[Elecraft] K3 AFSK TX Filter

[David Woolley]

Bottom quote below:

> In the case of the K3, the FSK mode uses DSP-generated IF frequencies in
> the 15 kHz range for the Tx, and simultaneously creates audio tones at

15kHz is audio frequency.

> In AFSK mode, the radio requires that you supply audio to a Mic
> connector or the Line In connector.  The AFSK Tx filter may optionally

That audio is really an IF of about 1kHz.  The same probably applies to 
the KX3.

The actual difference here isn't between audio and non-audio, it is 
between pure real and complex.  The algorithm that generates the IF 
level FSK signal within the DSP has access to both I and Q channels, so 
the image can be phased out when upconverting.  Although the PC may well 
use almost the same algorithm, it only has access to the I channel, 
which means that the DSP has to do what it does with audio, either 
generate the Q channel, using a Hilbert transform or equivalent, or 
filter out the image.

(Interestingly, whilst there is a big demand for I and Q outputs, there 
doesn't seem to be a corresponding one for I and Q inputs, although that 
is what one wants for efficient processing of data modes.)

-- 
David Woolley
Owner K2 06123


On 05/12/14 01:28, Lyle Johnson wrote:
> In the case of the K3, the FSK mode uses DSP-generated IF frequencies in
> the 15 kHz range for the Tx, and simultaneously creates audio tones at
> the selected pitch-pair.
>
> In AFSK mode, the radio requires that you supply audio to a Mic
> connector or the Line In connector.  The AFSK Tx filter may optionally
> be employed, but its use is not required and with later versions of DSP
> code is probably of marginal utility.
>
> In the case of the KX3, the tones are generated at baseband.
>
> 73,
>
> Lyle KK7P (who emphatically agrees these products are SDRs
>
>> On 04/12/14 18:20, Joe Subich, W4TV wrote:
>>
>>> The K3 does not generate AFSK internally. AFSK is only transmitted
>>> in response to external audio.
>>
>> But that is exactly how SDR architecture transmitters do generate FSK.
>> The only difference from what is popularly called AFSK is that the
>> tone starts as digital sine waves.
>>
>> The ones may be a bit higher pitch than one normally expects from
>> AFSK, but they are still audio tones.
>>
>> (As the K3 uses an I/Q structure, you could, theoretically, generate
>> the two tones as positive and negative tones at the same frequency,
>> but the problems with handling the DC crossover means that it is
>> unlikely that anyone would do that.)
>>
>> In case it is not clear, I consider both the K3 and KX3 to be SDR
>> designs.
>>
>
>