using the BITX on RTTY

Charles Darley

Hi all

I was at my local radio club the other night and was discussing the
possibility of using the BITX on RTTY .. This prompted the following
responce from one of the members ...

"The elegant way for RTTY is to have a carrier, and this is shifted
for one of the tones. A small amount of capacity added across an
oscillator is the usual way, switched by the RTTY signal coming from
(say) the teleprinter contacts - or today a computer output from the
com port. This results in a purer outgoing signal than two tones into
the mike socket, as the amplitude of the tones is identical, which
does not happen when tones in the mike socket is used. (Check the
output power of the TX on mark, and then space, when using audio in
the mike. Not the same!!"

So this prompted me to look for information on the varicap ...

I came across this site :-

I have printed off all of the data and this will be another display
project to show members ( or have them build it and then discover )
as stated in the article "any diode is a varicap diode".

Hans I wonder if you would like to confirm one or two items which
are not clear to me...
1. that the volts into the 4060 are 12V
2. that the 14MHz xtal could be changed for one of the 10MHz xtal
already obtained ( through you ) for the project ...
3. could I / a student use ugly construction for the test project ???

This whole BITX project is enabling so many teaching side lines that
the students will not realise that they are learning so much.

73 all

Charles G4VSZ

Hans Summers <Hans.Summers@...>

Hans I wonder if you would like to confirm one or two items
which are not clear to me...
No problem! My favourite varicap is a standard 5mm red LED. Have a look at
my 30m QRSS beacon project
which used one for the frequency shifting to produce DFCW and
slow-hellscreiber modes. Far too many people get stuck ordering components
for a project, when they can't find the specified varicap. Varicaps are
components which seem to go in and out of fashion quite quickly. In most of
those cases, they could easily substitute with something else, or even just
a simple diode. Don't forget even in the BITX20 we're using a 36V zener
diode for the VFO fine tuning! A "real" varicap diode will have more
precisely defined and reproduceable capacitance limits but in the majority
of homebrew QRP circuits, that isn't very important. One exception is if you
want a large capacitance, e.g. the BB212 varicap which seems to be
increasingly hard to find. Even in this case, many times the circuit can
easily be redesigned for a lower capacitance range.

1. that the volts into the 4060 are 12V
Errr.... No. I used 12V on the "varicap" but only a +5V supply to the 4060.
The reason I say "err" is that according to the datasheet, the "typical"
maximum clock frequency of the old 4000-series CMOS version of this IC is
7MHz on a 5V supply (no "maximum" maximum clock frequency is indicated).
Ooops. But it did work Ok for me anyway despite being violation of the
datasheet parameters. At higher supply voltages the 4060 has typical 16MHz
at 10V and 24MHz at 15V. So you would be better off using +12V supply anyway
;-) Even better, use the modern 74HC4060, at 5V supply you're looking at
over 30MHz for the max count freq. But you'd need 5V supply for 74HC.

2. that the 14MHz xtal could be changed for one of the 10MHz
xtal already obtained ( through you ) for the project ...
Sure, why not. Any crystal should exhibit the same pullable effect with
parallel capacitance. Although some are more "pullable" than others and as
you might expect, in general you will find that the higher the frequency the
greater the shift (in absolute KHz terms). From my "crystal penning"
experiments and those of Dave WA4QAL there appears to be some evidence that
physically smaller crystals are more shiftable, e.g. we obtained more shift
on HC49 style crystals than the larger slab in an HC6 case. The same might
apply to pulling with capacitors, but I don't know. The 14MHz crystal I used
was the common HC49-cased variety.

While we're on the subject of crystals, if you want to move them by more
than a bit of capacitance will accomplish, you can get many KHz lower by
removing the case and painting the crystal with ink! See my page about this: My 1-valve CW tx for
80/40m ( contains two
"penned" crystals: 3.558 (originally 3.579) and 7.010 (originally 7.030).
See also the first modification I made to the 30m QRSS beacon I mentioned
above,, which was
to add a 10.140 crystal. I guess that this crystal must have been custom
manufactured at some point but it wasn't very accurate, 4KHz to high. A
single ink dot on its surface was enough to bring it down to 10.142 which
was about perfect for what I wanted.

And if you want to go the other way and increase the frequency of a crystal?
Grind it and grind it, then grind some more. For this you need a larger
crystal (e.g. FT-243 style), the HC49 is just too small and you wouldn't be
able to remount if afterwards. See for some interesting grinding
experiments recently undertaken by Dave WA4QAL.

3. could I / a student use ugly construction for the test
project ???
Yes definitely! I used a scrap of plain matrix board but "ugly" is fine too!

73 de Hans G0UPL


Hans, would it be possible to pull the crystals in a filter using the LED or Zenor Varicap diodes? 

-Justin N2TOH 

Tom, wb6b

I noticed in one of the alternate versions of the uBITX software the WSPR mode was implemented by sending the frequency shifts to the SI5351 chip registers. Have not looked into if the SI5351 chip, and the related uBITX software, can reliably change the frequency at the rate needed for RTTY, but if so, adding a RTTY mode to the software could shift the frequency digitally using the CW key jack to make the mark/space frequency shifts. 

Bo Barry

Am I missing something? If you don't have a real RTTY machine why not use the free fldigi program like I do?
Missing my old Model 15 machine and CW rig with a varicap.


A 45 baud FSK RTTY signal changes frequency each 22mS.  The shift for RTTY is 170Hz. From my limited experience (not consulting the datasheet) I expect the si5351 would handle this easily.   The trick would be getting the timing of the changes right.   A software timer such as the Arduino delay()  may not work as it is pretty inaccurate with small arguments.   A hardware timer that generates interrupts should do the trick.  It would be easy to try. 

73 Paul VK3HN. 

John (vk2eta)

Hello Tim,

At 45.45 baud and 170 Hz shift it would not be an issue to do it all by software to meet the frequency and timing accuracy requirements.

Writing the frequency for the next RTTY bit to be sent can be triggered by a single byte write over I2C in less than 1/10th of a milisecond as only the last register write activates the clock change.

The CW keyer function of KD8CEC's software could be re-cast for sending canned RTTY messages/beacons.

73, John (VK2ETA)


Good morning,

Though not specifically uBitx related, the RTTY timing issues has been discussed ad naseum on the RTTY group which has recently rehomed to (RTTY) and IIRC the issues with timing are based largely in windows as it is not a real time OS. If you are like me, these discussions are, IMHO, very interesting and can be a time sink.

Arduino can and does a very nice job of handling RTTY with Andy, K0SM work creating TinyFSK, see which incidentally also functions as a CW keyer ala hat tip to Anthony, K3NG and his wonderful arduino CW keyer. Anthony maintains a group (radioartisan) on I do use David, G3YYD wonderful 2Tone software which has implementations for tinyFSK as your transmit source built in; see In a nutshell, I could not live without 2Tone when I play RTTY as it is truly fantastic and I routinely run several 2Tone receive windows along side Alex, VE3NEA Gritty software (print only - no xmit); see

The key is ensuring you grab the correct timing packages from Paul over at PJRC (Teensy), see

All my very best,

Mike N4EEV