From: "Dave Emery" <die@pig.die.com>
Date: Tue, 26 Dec 1995 06:55:52 +0800
To: lull@acm.org (John Lull)
Subject: Re: FH radios
In-Reply-To: <30de2109.16381795@smtp.ix.netcom.com>
Message-ID: <9512252206.AA10260@pig.die.com>
MIME-Version: 1.0
Content-Type: text/plain


> 
> On Sat, 23 Dec 1995 08:20:28 -0800, Steven Weller wrote:
> 
> > Thus in a frequency-hopping radio you can push the retuning (read RF
> > phase-locked loop) technology to its limit and build transmitters and
> > receivers around them. These typically hop in the order of 100 times a
> > second. The adversary has to find the uncorrelated signal very quickly
> > indeed *and* have PLL technology at least as good as yours to recover
> > anything from it. Finding the signal generally means listening to all
> > frequencies at once, requiring huge amounts of hardware parallelism and/or
> > realtime computing power. Once you throw ten or so radios onto the same
> > band, it's no longer any use looking for the strongest signal, making that
> > approach useless.
> 
> This is nowhere near the limit of the technology.  15 years ago, I was
> working on PLLs that would stabilize within a couple degrees of final
> phase within 3.5 microseconds.  That permits you to do useful work at
> 100,000 hops per second.
> 
	There is also a newer technology called direct digital synthesis
or DDS that works by accumulating phase (adding to the previous value) 
each tick of a high frequency clock in a register at a rate determined
by the contents of another register (the value here sets the frequency)
with the upper bits of the accumulated phase being used to address a
sine/cosine lookup table rom which in turn feeds digital output values
into a D/A converter.  The output of the D/A converter is a sampled 
approximation of a sine or cosine wave at a frequency set by the
increment register.  The sample rate is set by the high frequency clock
rate.

	DDS permits instanteous frequency shifts with phase continuous
output by simply reloading the phase increment register with another
value.  Unlike phase locked loop synthesizers a there is no transient
phase and frequency error after a frequency shift.

	The primary limitation of DDS is set by the speed of the 
rquired digital hardware (and various subtler considerations such as
clock jitter and output filtering) - current VLSI implementations work
up to around 100 mhz with .1 hz or better frequency resolution.

	And with a bit more sophistication the DDS principle can be used
to digitally generate vector modulation (BPSK, QPSK, QAM etc) and even
digitally filter the result with FIR filters to limit occupied
bandwidth.   There have even been some experiments with generating
broadcast FM stereo signals directly from digital music samples using
this technology.

	But to get back to the original point of this thread - while
such techniques are possible (as is full hard encryption), it is my
understanding that actual conusmer 900 mhz digital cordless phones
that use frequency hopping use a very limited set of frequencies
and a small set of fixed hopping patterns and don't hop very fast. 

	There is certainly little additional cost to building a
trully secure digital cordless phone given the dense ASIC technology
that is standard in this kind of product - but someone has to
persuade the manufacturers that there is a real need and find a way
to allow them to export the product.

	When the brand of cordless phones that most emphasizes security
from eavesdropping in its point of sale advertising display is the one that
uses open FM with simple speech inversion you know there is something
wrong, particularly when the company that makes it is a pioneer in
really secure digital speech over handheld radios (and a big governmeent
contractor).

						Dave Emery  N1PRE