Re: Cw only

Jerry Gaffke
 

Michael's suggestion of the SotaBeams audio filter for CW (and SSB) looks good, I assume his one:
    https://www.sotabeams.co.uk/variable-bandwidth-filter-modules-ssb-cw/
Seems comparable to the NESCAF:
    http://www.newenglandqrp.org/wordpress/nescaf/
For a one-off, that's  good deal, scratch building would cost a week.

Farhan's suggestion of using a 1.8432 mhz crystals for a CW filter in parallel with the 12mhz SSB filter 
is also a possibility:  https://groups.io/g/BITX20/message/36041  
Given the low crystal frequency and the narrow bandwidth required, we aren't too worried about ripple.
Should be easy to reproduce, I'd try to get crystals from the same manufacturer and match frequencies.
The switched cap audio filters would be far more flexible, but the crystal filter would do a better job
of rejecting strong adjacent signals, and quite a bit cheaper.  Might require a separate tap on the
impedance matching transformers.

We might think of creating an auxiliary board as a kit for the uBitx with a CW filter of some sort
and a bunch of other stuff.  Perhaps good AGC (and manual RF gain) and an inline SWR meter. 
Plus an AD8307 for use as an S meter and rf probe and sweeperino.
Glue it to the back of the uBitx main board.

The Bitx20A from Pacific Antenna has done away with relays, should give fast QSK:
    http://qrpkits.com/bitx20a.html
They snuff out any transmitted signal into the receiver using a string of series and shunt NFET's.

More typically, a PIN diode scheme is used, perhaps using 1n4007's as PIN diodes.
See post    https://groups.io/g/BITX20/message/32109
and the thread around it for a discussion on using PIN diodes for a transmit/receive switch.
Also of interest in the thread is the discussion about needing AGC prior to the audio pre-amp,
as that pre-amp is a dynamic range bottleneck in the Bitx40v3 and likely in the uBitx as well.
In the post, I suggest rectifying some of the transmit RF to create a DC signal to bias the pin diodes.
Somewhere, perhaps in this forum, I later saw a circuit diagram that implemented something like this. 
Saved a link to that circuit but have now lost it, does anybody have a pointer to that circuit?

The band switching relays for the LPF's can stay, they won't be switching around much.
But I'd get rid of the two transmit-receive relays using silicon devices.

The rx ptt pop can be solved by Simon's NFET to shunt the audio to ground:  https://groups.io/g/BITX20/wiki/AudioMute
the tx carrier burst perhaps by Raj's series cap.

With all that, this could be a very nice CW rig.

Jerry, KE7ER



Jerry Gaffke wrote:

Good fast QSK might be a bit more difficult to implement, but certainly possible.
I'd prefer to go without the transmit-receive relays, examples of this for the Bitx40v3 have been presented

No, you don't need to double the price of the uBitx to use it for CW.

Jerry, KE7ER

On Thu, Dec 21, 2017 at 12:57 pm, Michael Babineau wrote:
The K1EL K16 keyer kit is only $20 and you can get a 1Khz / 300 Hz DSP audio filter module (pre-assembled) from SotaBeams for about $30 ...
that is only $50 total.  You won't end up Ten-Tec QSK but it will still make a pretty decent little CW rig for < $200 total , which is hard to beat. 

Join BITX20@groups.io to automatically receive all group messages.