Frequency Hopping Spread Spectrum

Sat Dec 28 15:27:26 CST 2013

Frank,

FH is only what it says it is, it does not refer to or define a waveform,
however a waveform may define FH.  SS would be complimentary to FH and is
used in 802.11x waveforms as well.  The early frequency hopping HaveQuick
radios were analog, hopped very slowly and you lost the audio in the dwell
time when switching to a new frequency, so there were dropouts every time it
changed frequencies.  I don't recall the dropout time, but I'd guess ~100
ms, enough to lose at least one syllable.  Current radios can hop thousands
of times faster. 
As the story goes, a soldier says,"Don't Fire" and what was heard was
"Fire".
For Hams, using SSB, (unless you could drop the output power), you would
need to cut the audio before the VFO "swept" to the next frequency.
Current radios drop power and use a CVSD modulator/demodulator that
compresses and then expands the voice to fill in the dwell time gaps and
make it a seamless voice or data output.  This is done at the digital level.

To answer the question, I don't see how one could effectively do it without
using a data stream or preample in each transmission.  Data in that preample
is critical to syncing and knowing what the frequency sequence is and if one
is using FEC, whether or not a retry is needed.  

Jim - KR9U

-----Original Message-----
From: Frank Eliot [mailto:feliot at his.com] 
Sent: Saturday, December 28, 2013 11:33 AM
To: jbwolf at comcast.net
Cc: tacos at amrad.org; 'John Donaldson'
Subject: Re: Frequency Hopping Spread Spectrum

Jim, KR9U/All -

I was using "SS" as short for frequency hopping, which I believe is a form
of spread spectrum. I think the primary advantage of frequency hopping (FH)
for hams is for use when one or more of the many frequencies in the sequence
is covered by QRM. If the narrow-band modulation is analog e.g. voice, then
an occasional brief dropout would not prevent solid copy. The whole
discussion assumed a known FH algorithm and a pre-defined set of
frequencies, as I believe I explained. The question I was addressing was how
to attain the correct sync, or to quote your message, I wish to figure out a
cheap way for hams to attain "Only when the correct sync is detected, it
starts the FH sequence." If you are aware of a simpler way to do that than
what I suggested in my original message, then that would be useful.

To employ FH where the narrow-band modulation were digital, and in the
presence of QRM on one or more of the FH frequencies, would require the use
of a quite robust error correction coding technique to reconstruct the
dropouts. That might be an interesting experiment to run, once Andre gets an
on-the-air link set up.

Frank
W3WAG

On Dec 27, 2013, at 9:53 01PM, James Wolf <jbwolf at comcast.net> wrote:

> All,
> 
> I didn't see the presentation, but I'm somewhat confused by the semantics.
> "In this discussion, "SS" I thought might be Spread Spectrum, but it 
> seems to be used for Frequency Hopping -  Maybe I'm missing something.
> I assume we are talking digital communication?  If so, a preamble 
> containing the sync information, next frequency, etc. is typically 
> used.  To initially start the communications, a known FH algorithm or 
> a pre-defined set of frequencies is used.  Only when the correct sync 
> is detected, it starts the FH sequence.
> Are we discussing an FM or a SSB/AM mode?
> Sorry if I'm way behind in the discussion. 
> 
> Jim, KR9U
> 
> -----Original Message-----
> From: tacos-bounces+jbwolf=comcast.net at amrad.org
> [mailto:tacos-bounces+jbwolf=comcast.net at amrad.org] On Behalf Of Frank 
> Eliot
> Sent: Friday, December 27, 2013 8:26 PM
> To: John Donaldson
> Cc: tacos at amrad.org
> Subject: Re: Frequency Hopping Spread Spectrum
> 
> John -
> 	This is an interesting history, but it addresses a different problem

> from what Andre was proposing. The torpedo application was interested 
> in making it difficult for the enemy to jam or spoof the control 
> signals. It sounds like the torpedo would sync up with the transmitter 
> before launch using a rate offset until sync was achieved. If that was 
> to be the case, then finding a clear channel was not an issue. The 
> question I was addressing was how to sync the receiver to the received 
> SS signal in the real ham world with crowded band conditions.
> 
> On Dec 27, 2013, at 7:41 49PM, John Donaldson <johnab8yz at verizon.net>
wrote:
> 
>> Frank,
>>   SS technology goes back to WWII. A famous Actress came up with SS 
>> as a way to control torpedoes and keep the enemy from jamming the RF 
>> control signals. Synchronization was done with coded paper tape loop.
>> the torp would listen at the freq dictated by the tape and wait for 
>> the transmitter to catch up and then both would then start cycling 
>> thru the paper tape loop. The Navy did not use the system because 
>> they were not convinced that this Actress could be that smart and was 
>> trying to fool the Navy. Turns out she also held a Engineering 
>> Degree, so knew what she was talking about. LOL
>> 
>> John Donaldson
>> AB8YZ
>> 
>> On 12/26/2013 7:30 PM, Frank Eliot wrote:
>>> 	Earlier this month, Andre Kesteloot gave an interesting talk on
> frequency hopping, and demonstrated a transmitter driver he had designed.
> One aspect of the talk really interested me. I had often wondered how 
> commercial SS systems attained their initial synchronization. I had 
> assumed that there must be some characteristic that enabled them to 
> sync up quickly either periodically or on each transmission. I 
> couldn't figure out what it was, but just assumed it must be easy. In 
> the meeting discussion, it was brought out that there is no magic, and 
> that they do it simply by doing some parallel processing that is 
> probably not practical for hams. Much of the evening's discussion 
> centered on how hams could sync a frequency hopping system. For 
> discussion in what follows, assume, as Andre did, 127 points per cycle 
> from a shift register, known frequency channelization, and a one second
cycle.
>>> 	Andre's proposal for synchronization was to start off a 
>>> transmission
> 
>>> with the required NB FCC station ID on say, frequency one, with the 
>>> timing reference being a transmitted pulse or mike button release.
>>> Then the QSO could continue in SS mode. This would be cool, and a 
>>> great generator of bragging rights, but it seems to me that it 
>>> wouldn't be very useful in the wild. The reason is simply that, if 
>>> we are confident that frequency channel one, or any other for that 
>>> matter, is clear of QRM so that we can be heard sending this initial 
>>> timing signal, then we might as well simply sit on that NB channel 
>>> and not bother with frequency hopping. Frequency hopping yields no 
>>> processing gain in the absence of QRM. Its sole advantage, I 
>>> believe, is in the case where some of the frequency channels are 
>>> blocked by QRM. In that case, if you are transmitting information 
>>> that is still usable if brief snippets of the received time series 
>>> are blanked out, then frequency hopping will yield a succe
>> ssful QS
>> O,
>>>  whereas if you had unfortunately chosen one of those blocked 
>>> channels
> for your communication, you would be out of luck, and never get started.
>>> 	It seems to me that the best, and maybe cheapest, way to establish
> sync while taking advantage of the SS processing gain, would be for 
> the receiving station to temporarily offset his shift register clock a 
> little from its one-second cycle, and then either listen for the 
> transmitted sub-audible tone, as Andre suggested, or simply watch the 
> received S-meter that had a longish time constant. When the receiver 
> clock walks into sync, the tone will be detected, or the S-meter should
jump up to indicate sync.
>>> 	I have thought a lot about a related synchronization problem in
> connection with my work on coherent detection. This goes in a 
> different direction than frequency hopping. Its goal is to achieve 
> significant S/N processing gain on an open channel by pushing to very 
> narrow band transmission. Coherent detection requires that the carrier 
> phase as received be known at the receiving site. I had to figure out 
> a way to do that cheaply, but be able to establish that 
> synchronization under the same very weak signal conditions that I 
> wanted to communicate with after sync was achieved. I proposed such a
method at the AMRAD meeting a few months ago.
>>> 	In summary, I think that whenever a transmission format is proposed
> that attempts to overcome a limitation such as QRM, QRN, etc., then a 
> necessary requirement when suggesting it for ham use is that if the 
> protocol must employ an initialization procedure before communication 
> starts, then the only fair way to advertise the protocol is to define 
> the robustness of a "system" that employs both a synchronization step 
> and a communication step as that of the poorer performer of the two. 
> If the setup protocol requires a big signal or choice of a clear 
> channel in a crowded band, it seems like the wrong way to go for hams.
>>> 
>>> 73, Frank
>>> W3WAG
>>> 
>>> _______________________________________________
>>> Tacos mailing list
>>> Tacos at amrad.org
>>> https://amrad.org/mailman/listinfo/tacos
>>> 
>> 
>> 
>> _______________________________________________
>> Tacos mailing list
>> Tacos at amrad.org
>> https://amrad.org/mailman/listinfo/tacos
> 
> _______________________________________________
> Tacos mailing list
> Tacos at amrad.org
> https://amrad.org/mailman/listinfo/tacos
> 