Joe Rasmussen, Custom Analogue Audio, JLTi

[Zaphod Beeblebrox]

10 minutes ago, KenTripp said:

 

It's actually pretty easy to build a multi driver speaker that has a flat impedance curve.

 

There's two sides to that. Some designers prefer not to use conjugate networks to flatten the impedance curve of their speakers. 

 

10 minutes ago, KenTripp said:

 

The speaker systems you cited are nothing more than good examples of poor crossover design.

 

Some are, yes. Nonetheless, a good amplifier designer understands what is available on the market and addresses the market. 

 

10 minutes ago, KenTripp said:

 

One of them even has a widely rising impedance in the top end. Seriously. So easy to fix.

 

And how do you propose to 'fix' the ESLs?

[AudioGeek]

16 minutes ago, KenTripp said:

 

It's actually pretty easy to build a multi driver speaker that has a flat impedance curve. The speaker systems you cited are nothing more than good examples of poor crossover design. One of them even has a widely rising impedance in the top end. Seriously. So easy to fix.

 

A lot of this stuff is over my head as well. But, not having a speaker flat impedance curve means we can promote over-engineered super expensive amps to compensate for that shortfall.

Added bonus, if someone does not like said speaker you can simply point out that their amp is "inadequate".

 

ESL's are clearly a different ball game - solved by a Mr Sanders.....

 

 

Edited June 1, 2017 by AudioGeek
Spelling

[ArthurDent]

5 minutes ago, Zaphod Beeblebrox said:

And how do you propose to 'fix' the ESLs?

 

They shouldn't need fixing as rising impedance in a tweeter is due to voice coil inductance.

[Zaphod Beeblebrox]

Just now, KenTripp said:

 

They shouldn't need fixing as rising impedance in a tweeter is due to voice coil inductance.

 

And due to the capacitance of the ESL, HF impedance will fall. A falling impedance characteristic will not mate well with a high source impedance. 

[ArthurDent]

1 minute ago, Zaphod Beeblebrox said:

 

And due to the capacitance of the ESL, HF impedance will fall. A falling impedance characteristic will not mate well with a high source impedance. 

 

Are the "vast majority of speaker systems", ESLs?

[Joe Rasmussen]

52 minutes ago, Zaphod Beeblebrox said:

 

You're asking the wrong person. I ALWAYS ask for and cite output impedance. It is the only number that is of interest to me. 

Duh.

Of course. Amplifiers, unlike some humans, are not self-aware. 

 

Hi Trevor

 

> the only number that is of interest to me

 

Absolutely! And I have already noted that.

 

I am sure that there is a lot of things we would agree on, and based on one of your past questions, you asked for the output impedance of the amplifier I was mentioning using at the Melbourne gig.

 

So I did note that very carefully and sure we share that in common.

 

The only thing I would add, is that the output impedance is the only thing that matters and DF is something I never think about. 

 

At the alignment level, can we also agree that any added "Re" by the amplifier, means that the amplifier cannot make the alignment Q better, and only if it was zero could it make sure that it is not worse? Read my reply to Martin about the equation (the one that Richard Small chastised me about) and it says exactly that.

 

RS even suggested adding series resistance to a driver that I had just bought and I balked and blurted out "what about the damping factor." Now I know that Re is all that matters and that is the sum of the speaker, the cable, the choke in LF filter, and the output impedance - they are all the very same thing. They are all the same thing, they are all Re.

 

Using SoundEasy on my computer, this is very easy to demonstrate - you put a resistor in series that can be manipulated as a variable and you can put the anywhere and the response does not change if the sum of all of them does not change. I am hardly being a rebel in pointing this out.

 

BTW, I believe that I know where in Sydney you live, down the Shire way, and you would be warmly welcomed if you came down the Heathcote for a visit. I would love to demonstrate an amplifier with 270 Ohm o/p Z (even Nelson Pass F1J only achieves 50 Ohm) and chew the fat - most of which we would agree upon. That 270 Ohm design can be copied by anybody, because it has been posted on diyaudio.com and is actually very easy to make - all parts available from Jaycar and Altronics and good DIY for those who want to tackle it. Cheap and surprisingly sweet sounding too, but must be very careful about the speaker connected.

 

Now have to get some work done for the day... Cheers, Joe

 

 

[Zaphod Beeblebrox]

6 minutes ago, AudioGeek said:

 

A lot of this stuff is over my head as well. But, not having a speaker flat impedance curve means we can promote over-engineered super expensive amps to compensate for that shortfall.

 

As NAD proved with the 3020 (and subsequent models), amplifiers do not need to be 'over-engineered' nor 'super expensive' to deal with a non-flat impedance curve. The facts are: It is quite a simple job to deal with non-resistive loads in 2017. It just is not a problem. 

 

6 minutes ago, AudioGeek said:

Added bonus, if someone does not like said speaker you can simply point out that their amp is "inadequate".

 

ESL's are clearly a different ball game - solved by a Mr Sanders.....

 

 

Howso? Does Mr Sanders build ESLs that exhibit a resistive load? 

 

[Zaphod Beeblebrox]

4 minutes ago, KenTripp said:

 

Are the "vast majority of speaker systems", ESLs?

 

Nope and the vast majority of speakers do not exhibit a resistive load either. 

[Joe Rasmussen]

2 hours ago, Zaphod Beeblebrox said:

 

Trouble is, it is almost impossible to build such a speaker. Some get close, but most are a very, very long way from exhibiting a flat impedance curve. 

 

 

BTW, while not the norm, this was an actual speaker design and absolutely right, can be driven by any source impedance. It is down in Melbourne somewhere.

 

Below is an actual measurement, but not not just the flat impedance, but also the flat current phase angle.

 

Off to my work desk...

 

 

 

[sir sanders zingmore]

1 hour ago, Zaphod Beeblebrox said:

Howso? Does Mr Sanders build ESLs that exhibit a resistive load? 

Of course not. He builds amps that can drive ESLs (and other real-world speakers)

[Zaphod Beeblebrox]

1 minute ago, Sir Sanders Zingmore said:

Of course not. He builds amps that can drive ESLs (and other real-world speakers)

 

Which is what I figured. No different to NAD, Krell, Levinson, Rotel and many others. All will deal with most ESLs happily. And yes, I understand that Mr Sanders builds both amps and ESLs, which possibly places his expertise at somewhat of an advantage over the others. 

[sir sanders zingmore]

Just now, Zaphod Beeblebrox said:

 

Which is what I figured. No different to NAD, Krell, Levinson, Rotel and many others. All will deal with most ESLs happily. And yes, I understand that Mr Sanders builds both amps and ESLs, which possibly places his expertise at somewhat of an advantage over the others. 

 

 I know and like his amps but of course there are a number of other amps out there that drive ESLs well. 

 

I think the broader question is should speakers be designed so that they can be driven by any amp or should amps be designed so that they can drive any speaker?

 

the answer is probably a bit of both although I think apart from some niche ends of the market, the amp should be the more flexible component 

[Zaphod Beeblebrox]

Just now, Sir Sanders Zingmore said:

 

 I know and like his amps but of course there are a number of other amps out there that drive ESLs well. 

 

I think the broader question is should speakers be designed so that they can be driven by any amp or should amps be designed so that they can drive any speaker?

 

the answer is probably a bit of both although I think apart from some niche ends of the market, the amp should be the more flexible component 

 

Agreed. 

[Newman]

On 30/05/2017 at 7:57 PM, Joe Rasmussen said:

...It is interesting that I am the only one quoting numbers and maths here - from now on, address those. Or it will just be a merry go around... ho-hum.

 

So why did you immediately, below, introduce a quote that contains no numbers?

 

On 30/05/2017 at 9:28 PM, Joe Rasmussen said:

Andy Wehmeyer of Harman International:

 

Damping factor is hogwash, pure and simple. There is no appreciable cone control provided by low amplifier output impedance.

 

Here’s the short story. The preposterous argument for damping caused by amplifiers is based on the idea that the output impedance of the amplifier causes the energy stored in the inductive component of the moving speaker to be dissipated more quickly, because the current flow is greater from the speaker to the amplifier if the amplifier has a lower impedance. The problem with that BS theory is that the speaker’s DCR* is in series with the amplifier’s output impedance and is always MUCH greater than the output impedance of the amplifier. Therefore, the speaker’s DCR* is what controls damping in the electrical circuit and not the amplifier.

 

[End of quote]

 

That maths just don't lie - argue the maths: Qes = 2Pi*F*Mms*DCR/(Bl)^2

 

But it doesn't even stop there. The Voice Coil made of copper changes DCR with temperature at the rate of 0.4% per Centigrade.

 

If the DCR is 8 Ohm, then if the VC is elevated from 20C to 50C (not abnormal), the DCR will now be 9.017 Ohm and not 8 Ohm.

 

If the amplifier has an output impedance of 0.1 Ohm, it will now be swamped by 1.017 Ohm.

 

BY TEN TIMES !!!

 

Andy, you were right - hogwash!

 

https://www.stereophile.com/content/hot-stuff-loudspeaker-voice-coil-temperatures-page-2 : "So I strongly suspect that, for most hi-fi users—those who don't habitually wind the volume control to its highest position and indulge in PA listening levels—thermal compression is a paper tiger."

 

"Paper tiger" indeed.

 

On 31/05/2017 at 1:45 PM, Joe Rasmussen said:

So Richard Small was lying to me?

 

And Neville Thiele too?

 

 

I would rather see it in their texts than hearsay, if you don't mind, so if you could assist me in that respect, I would be grateful.

Edited June 1, 2017 by Newman
complaints

[ArthurDent]

 @Newman As you're the one making assumptions your comments are disrespectful and just plain rude.

[Marc]

I agree with @KenTripp, @Newman.

Perhaps you could engage in a civil discussion in a more respectful manner.

[Joe Rasmussen]

29 minutes ago, KenTripp said:

 

It's actually pretty easy to build a multi driver speaker that has a flat impedance curve. The speaker systems you cited are nothing more than good examples of poor crossover design. One of them even has a widely rising impedance in the top end. Seriously. So easy to fix.

 

You mean 99.999% of speakers out there are failures?

 

 

 

 

[Zaphod Beeblebrox]

1 minute ago, Joe Rasmussen said:

 

You mean 99.999% of speakers out there are failures?

 

 

 

 

 

Of course not. It's just that amplifier manufacturers from NAD to Krell have, for decades, been able to deal with the minor issues caused by a non-resistive impedance characteristic. It is a non-problem. 

[Joe Rasmussen]

1 hour ago, Newman said:

So why did you immediately, below, introduce a quote that contains no numbers?

 

https://www.stereophile.com/content/hot-stuff-loudspeaker-voice-coil-temperatures-page-2 : "So I strongly suspect that, for most hi-fi users—those who don't habitually wind the volume control to its highest position and indulge in PA listening levels—thermal compression is a paper tiger."

 

"Paper tiger" indeed.

 

You assume that you heard and recalled them exactly correctly, plus that they heard you exactly correctly, plus that all contextual comments and assumptions in their minds were correctly represented in this casual conversation.

 

In short, if they haven't put it in a book, don't build your career around it.

 

What a strange post.

 

But then your by-line seems to say a lot about you.

 

[ArthurDent]

38 minutes ago, Joe Rasmussen said:

 

You mean 99.999% of speakers out there are failures?

 

 

 

 

 

Of course not Joe, it's just that some of us strive for perfection

[Martykt]

 

5 hours ago, Joe Rasmussen said:

 

Martin, I give you credit, you are the only one so far who has been willing to tackle the maths, seems others here are shying away from it.

 

But first, I want to tell you, I am not on some kind of crusade against DF, I doubt I will ever win that, the idea has become so entrenched that a little bloke like me can do only so much.

 

These days, I only think in terms of the added series elements/resistances/impedances, the sum of everything is in the Re of the equation you have finally acknowledged, and good on you.

 

The thought experiment I devised is dastardly wicked and shows up the fallacy beautifully - and the purpose is to show that the reality is in the equation, and you have in your own way recognised it because, you started to add the Re elements together in the equation.

> According to the DF equation it will be 100 200 & 50 respectively for what it's worth.

> According to the 'theory' that is the correct answer. But the following means those numbers have little value.

> Moving on firstly the Qms will remain constant for each driver.

 

Agreed.

 

> Working out the Qes, according to wiki the "output impedance should be added to Re for calculations involving Qes."

 

Bingo!

 

There is the key right there, bravo! The key word here is added.

> The Qes for the single 8 ohm driver would then be

> Qes = 2pi x Fs x Mms x 8 + 0.08 divided by (Bl)squared

> For the Qes for two drivers in series this would be 

> Qes = 2pi x Fs x Mms x 16 + 0.08 divided by (Bl)squared
> (The output impedance of the amplifier is the same ??)

 

And here you are doing the actual exercise and getting your head around the equation.

 

For simplicity, I will give you the results of our three scenarios that gave us respectively a supposed DF of 50, 100 and 200.

 

A:   4 Ohm Parallel, the Qts changes from 0.707 to 0.721

 

B:   8Ohm Single, the Qts changes from 0.707 to 0.714

 

C:  16 Ohm Series, the Qts changes from 0.707 to 0.710

 

The fact that we can put the two speakers in both series and parallel and the fundamental Qts stays the same, while DF varies enormously, shows that DF is not a number that describes what really goes on.
 
But that equation you tackled shows the Re is the only value that matters. Just wiki said. They got 100% right.

 

And you have yo add ALL ELEMENTS to get to Re and then apply the equation.

 

And you did that, you added the right numbers together.

> Also does the Bl figure double with two drivers ??

 

No, it stays the same. It makes no difference whether in series or parallel.

 

> If this is correct with all the additional figures required it would mean there would be a small addition to Qes however never a decrease.

 

You are right, never a decrease. Because Re can only be increased.

 

And the amplifier can only increase Re and make the damping worse, so if that is the limitation of the amplifier, then how can it add to damping when it can only increase Re?

 

The amplifier cannot add damping period, it can only decrease damping.

 

---

 

What else is revealed by the equation is of paramount importance:

 

If Re is increased by 1%, the Qts goes up by 1%.

 

If Re is increase by 10%, the Qts goes up by 10%

 

If Re is increase by 24.349%, the Qts goes up by 24.349%.

 

Re and Qts are entirely proportional, change one at one rate and the other changes by the exact same rate.

 

> Am I on the right track

 

YOU ARE INDEED!

 

 

Thank you @Joe Rasmussen.

 

It's not easy to get your head around this stuff and there's a lot of elements to figure out in the Thiele/Small Parameters to get a full understanding of this subject.

It shows just how amazing and accomplished Neville Thiele and Richard Small were to figure all of this stuff out.

We should never forget the contribution they have made to this field.

 

While there's still a lot for me to learn and figure out in this subject I've definitely gained some knowledge here which in my opinion is always a good thing.

This thread has definitely been worth the effort.

[acg]

On 30 May 2017 at 9:40 AM, mwhouston said:

This is true and I'm not sure where the idea came from that they don't. It may have been that originally they were low power while they worked out the design. I have built 100, 150, 180, 250, 500 and about to build a 1500W Class D and they all have excellent bass performance. It also maybe that most Class D amps use SMPS. I think this is where Class D gets its bad name. I have only ever used rugged linear PSs with big power trannies and lots of heavy duty power caps. My Class Ds sound warm and very in control.

 

Maybe Class Ds with SMPSs don't do good bass. Ill never know because Ill never build one with that type of PS.

 

Every Class D that I have heard, including those with beefy LPS, sound OK in the bass but more and more ordinary as the frequency increases.  Have not heard your Tripaths though.

[alistairm]

Guys

 

There has been a report about this thread and it is on moderator watch. I'm sure the discussion is fascinating, but I honestly can't be bothered reading it, so am going to do with a quick general note here. Be nice or we will just close the thread. That would be a shame, as you all seem to be enjoying it. 

[Newman]

2 hours ago, KenTripp said:

 @Newman As you're the one making assumptions your comments are disrespectful and just plain rude.

 

The only assumption I made is that memory is not perfect -- which is more of a fact than an assumption -- but my wording was a little harsh, so I have modified it. Thanks for pointing this out.

 

Personally, though, I think that this thread is somewhat obsequious, starting with your thread-starting post. The person at the centre of discussion, I think I can safely say, holds audio-related views that are, um, ah, how shall I put this, a little unconventional, and thus I don't think that it is in any way inappropriate or disrespectful to strongly challenge those views or opinions, and the burden of proof (or at least evidence) should lie with the unconventional claimant rather than with the challenger.

 

I'm pretty sure that the report about this thread was you reporting me, which is fine BTW as I was a bit harsh, but FWIW I am very amicable to a friendly PM as an alternative way that doesn't get your thread put under threat of immediate closure. However, regarding showing respect, JR as a person has my fullest respect, I'm sure he is a lovely lovely guy, but if I think his audio ideas are not deserving of respect, I expect to be free to tackle them head on without being called disrespectful. So bear that in mind -- I don't respect his ideas in the same way that you seem to, if your posts so far are anything to go by.

 

Also FWIW, I think that JR's personal participation in this thread starts to get into a little bit of grey area with the commercial interests policy of SNA, where IMO some of the long, um, lectures already posted look a lot like selling the philosophies that are embedded into his products for sale, and hence promoting the products. Obviously the policy is not being transgressed or there would have been intervention by now, but I am uneasy about it. How is a commercially interested participant going to take a backward step in a discussion of his opinions that are embedded in his products for sale? It could lead to 'loggerheads' debates, and I fear that the wrong person may be accused of stubbornness (and reported). That is why there is a commercial interests policy.

 

cheers

[Joe Rasmussen]

 

50 minutes ago, Newman said:

 

The only assumption I made is that memory is not perfect -- which is more of a fact than an assumption -- but my wording was a little harsh, so I have modified it. Thanks for pointing this out.

 

 

And Newman, hopefully I speak for others here too, good-on-ya!