[RTTY] RTTY crosshair pattern

[Kok Chen]

On Feb 3, 2013, at 12:21 PM, Lee Roberts wrote:

> Anyway, I modified my Kam+ several years ago (a 
> Kantronics approved mod) so that I can hook up the scope and get a cross-hair 
> pattern rather than using the LED panel on the front panel for tuning. But, it 
> always displays 2 ellipses on the scope that I can never get to be at a 
> perfect or even near perfect right angles. I don't know if it's possible to 
> get 2 straight lines instead of the ellipses.

Lee, there are three things involved when it comes to cross ellipse displays, the first is the center frequencies of the Mark and Space filters for the scope, the second is the bandwidth of the filters, and the third has to do with the phase shift relationship of those filters.

These three together, determines the appearance of the crossed ellipse.

To get a thin cross, the filters have to be very narrow and centered exactly at the Mark and Space tones.  I.e., the Mark channel of the oscilloscope cannot receive any space signal.

With hardware TNC and TU, the scope tap often comes from the Mark and Space demodulation filters, instead of a specialized circuit designed just for tuning.

The KAM Plus shows the scope taps in their schematics, and it is also on the PC board trace, they just did not bring the two wired out to the DB-25 connector pins.  The "mod" that you mention is simply wiring those two PC board traces to the DB-25 connector.  Like most cheap scope implementation, the tuning signals in the KAM Plus originates from the Mark and Space (switched capacitor) filters.

Recall that the KAM Plus was designed for the data bandwidths of Amtor and Pactor.  Unless you have modified the resistor values of the switched capacitor filters, they are wider than a barn door when it comes to RTTY data rates (that is why the KAM Plus are also so poor for RTTY reception).

As such, the Mark filter will let through a very substantial amount of the Space signal, and vice versa.  

So, the best you can get is an oval or ellipse (also fondly called bananas by RTTY guys :-).  I still call them crossed bananas informally.

Now, recall Lissajous patterns from high school Physics classes.  When you inject the same audio tone, into the X and Y axes of an oscilloscope -- you get a tilted oval.  This is in fact, the age old way of determining the phase shift between two signals.  I.e., if they are at 0 or 180 degree phase shift, you get a 45 degree line, and if you have 90 degree phase shift, you get a circle if the two signals have the same amplitude and a vertical or horizontal oval if the amplitudes are not the same.

So, if you have two ovals, one horizontal and one vertical, the only way to produce no tilt from the horizontal and vertical is if 1) the filters are perfectly centered on the two tones, and (2) there is 90 degree phase shift in the Mark filter when it is receiving a Space tone.  You can get the first condition with the KAM by choosing 170 Hz shift, but the second condition is far from possible.

You can actually adjust for different Mark and Space centers to get an orthogonal cross ellipse, but the KAM will then decode even worse than it normally decodes RTTY.

The short answer is that you can neither get a thin cross from the KAM Plus' scope output, but instead ovals, and furthermore, the ovals from the KAM Plus are not even at right angles to one another when you are perfectly tuned.

As some folks may recall, you can pretty easily get right angle ovals by feeding a little of the the X channel signal into the Y channel and vice versa, to cancel the unwanted Mark and Space leakage.  You need a couple of potentiometers and some op amp buffers (since you may need to negate the small canceling signal).  Be sure to buffer or the Mark and Space detectors in the KAM will see this mix too :-).  In the graphics software world, this is called coordinate transformation and is a very convenient way to get right angled ovals from filters that have no phase shift.

Some of you may also remember that DJ0OT (ex-F5FC) had published the circuit of a tuning meter that draws thin lines instead of ovals (the tuning meter itself, constructed with op-amps, is more complex than the KAM's RTTY demodulator, HI).  The way she achieved the thin lines is that, instead of using band pass filters, she had use notch filters.  I.e., instead of trying to extract Mark from a bandpass filter and send it to the X channel of a scope, she instead notches off the Space for the X channel.  Quite ingenious really, since a notch filter can really knock out the tone you don't want.  Claude still has the circuit and pictures here:

http://www.mydarc.de/dj0ot/discri.pdf
http://www.mydarc.de/dj0ot/

For green keys nuts, you can see a picture of her old RTTY station here

http://www.mydarc.de/dj0ot/F5FC.jpeg

73
Chen, W7AY