 60Ω cause a much higher increase than +2dB?

 

[ToTo Man]

I came across this interesting thread on SHF about selecting the optimal output tap for your speakers and am somewhat confused by the advice given by Mr Hoffman himself:

"What I do (oft told): Take the left channel of the stereo Beatles I FEEL FINE (the drums, bass side) and compare. Which tap sounds better on this analog compressed song? Try the 8, try the 4, go back to the 8, etc.

Using a song with reduced dynamics is the best way to tell which tap is giving you the most bang for your buck."

Can someone please explain to my why using a song with limited dynamic range is the best way to determine which tap facilitates the best dynamics from your speakers? Or have I misunderstood his point?

 

[Arkless Electronics]

I came across this interesting thread on SHF about selecting the optimal output tap for your speakers and am somewhat confused by the advice given by Mr Hoffman himself:

"What I do (oft told): Take the left channel of the stereo Beatles I FEEL FINE (the drums, bass side) and compare. Which tap sounds better on this analog compressed song? Try the 8, try the 4, go back to the 8, etc.

Using a song with reduced dynamics is the best way to tell which tap is giving you the most bang for your buck."

Can someone please explain to my why using a song with limited dynamic range is the best way to determine which tap facilitates the best dynamics from your speakers? Or have I misunderstood his point?

Almost everything in that thread is rubbish from a technical POV...(what a surprise! On a hi fi forum!?).

Ultimately you'll have to decide what sounds best to you

 

[davidsrsb]

4 Ohm taps will have the highest turns ratio, so a valve amplifier can easily sound different with that tap than into an easier load like an old LS35/A at 15 Ohm

 

[Arkless Electronics]

So many factors come into it that ultimately and generically you have to use your ears....

eg, where is the feedback taken from? (N/A for most SET's). Some take the feedback from usually the 16R tap and then the other impedance's are just tapped down in autotransformer style from there, like the Stereo 20.. It should theoretically be at its best set to 16R and obvs driving a genuinely 16R speaker then...

Others switch in different NFB networks for each tap.... (better in theory)

and the likes of the STA25 take it all the way and the various secondary windings are rejigged by the very multiway selector to connect secondaries in parallel or series etc to present the lowest intrinsic impedance and most "most amp-turns in the winding window" AND switch the NFB to be optimum for each setting!

TdP however in most of his designs takes the feedback from the primary only... which goes against conventional wisdom but if you can make the output transformers good enough has it's attractions as you could I guess get away with more NFB... he erm.. seemed to have the knack, as it were, on this! In fact he used multiple feedback loops, some from tertiary windings, and bootstrapped the supplies to the drivers via more windings! He was a VERY clever guy