• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Master Thread: Are measurements Everything or Nothing?

So you'd need an idea of the "average' output impedance of the tube amp. I've no idea what magnitude this typically has - this thread suggests in the region 1 to 4 ohms.

So let's take 2ohm.

So you are going to need a 2 ohm resistor. Lets say you are dumping up to 50W into an 8 ohm speaker. That will be about (I = sqrt( P/R)) - about 2.5A

That 2.5A through 2 ohms will be 12.5W (unsurprisingly in the ratiio of the two impedances) . This is a minimum - I'd go for 4 x that to stay well on the safe side, so 50W

So you are going to need a 50W 2 ohm resistor in series with the speaker. All you need to do is solder one of the speaker wires to one end of the resistor, and connect the other end of the resistor to the speaker terminal using a short length of wire. At the amp end or at the speaker doesn't matter. Suitable attention to prevent shorting of bare resistor terminals to nearby kit should be taken.

With the (potential) power you are talking about, you'll probably need a metal clad resistor (example), bolted to a heatsink. Though you could try it without a heatsink first assuming you are listening in the "first watt" range.

Bear in mind these types of resistor are wire wound, so will have some inductance of their own. Ive no idea if this is enough to have an audible impact.


Also bear in mind. The output impedance of an amp will not be a fixed resistance. It will vary with frequency like the impdeance charactersitic of a speaker. Again I have no idea how much - probably much less than a speaker does. But it means a resistor won't provide the exact same effect.


EDIT - looking at your amp measurements above, I think around a 3db variation was seen in the frequency response curve.

If we assume the 8 ohm speaker here varies from say 5 ohm, up to 20ohm, then with a 2 ohm series resistor, you would get voltage variations from 70% (at 5ohms) to 90% at 20 ohms. This would be a 2.2dB variation. If the speaker drops as low as 4 ohms, then the variation would be 2.7dB. Alternatively changing to a 3ohm resistor would result (in the 5 to 20 ohm case) in a variation of about 3dB

Thank you for taking the time to write a clear and detailed description.

Given what is involved, I can see that’s not something I would endeavour to do in my system.

But I’ll be passing this information on to my friend.

Thanks again!!!
 
I’ve mentioned elsewhere my anecdotal experience of listening to a pair of small floorstanding speakers at my friends place, powered by a 300 B tube amp. I found myself intoxicated with the sound. In a way that I did not get from when they were hooked up to a Hegel solid-state amplifier.

I see that the amp has been measured, and since I’m not good enough at this stuff to confidently derive any conclusions, does anybody want to take a look at the measurements?

What I’m wondering is whether the measurements suggest that the amplifier could indeed result in audible colorations on a loudspeaker versus a good solid state amplifier.




Excerpt:

—————-

Here we can see significant deviations of about 4dB from 4 ohms to no load through the flat part of the audioband (100Hz to 10kHz), reaching as high as about 6.5dB at 20Hz. This is an indication of a very low damping factor, or high output impedance. The variation in RMS level when a real speaker was used is about 3.5dB through most of the audioband.

To expand on the Lybra’s frequency response when using a real speaker, the chart below . .
.View attachment 417756



. . . shows the frequency response (relative to 1kHz) using a continuous sweep for the Focal Chora 806. Again we see deviations of up to 3.5dB within the audioband. It’s important to mention that deviations of this magnitude would be clearly audible, giving the Lybra amplifier a “sound” that would change based on the characteristic impedance curve of the speaker it’s connected to.

————-

Do folks here concur?
The change in frequency response of a real loudspeaker because of the amp's high output impedance will surely be audible at this level. It will be different with every speaker though, depending on the speaker's impedance.

The distortion level looks like it may be audible to me too based on my crude self test.

My theory has been for some time that amps with a high output impedance audibly changing the frequency response of the speaker it is driving may well be interpreted as "better sound" if the device making the change is expensive enough.
So many people think price is the most important indicator of quality :(
 
The change in frequency response of a real loudspeaker because of the amp's high output impedance will surely be audible at this level. It will be different with every speaker though, depending on the speaker's impedance.

The distortion level looks like it may be audible to me too based on my crude self test.

My theory has been for some time that amps with a high output impedance audibly changing the frequency response of the speaker it is driving may well be interpreted as "better sound" if the device making the change is expensive enough.
So many people think price is the most important indicator of quality :(
Different is better ;) , due to many factors we simply have no good references , we are running around in the circle of confusion ...
 
Different is better ;) , due to many factors we simply have no good references , we are running around in the circle of confusion ...
Different can be better it depends to an extent on your priorities.

Take as an example Klipsh Cornwall. I've heard them sound fine except for that blatant midrange colouration plastered over everything. I could not live with that.

But the people who have them don't care about that at all - the speakers make everything into a live, exciting sound, and that's what they prioritise.

For many years I assumed that people at show demos who raved about the sound when I thought it was some of the worst quality playback I'd ever experienced, must never have heard what I consider to be a good speaker system. But fact is most of them have. It just isn't what they're looking for.

So I support Matt Hooper's stance on this most of the way. But there is a sub-set of audiophiles with systems even they don't like. They just don't know it. They complain a lot about recording quality, is how to spot them. They will also tell you that the better a system gets, the more recordings that become unlistenable.
 
But there is a sub-set of audiophiles with systems even they don't like. They just don't know it. They complain a lot about recording quality, is how to spot them. They will also tell you that the better a system gets, the more recordings that become unlistenable.

That is an amazing insight, and it is so true. They build strange systems with "fun" sound with all sorts of nonlinearities. They widen Toole's circle of confusion instead of narrowing it. The result is: recordings become unlistenable.
 
Different can be better it depends to an extent on your priorities.

Take as an example Klipsh Cornwall. I've heard them sound fine except for that blatant midrange colouration plastered over everything. I could not live with that.

But the people who have them don't care about that at all - the speakers make everything into a live, exciting sound, and that's what they prioritise.

For many years I assumed that people at show demos who raved about the sound when I thought it was some of the worst quality playback I'd ever experienced, must never have heard what I consider to be a good speaker system. But fact is most of them have. It just isn't what they're looking for.

So I support Matt Hooper's stance on this most of the way. But there is a sub-set of audiophiles with systems even they don't like. They just don't know it. They complain a lot about recording quality, is how to spot them. They will also tell you that the better a system gets, the more recordings that become unlistenable.
I tend to agree in the case of loudspeakers, and I really think we like what we are used to (though that could be the Asperger's me not liking change)

In the case of the amp with such a high output impedance that it audibly changes the frequency response of the speakers it is driving the changes it makes will be dependant on the speakers particularly between a 2-way and 3-way which will probably have different sized peaks at different frequencies, so unpredictable.
 
That is an amazing insight, and it is so true. They build strange systems with "fun" sound with all sorts of nonlinearities. They widen Toole's circle of confusion instead of narrowing it. The result is: recordings become unlistenable.
I've always thought this was the explanation for, for instance, Zu, or possibly Tektron and certain tube amps. I haven't heard either of those speaker brands, tho, just marveled at their measurements.

Whereas the other intentional distortion is 'showroom treble' so the speakers seem to have more 'detail' in the showroom (B&W and Paradigm being offenders I HAVE heard, and found irritating after more than a minute or two).
 
Whereas the other intentional distortion is 'showroom treble' so the speakers seem to have more 'detail' in the showroom (B&W and Paradigm being offenders I HAVE heard, and found irritating after more than a minute or two).

I find the B&W showroom sound to be very obvious (doesn’t seem to need a blind test to notice the boosted upper frequencies).

But I still find it sort of smooth sounding, as in not coarse sounding, and so they don’t seem to bother my ears. At least not quickly. That may be totally different if I tried to live with them over time.

I auditioned the paradigm persona speakers when I was looking for a replacement for my Thiel 3.7s

I adored the sound of the Thiels but their cabinet was too deep and caused ergonomic problems near my room entrance.

My first impression of the persona speakers was that they sounded very similar to the Thiels. Except that over the audition I found the highs fatiguing. I just wanted to stop listening after a while.

Not surprising given how they measure in the highs.

That said, If you compare the Stereophile measurements of the paradigm persona versus the Thiel 3.7 that I owned, you can see they’re pretty similar in someways in the on axis frequency response.



Paradigm:

1737595480466.png

1737596074531.png



Thiel 3.7

1737595509779.png

Though with the Thiels, the in room measurements at the reviewers showed a sloping down response in the highs:

1737595897528.png

Which I expect would help explain why I did not find the Thiels right or fatiguing to listen to in my room at all. And it’s possible my tube amplifiers and well-controlled room is doing some work there as well. (I had an audiophile buddy who put off coming over to listen to my system because he’d had previous “ so bright they drove me out of the room” experience with Thiels. When he eventually settled down for a long listening session at my house, he was very pleasantly surprised to find they sounded very smooth and easy to listen to, while not sounding “ dark” at all.
 
So you'd need an idea of the "average' output impedance of the tube amp. I've no idea what magnitude this typically has - this thread suggests in the region 1 to 4 ohms.

So let's take 2ohm.

So you are going to need a 2 ohm resistor. Lets say you are dumping up to 50W into an 8 ohm speaker. That will be about (I = sqrt( P/R)) - about 2.5A

That 2.5A through 2 ohms will be 12.5W (unsurprisingly in the ratiio of the two impedances) . This is a minimum - I'd go for 4 x that to stay well on the safe side, so 50W

So you are going to need a 50W 2 ohm resistor in series with the speaker. All you need to do is solder one of the speaker wires to one end of the resistor, and connect the other end of the resistor to the speaker terminal using a short length of wire. At the amp end or at the speaker doesn't matter. Suitable attention to prevent shorting of bare resistor terminals to nearby kit should be taken.

With the (potential) power you are talking about, you'll probably need a metal clad resistor (example), bolted to a heatsink. Though you could try it without a heatsink first assuming you are listening in the "first watt" range.

Bear in mind these types of resistor are wire wound, so will have some inductance of their own. Ive no idea if this is enough to have an audible impact.
Those styles of wire wound load resistors can have an appreciable increase in impedance over the audible frequency range.

"Non inductive" thick film resistors in a TO-220 style packaging tends to be superior and often do not have an appreciable change into the MHz region. Their accuracy and temperature coefficients can also be excellent. My experience has been in test and measurement systems, or EMC testing. I can't attest to their 'sonic purity' :)

Caddock's MP9100 line is pretty commonly used. https://www.digikey.com/en/products/detail/caddock-electronics-inc/MP9100-2.00-1-/1284436
 
Those styles of wire wound load resistors can have an appreciable increase in impedance over the audible frequency range.

"Non inductive" thick film resistors in a TO-220 style packaging tends to be superior and often do not have an appreciable change into the MHz region. Their accuracy and temperature coefficients can also be excellent. My experience has been in test and measurement systems, or EMC testing. I can't attest to their 'sonic purity' :)

Caddock's MP9100 line is pretty commonly used. https://www.digikey.com/en/products/detail/caddock-electronics-inc/MP9100-2.00-1-/1284436
Yes, wire wound resistors do have measurable inductance. I grabbed a few 20 and 50 Watt, different manufacturers, different resistances. Here are the resulting impedances divided by the DC resistance.
1737616787585.png

Would you want to use these in MHz applications? Depends on the circuit.
Will this have an effect on an audio circuit? The worst example I could find from the parts bin has 0.3% increase at 20kHz. This is an inaudible change in impedance, in a region where we have little remaining sensitivity.
 
Those styles of wire wound load resistors can have an appreciable increase in impedance over the audible frequency range.

"Non inductive" thick film resistors in a TO-220 style packaging tends to be superior and often do not have an appreciable change into the MHz region. Their accuracy and temperature coefficients can also be excellent. My experience has been in test and measurement systems, or EMC testing. I can't attest to their 'sonic purity' :)

Caddock's MP9100 line is pretty commonly used. https://www.digikey.com/en/products/detail/caddock-electronics-inc/MP9100-2.00-1-/1284436
Sure - but seeing as we are "emulating" the output impedance of a tube amp - which will also have significant reactive characteristics - also interacting with the speaker impedance curve - I don't see it as a huge problem.

As I've said upthread - a far better way than using any resistor of any accuracy would be to measure the specific impedance characteristics of the amp and of the speaker, and emulate the resulting frequency response with DSP.
 
Different can be better it depends to an extent on your priorities.

Take as an example Klipsh Cornwall. I've heard them sound fine except for that blatant midrange colouration plastered over everything. I could not live with that.

But the people who have them don't care about that at all - the speakers make everything into a live, exciting sound, and that's what they prioritise.

For many years I assumed that people at show demos who raved about the sound when I thought it was some of the worst quality playback I'd ever experienced, must never have heard what I consider to be a good speaker system. But fact is most of them have. It just isn't what they're looking for.

So I support Matt Hooper's stance on this most of the way. But there is a sub-set of audiophiles with systems even they don't like. They just don't know it. They complain a lot about recording quality, is how to spot them. They will also tell you that the better a system gets, the more recordings that become unlistenable.
"They will also tell you that the better a system gets, the more recordings that become unlistenable."
Is that not the contention of all pro monitors manufacturers?
I share your dislike for the Klipsh Cornwalls, as with the majority of speakers I have auditioned at shows, they are neither Hi Fi nor accurate pro monitors. The lively sound you mention is easily attained with Highly efficient light weight paper cone drivers costing a fraction of the price of Klipsh The horns are made by electrovoice, not the greatest.
It should be clear to all audiophiles by now, that there are HiFi speakers and monitoring speakers, make your choice and live with it.
 
I haven’t found that at all, ‘neutral’ loudspeakers just sound transparent, which as transducers that is exactly what they should be.
Keith
 
I haven’t found that at all, ‘neutral’ loudspeakers just sound transparent, which as transducers that is exactly what they should be.
Keith
Good for you, but in my many years in the business and as an audiophile, I have never attempted to make a sale by mentioning the adjectives neutral or transparent.
 
Perhaps you should have.
Keith
 
Good for you, but in my many years in the business and as an audiophile, I have never attempted to make a sale by mentioning the adjectives neutral or transparent.
If that is the case, it is probably down to the audiophilliac Stockholm syndrome of your customer base.
 
If that is the case, it is probably down to the audiophilliac Stockholm syndrome of your customer base.
The audiophiliac Is still schlepping, bs-ing and targeting gullible ignorant customers, I repented decades ago and I am now a born again audiophile.
 
Yes, wire wound resistors do have measurable inductance. I grabbed a few 20 and 50 Watt, different manufacturers, different resistances. Here are the resulting impedances divided by the DC resistance.
View attachment 423211
Would you want to use these in MHz applications? Depends on the circuit.
Will this have an effect on an audio circuit? The worst example I could find from the parts bin has 0.3% increase at 20kHz. This is an inaudible change in impedance, in a region where we have little remaining sensitivity.
This means that when we have a 4ohm band tweeter and want to attenuate that say... 3.8dB using the 'worst' resistor this comes down to 0,1dB treble roll-off at 20kHz compared to a non-inductive one.

Wirewound can also be bifilar wound and would have lower inductance.

Agreed... no audible consequences other than in one's mind (knowing something non-inductive is being used). You can certainly 'hear' that.
 
Back
Top Bottom