Hi,

On Wed, 19 Sep 2007 05:55:10 -0700, Telstar Electronics wrote:

The SkyWave 2879ABTC was discontinued because of a poor profit margin.

Raising the price was not an option?

Cheers,
__
Gregg

Distortion is more than harmonic distortion.

If any output does not replicate the input in amplitude and frequency response,
there is distortion ... period!

If a 1 Vp-p pure sine wave swept between 300 Hz and 3000 Hz goes into a device
and produces an output of 2 Vp-p pure sine wave between 300 Hz and 3000 Hz there
is a uniform gain of 2 and NO distortion.

If a 1 Vp-p pure sine wave swept between 300 Hz and 3000 Hz produces 2 Vp-p at
300 Hz, then rises to 2.5 Vp-p at 1000 Hz and further rises to 2.8 Vp-p at 3000
Hz, the signal is distorted! The output is NOT a constant multiplier of the input.

Run the test I posted. Or run it against a spectrum analyzer. Or, simply admit
the output does not replicate the input. Lack of replication, in amplitude,
frequency response or internal non-linearity, is distortion.

A-D is specifying harmonic distortion ONLY.

Speech compression is the deliberate introduction of distortion [to provide some
desired result]

In the recording industry, when analog recording was the norm, the recording
studio distorted the recorded signal by the addition of 'pre-emphasis'. That is
shaping the frequency response to compensate for known frequency variations in
the recording media. The playback electronics may have had 'de-emphasis',
depending upon recording media, to remove the effects of pre-emphasis and media
distortion so that the output replicated the input. That is controlled
distortion to achieve a desired result.

Speech compression in radio-telephony is intentional distortion to achieve a
perceived desired result.

In mathematical terms TD = THD + dA/dF + dA/dT + dA/dV + d(Af(F))/dF ... should
I continue?

TD = Total distortion
THD = Total Harmonic Distortion
dA/dF = Intentional Amplitude variation as a function of frequency
dA/dT = Amplitude variation as a function of temperature
dA/dV = Amplitude variation as a function of bias voltage [AKA common mode, bias
effects]
d(Af(F))/df = Amplitude and gain variations as a function of gain roll off of
the active device [a function of device gain rolloff [AKA gain*bandwidth product]].

I can add some more terms if you like.

As a Chief Engineer I know and understand exactly what I am stating.

A speech compression circuit deliberately produces a dA/dF, a variation in gain
as a function of frequency, to achieve an intended result. It is distortion!

As a company advertising and marketing a device, you should be precise in your
language or define your terms explicitly.

As far as I'm concerned this topic is dead, the funeral has been held, and the
grave has been covered. I have no need to defend myself.

Deek




Telstar Electronics wrote:

On Sep 19, 12:58 pm, Deek wrote:

I'm sure they will acknowledge that it introduces controlled distortion.


I'm surprised at you... being a chief engineer and all... you should
realize that in this audio application we're talking strictly about
harmonic distortion. After all, that's what you can hear. In the
Analog Devices datasheet (page2)... this is given for the SSM2166 as
"Total Harmonic Distortion including internal chip noise" of typical
0.25%... and a maximum of 0.5%. These figures rival the finest audio
equipment!
www.telstar-electronics.com

The SkyWave 2879ABTC was discontinued because of a poor profit margin.
We hated to do it... but with the temperature compensating biasing...
and all the premium components and chassis... it was just too costly
to manufacture.

Why on earth would you put temperature compensating components
into this product. As far as I know no other doorstop on the market
uses them.

Wes Stewart wrote:

SNIPPED

So a bandpass filter for example is a distortion generator?

Absolutely! The output does NOT replicate the input. That is distortion!

Your definition of distortion is limited to harmonic distortion. There are other
forms of distortion that have been used throughout the industry for years.

When the output does not faithfully replicate the input, including any gain
factor, it is distortion. So, a bandpass filter absolutely introduces distortion
when it is used to exclude components of a broad band of signals.

Better
tip off all of those BC stations and recording studios that are using
equalizers.

recording studios know exactly what they are doing. they are selectively
distorting the audio to achieve a desired effect. See my other post on this topic.



A brief course in Fourier transforms will convince you that a distorted sine
wave signal is still a series of sine waves, the fundamental, the 2nd, 3rd, 4th
harmonic, etc. The existence of those harmonics is by definition DISTORTION.


I see. So a perfect sine wave run through a perfect clipper will
generate a 2nd harmonic?


In the general cse the distortion products for all frequency components have to
be considered in the analysis. The coefficients for any one or more frequency
components can have a zero value.

A perfect sine wave running through a perfect clipper will produce a square wave
that is harmonically an infinite series of odd valued frequency components.

A device that cause an amplitude variation as a function of frequency does not
replicate the input signal. By definition that is distortion [I naver stated
that it is harmonic distortion. Your definition of distortion is incomplete.]

On Wed, 19 Sep 2007 22:00:37 -0400, Deek wrote:

Wes Stewart wrote:

SNIPPED

So a bandpass filter for example is a distortion generator?

Absolutely! The output does NOT replicate the input. That is distortion!

Your definition of distortion is limited to harmonic distortion. There are other
forms of distortion that have been used throughout the industry for years.

Don't put words in my mouth. I asked *you* whether *you* thought a
BPF was a distortion generator. Now we know your position. Any
device that doesn't have DC to daylight frequency response and a
perfect, infinite headroom, linear transfer function causes
distortion. Gotcha.


When the output does not faithfully replicate the input, including any gain
factor, it is distortion. So, a bandpass filter absolutely introduces distortion
when it is used to exclude components of a broad band of signals.

Better
tip off all of those BC stations and recording studios that are using
equalizers.

recording studios know exactly what they are doing. they are selectively
distorting the audio to achieve a desired effect. See my other post on this topic.

Uh huh. So recording studio personnel know what they are doing, but
when I said earlier:

"Since the purpose of a communication system (unless you're just
peddling hardware) is to communicate, then one has to look at the
trade-off between distortion and intelligibility. If purposefully
"distorting" the signal by frequency shaping, compression, clipping or
any combination thereof improves the intelligibility at the other end
of the circuit then "distortion" is a good thing, semantics aside."

I didn't know what I'm talking about. Right?


A brief course in Fourier transforms will convince you that a distorted sine
wave signal is still a series of sine waves, the fundamental, the 2nd, 3rd, 4th
harmonic, etc. The existence of those harmonics is by definition DISTORTION.

And of course you believe that you are the only one in this discussion
that understands Fourier transforms.



I see. So a perfect sine wave run through a perfect clipper will
generate a 2nd harmonic?


In the general cse the distortion products for all frequency components have to
be considered in the analysis. The coefficients for any one or more frequency
components can have a zero value.

Why are you slipping into the "general case" when everything else you
have put forth is in absolute terms?


A perfect sine wave running through a perfect clipper will produce a square wave
that is harmonically an infinite series of odd valued frequency components.

Which is different from "a series of sine waves, the fundamental, the
2nd, 3rd, 4th harmonic, etc."


A device that cause an amplitude variation as a function of frequency does not
replicate the input signal. By definition that is distortion [I naver stated
that it is harmonic distortion. Your definition of distortion is incomplete.]

You are "distorting" what I said, I never offered any definitions.

On Sep 19, 8:53 pm, wrote:
Why on earth would you put temperature compensating components
into this product. As far as I know no other doorstop on the market
uses them.

Not temp compensating components... temp compensating bias.
www.telstar-electronics.com

On Sep 19, 12:22 pm, wrote:
i do truly admire the way you have these clowns helpsing advertise
your product
My hat is off to you.

Funny how it works that way... lol
www.telstar-electronics.com

On Thu, 20 Sep 2007 02:54:33 -0700, Telstar Electronics
wrote:

On Sep 19, 8:53 pm, wrote:
Why on earth would you put temperature compensating components
into this product. As far as I know no other doorstop on the market
uses them.

Not temp compensating components... temp compensating bias.
www.telstar-electronics.com

Oh...your bias is magical? It uses no components for compensation?

I always suspected your design to be that of an imaginary nature.

On Sep 20, 3:03 pm, wrote:
Oh...your bias is magical? It uses no components for compensation?
I always suspected your design to be that of an imaginary nature.

Not magical... it used the base-emitter voltage drop of an MJE3055T
(mounted directly to heat sink) to track with the 2SC2879 power
transistors. Since it used a transistor with gain... it worked so much
better than just a plain old diode on the sink for tracking. It worked
fantastic... I was able to set set class AB bias ( I chose about
200Ma) on the power transistors... and hold +/-5% collector current
changes from -40 to +100C heat sink temp during amp operation.
www.telstar-electronics.com

On Thu, 20 Sep 2007 17:08:23 -0700, Telstar Electronics
wrote in
. com:

On Sep 20, 3:03 pm, wrote:
Oh...your bias is magical? It uses no components for compensation?
I always suspected your design to be that of an imaginary nature.

Not magical... it used the base-emitter voltage drop of an MJE3055T
(mounted directly to heat sink) to track with the 2SC2879 power
transistors. Since it used a transistor with gain... it worked so much
better than just a plain old diode on the sink for tracking. It worked
fantastic...


How well it works depends on how well you hacked the design from this
link I posted (while pointing out some of the many problems with your
earlier splatterbox):

http://www.ifwtech.co.uk/g3sek/tr-bias/tr-bias1.htm


I was able to set set class AB bias ( I chose about
200Ma) on the power transistors... and hold +/-5% collector current
changes from -40 to +100C heat sink temp during amp operation.


That's interesting. Previously you said your bias scheme can "hold the
500mA bias to 10% from -30 to +85C". It's funny how many times you
have changed the specs. I especially enjoyed the time when you changed
the power output specs three times in one week after you were caught
fabricating them instead of actually measuring them. Oh well, such is
the life of an internet-educated hack.

But alas, there's still one unresolved issue here, Brian.... where's
the schematic?