View Single Post
  #26   Report Post  
Posted to rec.audio.high-end
Trevor Wilson[_3_] Trevor Wilson[_3_] is offline
external usenet poster
 
Posts: 139
Default Introducing a New Horse to the Stable

On 19/09/2019 8:00 pm, wrote:
While much of what you say below is, strictly speaking, technically
correct under some very specific cases, it's misleading and could well
not be applicable in reality.

I've not got a lot of time to spend on this, so let me just
take on a couple of your points.

On Monday, September 16, 2019 at 5:52:54 AM UTC-4, Trevor Wilson wrote:
On 16/09/2019 7:49 am,
wrote:
On Saturday, September 14, 2019 at 9:58:44 AM UTC-4, Trevor Wilson wrote:
So, a little Ohm's Law should tell you if you are demanding more current
than the output devices are capable of delivering. 14 Amps is, by high
end audio standards, a relatively modest current ability for a (say) 100
Watt @ 8 Ohms amplifier. Provided the driver impedance is relatively
benign, you should be OK.

Hmm, that's not what a little Ohm's law tells me.

100 Watts into 8 Ohms is a tad over 3.5 amps. Let's say it's a VERY
robust 100 watt amplifier, delivering 200 Watts into 4 Ohms requires
about 7 amps, and, let's pretend it has essentially ZERO output
impedance and an effectively limitless power supply, you're not reaching
14 amps until you're driving 400 watts into 2 ohms.


**Well, no. The RMS current is certainly 3.5 Amps, but output devices
only 'care' about PEAK currents. The peak current is, of course, 3.5 X
1.414 ~ 5 Amps.

With a 4 Ohm load, the peak current required is 10 Amps. For 2 Ohms, it
is 20 Amps.

Assuming a 100 Watt amp. For a (say) 200 Watt amp, those peak current
figures become 7 Amps, 14 Amps and 28 Amps respectively. WAY past the
ability of two pairs of old Hitachi MOSFETs to deal with.


All of your calculations ASSUME several things, none of which are
likely to be true in real usage.

1. Your continuous-to-peak current calculations using a factor of
sqrt(2) ASUMMES that the excitation is pure sine. That's surely
not the case in real life, I'm sure you'd agree. Yes, the actual
crest fact may be greater than 3dB, but, again, you ASSUME that
in actual practice those peak current are REQUIRED I suggest
they are not, in absence of any supporting evidence they are.


**Whoaa. Hang on a sec. Let's get down to basics. The peak to average
figure available in music is not the issue here and is certainly not
what I am talking about.

Let us assume, for the sake of simplicity, that the music played IS
close to pure sine waves (maybe the listener is a pipe organ lover). Now
I certainly recognise that 99.9999% of music is not sine waves
(apologies to Fourier), but I need to make an important point. Let's say
we have (again, for simplicity and ignoring SOA considerations for a
moment) an amplifier that uses the most excellent Toshiba
2SC5200/2SA1943 output devices. Let's say an RMS output current of 15
Amps is required. This would not be healthy for the output devices, as
their maximum current rating is 15 Amps (peak). Yes, it can
(theoretically, SOA considerations aside) deliver a DC current of 15
Amps. However, a sine wave of 15 Amps means that the peak current will
be in excess of 21 Amps. Examine the SOA curve of the 2SC5200:


https://toshiba.semicon-storage.com/...l.2SC5200.html

You may care to note that even if the current flowing through the device
exceeds 15 Amps for as little as ONE MILLISECOND (in truth it is far
less), the device will be destroyed.

It is for this reason that peak current MUST be taken into account when
designing any solid state amp.


2. The comprehensive list of impedance curves is, yes, useful but
the interpretation of them as applied to your case ignores the
VERY important details and thus is overly simplistic and mis-
leading. Let me take just a couple of example from the list:

a. Westlake BBSM-6F
Yes, the impedance gets to 2 ohms, but look at the
broadband sensitivity 92dB/2.83V: less power is needed to
achieve a given sound pressure level, continuous, peak or
otherwise. It illustrates that you simply can't look at one
particular measurement in isolation.


**Indeed. I can assure you, however, having owned a pair for some years,
that they are utterly ruthless in exposing an amplifier's ability to
deal with tough loads. Even at modest listening levels.


And, not that it may be relevant, this is specifically
marketed towards studio use at high (deafening?:-) level.


**I can assure you that I do not and have not, for many years listened
to music at silly levels. Back when I was in my teens and early 20s,
yes. But not now and not when I owned the Westlakes.


b. The acoustat impedance curve you posted on the RageAudio
site: indeed, the impedance drops WAY down to under 1 Ohm.
BUT it does it over a VERY narrow bandwidth, and it does
it at 15 kHz. It's above 4 ohms over the entire audio range
from 10Hz to 9kHz. Exactly what kind of musical material
would require one to dump a LOT of power between 9 kHz and
20+kHz? Over the majority of the audio bandwidth, the impedance
is 6 ohms or higher. Even if you assume (quite unrealistically)
that the energy is distributed across the spectrum, equal
energy per octave, your impedance problems ar confined to about
1 octave out of 10, suggesting that if you need 100 watts in that
one octave, you'll need another 900! for everything else.

Ah, but what about short-term transients, you ask. Look at the
spectral distribution of such in actual music: sorry, you're
still NOT generating a lot of power over such a narrow bandwidth
(Sorry, Mr. Fourier, you can lay back down, we shan't be needing
you just yet).


**Oh, I agree, you won't be generating much power. However, VI limiting
systems in most amplifiers don't care about such things. They just
operate anyway. Mostly. And to the detriment of sound quality. And, more
seriously, when using almost any valve amplifier with the Accustats,
will guarantee a suck-out at the impedance dip. That is likely to be
audible.




So, a couple of questions that Mr. Ohm may ask; what kind of loudspeaker
presents a broadband 2 ohm impedance or, conversely, what kind of
musical content would generate that kind of power requirement over
the pretty narrow band of frequencies where a loudspeaker has the
kind of pathological impedance curve that would dip to as low as
2 ohms.


**I have a few here that are tougher than that. Some of the Peerless
XXLS drivers dip to the low 2 Ohm region. Most ESLs fall lower than that
at HF.


Yes, over VERY narrow bands and at high frequencies, where your
power requirements are not anywhere near as large as at significantly
lower frequencies.


**The Peerless XXLS drivers are LF drivers, requiring prodigious currents:

http://www.loudspeakerdatabase.com/P...LS-P835017.pdf

I have passive radiator based subwoofer in the workshop right now. I
will run a real world impedance plot for the system when I get some time.


--
Trevor Wilson
www.rageaudio.com.au

---
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus