Re: How does QDX implement the raised cosine ramp in the FT8 spec? #qdx

Hans Summers

Hello Paul 

Yes. Constant amplitude. But my understanding of section 5, on pages 9 and 10, is that this is achieved by applying a Gaussian shape to the individual rectangular pulse making up symbol. The transmitted signal is the summation of these pulses. At the transition between symbols (audio tones) there is a summation of the preceding and succeeding tone, which are Gaussian shaped such that the summed amplitude is in fact constant.

So each symbol (tone) is a shaped pulse; but adding adjacent Gaussian smoothed pulses gives in effect a constant amplitude. It is also equivalent to a gradual frequency slide between the two tones. So there is no conflict with the "single tone FSK" theory, which is why we don't need a linear modulator and amplifier. 

And then on top of that, a raised cosine at start and end. 

That was my understanding but it's quite possible I misunderstood. 

Either way - QDX doesn't implement shaping at the start and end, but it does track the frequency slides so it minimizes trnamsitted bandwidth during the transmission itself. 

73 Hans G0UPL

-------- Original message --------
From: Paul Christensen <w9ac@...>
Date: Sun, Mar 27, 2022, 6:37 PM
Subject: Re: [QRPLabs] How does QDX implement the raised cosine ramp in the FT8 spec? #qdx

In FT8 Software.  From Joe Taylor’s QEX article:


“FT4 and FT8 waveforms have constant amplitude except at the very beginning and end of a transmission. FT8 signals are ramped up

gradually over a transition interval T/8, or 20 ms, at the beginning of the first sync symbol. The ramp function is a raised cosine.”


Paul, W9AC


From: <> On Behalf Of TomCircuit
Sent: Sunday, March 27, 2022 11:27 AM
Subject: Re: [QRPLabs] How does QDX implement the raised cosine ramp in the FT8 spec? #qdx


This is an interesting discussion to me. I work as an application engineer for a semiconductor company and define/support UHF ISM band mW output wireless sensor ICs. Most applications use FSK or GFSK to send short (~10ms duration) telegrams every minute or so. Smoothing the carrier envelope on/off edges wasn’t a concern as it wasn’t required to obtain government approval in these bands of operation. What also helped was that the antenna matching Q was reasonably high enough to prevent excessive splatter. Several years ago, S Korea added a challenging Occupied Bandwidth requirement to their RF regulations, and we had to redesign the IC to include circuitry to control the carrier envelope transitions. It added a surprising amount of complexity to the modulator, pre-driver, and driver stages to accomplish. In the QDX I suspect it would be even more challenging to implement because it is given no ‘warning’ as to when the carrier will need to be ramped up or down. 

If only the PC-side software would have the good sense to pass the frequency and amplitude information numerically via USB… :-)


Join to automatically receive all group messages.