Very interesting ideas.
On the subject of VFO's, one addition I am going to make at some point on my
rig is a Huff Puff stabiliser. You can see my version, and read much of what
has been written about the Huff Puff stabiliser technique since it was
developed by PA0SDB in 1973, visit my page
. A year or two ago I
spent a day in the Science Museum / Imperial College library in London (I am
an ex-IC student in fact!). I had all the RSGB RadCom's bought up from the
archives on a trolley, and went through and photocopied all the original
articles and follow up. RSGB have given me permission to reproduce them on
my website. All that remains is to find the time to scan them etc. In any
case, these early articles are of historical interest but in practice you
will want to read the more recent articles if you are intending to build a
stabiliser. Again, the components are simple logic IC's and should be
reasonably easy to source.
On the BITX20@..., you will see that
Hans Summers (www.hanssummers.com) has a binary counter made
of two 74HC4040s. This seems ideal to me. I have avoided it
from the basic design to keep the construction challenge low.
I personally use my lab frequency counter along with my
transceiver for frequency spotting. I keep the counter
separate because i found that unless they are well sheilded,
the counter tend to generate a lot of noise that back
couples from the VFO into the receiver front-end.
My counter uses one 74HC4060 and one 74HC4040, not two 74HC4040. The
74HC4060 is used for the timebase because of its onboard crystal oscillator.
The URL is: . All these
components should be easily obtainable anywhere. Here in the UK the
component cost is about ?2. The old CMOS 4060 and 4040 should work just as
well, but the maximum count frequency will be lower. At the BITX20's VFO
frequency, old CMOS implementation would be fine.
I am in frequent correspondence with Onno PA2OHH who designed the original
3-chip version of the counter , which
inspired my modified and even simpler version. We have been sharing ideas on
the counters, a nice collaboration. He is currently performing some very
interesting experiments with a 74HC390 dual decade counter instead of the
74HC4040. With the 74HC4040 the LED's indicate 64, 32, 16, 8, 4, 2, 1,
0.5KHz. You get good practice at reading binary. With a 74HC390 it is easy
to obtain a different sequence: 80, 40, 20, 10 and 8, 4, 2, 1. The LED's can
be arranged in two columns of 4, one to indicate 10's of KHz and one to
indicate KHz. The mental arithmetic is faster and easier, but the resolution
of the counter worsens from 0.5KHz to 1KHz.
Note though that the intensity of the least significant LED acts as a kind
of analogue frequency indication, which makes it possible to estimate the
frequency to better the resolution by a factor of 2. For example in my
design if the LED is half brightness, it is possible to estimate that the
frequency being measured is xx,xxx,250 Hz.
Another advantage of the 74HC390 version is that it does not require the
100-count detect gate (3 diodes + resistor), nor OR'ing this 100-reset
signal onto the reset pin (2 diodes + resistor). Thereby eliminating a
further 7 components from the design (5 diodes, 2 resistors).
I have installed my Mk1 simple 2-chip frequency counter in the front panel
of my BITX20 and marked the tuning dial 14.0, 14.1, 14.2, 14.3 and 14.35. My
readout is now accurate using the dial for 100KHz accuracy and the counter
to read 0 to 99.5KHz. I have not yet debugged my rig so I can't yet be sure
how much effect it has on the rig. There is an on/off switch for the counter
so that it could be used to read the frequency then disabled. At the present
time I hear hiss in my headphones and a faint tone at about 1KHz (the
frequency of the pulse-width-modulated LED's) when the volume is at full.
But the counter has no shielding yet. I will report further results as and
when I have time to progress the project.
Another interesting observation: In my 80/40m polyphase receiver
I have installed my
8-digit frequency counter
as well as a 24-hour clock
. The counter
and clock are both built in screened boxes made from PCB stock, and they
each have an inductor in series with their power supply. Both of them
pulse-width modulate the LED displays to eliminate LED series resistor. Duty
cycle is 1 in 6 if I remember correctly. Both the clock and the counter have
their own on/off switches because I was worried about digital noise getting
into the receiver. But during extensive use I have never found it necessary
to switch off the counter or the clock, since they appear to contribute
nothing at all to the receiver noise floor apart from possibly the
occasional and very faint birdie. But even this I have been unable to
conclusively prove to myself.
Use of low current LED's in the 2-chip simple frequency counter and lower
duty cycle (I use 1 : 64) helps noise, because it means the power supply
smoothing is much more effective (all other things being equal). I am
currently trying (with only limited success) to source a 16 or 32KHz crystal
(yes KHz not MHz) to replace the 4.096MHz crystal in the counter, which I
hope will reduce the power consumption even more. For the fun of it I am
going to build a miniature version using one of the tiny 6mm long
cylindrical 32KHz crystals and surface-mount 74HC4060 and 74HC390.
Surface-mount, but "ugly" not on a PCB. I hope it will be about the size of
a postage stamp, and a few mm thick, with current consumption < 1.5mA.
Rather an extreme project and not recommended. But I do definitely recommend
the counter design, use more real-world size components!
73 Hans G0UPL
