[RTTY] 200 hz filter on RTTY

[Kok Chen]

On May 31, 2010, at 11:37 PM, Bernd DC3HB wrote:

> 1) Do you know any software with MO/SO other than ritty?

cocoaModem has them.  See "Mode" buttons in Figure 10 here:

http://homepage.mac.com/chen/w7ay/cocoaModem/UsersManual/RTTYPage/DualRTTY/DualRTTY.html#Fig10

> 2) How do you decode a weak signal with the 500 Hz Filter if the weak
> signal is near a very strong one? The AGC in the receiver kills the weak
> signal.

You would use a narrow filter of course if there QRM close by, and forgo the extra 1 or 2 dB of sensitivity that you get from a matched filter.

Also, you might not be able to hear the weaker signal by ear, but it is there for the demodulator to decode if the weaker signal is within 80 dB or so of the stronger signal.

I always imagine the I.F. filter of a receiver to be a roofing filter for the rest of the receiving chain, including the sound card.  

As far as the sound card goes, the I.F. filter is there only to keep the sum of all signals in the passband from saturating the full scale of the sound card.  Rigs such as the K3 has a dynamic range of over 95 dB, and so do pretty much all the modern sound cards except for the cheapest ones.  The best sound cards (Asahi chip based) have 123 dB of dynamic range and even the E-MU 0404 that is widely used in the ham community has over 110 dB of dynamic range.  Even the 16 bit converter in the microHAM digiKeyer has 96 dB of dynamic range when I measured it.

DSP filters can have very high stop band rejection as long as the sound card is not clipped and it is especially easy to achieve with floating point computation.  Blackmann-Harris window based filters has ultimate rejection in excess of 95 dB, see here:

http://en.wikipedia.org/wiki/File:Window_function_(comparsion).png

So, 95 dB of dynamic range through the entire demodulation chain (from the antenna all the way to the Baudot decoder) is very practical today.

A selectively fading signal has to be about 10 dB to 15 dB above the noise floor to have good copy.  See VE3NEA's curves here http://www.dxatlas.com/RttyCompare/ .  Alex uses a 3 kHz noise bandwidth, so a 0 dB SNR in his curves correspond to +10 dB SNR for a 300 Hz passband.

This means that if you have two signals in the passband, and the dynamic range of the receiver chain is 95 dB, the weaker signal can be 80 dB to 85 dB below the louder one and still be decoded reliably, assuming that the gain of the receiving chain (including the sound card) is set up correctly.  At 6 dB per "S" unit, the weaker signal can therefore be 13 to 14 "S" units weaker than a second stronger signal in the passband and still print.

That being said,  IMD does not tell the complete story.   I have found that the receiving chain's harmonic distortion can be a bigger problem than its IMD when you have more than one digital signal in the passband.  With PSK31, the second harmonic a strong signal at 1 kHz offset could be clobbering a weaker signal at 2 kHz offset even before any problem from intermodulation distortion comes up.  The same is true with an RTTY signal with a tone pair at 1 kHz clobbering an RTTY signal with a 2 kHz tone pair when you open up the passband to 3 kHz.

Also, while modern sound cards and software filtering can have very good dynamic range, some rigs still have dynamic ranges that is worse than 95 dB.  Rob Sherwood's table is most useful when selecting rigs: http://www.sherweng.com/table.html .  Rigs could be the weakest link today.  You can also see from Sherwood's table that good dynamic range do not always correlate with price :-). 

73
Chen, W7AY