[GreenKeys] Dtc

[Gerry Block]

 .... to solve that problem, Dovetron up-converted (mixed) to two IF chains and then used two identical discriminator channels in the two IF chains.  .....As far as I know this was all done so that the Xmas Teletype Art featuring Marilyn Monroe could be received error free....


    On Thursday, March 17, 2022, 10:13:10 AM PDT, Bob kb8tq <kb8tq at n1k.org> wrote:  
 
 Hi

If you dig into it, your “FSK” signal is really two ASK signals 
chugging along. In an ideal world, you could simply demodulate
one or the other and get perfect copy. ( no don’t bother, this is 
not an ideal world …). 

Your AM signals are square wave modulated and have the pretty
well know sidebands at the appropriate frequencies. To get good
demodulation, you want the delay on these sidebands to be reasonably
close. A typical BPF of Q “something” is going to have a delay function
that changes pretty quick as you get off center. There are a number
of DSP things you can do to fix this. One is an FIR with a flat delay
design, another is a delay equalizer cascaded with a more conventional 
filter. There are other more exotic things you could do. 

As transmitted, each ASK goes from 100% to zero as you go from a 
1 to a 0 on that channel. As received noise sums in and you have a 
range like 100% to 20%. Both numbers will vary with time. Working 
out a decent averaging approach for each level is a bit interesting. 
Assuming you can come up with exact numbers ( good luck …. ) then
the threshold on each channel would be (high + low) / 2. 

You now have two “useful” signals and can decide what you want to 
do with them. You also have the more conventional “which one is bigger?”
decision threshold. Since you have a fancy computer involved, you can
decide on the fly which of these three things you want to go with. 

Yes, this is just the start of wandering down the rabbit hole. There 
are lots of zigs and zags you can take. 

Bob

> On Mar 17, 2022, at 12:42 PM, Harold Hallikainen via GreenKeys <greenkeys at mailman.qth.net> wrote:
> 
> 
> 
> On Wed, March 16, 2022 10:46 pm, Gerry Block wrote:
>> Harold
>>  For your project you might look up the old patents held by Frederick
>> Electronics regarding the design of the DTC known as the decision
>> threshold computer. This was a rather complex rc circuit which tracked
>> the peak (Dc) amplitudes of the mark and space channels and biased the
>> output comparator (“slicer�) threshold to provide the most
>> advantageous switching  point. In a bipolar design this would be 0v in
>> perfect conditions.  But during selective fading it would move the
>> threshold to as far as  .5 x the working channel peak amplitude.
>> They did a lot of real research work in this.  Irv hoff and Keith
>> Petersen took this design (which I think came from Vic Poor) into the TT/L
>> design  and I think there are also some descriptions in issues if RTTT
>> journal.
>> Best
>> Gerry
>> AD6MC
>> 
> 
> Thanks! I'll do that. So far, my DSP TU works great on ITTY, but not that
> great on over the air signals. I've recorded some over the air signal at
> https://w6iwi.org/rtty/audio/ . I wonder how much selective fading there
> is with 170 Hz shift. I'm going to calculate how far down one tone is when
> a single reflection results in a complete null of the other tone. I also
> plan on recording a two tone signal (mark and space simultaneously)
> received on a web SDR and see how the tones track with fading. I'm also
> going to add high speed UART output so I can get samples of various
> signals inside the chip. I'll look at the patents on dynamic threshold
> control. It SEEMS like perhaps the discriminator output could be run
> through a low pass filter and divided by two to get a threshold. With hand
> typed data, where the data spends most of its time in mark, the threshold
> would tend to be the mark level divided by two. I did a closed captioning
> system for movie theaters where I was transmitting 10 kbps data
> continuously on FSK over IR. There, I used a couple diodes, resistors, and
> capacitors to get the positive and negative extremes out of the
> discriminator, took the center of those two voltages, and used that as the
> threshold. That would be similar to the above LPF, though it used peak
> positive and negative instead of the average the LPF would give, and the
> continuous data was better than "hand typed."
> 
> Besides selective fading, there is also the transmit and receive bandpass.
> This can result in the space tone having a lower amplitude.
> 
> I note also that my TU-170 uses several 2 pole BPFs in series. So far, my
> DSP TU has just a single 2 pole BPF for each tone (with a Q of 50).
> 
> So, lots of stuff to play with!
> 
> Harold
> 
> 
>> 
>> 
> 
> 
> -- 
> https://w6iwi.org
> ______________________________________________________________
> GreenKeys mailing list
> Home: http://mailman.qth.net/mailman/listinfo/greenkeys
> Help: http://mailman.qth.net/mmfaq.htm
> Post: mailto:GreenKeys at mailman.qth.net
> 
>>>> Jordan Spencer Cunningham's GreenKeys Search Tool: https://teletype.net/gksearch
>>>> 2002-to-present greenkeys archive: http://mailman.qth.net/pipermail/greenkeys/
>>>> 1998-to-2001 greenkeys archive: http://mailman.qth.net/archive/greenkeys/greenkeys.html
>>>> Randy Guttery's 2001-to-2009 GreenKeys Search Tool: http://comcents.com/tty/greenkeyssearch.html
> 
> This list hosted by: http://www.qsl.net
> Please help support this email list: http://www.qsl.net/donate.html
> Message delivered to kb8tq at n1k.org

  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.qth.net/pipermail/greenkeys/attachments/20220317/2e519d82/attachment.html>