If you note the post from you I quoted there is no missing context. In that you put in a blanket statement which I quoted in it's entirety. In other posts you may have mentioned it but not here, which is why I picked it up
I hate people who try to be a smart-*ss and don't read the information and papers that have been shared.
That's like talking to a wall.
I don't think we are talking about the same thing.
An high mechanical Q (Qm) , is the result of high mechanical losses in the speaker.
Mostly in the Rms. Which can result in more distortion (not always true).
It most certainly means a lot less efficient speaker.
So choosing a speaker with a low Qm is a rather strange approach, just to meet the needs for a CC amplifier.
That's like fixing one problem and get a couple of extra problems in return
Right, that's why reading a couple of comments and/or pages is important to understand context.
I said right and very clear from the beginning that I was talking about the lower frequencies.
Communication is rather hard if people don't put in the effort of reading properly.
Misreading or miscommunication happens sometimes, in that case ask for clarification, instead of just bluntly assuming and judging certain things.
A loudspeaker is not a linear impedance. The voice coil heats up and it has a substantial temperature coefficient, the voice coil self inductance gets modulated as the cone moves back and forth and you indeed have the back-EMF that is neither linearly dependent on the cone motion nor on the applied current.
This is a test , to prove which is the main limiting factor, and that behind the "visible" current source we actually have a voltage source: the power supply.
That makes it quite relevant.
Lamented Enzo once said: "people make great fuss around minute amplifier details, while actually the amplifier is just that thingy between the power supply and speakers"
If the damping is not provided by the driver it needs to be damped electrically. No getting away from that.
dave
Right we are not talking about the dampening.
An high Q means very little mechanical resistance.
That is actually something we want!
The higher this Q, the better.
Anyhow, I am a little confused why this is still a thing?
Constant current amplification doesn't work well for when the speaker works in the controlled system region anyway.
Which has been shown in quite some papers and literature now.
Only maybe in some very specific applications.
So why bother going that route? Just for the novelty of it?
Especially when other techniques like motion feedback do work extremely well there!
Or just get a better performing woofer for the low-end
Wishing we could leave the low-end out of the discussion. It's a different animal. They say, with posts titled "Music lives in the mid-range" and if there's a speaker-amplifier coupling technique for reducing the nasty sounding distortions in that frequency range, I'm all for learning about it. (Maybe even doing something about it one day)
Everyone knows some speakers are "Bi-amp" or "Bi-wire" capable. Why would you want to run two sets of wires from your voltage amp to your speakers? Separate the working current paths between the woofer and the mid-tweeter? IMHO, such wouldnt be offered commercially if it didnt have significance as a selling point; if customers didnt know what to do with that connection opportunity. Such as using different sounding amplifiers per section, or different sounding cables.
Back when speaker cables were branded "Monster", they said it was about control. If you're using thinner guage than jump-start cables, your woofer will lose control! So they were touting putting the woofer right up against that milli-Ohm amplifier output Z. Certainly understandable as a sales pitch for some scenarios.
Prior to that, tube amplifiers sometimes featured the "damping factor" control, which was basically a feedback level control, which was basically an output impedance control, that was supposed to be used to tighten up the bass response - if necessary. Who knows what it did to the rest of the audio spectrum's sound, in terms of the variable output impedance of the amplifier driving mid-tweeters. If on-line forums existed in the 1960's would people have written "I turned down my damping factor and the mids lost all that grainy sound"?
In addition, there's the huge space of opinion around just what an amplifier's feedback level should be, what it should be when driving a speaker versus a test load; which, I assume for the case of a tube amp, is just the straight forward way of controlling the amplifier output impedance. Which I assume ranges from sub-Ohm to multiple Ohms as feedback varies out to nothing (Zero). As we've probably all heard, some of the most musical sounding systems use zero feedback triode tube amplification. Which sounds like leaning a lot toward current source - high Z - speaker drive.
Could there be a fit of these ideas into the discussion here? For the mid-tweeter part of the spectrum, please?
I agree. I never claimed the impedance of the voice coil was real (i.e., purely resistive) or linear. Ohm's Law applies to nonlinear impedances just as much as it does to real/purely resistive ones. The math just gets more complicated once the nonlinear effects are factored in.
Tom
Not at all. If you look into data sheets, PP pentode amplifiers typically have much less distortion at the same power than PP triode amplifiers. This is in straight circuits without feedback. For example, Class AB1 2A3 can produce 15 W at 2.5% THD, and 807 70 W at 2.0% THD. Superiority of pentode, particularly 6L6, has been discussed in technical publications, for example Kenyon bulletin below.
I am familiar with this work, and other articles on the subject, like Crowhurst. They discuss problems of pentodes driving reactive loads and frequency-dependent impedances, and conclude that triodes are more suitable drivers for real world speakers. This is all true, but there are types of speakers that present almost purely resistive, frequency-independent load to amplifier output.
"High fidelity" concept has been established long before Williamson. Williamson' s followers hijacked this concept and assigned their own definition to it, namely bandwidth in excess of 20-20,000 Hz and THD in a fraction of percentage point achieved through the use of global negative feedback.
It was, unfortunately.
Last edited:
Hmmm.... Good point. But where does the back-EMF go then? I don't buy Dave's claim that it becomes zero – or if it does I don't understand how and would like to know – but I also doubt it goes into the infinite output impedance of the ideal current source and results in infinite voltage across the speaker. Maybe we should recommend that those who use current output amps install a spark gap across the speaker. Clearly there's a market for cryogenically treated, rhodium plated spark gaps now...
Tom
The same functionality as the SS TransAmps Daniel built. VERY revealing/enlightening swapping speakers around and playing with the output imedance dial (damping).
dave
Repeating what i learned reading Esa. What is the voltage/current thru an infinite resistor? ie the ideal current amp output R.
dave
Thank you for your brief intro to the subject. I really appreciate it. In case it isn't already obvious, I am not a speaker designer and I won't even play one on TV.
Tom
The (non-linear) distortion becomes dependent on the type of drive: voltage, current or something in between.
Just to be absolutely clear: when I write non-linear, I don't mean conplex but I mean non-linear, that is, not complying to superposition, voltage not proportional to current.