Re: Question For Craig Markwardt re Pioneer 10 Data

"ralph sansbury" <[EMAIL PROTECTED]> wrote in message
news:[EMAIL PROTECTED]
>
> "George Dishman" <[EMAIL PROTECTED]> wrote in message
> news:[EMAIL PROTECTED]
> >
> > "ralph sansbury" <[EMAIL PROTECTED]> wrote in message
> > news:[EMAIL PROTECTED]
> > >
> > > "George Dishman" <[EMAIL PROTECTED]> wrote in message
> > > news:[EMAIL PROTECTED]
> > > >
> > I think it is entirely appropriate where you treat a wideband
> > amplifier designed with the primary aim of having equal gain
> > at all frequencies as a resonant LC circuit whose purpose is
> > to boost one unique frequency above all others.
>
>    The relevant point here is filtering out other frequencies
> outside the range of the expected frequency. The details
> needed to smooth the reception you obviously know more about than
> I do.

Indeed I do, so when I tell you that treating a wideband
flat filter as a resonant circuit will give you entirely
the wrong conclusions, you should perhaps listen.

> > >> I'm not sure of that. Did you understand that "dual gate
> > transistor" meant simple arithmetic multiplication, not
> > superposition?
>
>  Pehaps I misunderstand the dual gate transistor: the
> voltages that arrive at the gate at the same time
> add so that a high positive voltage for example would
> be cancelled by an equally low negative voltage and two
> equally high positve voltages would be double.
>    And these combined effects would influence
> the collector emitter output of the transistor

Bipolar transistors have base, collector and emitter.
Field effect transistors have gate drain and source.
The relationship between voltage and current in each
is entirely different and neither is linear.

> which
> is then passed to different inductive capacitive circuits
> to output the difference frequency or the sum frequency or
> the reference frequencies.??

Simply adding doesn't produce sum and difference frequencies.
For a dual gate FET, the current depends on the square of the
voltage so adding the two signals can then produce a current
that includes the product because:

  (a + b)^2 = a^2 + b^2 + 2ab

The a^2 and b^2 terms can again be filtered off if the
frequencies are carefully chosen to leave just the a*b term.

> > >    Do you use an analogue digital converter to change the analogue
voltages to
> > > digital values and multiply them and then convert this back to an
> > > analogue value?
> >
> > No The signals are multiplied by analogue
> > circuits (continuously varying voltage) and the reference and
> > sum frequencies discarded to get the band of interest down to
> > a frequency range low enough for an ADC to handle it.
> >
>     Yes I understand the result but I dont understand the
> multiplication by analogue circuits???

There are several methods available, dual gate FETs (explained
above) or diode mixers for example. The key though is that the
principle of a mixer is that it produces an output containing
the product of the signals.

> > I think you misunderstood. You said ".. or perhaps striclty
> > capacitive high and low pass filters." and I said "That is
> > correct ..". You are probably correct in that it will use
> > separate high pass and low pass filters but there are no
> > details in the DSN documents and I don't intend to guess.
> >
>      But it is reassuring to know that elementary capacitive and inductive
> circuits can filter out higher and lower frequencies and leave
> the range of interest at relatively smooth voltage level over the
> whole range. And so the procedure can give reliable results.

Certainly, the only point of disagreement was that you
used to insist on treating it as a tuned circuit rather
than separate high and low pass. The difference is the
flat bit inbetween.


> > >    I snipped your diversion. It was not appropriate. An apology
> > > would be appropriate.
> >
> > If by "the single difference frequency" you meant the
> > entire 110MHz band then yes, I think you should apologise.
> >
> Ok I apologize.

Fair enough I apologise for getting a bit shirty about it too.

> > You seem to be inventing your own jargon. "Commands" are
> > instructions sent to the craft. They are not "modulations
> > of the carrier" but binary patterns.
> >
>    The binary patterns are modulations of the carrier, right?

More accurately, each bit is carried as a pattern of modulation
of the sub-carrier. The sub-carrier is then modulated onto the
carrier. There are separate stages with confidence tests at each
stage.

> > If you want to get technical, all information, be it "commands"
> > or scientific data returning is sent as modulations of the
> > sub-carrier which is a 16kHz sine wave that in turn modulates
> > the carrier.
> >
>     Which since modulation is transitive this implies my jargon
> was correct.
> But it is good to spell this out as you do here.

"Jargon" often means use of common words in a specialised
meaning. In that sense your jargon was incorrect but the
underlying ideas are right. However, when it comes to
assessing the confidence in the method, it is all about
details.

> > We are talking about varying voltages. Again going into
> > how charges moving through a resistor produce a voltage
> > is excessive detail.
>
>     Since I dont deal with these matters on a daily basis like
> you I like to keep in touch with the specific details.

Fine, but again you need to be sure the details are correct.
When a signal is applied to a transistor, it must be DC so
a base bias is added so the charges are always flowing in
the same direction but with a speed that depends on the
voltage.

> > >     But ultimately this is reduced to the sequence of 1s and
> 0s right?
> >
> > Eventually but not at this stage. There is the FFT that finds
> > the signal, the PLL that locks on to the carrier, the second
> > PLL that finds the sub-carrier then the modem before you get
> > to binary data. Then you have the error correction scheme
>
>     Yes this is good to spell out and what is the error
> correction scheme in a little more detail?

Very complex! This is the area where I have used simple
schemes and can appreciate what is being done in the more
complex ones but don't understand the details myself. It
needs a purely mathematical analysis to understand it.
You would need to study a lot and ask in a different
group for that.

The DSN document 208 give a lot of detail but I don't know
which scheme Pioneer used. The important thing to note is
that the data is broken in blocks and the error correction
scheme can detect corrupt bits within a block _and_ fix them
without first know what the transmitted data was. It allows
a direct measurment of the Bit Error Rate, the BER that you
can see on a meter on the Pioneer home page.

George