[GreenKeys] 170-Hz discriminator?

[Bob Camp]

Hi

Once upon a time I had a log book with a *lot* of data on the 
properties of 88 mHy toroids. Seems to be one of the many things that I 
have misplaced over the years. To the best of my recollection an 88 mHy 
toroid has a Q around 100 to 120 or so at 2 to 3 KHz. Q is slightly 
better at 2 to 3 Kcps ...

A Q of 100 should give you a bandwidth around 20 Hz at 2 KHz. That's 
plenty narrow enough for a 170 shift discriminator. A reasonable 
bandwidth might be in the +/- 60 Hz range. That would put the loaded Q 
down around 33 or so.

Since you are trying to make the bandwidth (and delay) of the two sides 
of the discriminator the same you actually resistively load one side 
more than the other to lower it's Q and broaden it's bandwidth. The 
other common approach was to pull a couple of turns off of one of the 
toroids to equalize the bandwidths with equal resistive loading on each 
side of the discriminator.

One thing that a lot of us miss is that the discriminator coils are 
loaded both by the driving resistance *and* by the DC load resistance 
on the other side of the detector diodes.

The TI rep was by the office this morning. Some of their low end DSP's 
appear to sell for less than $10 these days. I suspect that with some 
work you could make a pretty good FSK TU with one.

	Take Care!

		Bob Camp
		KB8TQ




On Sep 23, 2004, at 10:17 AM, AI2Q wrote:

> Looking over some old ARRL handbooks and RTTY magazine articles, I see 
> a
> number of TUs using the ubiquitous 88-mHy toroid filters that I knew 
> and
> loved as a youth. However, most circuits use 2125/2975 "tones," 
> although
> there are a few references here and there about putting in an extra
> capacitor switch on one filter to permit operation at 170-Hz shift, 
> too.
>
> However, for the life of me, I don't see how that's possible. I've 
> swept
> some of the single-toroid/single-cap 88-mH L-C filters that I have 
> here, and
> most are quite broad. I find it hard to believe that a simple resonant
> filter with a 300-Hz-plus skirt is going to discriminate between a 
> 2125-Hz
> tone and a 2295-Hz tone.
>
> I must be missing the point somewhere.
>
> (Not really relevant to this thought is that my application is to pick 
> off
> AFSK that my present TU generates and decode it for driving my local 
> loop. I
> need to do that because my homebrew TTL logic for driving the selector
> magnets on my Model 28 on Rcv is derived from a switched-capacitor 
> filter
> bank in an old Kantronics UTU, and the clocking of those filters is 
> turned
> off during Xmit. Otherwise, I'd pick off the AFSK from the Rcv string 
> in the
> switched-cap circuit when transmitting).
>
> Does anyone have a circuit for a 170-Hz audio 
> filter/detector/discriminator?
>
> Thanks in advance.
>
> Vy 73, AI2Q, Alex in Kennebunk, Maine
> http://users.adelphia.net/~alexmm/ai2q.htm
>
>  .-.-.
>
>
> _______________________________________________
> GreenKeys mailing list
> GreenKeys at mailman.qth.net
> http://mailman.qth.net/mailman/listinfo/greenkeys
>