Remco.org

Seductive serendipity / Verleidende serendipiteit

April 3rd, 2016

Yaesu FT-780R revitalisation

Work in progress… read on …

A priori: As always, click on images to enlarge in new tabs.

Introduction.
I worked at the Dutch Radio Communications Agency and periodically
administrative obsolete equipment was offered to staff members before it was destroyed.

Nowadays, due to governance issues this seems impossible … anyway …

The procedure was co-workers could subscribe to a list. After a while
you were informed whether you wanted to buy the item for a scrap price.

Around 1998 I got my Yaesu FT-780R for around 10 Hfl (ca. 5 Euro nowadays).

This FT-780R was used in our monitoring station (NERA) ‘to enforce amateur satellites’
(I was told). I was also told this FT-780R was ‘custom modified’ so that it was not able
to transmit, in order not to damage other sensitive monitoring/receiving equipment.

After all, NERA was a monitoring station, not a transmitter site ! ; -)

When I got this FT-780R the receiver worked okay, but when you pressed PTT
the processor crashed, resulting in ’8888888 88′ on the display.

For whatever reason I left this FT-780R in a box for more than 15 years . . . until recently.

Below the bottom cover of my FT-780R is depicted.
I think not many radio amateurs use equipment used and owned by their enforcement agencies ; -)


Rationale.

A few months ago some friends in my neighbourhood decided to build a linear transponder using
2320.7 MHz in and 432.7 MHz out, BW = 15 kHz. I own two IC-402′s but these lack crystals for 432.7 MHz.
Despite I have a FT-857 I thought it would be a nice idea to have a dedicated rig for this transponder
in conjunction with a 2320 <–> 432 MHz transverter.

So… I fetched the old FT-780R from my storage box and decided to ‘remodify’ it.

First I looked up the circuit diagrams on the internet in order to investigate these ‘secret
non transmit’ modifications. I found a ‘user manual’ PDF, but circuit diagrams were split.

I ‘reassembled’ the circuit diagrams into one piece with a pair of scissors and took
pictures of the results as depicted below.


‘Remodification = repairment’

My assumption this FT-780R was modified in order NOT to transmit seemed valid at first glance.
My connotation of the word ‘modification’ involves (some degree of) reversibility.

Along the PTT line an ‘extra wire’ was hooked up to the processor board.
Removing this wire eliminated crashing after PTT.

RF output was absent but . . .  I could hear myself on a nearby receiver.  Promising!

Optimistically I started to inspect the final stage, consisting of a Mitsubishi M57716 module.
The relevant part of the circuit diagram is depicted below.

In/nearby the final amplifier stage I noticed three ‘issues’ (refer to right picture above):

1. The antenna relay did not switch per PTT because the ‘RL’ wire was dismantled.
2. Power supply leads of the last amplifier stage were removed both inside and outside.
3. Q2 (2SD235Y) was missing (??) and bridged so PO CONT is forced to 13.8V (??)

Issues #1 and #2 were solved as depicted below.


Issues #1 (left) and                                  #2 (right) solved.

Issue #3 is not a real issue concerning output. It  eliminates the function of the LO/HI power
button on the front. Of course I hadn’t a 2SD325Y but inserted a BD139 in the small pertinax board
next to the 7808 (Q1).

Anyway,  issue #3 is a very queer ‘modification’ in order NOT to transmit . . . (?)

After solving these ‘issues’ I pressed PTT in FM mode . . . .  NO output.
Perhaps the RF module was damaged or received no drive?

Indeed, I measured no drive, so . . .  further investigation was necessary.

The driver for the M57716 resides inside the PLL unit.
Relevant part of the circuit diagram is depicted belowt.


.                                           FT-780R M57716 driver stage.

First inspection of the driver stage didn’t reveal something strange.
Relevant power supply voltages (13.8V and TX 8 Volt) were there, but no drive output.

After careful inspection I couldn’t believe my eyes . . . .  Q05 (2SC2026) was ‘missing’ !! ?

Remember my connotation for the word ‘modification’ ?
For me a ‘modification’ owns a certain degree of reversibility.
In my perspective one of my ex colleagues from the technical department stripped
Q05 from the PCB, and very likely landed in the waste bin !!

I know lots of ex colleagues read and enjoy this blog.
So. . . when you read this and it was you, or you know who it was, contact me?

Being ‘in full swing’ I dismounted the PLL unit for inspection and insert a new Q05.

Below pictures of the ‘missing’ Q05 and dismantling of the PLL unit are presented.

I could have soldered a new Q05 on the top side of the PCB but I wanted a ‘clean repair’.
Of course I hadn’t a 2SC2026 so I chose a good old BFR90 instead. And old it is, 41 years !
Below pictures of the bottom side of the PLL unit are presented. I reckon very few people have seen this side ; -)

After reassembling the PLL unit I measured 5.5Vpp RF @432 MHz over 46.4Ω with a decoupled OA91 germanium diode.
This means corrected around 5.8Vpp, resulting in (5.8/√2)² / 46.4 = 362 mW drive (which is too much btw).

I reconnected the drive cable to the M57716 unit and gave PTT in FM mode . . . . NO output : -(

Thus, very likely also the M57716 module is defect! See pictures below.

At this moment a decision had to be made to replace the M57716 module. In conjunction with
a 2320 MHz transverter replacement is not really necessary. I can route the drive signal from the
input (pin1, right) to the output (pin5, left) of the module with a small coax.

On the other hand, it is elegant to restore the FT-780R for standalone work. So, I ordered a M57716 from Ebay.

Awaiting its delivery . . . more to come, stay tuned!

 

 

March 20th, 2016

Bulgarian yoghurt on 23cm

In 1992 I built a 23cm transverter consisting of two units:

1. UEK3 clone (we had the PCB layouts then ; -), i.e. 1152 MHz LO with 1296 –> 144 MHz RX converter
using a MGF1302 GaAs fet preamp and CF300 mixer.

2. DJ9HO (sk) ‘UHF Unterlage’ or DD9DU style 144 –> 1296 MHz TX converter (output 500 mW @23cm).

Below pictures of my homebrew units are depicted (click on images to enlarge in new tabs)


UEK3 23cm RX converter (by PA3FYM)       23cm DD9DU ‘UHF Unterlage’ TX converter (by PA3FYM)

This transverter served and worked well for some years with a Mitsubishi M57762 module, delivering ca. 17W out on 23cm.

However, when I moved to Groningen in 1997 the transverter landed in a box and remained there for almost 20 years.

Recently I was philosophizing about a long term project: building a 23cm EME station.
I held the transverter in my hands and thought . . . isn’t there something more state of the art, after ca. 25 years passed by?

Let’s face it  . . . over the past 20 years the world changed. GSM, DCS1800, UMTS and now LTE/4G is common practice.
These applications introduced new lines of RF technologies ‘around 1 GHz’, and here I’m standing with old skool gear in my hands?

In December 2014 Arie PAoEZ (sk) told me he bought a ‘Bulgarian 23cm transverter’.
When he told me I didn’t took much notice because I was more concerned about his health.

Around two months ago my friend Hans PE1CKK confessed he bought the same Bulgarian 23cm transverter and told me:
“Remco, quit the old stuff we used to build long times ago. It’s a nuisance, just buy this Bulgarian transverter and focus your
efforts and attention on other stuff.” <– period.

Google led me to the 23cm transverter from LZ5HP. Just one box with everything in it!
RF output on 23cm is around 2W. LO is generated by a PLL locked oscillatorwith several LO frequencies
and . . .  repeater TX shift (-28 or -6 MHz) !

After pleasant emails with Hristiyan LZ5HP I payed 164€ via Paypal and
received the 23cm transverter after three days by mail (with track & trace).

Below a close up picture of the transverter is depicted (click on image to enlarge in a new tab <- worthwhile! ; -)


LZ5HP (sg-labs.com) 144 <–> 1296 MHz transverter.

The transverter is versatile and works fine. After five (5) minutes you’re QRV on 23 !

At this moment of writing the current transverter version is v2.3.
It has the possibility of injecting an external 10 MHz reference signal instead of the internal 26 MHz TCXO.
I have tested this with my 10 MHz Rubidium standard and it works flawlessly.

 

January 8th, 2016

PS-57 balloon tracking

My homebrew WSPR receiver is placed at a (for The Netherlands) quiet location
in a nearby forrest and connected to a ca. 100m long open ‘bent’ Beverage.

The performance of the setup is quite remarkable and processes every electron ‘passing by’.
It delivered me a first place in the world wide WSPR challenge one day.

In the forrest I received WSPR traces from VK3YT and first thought his data was
somewhat mangled. His locators were in the middle of the Indian Ocean (?)

Bob ZL1RS noticed I was one of the few EU stations receiving VK3YT’s traces
and sometimes the only station in the northern hemisphere copying ‘him’.

It appeared that ‘VK3YT’ was/is a balloon (!!?) with designator PS-57 and
transmits WSPR and JT9 packets at least twice every hour at xx.00 and xx.30.

The output power of the balloon is only 25 mW,
making his traces almost 460000 km/Watt (!)

During an email conversation Bob seduced me to try to receive PS-57
with JT9 too. Balloon telemetry is embedded inside these JT9 packets.

I needed to install a special version of WSJT-X, modified to
upload JT9 telemetry packets to the tracker.habhub.org site.

Yesterday afternoon I went to the reception site and installed the
necessary software to process PS-57 JT9 balloon packets.

After 72 minutes I received my first PS-57 JT9 packet (!)
This was the first JT9 decode in my life! One hour later
the receiver decoded another one. See below (click to enlarge in new tabs).

Investigation revealed I also received one PS-58 WSPR trace.

Although this southern hemisphere balloon hunting from the northern hemisphere is
considered to be ‘notoriously difficult’, for me it is ‘by-catch’.

However, I let the contraption also capture balloon JT9 packets
hoping it may be of use to the world wide HF balloon tracking community.

December 14th, 2015

SSB phase method 30m WSPR receiver

In order to improve the SNR performance of my previous WSPR receivers/grabbers
I decided not to reinvent the wheel and used Onno’s PA2OHH design with some tweaks.

By Onno’s knowledge I was the first to use his SSB design. Why …. ?

Design
It’s a direct conversion (DC) receiver with LO at half frequency using subharmonic or Polyakov
mixers. The unwanted lower sideband (LSB) is suppressed using the phase shift method.
Theoretically this should increase SNR with +3dB (<– a difference between a ‘yes’ or ‘no’ WSPR decode!)

Of course you also need +3dB more components in the detector ; -)
Fortunately this design uses common junkbox components.

To improve the audio response (read: selectivity) I deployed some filtering
after the ‘adder’ <– the two transistors adding the ‘I & Q signals’.

Although it is stated that additional filtering is not necessary for WSPR, I adhere the credo
rubbish in = rubbish out. Furthermore, the receiver is also used for QRSS reception.

Filtering is done by a band pass filter with Q = 10, Fc = 1500 Hz and G = 100 (at Fc),
followed by a gyrator also with Fc around 1500 Hz.

Why a gyrator? Well, it’s cool to say that you’ve a receiver with a gyrator ; -)

Seriously, the gyrator was also inserted with another feature in mind: it is designed as
emitter follower, having a low output impedance (Z). Low Z out reduces hum and/or noise.

A gyrator simulates a parallel LC circuit with f = 1 / (2π√LC) as depicted below.

Capacitance multiplier
Lots of (home brew) power supplies provide stable and accurate output voltages.
However, there is something very important to consider, namely noise.

E.g. very popular 78XX voltage regulators are not bad, but they are noisy.
A very simple and effective trick to reduce power supply noise and hum is to insert a
capacitance multiplier, consisting of only three components (see below).


Capacitance multiplier.

The transistor ‘isolates’ the receiver from the power supply and its base
capacitor value (CFilter) is multiplied by its current amplification factor (hfe or β).
In this way very large capacitors can be created, resulting in effective noise damping.

Receiver circuit diagram
The circuit diagram of my receiver is shown below (left), as well as the ‘finished’ prototype (right).
(click on images to enlarge in a new tab. Big pictures! ; -)

Note: lots component tolerance is allowed, except for the phase networks.
E.g. 511Ω also may be 560 or 470Ω, 442K may be 390 or 470K, etc.
It depends how diverse and large your junkbox is : -)

Initially I used a NE5532 low noise dual opamp as adder, but for whatever reason it kept oscillating.
To save time I went back to the original design with two transistors.

Some measurements
Injecting -47 dBm (1mV) 10.140200 MHz into the receiver delivers a ca. 3 Vpp
beautiful non distorted 1500 Hz sine wave on the oscilloscope.

So, overall gain of the receiver @10.140200 MHz is 20*log(3/0.001) = ca. 70 dB
(assuming 50Ω RF in and 50Ω audio out).

Note: 10.140200 – 10.138700 = 1500 Hz and 10.140200 MHz is the middle of the WSPR band.

Injecting -47 dBm 10.137200 MHz (thus ‘-1500 Hz’, i.e. the lower sideband) after adjustment revealed a
noisy (estimated) 15 mVpp, making the overall lower side band suppression 20*log(3000/15) = ca. 46 dB (!)

Receiver circuit simplicity considered (and ‘standard’ junkbox components) this performance is remarkable !

Audio response / selectivity
With Audacity frequency response was recorded during night (left picture below).
In the spectrogram below (right), recorded around 2230 utc, sensitivity around 10140200 Hz is clearly visible.
(click on images to enlarge in new tabs)


Audio response SSB 30m WSPR receiver.     (Noise) spectrogram of 10.138700 – 10.144500 MHz

As can be derived from the left picture, ‘audio power’ is around -6dB @1000 Hz and @2200 Hz.
I tried to find out whether the Audacity shows ‘voltage’ or ‘power’ dB units … with no success.

Anyway, the audio response is quite sharp and centered around 1500 Hz, the middle of the WSPR band.

Despite its simplicity, this receiver performs very well and competes easy with more complex
and expensive colleagues.

November 4th, 2015

Experimental 30m QRSS grabber (2)

A priori: when operational, my 30m grabber can be seen here (<- click to open in a new tab)

Introduction.
I decided to write a new post, otherwise the previous posts become too long,
knowing that the average reader is too impatient and wants his information quick! ; -)

For the context of this post, please read this and this post first.

During the last days I tried to improve my 30m QRSS receiving contraption in
several stages. The estimated overall SNR improvement from the first setup
is around +32 dB, consisting of the following measures:

1. changed 5m of wire to 7.5m wire + central heating system as counter poise (+12 dB)

2. changed laptop power supply + isolated antenna coupling loop + RFC’s  (+14 dB)

3. isolated audio + DCTL antenna (+6 dB)

More modifications.
Yesterday Peter PA3EXL gifted me a 10.140 MHz crystal from an old (1978) CB-radio.
I changed the input circuit of the receiver by replacing the air wound coil + trimmer
with a modified small adjustable inductor.

I found this inductor at my radio club. It was rated 4 – 8 uH. The metal can was temporarily
removed. Using a magnifier the 22 original windings were reduced to 16 windings with
a tap at 3 windings. With the remaining wire I wound a 2 windings isolated input coupling
loop at the ‘cold’ side of the primary coil.

Measured inductance of the primary coil was ca. 3.4 uH. With 82 pF parallel ca. 3 Vpp
was obtained with -47 dBm 10.140200 MHz injection (see also previous post).
More than 1 Vpp improvement! Of course this is due to the higher Q of the new input circuit.
The previous input circuit needed around 200 pF for resonance at 10 MHz.

Subsequently the 10.140 MHz crystal was mounted and with a series trimmer trimmed
tuned tothe desired pass band. The resonance peak looked sharp and the optimal series
capacitance  was measured 21.7 pF. The trimmer was replaced by a fixed 22 pF capacitor.
With -47 dBm reference signal I measured ca. 2 Vpp. A loss of 1 Vpp, due to the crystal and series cap.

A picture of the current setup is depicted below (click to enlarge in a new tab).

Microphonics (?)
The receiver audio timbre was totally different. Different in a sense it was more quiet
(i.e. less noise) but also sharp high tones were audible. Just like in my young days when building my
first crystal receiver. So, could it be possible that I may have some AM detection?

Perhaps it’s still a good idea to add the balance potmeter in the Polyakov mixer?
Well, I added a 100 Ω potmeter and it makes no (measurable) difference.

While trimming the balance potmeter and listening to the receiver audio I initially couldn’t believe my ears.
With the antenna connected and tapping with the screwdriver on the PCB board I could hear ticks!

I went to the kitchen, got the kitchen towel roll and shouted almost my lungs out
(which is VERY loud! ; -), through the roll over the receiver.

The recording (somewhat filtered and amplified) can be downloaded here.

Besides shouting ‘testing 1,2,3′ and ticking you hear two bursts, but also IK3NWX/B on 10.1373 MHz.
I believed attenuating that significantly with the 10.140 MHz crystal?

Apparently not enough . . .

Perhaps the overall SNR performance is too good and the receiver is therefore susceptible
to microphonics? The only ‘mechanical part’ in the receiver is the crystal in the oscillator block.

Btw, if you read this and use a 5.0688 MHz oscillator block or crystal too, try if you detect microphonics?

Results.
After a while I was able to identify GM4GKH’s traces with the 7.5m wire + central heating counterpoise.

With the DCTL antenna these traces became more clear, however I also had around 6 dB less audio
and got the feeling that other signals were weaker. With crystal modification and DCTL antenna MIC
volume settings now have to be MIC boost = 20 dB and slider around 60 to have around 3dB ‘WSPR noise’.
If I may believe the dB scale in Windows, the overall noise reduction is also confirmed by audio volume settings.

The current result is depicted below left, judge for yourself.
Below right is a LA5GOA ‘benchmark lopshot’ in the same time frame.
Warning: The more you watch these pictures the more you see which may not be there! ; -)

That evening I copied several US stations and saw my amount of unique 30m WSPR spots/24hr increasing
even more. Two days ago I was on par with LA5GOA/RX2, yesterday before the crystal mod around +10.
At this moment of writing (1200 UTC 5 Nov 2015) the difference is around +20 uniques (!)
With this number I seem on par with LA9JO.

My next goal is PI4THT, ON7KB or ON7KO , at this moment of writing I have averaged -15 uniques.
Other benchmarks are GM4SFW or DK6UG having +40 uniques. I reckon for this result I have to
move the receiving setup into a more quiet environment.

My overall conclusion after a few days of experimenting and measuring: this super simple receiver,
together with all noise reducing measures, in a suburban environment with lots of noise sources,
is a good performer and is competing in the EU top of 30m WSPR receivers.

 

 

 

October 31st, 2015

Experimental 30m QRSS grabber

A priori.
When operational, my experimental 30m QRSS grabber can be seen here. (<– click to open in new tab)

Introduction.
The receiver from my previous post can be used for other weak signal modes like QRSS.
QRSS is transmitting information at very low speeds. At the receiver side this information
can be integrated for long periods, increasing (weak) signal to noise ratio (SNR) significantly.

It’s a public secret that QRO (high power) guys use low power (QRP) techniques
to optimise their top (contest) stations. Besides having sufficient output power it is VERY
important you’re also able to receive low power.

The quest therefore is to improve the RX SNR of your (contest) station.
This is where WSPRP/QRSS comes in.

QRSS grabber?
In order to experiment with the super simple DSB subharmonic receiver I installed
a grabber. A grabber is a piece of software analyzing the audiospectrum using Fourier
transform techniques (FFT).
This allows you to visualize the weak signals because you can’t hear them.

Onno PA2OHH wrote LOPORA (LOw POwer RAdio) grabber software in Python.
After installing python 2.7 and some fiddling I managed to get it running.

First results.
While eagerly awaiting the first ‘lopshot’, results were disappointing. Besides some weak
WSPR signals I hardly couldn’t see anything with my ‘quick & dirty’ 30m antenna
consisting of 5m wire running through a window into the garden around 2m above ground.

Because I live in a very noisy environment, my first action was to minimise noise from
the receiving contraption itself. The receiver is connected to a SMPS (I know, not ideal)
and a computer. All their (ground!) connections were fed through 6 hole RFC’s.

Subsequently I decided to lengthen the antenna wire to around 7.5m (1/4λ on 30m)
and use my central heating system as counterpoise.

These simple actions resulted in a dramatic SNR improvement!

It almost resembled my first experience listening to Beverages on 160m. At first
glance you think your receiver is broke because you think you don’t hear anything.

Below is the difference between the 5m wire and the temporary 7.5m wire + counterpoise.
(click on image to enlarge in a new tab)

Believing the SNR algorithms of WSPR, SNR improvements were around +12 dB (!!)
Although the receiver now sounds very quiet (I still here my antenna btw), my amount of WSPR
spots increased flabbergastingly and am now spotting a new league of WSPRers.

I also had to fiddle with the audio level, FFT settings, contrast and brightness levels in LOPORA.
The result was opening of a new 30m QRSS world. When a (WSPR) signal appears very bright,
its SNR is mostly around -6 dB. Most signals are between -20 – 25 dB SNR, or even lower.

Future improvements.
My benchmark is the 30m grabber of Steen Erik LA5GOA. Click on this link and see why (of course it depends
on the time of the day, try between 11 – 18 UTC).

Apparently Steen Erik lives in a very quiet environment and must have a good take off, also due to the
nearby sea (salt water!). It’s almost incredible what he’s able to receive with his PA2OHH style
DC receiver (I reckon also with a subharmonic mixer) with ‘own adjustments’.
I mailed him and learned he uses the same setup as Joachim, or vice versa.
LA5GOA’s antenna is a dipole directed E/W.
Below is a picture of Steen Eriks receiver (click to enlarge in new tab).

Less receiver noise.
One of the first things I’ve to do is decreasing the intrinsic receiver noise. Thus, like Onno did,
surpress the unwanted lower sideband. Theoretically this results in +3dB SNR improvement.

Secondly, increase the selectivity using a 10.140 MHz crystal as band pass filter.
Onno measured 800 Hz crystal filter band width. Bandwidth of the usable audio now is around 7 kHz.
Narrowing this to 800 Hz theoretically increases SNR with 10log(7000/800) = ca. +9 dB.
Btw, my WSPR/QRSS audio is around 4.5 kHz due to the LO frequency of my receiver.

In other words, if these two measures are carried out another +10 – 12 dB SNR improvement is possible!

Increasing frequency stability.
The receiver now lies open on the living room table without measures to stabilize it.
E.g. it’s not temperature compensated and I notice around 2-3 Hz/°C frequency drift.

When one of my cats lies next to the receiver (for whatever reason she wants to) LO frequency goes up,
and when she leaves LO frequency goes down ; -)

Antenna improvement.
At this moment of writing 7.5m wire with the central heating system counterpoise is used.
I did not measure the antenna impedance, but from 40m I know that such antennas are noisy.
Also the radiation pattern is lousy due to its low height above ground. It’s looking up to the clouds (NVIS).

Building a (vertical) deltaloop introduces two assets: a) the antenna is a closed loop <– less noise,
b) the take off angle is relatively low <– less interference from (strong) nearby signals and good for DX.

This may result in at least additional +3dB, but more likely +6 dB SNR improvement.

Receiver QTH.
Last, but not least, try to look for a nearby quiet place to install the receiver. I live in a busy city
with lots of interference like Power Line Communications (PLC), LED-lights, plasma screens, etc.

When the receiver is installed in a quiet environment +15 dB SNR improvement will be a (very)
conservative estimation. It might be an idea to use Beverages (or BOGs) there, but a preamplifier
is inevitable then.

Let’s wait and see . . .?

Update: I made a DCTL antenna for 30m but it initially seemed no success.
That is, no SNR improvement was visible, only lower signal levels.

I discovered that the power supply of my sampling laptop generated some noise.
Replacing the power supply with another one resulted in less noise (around -6 ‘WSPR’ dB).

Antenna input coupling was changed to an isolated coupling loop, just like Onno did.
I still could discern antenna noise and had to increase the ‘audio input slider’ in Windows.

The microphone audio input is used and is now 100% (without ‘MIC boost’ or AGC),
resulting in around 4 dB ‘WSPR noise’.

Below the temporary input coupling loop is depicted. I experienced less audible noise when
the ‘hot side’ of the antenna (yellow clamp, green clamp is GND) is connected to the
‘cold side’ of  the coupling loop, i.e. that side which is more near to the cold side of the input coil.

The other way around more noise was audible. With my temporary antenna I tend to believe:
“Less audible noise = better SNR” ;  -)

Before these two modifications (power supply & coupling loop) I had +16 – 20 ‘WSPR’ dB noise with
100% MIC volume. Above on the right a ‘lopshot’ showing more and more signals.

If I may believe the WSPR SNR algorithm my SNR further improved with 12 – 16 dB so the
(sub) total SNR improvement since the grabber is up amounts 24 – 28 dB (!?)

That evening I was one of the few EU stations copying ‘early’ US stations and was ‘competing’
with some EU WSPR ‘big guns’ on receiving several US and Asian 30m stations.

Promising? Yes and no. For example, I discovered PI4THT (Twente WebSDR using a Miniwhip
antenna) was able to receive the same stations with sometimes 20 dB higher SNR’s (!!)

The next day I took some additional  measures:

1. Tried the experimental DCTL again.

2a. Connect both L + R channel of the soundcard to the receiver as I am not sure whether WSPR
and LOPORA sample ‘in stereo’. If this is the case, then on one channel only (audio) noise is present
which adds to the so called quantization noise of the soundcard.

2b. Isolate the audio path with a 1:1 audio transformer.

The results are visualized below (click on images to enlarge in new tabs).


Measure 1. Switch between wire and DCTL.  Measure 2. Audio -> mono and isolate audio with xfmr.

In the above left picture Hell Schreiber traces of GM4GKH IO77WL become visible with the DCTL.
Apparently the overall SNR of my receiving contraption was not good enough when I tried the DCTL
the first time?

Reconnecting the 7.5m wire with counterpoise delivers more signal but results in lower SNR,
to such an extend that GM4GKH’s Hell traces disappear.  These are the ‘highlighted’ parts in the spectrum,
in which the signal of my GPS locked signal generator @10.140000 MHz are also visible.
In this snapshot the reference signal may look wobbly,  however this is the receiver LO, not my signal generator!

The influence of 2a + 2b may be seen at first glance. The right picture above looks darker (click to enlarge).
Audio settings between the left and right picture were equal.

Judge for yourself, but I reckon the overall SNR improvement of 1. and 2. combined is at least around 6 dB.
I.e. from seeing nothing (no GM4GKH with 7.5m wire) into seeing something (+ 3dB)
and being able to identify it (another 3dB) = 6 dB.

It could be that the difference in radiation patterns of the wire + counterpoise vs. the DCTL are
responsible or perhaps propagation. However, after these two measures I’m able to see GW4GKH’s traces
and other signals still look ok.

Therefore I estimate the total SNR improvement since the grabber is up to 6 + 24 – 28 = +30 – 32 dB !!

Bear in mind this is still the DSB receiver, i.e. no 10.14 MHz filter crystal and no Weaver SSB demodulator.

Another interesting fact is that at this moment of writing I am six unique WSPR spots
ahead of my benchmark (LA5GOA/RX2) in the last 24 hours.
</update>

 

 

 

 

October 27th, 2015

Junkbox 30m WSPR receiver

Introduction.
WSPR (pronounced as ‘whisper’) stands for Weak Signal Propagation Report and
is invented by Nobel laureate Joe Taylor K1JT.

WSPR algorithms are able to look ca. 30 dB deep into the noise. The result is that with
very simple and low power equipment world wide communication is possible.

Having built a standalone WSPR beacon with an Arduino and AD9850, I thought
it was time to build a simple 30m WSPR receiver.

The following isn’t Nobel prize worthy nor state of the art, but can be considered as an example of how
lots of fun is realized with a minimalistic approach and simple components.

Minimalistic approach.
Google found some simple 30m receivers. However, as fas as I could ascertain the ’30m root design’
is from Onno PA2OHH using a subharmonic mixer, also known as the Polyakov or ‘Russian’ mixer.

A mixer is actually nothing more (or less) than switching a given signal, in this case the received 10 MHz signal,
at a certain rate. This rate is called the ‘local oscillator’ (LO). The (mathematical) result in an ideal world
is that the LO signal is nulled out and two side bands mirrored around the (disappeared) LO signal appear.

Although being not super ideal, the elegance of the Polyakov mixer is that switching occurs at
both the ‘positive’ and ‘negative’ peak of the LO (sine) signal, thus having two ‘switch opportunities’
during one sine period.

In practice this means the LO frequency can be halved, which is advantageous in a direct conversion receiver approach.

I happened to have two 5.0688 MHz block oscillators in my junkbox and went for a similar approach as Joachim PA1GSJ
but did not include the 10.140 MHz ‘band pass’ crystal as I didn’t found one in my junkbox.

It is stated that subharmonic mixers do not work properly with square wave forms because
there are no clear ‘switching points’ for the diodes. As I hadn’t a 5.0688 MHz ‘filter’ crystal available,
reshaping the square wave of the block oscillator was done with one RC-network.
The result is a ‘triangularish’ wave form with supposedly enough discernable thresholds for the diodes.

The receiver was build with junkbox material and the circuit diagram and prototype are depicted below
(click on images to enlarge in a new tab).


Figure 1. Circuit diagram                            Figure 2. One hour later ; -)

Circuit description.
The circuit diagram in fig. 1 is straight forward.
L1 provides some selectivity and a ‘Josti Kit’ microphone amplifier boosts the received
10 MHz signal, which it is fed to the Polyakov mixer, consisting of two anti parallel diodes.

I did not use a potmeter to ‘equalize’ the diodes. The output of the mixer is fed through a RC filter (2k15/2n7),
selecting the audio spectrum (f-3dB = 1/(2πRC). This is fed into another Josti Kit amplifier with
another RC low pass filter. The resulting audio level is more than enough for standard (internal) PC soundcards.

The exact LO frequency was determined with a GPS locked signal generator by determining
the zero beat frequency. With the oscillator block used this appeared to be 10.135.538 Hz.
Apparently this block oscillates somewhat lower than 5.0688 MHz, namely around 5.0678 MHz.

Inserting -47 dBm 10.140200 MHz into the receiver yielded, after tuning the input circuit (60 pF trimmer)
and 4k7 LO potmeter, a nice sine wave with ca. 1.6 Vpp on my oscilloscope, see below (click to enlarge).

Update: After my UK trip (read below) and while fiddling and trying to improve the receiver,
i.e. maximizing the audio level relative to the injected 10.140200 MHz (Pref = -47 dBm),
I discovered that the waveform of the LO signal is more important than I initially thought.

Increasing the 100 pF capacitor to 220 pF in the LO RC-network resulted in a ‘cleaner’ LO waveform
but with lower amplitude. This is not surprising as f-3dB of the RC-network is now around 2.2 MHz, thus
‘damping’ the 5 MHz signal significantly. An alternative could be two RC-networks, but I was too lazy for that ; -)

Replacing the mixer diodes (initially 1N4148, BAW62 etc etc) with low barrier diodes like BAT86′s resulted in a
‘smoother’ adjustment of the LO level with a sharper maximum in the audio response.
The audio level increased around 10% to ca. 1.8 Vpp.

I tried a common base amplifier behind the mixer as it is said that a Polyakov mixer needs low Z termination.
Besides I had less less audio (of course), I didn’t experience less receiver noise.
On top of this, the minimum discernable signal (MDS) was higher compared to the original setup.

I also tried KE3IJ’s ‘double’ Polyakov mixer. It indeed results in more audio, but during that
evening it was an ‘AM hallelujah’ with the BBC world service competing with other broadcasters who
would be stronger than the desired WSPR signals.

So, I went back to two antiparallel diodes trusting they are ‘balanced’ enough.
</update>

Receiving WSPR signals.
By default WSPR software looks into the 1400 – 1600 Hz part of the audio spectrum.
The user has to tune his SSB receiver to the ‘USB dial frequency’, which is 1500 Hz
lower than the actual transmitted 4-MFSK WSPR signal. On 30m 10.140.100 – 10.140.300 Hz
is allocated for WSPR. Thus the ‘USB dial frequency’ normally is 10.140.2 – 1.5 = 10.1387 MHz.

However, this receiver has no ‘USB dial’, it uses a single frequency LO, which happens to be
lower than 10.1387 MHz, in my case 10.135538 MHz. This ‘USB dial’ difference is 3162 Hz.

In WSPR 2.xx the I/Q-option (for SDRs) can be (mis)used to compensate for this frequency difference.
Just fill in 3162 as ‘Fiq’ in the appropriate settings window.

How does it work?
Summarizing: the receiver works flabbergastingly well!
Especially considering its simplicity and the fact you receive both sidebands.
I experienced no AM ‘bleed through’ and it takes around 5 minutes before the LO is stable.

During a short holiday near Rye (England, JO00JW) I tested this receiver using only 5m of thin wire
as antenna, fed through a window of our camper. The received unique 30m spots can be viewed here.
Perhaps you are among them? It should be noted that the RX environment was very quiet.
MUCH better than here in Holland with lots of PLC equipment and LED lights polluting the radio spectrum!

This experiment also demonstrates that a clean radio spectrum is of significant importance
and not specifically receiver/conversion gain. As long as you can hear antenna noise it’s OK.
If you hear more than antenna noise, like computers, plasma screens, LED lights, PLC etc. more expensive
(and better?) receivers have no added value.

I was pleasantly surprised repeatedly receiving my own 30m WSPR signals from Holland!
Output there is around 150 mW and the antenna is a 3m long curtain rail (with curtains ; -)

During the evening and night lots of US stations were logged with good signals, despite the small RX antenna.
The fact that I was located near the sea certainly may have helped.

Below some pictures of the setup in and around the camper (click to enlarge in new tabs).

 

 

September 16th, 2015

Folded counterpoise (FCP) experiment on 40m (2)

In my previous post I tried to elaborate on the ‘magical properties’ of the
folded counterpoise (FCP).

My first conclusion is/was there is no such thing as ‘current cancellation’
in a counterpoise. Fields cancel, not the current, otherwise
the counterpoise never can act as a counterpoise.

I built an inv-L version for 40m in Germany and was forced to
conclude that it performed overall better than a 11m high inverted-V.

Secondly, there is no need for an ‘isolation transformer’ or other fuzzy materials.
A FCP based antenna can be made of simple (yes, also insulated!) wire.

The trick (as always!) is to build an antenna from a defined starting point and
in this case make the FCP-antenna resonant ( j = X = 0 ). From there the antenna
may be tweaked to the desired impedance ( Z = 50 + j0 Ω), if necessary.

My initial observations: despite all discussions a FCP as counterpoise seems reasonable,
i.e. not too good, but also not too bad. Especially concerning the FCP footprint.

Bear in mind, the radiation resistance (Rrad) of a 1/4λ (‘full size’) vertical above
ideal GND is around 36Ω. So if you measure an impedance of 50Ω there have to be
(earth + construction) losses involved.

Measuring VSWR doesn’t say anything, except that your transmitter feels it’s pushing
all his energy into the antenna contraption. Whether all this energy will be converted into
electromagnetic / radiated energy is the heart of the matter.

From the transmitters perspective a 50Ω dummy load is the ideal ‘antenna’!
However, this ‘antenna’ does not radiate (if everything is well ; -).

But how ‘good’ or how ‘bad’ is a FCP antenna, from a slightly more quantitative perspective?

Nike’s EU headquarters is around 1km from here, therefore:

The proof of the pudding is in the eating!

So, I built a new FCP with full size radiator (1/4λ) and compared it with
an antenna that is generally considered to be ‘reasonable’ and beyond suspicion:
A full size (1/4λ) vertical with two 1/4λ elevated sloping radials, defined as the ‘reference vertical’.

The corresponding NEC file for the reference vertical can be downloaded here, the FCP vertical here.

I made a movie of the experiment, see below. The aggregated RBN data of the experiment is here.

I’m not a super spreadsheet wizard, below is one of the RBN RX sites with perceived SNR’s.
Distance is ca. 1900 km. (Click on image to enlarge in a new tab)

A more detailed analysis of the RBN data will be published soon.

Measured relative field strengths in the movie have to be corrected for the distance.
This can be done easily because distances between measuring points and antennas
are depicted in the movie.

After correction my calculations reveal that the FCP (PA2FCP) and reference antenna (PA3REF) generate
almost equal field strengths at the two measurement locations.

The deltaloop seems an outlier, however its radiation pattern (‘egg shape’) has to be considered.

Update: I received quite a few emails on how I constructed the 40m FCP vertical in the movie.

Recipe
With a ground drill (diam. ca 8cm) a 40cm deep hole was drilled.
One of my 12.5m long fibreglass poles was inserted in this hole.
However, it’s also possible to mount the pole to a caravan adze ; -)

The modeled dimensions were reduced with the velocity factor (vf)
of the used wire. Bases on my experience and measurements the vf of my wire was 0.95.
My wire: PE insulated, 3mm OD, ca. 1.75mm ID copper.

In total I needed  (1067 * 0.95) + [ 5 * (279 * 0.95)] + ( 2 * 10 ) = 2360cm wire.
(see this NEC file or picture below)

Spacer material was ‘electricity’ PVC tube, 88cm total, divided in three (3) pieces
of 22cm, and two (2) pieces of 11cm. Thus, a spacer every 132.5cm

Holes for the wire were drilled somewhat skewed so that the spacers fixate themselves
when the FCP is ‘stretched’.

FCP ends were connected with rope to bamboo poles, which were guyed with
rope to strengthen and fixate the contraption.

Alignment/measurement
After building the contraption accordingly my MFJ-269 measured resonance (j = X = 0)
around 6.9 MHz. R was around 52Ω. At 7020 kHz Z was  55 + j15 Ω (iirc).

The FCP end was shortened with 15-20 cm and a (relatively broad) X = 0 dip was measured
at 7020 kHz with R = 54Ω. Good enough.

When (at resonance) R is too high, increase the FCP height. However, the contraption
has to be trimmed back to resonance by either tweaking the vertical part
( = ‘nuisance’ because the pole has to be dismounted) or the FCP end.
This may require some iterations.

The feedline contained a CM choke of 10 μH, consisting of 5 TDK clamps of 2 μH each, XL @ 7MHz = 440 Ω.

Below some pictures to visualize the aforementioned. (Click on each to enlarge in a new tab)

 

August 27th, 2015

Folded counterpoise (FCP) experiment on 40m

A priori: please read the whole (long) post because my context and experimental boundary conditions are explained.

Shortcut: the NEC input file of the experimental 40m FCP vertical is here.

Antennas are molecules
I own a PhD in chemistry and have experience in modeling molecules with ab initio quantum chemical
calculations
using e.g. GAMESS(-UK) (and other programs).

One of the hard core participants in the GAMESS project worked at ‘my’ faculty.
He introduced me, as one of the few students, into supercomputing in the mid and late 80′s.

Thanks to him I entered the world of worldwide computer networks such as BITNET.
It also opened doors to other supercomputing intensive chemical areas such as X-ray crystallography.

Best of all, it gave me access to -what became known as- . . . the internet!

During this time I experienced tensions between ‘practical’ and ‘theoretical’ oriented chemists.
The latter category was able to ‘synthesize’ molecules inside a (super)computer.

Although some outcomes were chemically/physically correct, it simply was impossible to synthesize these
outcomes with (at that time) current synthetic organic chemical methods.

Thus, from a practical perspective these ‘theoretical’ molecules didn’t ‘exist’.
Later … with new preparative methods some of these molecules actually have been synthesized in laboratories.

I also learned from intense discussions concerning ‘basis sets’.
A basis set is a combination of wave functions that are used to describe orbits of electrons in atoms.

This is a complex (quantum mechanical) area and not discussed here because I forgot almost everything ; -)

For example, theoretical chemist Dr. Y believed that calculating H2 properly was only possible with 6-31+G* basis set,
Dr. X stated that STO-3G is enough (because he was pissed off on Dr. Y using too much load/time on the super computer ; -).

From that moment I decided to become an ‘end user’ of this quantum chemical knowledge and stay out of these discussions.

Just look at these orbitals as the ‘radiation pattern’ of electrons. See below.

During my PhD period I used my experience in computational chemistry in order to synthesize,
isolate and crystallize molecules in reality. The combination of practical methods and computers delivered me a few
‘firsts’ (like this one) in organometallic chemistry and peer reviewed academic publications at that time.

Therefore, antennas are molecules. They follow the same or perhaps slightly different laws of quantum physics.
You can build antennas in reality and/or model them with software. Like in chemistry, I (try to) do both.

In university I also learned an important lesson for life:
The fact that I don’t know how and why something works and/or exists, does not mean it doesn’t work and/or exists!

Folded counterpoise (FCP)
Around 10 years ago Guy Olinger K2AV published a solution for owners of small lots to become active
on 160m relatively efficient. He named his solution the folded counterpoise, or FCP.
Since then the FCP is subject to heavy debate.
At that time I also read his results and reports and categorized K2AV’s solution as ‘Tell Sell Amazing Mike‘ (SK).

One of my reasons to classify the FCP as such, was that it needs very specific conditions to ‘perform’.
Due to ‘research and development efforts’ it is firmly stated that the FCP only works in conjunction with a very special
‘isolation transformer’ (sounds to me like the ‘Flux Capacitor‘ from the movie “Back to the Future” ; -).

This ‘isolation transformer’ must be made with one specific #2 material powdered iron toroid core (Amidon T300A-2)
having 20 bifilar turns of teflon sleeved double polyimide insulated AWG14 wire.
Secondly, it is stated that the FCP only works with blank (i.e. non insulated wire).

Wow!

For me it’s suspicious when requirements are mentioned so precisely.
It resembles the commercially well known slogan:

“Due to intensive R&D <insert productname> with compount XYZ, <insert productname> is very effective”.

Subsequently customers testify how excellent the product is.
This is exactly what happened with the FCP.

By chance I recently read some more/new information -and discussions!- on the FCP (google on it to find out).
Summarizing there are two categories:

1. people who firmly believe the FCP works
2. people who firmly don’t believe the FCP works

These two categories try to convince each other with orthogonally different statements and theories,
built on their perspectives on laws of physics and mathematics. Subsequently these perspectives are discussed, etc.

W8JI is a well known 160m guru and belongs to the 2nd category. On this page he elaborates on FCP systems.
I repeated his calculations with 4NEC2 and obtained similar results (vide infra).

DM9EE (ex-DL2OBO) is an example in the 1st category and the first who reported successfull use of FCP’s in a 80m 4-square.
KN2M also constructed FCP verticals and 4-squares. Other people report using FCP verticals during 160m contests
and state that disproportionally good scores were obtained. W4KAZ is such an example.

It appeared to me that people who actually built antennas with FCP’s according to K2AV’s instructions fall in the
first category. People who didn’t construct antennas with FCP’s and model antennas fall in the second category.

Now … what is ‘true’ and/or ‘reality’?

One way to find out is ‘just do it’ and build a FCP system myself . . .

New incoming mail
(Very) recently two ingredients caught my attention.

1. K2AV published his FCP in the ‘National Contest Journal’ (NCJ) please download and read this (later)
2. HC1PF uses an inverted-L with FCP

Ad. 1 The ARRL is an esteemed organization with benchmarking magazines and publications like QST, QEX and NCJ.
If the FCP is all nonsense or a hoax editors of QST, QEX and NCJ would not have routed K2AV’s paper towards NCJ.
It may seriously damage the image of NCJ.

Ad. 2 Perhaps more interesting . . .   HC1PF is Luis IV3PRK. Luis went to Ecuador in 2014.
However, due to ‘some living problems’ he wants to return to Italy soon.

IV3PRK is an avid 160m experimentator and often referred to as ‘Radio Italy’ on 160m.
Luis is a HAM who wants to get the most out of his 160m station within his specific boundary conditions.
Like me, Luis  is open to new things and investigates them.
If it works and measurements confirm . . . it works. Period.

There is virtually nothing Luis didn’t measure and/or calculate in order to improve his 160m station.
If there may be a better alternative or technique, Luis is the first man to investigate, try it,
and share his findings with you and me.

Moreover, when Luis is not able to reproduce measurements/results others obtained with similar equipment
he is the first to be in your mailbox with in-depth and intrusive questions.

So . . . why on earth Luis HC1PF ended up using a FCP in conjunction with his radiating element
(inverted-L) after several decades of successfull results, experiments and experience on 160m ???

Several YouTube videos are available in which Luis HC1PF is recorded with good signals all around the world.

All this information fired up my curiosity  . . .

Antennas are molecules
As already stated, I consider antennas as molecules. I prefer to calculate and build them.

The combination of theoretical and practical skills gives insight in when a converged solution of a computer model is ‘real’ or not.
I.e. being able to establish if a molecule -although perfectly in line with ‘theory’- ‘exploded’ inside the model.

The same can happen witn modeled antennas.

For some years I use 4NEC2, a free package brought into the public domain by Arie Voors, a friendly fellow Dutchman.

Similar to quantum chemical models, I am an end user and will not go into discussions concerning validity and/or correctness
of e.g. ground parametrization in NEC2 vs. NEC4 or that kind of stuff.  I have the NEC2 engine and have to deal with that.
If someone is willing to sponsor me the NEC4 engine, please do ; -)

Having said this, this antenna modeling software is not written by morons, but by very capable and highly skilled scientists
working in well esteemed institutes, and perhaps know Maxwell’s Laws better than Maxwell did himself!

Last, but not least, I have no degree in RF-design or electrodynamics but have good experiences with 4NEC2.
4NEC2 prevents ‘surprises’ in practice and yields insight on e.g. degrees of freedom in antenna designs and their behaviour.

Convergence criteria yield information on (relative) sensitivity of variables and more practical knowledge concerning
performance, design and actually building antennas in practice. It’s similar with synthesizing molecules behind the fume cupboard.

While actually building antennas from modeling experiments I learned a lot from the conversion from ‘theory’ into practice.
As of today 4NEC2 never let me down and helped me getting good (and sometimes excellent) performing antennas.

First real 40m FCP experiment
A good HAM friend of mine is captured in the story of his life:
For whatever reason he always produces weak field strenghts.

Last week he invited me for a BBQ at his campsite at a HAM fest, known as ‘Bentheim‘ (Germany).

The campsite was saturated with antennas, hung between scarce trees.
He had put up a 40m inverted-V above his caravan with the feedpoint around 11m high.
However, several (40m) dipoles of other HAM friends crossed or paralleled his antenna.

The day before I worked him in the evening from my home town and his signal was, as ever … weak.
Distance was around 150 km and an inverted-V with 11m apex has to act as NVIS antenna.

I tried to explain him that his environment was an illustrative example of a ‘recipe for disaster’ as (resonant) antennas couple.
Destroying the radiation pattern and performance is the result.
The chance that all these randomly hanged out ‘resonant’ 40m dipoles form a passive array is zero.

Now the reason why people from this HAM fest were so weak became manifest to me.
My friend demonstrated that the VSWR50 of his antenna was around 1.3, which was a miracle to me anyway . . . ; -)
His conclusion was, due to the VSWR50, his antenna ‘was good’. . .

Last year we built a good performing vertically polarized 40m-deltaloop (in identical environment)
which annoyed the owner of the camping site as well as other HAMs/guests.
They almost broke their necks by tripping over the horizontal wire above the ground.
Secondly, there were more guests than last year, which meant less space. Therefore a deltaloop was a ‘no go’.

I decided to skip the beers and watching movies, and suggested him to build an antenna that was ‘different’
and also told him this antenna would be a real experiment. That is, any outcome would be new to me.

Quick & dirty googling didn’t result in 40m FCP antennas. So, would we be the first ones to try a FCP on 40m . . . ?

While others enjoyed drinks in the sun, I enjoyed modeling a 40m FCP vertical in the same sun on a laptop.
I decided to follow basic knowledge in conjunction with my own ideas, superimposed on K2AV’s counterpoise shape.

Followed procedure
1. Starting point was a vertical part for 1.83 MHz, resonant above ideal GND -> Z = 37.2  + 0 Ω and length = 39.96m.
This results in an electrical length of (39.96/163.8) * 360 = 88°. Enough in line with theory.

2. Knowing that in this case Z equals the radiation resistance (Rrad) the vertical part was raised to K2AV’s height (2.5m or 8 ft).

3. K2AV’s FCP dimensions were entered into the model.

While W8JI used a FCP height of 10m (30 ft) and got Z = 34.5 – j196 Ω, I got Z = 31.2 – j206 Ω. Fair enough, similar trend.
It may be clear the antenna is far from resonant (j ≠ X ≠ 0). The calculated phase angle φ = -81.4°, it almost can’t get worse . . .
Similar to W8JI I noticed that varying the size of the FCP brings the contraption into resonance (j = X = 0).

Doing so resulted in Z = 33.5 + j0.1 Ω at a ‘half length’ of around 12.26m (40ft), also compliant with W8JI’s results.
However, the real part of my impedance result is significantly lower than W8JI’s result, 33.5 vs 39.6 Ω respectively.

Anyway, I was able to reproduce his trend.

The claim K2AV obtained a match with his very specific ‘isolation transformer’ attributes W8JI to ‘considerable flux leakage‘ (?)
of the toroid core material used in conjunction with the reactance of the 20 bifilar windings. It magically adds around +j200.
A ‘resonant match’ on 160m in conjunction with only K2AV’s FCP dimensions and material is the flabbergasting result.

HC1PF uses this magic ‘isolation transformer’ too. However, he had to use a trick I refer to as overloading.
He enlarged the horizontal part of his inverted-L to obtain R = 50 Ω on 1825 kHz, as measured with his VNA.
A series capacitor was used to tune out the resulting +j reactance.

I deliberately decided NOT to use an ‘isolation transformer’ for two main reasons:

a. I don’t believe in ‘Flux Capacitors’ and don’t trust ‘flux leakage’
b. There was no ferrite toroid core available ; -)

The NCJ article states FCP ‘half length’ amounts 10m (33 ft) on 160m and 5m on 80m with 10cm (4″)
spacing between the wire. So I bluntly scaled the original FCP ‘half length’ for 40m.

4. The 40m FCP was ‘scaled’ to 7.075 MHz with an initial ‘half length’ of 2.5m and empirically placed 1.5m above
‘Average Real Ground’ (conductivity = 5 mS/m, ε = 13). FCP wire spacing was 10cm.

5. My friend said he had a spare glassfiber pole of 12.5m length. Thus, the max. vertical length amounts 12.5 – 1.5 = 11.0 m

With this data the model calculated Z = 48.2 -j 78.6 Ω , |Z| = 92.2 Ω  . . .  interesting.

6. Converging the vertical part of my ‘molecule’ to resonance resulted in Z = 62.4 Ω with a 13.59 m long vertical.
Considering the available (spare) pole this was too long . . . never mind. Mechanical details can be solved anyway.

This resonant impedance didn’t surprise me, as reducing radial lengths in verticals, increases R (and losses!).
Let me make this clear, the real part of Z (R) is not always Rrad!

Bluntly assuming 37.2 Ω from the original Rrad (although the vertical part is longer) the model ‘added’ 62.4 – 37.2 = 25.2 Ω.
Not considering structure losses and assuming the ‘extra resistance’ (R) is due to ‘earth losses’
the efficiency amounts 37.2 /62.4 = 60%. This is not very good, but also not very bad . . .

7. Returning to the original vertical length (11m), FCP height of 1.5m, 10cm wire spacing, converged the ‘half length’ of the
FCP to 277.5 cm with Z = 48.8 + 0.1 Ω. Even more interesting . . . from a matching perspective.

Trusting the ground model in NEC2, antenna efficiency (assuming Rrad = 37.2 Ω and no structure losses)
now increased to 37.2/48.8 = 76%.

NOT considering debates and fights concerning the ‘crappy’ ground model of NEC2, this is not too bad at all . . .

From my experiences 4NEC2 does not perform too bad in practice. Taking velocity factors of wires/conductors into
account when building antenna constructions minor/no fiddling is necessary.

‘Real Average Ground’ as ground model seems to work for me in most cases.
Also in these cases where I (mis)use the ground to ‘force’ the impedance of an antenna to a ‘desired’ value.

I succesfully use this trick for several years to ‘force’ my 40m vertical deltaloop implementation to 200 Ω,
so it can be matched with a 4:1 UNUN.

Based on my modeling experiences the 7. antenna (see above) would be a nice candidate for our experiment.

Cancelling currents
Now . . . let’s go back to K2AV’s statement on ‘cancelling currents’ in the counterpoise, giving the FCP its ‘unique features’.
Fig. 1 depicts the current distribution around the feed point in my 40m-FCP computer vertical. Click on the picture to enlarge in a new tab.


Figure 1. Current distribution FCP                     Figure 2. Current phase distribution FCP

Figures 1 and 2 are zoomed to give a more precise look at the counterpoise, knowing that the current distribution in the vertical part
is well known (and predictable).  By the way, it doesn’t matter whether the counterpoise is exactly 5/16λ long (cf. K2AV or NCJ article)
or its length (≠ 5/16λ) converged to resonance (j = 0).

If you don’t trust, just fill in the appropriate values for 160m into the NEC model (vide infra).
You’ll see it doesn’t matter.

What doesn’t matter? The current distribution in the FCP!

It can be clearly seen that the model didn’t null out or cancelled currents, as suggested by K2AV in figure 1D of the NCJ article.

Now let’s look at the current phase distribution inside our 40m candidate in figure 2.
Bear in mind this antenna is calculated to be resonant, so φ = 0°.
If you had calculated a non resonant antenna, the ‘starting’ phase angle (of course) differs.

It can be clearly seen in fig. 2 that the radiator and counterpoise are homogeneous 180° out of phase.
There is no current phase difference visible. Just as it should.

Note: The phase angle flips between 180° and -180° in the first segments of the FCP. This is due to the used segment size in the model
but 180° and -180° are equal. If, for your psychological convenience you want to eliminate this, fiddle with the FCP segment sizes
or bring the antenna out of resonance somewhat. At resonance there is a singular point.

If I remembered Kirchoffs Laws from school properly current cancellation within the same conductor is impossible.

In other words, if ‘current cancellation’ takes place in a FCP, a FCP is not a counterpoise.

Note: NOT considering TEM lines like coax and e.g. 1/4λ (coax) baluns.
In the latter current can flow on the inside and outside of the braid.

Presumably K2AV refers to the phenomenon that e.g. having more (elevated) in line or symmetrical radials,
the current in each radial is divided by the amount of radials. Resulting in less current per radial because there is less flux density,
i.e. the current is smeared out on a larger surface with possible reduction of earth losses as a result.

Or . . . perhaps the difference in radiation patterns between a T-antenna, having two in line top loads, and an inverted-L.

Anyway, if I remembered physics class in high school well, the fields ‘cancel’, not the current (cf. corkscrew rule)

From my perspective, ‘current cancellation’ in a FCP is busted and so is the need for an ‘isolation transformer’.

Back to my 40m FCP experiment
The results from 7. nevertheless revealed a possibility to ‘synthesize’ my modeled ‘molecule’ from scrap material
with a small foot print on my friends HAM fest lot. It unfortunately appeared that his 12.5m spare glassfiber pole was damaged.

Max. 8.5m could be saved and erected. He was located at an edge of the HAM fest terrain.
This left one option open: an inverted-L with sloping topload. By chance my friend had plastic tubes with two 4 cm separated holes.

These boundary conditions were entered into the model. Practical values were produced after optimalization,
of which lowering the FCP was one of them. It may not be the most optimal solution from a combinatorial perspective,
but from a practical perspective it was. The resulting NEC file is here.

A visualisation of the contraption showing current distributions is depected in figure 3.
The calculated radiation pattern is shown in figure 4. Feed point information is given in figure 5.
Click on each picture to enlarge in a new tab.

Figure 3. 40m inverted-L with FCP               Figure 4. Hor + vert radiation pattern      Figure 5. Feed point information.

The calculated radiation pattern (fig. 4) is OK. I will not go into detail concerning the claimed ‘gain’ of the calculated antenna,
otherwise I will end up in ‘basis set’ discussions.

Because the vertical part is <1/4λ 50Ω matching is required.
We needed a match for the lower SSB portion of the 40m band (7060 – 7100 kHz).
Given the available (scrap) material the good old hairpin method (matching against admittance, Y = 1/Z) was selected.

The trick is to make such a short antenna resonant (converge to j = 0) [on a somewhat higher frequency]
and then shorten (in this case) the topload until Zp reaches 50Ω at the desired frequency.

In figure 5 it can be seen that @7.075 MHz Zp = 52.2 // -j 85.5.
Only one coil of around 2 μH (+j 85.5) parallel to the feed point is required for a match.

Perfect.

It seemed that my friend only had PE-insulated 2.5mm² copper wire.
From my experience, using a velocity factor (vf) of 0.95 for this type of wire is a good starting point.

Oh oh, another deviation from the original FCP recipe . . . insulated wire! ; -) Also busted . . .

All dimensions were multiplied with 0.95 and I decided to multiply the ‘half length’ of the FCP with the used vf.
An alternative could have been applying the vf for the total length of the FCP wire.

After building the contraption accordingly I wanted to measure its impedance with a VNA.
Unfortunately several nearby fellow HAMs only got MFJ-259 (so not 259B) devices.
This left me with only ‘SWR’ and R values, no X . . .

The MFJ was connected through a 50cm short cable to the antenna feedpoint. The antenna ‘dipped’ around 6.85 MHz.
I decided to reduce the open end of the FCP with around 60cm.

After this, I measured a dip at 7.1 MHz with R around 32Ω, and SWR was roughly around 1.8.
I couldn’t believe my eyes . . . very suspicious! (instead of promising . . .)

The vertical part of the inverted-L was 8.5 – 1 = 7.5m. This results in a monopole length of (7.5/42.37)*360 = 64°.
The radiation resistance of a 64° thin monopole amounts ca. 30Ω (cf. 5th ed. “Low Band DXing”, ON4UN, fig. 9-8B).
Next, the topload had a sloping angle of around 65°, reducing the electrical height (l/λ) thus reducing Rrad more.

Now what? I asked for another MFJ-259 and reproduced the first MFJ measurement.
Queer . . .

Because nobody else could help me out with yet another (type of) VNA I decided to follow Luis HC1PF’s approach:
“If this is what I measure, this is what I measure.”

For my own convenience I decided to conclude that my friends caravan and other nearby 40m antennas were ‘assisting’ me.
Given the circumstances I couldn’t imagine a more plausible explanation.

A 2 μH hairpin coil was wound on a small plastic ‘smoothy’ bottle and connected to the feed point.
After this, the MFJ-259 displayed a SWR of nearly 1 with R around 50Ω. Fiddling with winding spacings on the bottle forced
the MFJ-259 to keep the SWR needle rock solid in the left corner @7.08 MHz.

An improvised ‘ugly balun’ choke was made of several feedline windings.  The antenna was connected to
my friends FT-857 inside his caravan tent. There the good match was confirmed.

Below some pictures of the contraption. Click on each image to enlarge in a new tab.

First on-air experiences with the experimental 40m FCP
My friend had a coax switch (not a hoax switch ; -) and we decided to compare his inverted-V and the freshly built inv-L FCP antenna.
His immediate observation was that, compared to his inverted-V, ‘different’ stations could be heard with the FCP vertical.

Very nearby stations (< 70 km) were stronger on his inverted-V, all others had significant better SNR’s + ‘displayed’
signal strenghts. Sometimes up to four (4) ‘S-points’ on his FT-857 . . . (besides a ‘bar’ the FT-857 also displays e.g. ‘S7′).

Of course the intrinsic difference between an inverted-V and this vertical (e.g. take off angle) is obvious.
However, at first glance the results were remarkable, also at medium distances  (150 – 400 km).

<Tell Sell Amazing Mike (SK)>
My friend is no DXer and it was getting dark so I decided to scan the band. I worked UN7AR after a first call through an EU-pile up
on 7094 kHz with reasonable signals (57-58 or so) with 65W PEP.  Path length is around 3400 km. When switching to
the inverted-V UN7AR disappeared into the noise/QRM, so I didn’t bother to disturb his EU-pile up with antenna experiments.
</Tell Sell Amazing Mike (SK)>

After this promising but peculiar result I took my first beer and enjoyed a barbecue with fellow HAM friends and our partners.

Qualitative (not quantitative) experiment
After the BBQ, my friend was very curious if his new antenna would improve his signal in The Netherlands.
During summers there is a Dutch 40m ‘holiday round’ (‘Gooische Ronde’) with people on vacation spreaded all over Europe.
Also lots of Dutch HAMs located in The Netherlands join in.

When net control asked for subscribers I called in with the vertical, deliberately not signing /DL.
My friend eagerly wanted me to inform others that we were testing his new antenna.

I told him that this would screw up the experiment.

After a short while it was my slot.

During my slot I ‘randomly’ switched between the vertical and the inverted-V. ‘Talk punch’ was kept equal.
When finished, the immediate response of net control was that there was something wrong with my setup.

The conclusion was a loose contact somewhere because my signal fluctuated severely. I was asked to try it again.

I shortly came back on the inverted-V, asking if ‘this’ was better. Net control and others replied the signal was
weak and that I should fiddle a little. Subsequently I switched to the vertical, said I fiddled a little and asked if ‘this’ was OK.

The unanimous response was ‘that everything was OK now’ and that I shouldn’t touch anything and leave it this way.
Some (relatively nearby) people remarked they weren’t able to hear me properly on the inverted-V . . .  ?

In my second slot I confessed building ‘a vertical’ and I was not in PA but in DL instead,
and that I switched between the two antennas during my initial slot.

After this confession more people called in and agreed that the signal on the vertical was significantly better.

As expected, there was a relation between signal improvement and distance.
One participant in Italy perceived +10 dB (from ‘S9′ to ‘S9 +10′) in favour of the vertical, which could be plausible.

One Dutch station, we never heard before, suddenly reported being /MM on a yacht in a harbour in
West Scotland (distance around 1100 km) in salt water.

He said he ‘had’ to comment on this ‘remarkable’ signal difference. For him it was a difference between ‘day and night’
and reported even a larger perceived difference of +20 dB in favour of the vertical . . . (?)

Stations in the Mid/West Netherlands reported around one S-point difference or, a difference between
‘mediocre’/'no’ and ‘good’/'comfortable’ copy.

Although I initially didn’t tell I was using an experimental FCP vertical, who bothers anyway ; -), the overall and
summarizing conclusion from all the participants, regardless from their distance, that evening was:
“That vertical works better than the inverted-V”. I can’t deny to confirm this was reciprocal.
I experienced the same trend in reception while switching between the inverted-V and FCP vertical.

Conclusions and remarks
This experiment was only one experiment. Therefore it has no use to draw conclusions, other than that the
experimental 40m FCP vertical in this particular situation seemed to work better than the inverted-V @11m high.
In fact, it’s ‘comparing cows and horses’.

My conclusion is that there is no such thing as ‘current cancellation’ in a FCP.
An ‘isolation transformer’ is unnecessary when making the antenna resonant.

Nevertheless, having had enough experience in building verticals (with ground nets, (tuned) elevated radials, deltaloops etc)
I have to admit that the experimental contraption performed remarkably well that evening.
Especially considering the ‘minimal’ and small earth net/counterpoise in conjunction with the small antenna foot print.

However, there have to be (significant) earth losses, reducing antenna efficiency, involved . . . but this wasn’t confirmed
with the ‘quick & dirty’ VNA measurements.

Therefore, the measured impedance of the FCP antenna is a miracle to me and I certainly want to build another one to thoroughly
investigate it with proper equipment. I also want to compare its performance with an optimized deltaloop
and/or ‘reference 40m vertical’ in the (near) future by measuring (calibrated) field/signal strenghts.

One final conclusion may be that this experiment was fun, we had a great day, nice BBQ and lots of laughs : -)

 

March 15th, 2015

Noise/interference enduring UTC DCF77 clock

Warning: This is a long post. So fetch a coffee, beer or nice glass of wine ; -)

For a long time I want a UTC clock locked to DCF77 as a separate unit with six 7-segment LED displays.

Why? Perhaps because almost nobody has one, perhaps because it’s easier to make a UTC clock
with a GPS receiver.

The picture below displays a quick and dirty prototype of my UTC DCF77 clock (click to enlarge in a new tab).

DCF77 broadcasts Central European (Summer) Time (CE(S)T) which is one or two
hours ahead of UTC during wintertime or summertime respectively. Although very
cheap DCF clocks are available, I never found a DCF77 clock with a ‘UTC button’.

So the remedy for my desire was to build such a clock myself.
Nowadays this is relatively easy compared to a few decades ago.

Google found several DCF clock designs. However, a lot of them lack the seconds due to hardware restrictions,
like not enough I/O pins on a microcontroller chip to multiplex six 7-segment displays.

I had some Arduinos lying around and found out a very elegant solution for the displays: the MAX7219 multiplexer.
On Ebay I found cheap 8 digit displays with this chip. From my NTP experiments a while ago I discovered a
Conrad DCF77 receiver module in my junk box, so I had all the parts to start.

Google almost immediately led me to the site of Thijs Elenbaas. Thijs is the author of the popular
Arduino DCF77 library and states he added ‘UTC support’ in April 2012.
As I am not a skilled C++ programmer, this was exactly what I needed!

Thijs’ code was patched to drive the MAX7219 display multiplexer and in no time my UTC DCF77 clock
was up and running by calling the DCF.getUTCTime() routine. Looking at this routine I think there is a bug
concerning converting CE(S)T to UTC but I didn’t had the patience and time to verify.

Because …

After additional googling I stumbled upon the blog of Udo ‘Blinkenlight’ Klein. Udo did some magnificent and
complicated work on decoding DCF77 signals with low SNR’s and/or interference.

From earlier NTP experiments I remembered that e.g. when there was a thunderstorm in Europe my NTP
server had difficulties to decode DCF77 time stamps properly. Glitches on the signal resulted in losing lock.
So, Udo’s page was interesting for me and I examined his experiments.

Although DCF77 has a good signal here, I started to implement his exponential filter first.
This filter was integrated in a sketch based on Thijs’ DCF77 library. Udo states this filter results
in +15 dB (perceived) SNR. As the DCF77 signal arrives here strong enough, I had to decrease the perceived signal
strength by pointing the ferrite rod antenna in DCF’s direction and detune the receiver coil with a trimmer.

The result was indeed a better decode with lower signal strengths compared to Thijs’ original code. Promising!

In figure 1 (shamelessly stolen from Udo’s blog) the effect of the exponential filter is visualized when 60% noise is
super imposed on the DCF signal. Udo added a Schmitt trigger function to shape the signal.


Figure 1. Shaping a noisy DCF / pulse signal with exponential filter + Schmitt trigger (courtesy of Udo Klein).

The net result from this filtering is that it results in lower perceived band width, which means that signals
can be detected/demodulated better as (thermal) noise is directly proportional with band width (B): N = kTB

Although this approach results in cleaner pulses at lower SNR’s, there are a few side effects.

First, it can be seen that the pulse width after this filter highly depends on the Schmitt trigger threshold value.
In figure 1 this threshold amounts 50%, and magically results in suitable pulse lengths for a ‘normal’ DCF
interrupt driven decoder, like Thijs’ decoder. But when SNR’s become lower this kind of filtering, combined
with the Schmitt trigger option, may create wrong results.

Secondly, such a filter introduces additional phase delay… which is delay in Time (time with a capital T),
not considering the necessary time to execute the Arduino code.

Having read the DCF77 protocol a long time ago, I knew that DCF uses PSK on its signal as an additional
asset to increase the accuracy of the received time signal. However, in order to use this PSK option a
significant more complex (hardware) receiver is necessary. When around 2005 I used DCF77 for my NTP server
no home brew PSK receivers seemed publicly available. I just googled and found that in January 2012 such
a receiver was published in Elektor Magazine.

Udo’s second published experiment involves ‘phase detection’ of the DCF signal. At first glance I thought he
found a way to demodulate the pseudo-random phase (PSK) signal with cheap DCF77 receiver modules.
If this was true that would be wonderful and could result in 50 ns accuracy!

However, Udo has another perception of phase than I initially had. He considers ‘phase’ as relative to the
second epochs of DCF. On other words, he considers DCF77 as a PPS source, just like I did in 2008 with ntpd
in conjunction with the ATOM (22) and PARSE (8) driver. The latter driver is necessary to obtain Time.

Here and here are my posts from 2008 and 2011 respectively. Connecting the DCF signal to the DCD pin of a
RS232 interface resulted (and results) in ca. 50x lower PLL offset compared to the PARSE (8) driver alone.

What Udo basically programmed, is a software PLL where the time stamps of the Arduino are compared
with the second epochs of DCF. Another use of a PLL is that it can be used as a very effective filter,
depending on the perceived phase (time) difference and loop filter time constants.

When such system is in lock, band widths of fractions of a Hertz are achievable, resulting in detecting
lower and lower SNR’s. However, it takes some precision and time to get there.

Although Udo claims that running his code on an Arduino with a ceramic resonator is useless, I gave it a try
with an Arduino Nano … with a ceramic resonator.  I ran Udo’s DCF77 Scope and concluded quickly that the
clock source was not stable enough. Fiddling with the time constants in the software didn’t result in a lock.

I inspected the Nano and regarded the space around the resonator too small to insert a HC49(/U) crystal.

I was impatient and considered building an Arduino myself with a DIL28 ATMega328. Unfortunately one
of my nearby friends used his last DIL28 ATMega328 in one of his projects, so I looked on Ebay
and forecasted that my project would be delayed for at least two weeks…

At first glance my Ebay search seemed ‘polluted’ with Arduino Mini’s and I skipped them. Suddenly I saw a
picture of one of these Mini’s with a crystal and apparently I clicked on this particular link earlier!

It took me some time to realize I ordered some Mini’s from the same Ebay shop a while ago for ‘just in case’.
They were lying unopened in my shack as I planned to use the ‘old’ Mini’s first.

Presto! With great joy I saw that the last ordered Mini’s had a 16 MHz crystal clock source!

See picture below (click to enlarge and open in a new tab).

Immediately I rewired the Nano contraption to the Mini and ran DCF77 Scope. As you can see
below, the stability of the clock source is almost perfect for Udo’s software and/or library, albeit
that the signal may be ‘too clean’ to fully recognize the claimed capabilities of his DCF77 decoder.


32, +———+2XXXXXXXXX5——–+———+———+———+———+———+———+———
33, +———+8XXXXXXXXX7——–+———+———+———+———+———+———+———
34, +———+5XXXXXXXXX6——–+———+———+———+———+———+———+———
35, +———+6XXXXXXXXXXXXXXXXXXX2——–+———+———+———+———+———+———
36, +———+3XXXXXXXXX4——–+———+———+———+———+———+———+———
37, +———+4XXXXXXXXX2——–+———+———+———+———+———+———+———
38, +———+5XXXXXXXXXXXXXXXXXXX1——–+———+———+———+———+———+———
39, +———+9XXXXXXXXX6——–+———+———+———+———+———+———+———
40, +———+5XXXXXXXXXXXXXXXXXXX———+———+———+———+———+———+———
41, +———+3XXXXXXXXX5——–+———+———+———+———+———+———+———

….

2381, +———+—–5XXXXXXXXX3—+———+———+———+———+———+———+———
2382, +———+—–7XXXXXXXXXXXXXXXXXXX—-+———+———+———+———+———+———
2383, +———+—–6XXXXXXXXX—-+———+———+———+———+———+———+———
2384, +———+—–7XXXXXXXXXXXXXXXXXX8—-+———+———+———+———+———+———
2385, +———+—–5XXXXXXXXXXXXXXXXXX8—-+———+———+———+———+———+———
2386, +———+—–6XXXXXXXXXXXXXXXXXX8—-+———+———+———+———+———+———
2387, +———+—–5XXXXXXXXXXXXXXXXXX7—-+———+———+———+———+———+———
2388, +———+—–7XXXXXXXXXXXXXXXXXX5—-+———+———+———+———+———+———
2389, +———+—–7XXXXXXXXX—-+———+———+———+———+———+———+———
2390, +———+—–4XXXXXXXXX2—+———+———+———+———+———+———+———

This particular Arduino Mini deviates 50 ms in 2390 seconds, which is 0.05/2390 = 21 ppm or 335 Hz @16 MHz during the test period.

The estimated 16 MHz clock source offset is used for -what Udo calls- ‘auto tune’ the internal clock of the Arduino.
Normally you would adjust the frequency of an oscillator with the outcome of a phase loop but that is not possible here because the
Arduino 16 MHz clock source has no external means (e.g. a varicap or VCO pin) to adjust its frequency.

Udo used a ‘software varicap’ by manipulating internal timer (CTC) prescaler and division constants. Knowing that the DCF epochs are
1s or 1000 ms by definition, and knowing that the local clock source is approx. 16 MHz one can simply count cycles with a CTC.
In his interpretation one of the internal CTC’s is prescaled with 64, resulting in 16E6/64 = 250000 counts (needed for interrupts).

When 1ms resolution is required, these 250000 counts can be divided inside the CTC with 250. During a certain period the amount of
counts is collected. If this amount exceeds or deceeds the theoretical value, the CTC division factor is (temporary) in- or decreased.

In other words, it’s a PLL the other way round: not changing the oscillator frequency with a fixed division number (N), but change
N in relation to the oscillator frequency, similar what dual modulus prescalers do.

The reason to ‘tune’ the internal clock source is when the Arduino loses its DCF lock, the clock must run on its internal clock
as accurate as possible. This accuracy is important to retrieve DCF lock, especially in noisy or interfering environments.

The next thing Udo did to obtain Time is using a mechanism known as convolution.
Convolution is known for several years in programs as JT65 and WSPR, brought into the HAM radio community by K1JT.
As far as I could ascertain Udo is the first to deploy convolution for DCF77 signals.

I try to explain it in my own words. When the local clock source is locked with DCF a few things ‘happen’.
The software starts searching for a known a bit sequence. Udo selected some bits of which the ‘sync mark’ during
the last second is a member.

During the sync mark the DCF carrier is not reduced, which means full DCF carrier for at least 1 second.
This occurs only once every 60 seconds, except when a leap second is inserted (which is very rare).
Combined with other known bits, like ‘start of second bit’ = DCF bit20, the (noisy/distorted) signal is convoluted
with the known bit sequence to -summarizing- produce Time.

In the figure below (shamelessly stolen from Wikipedia) the mechanism behind convolution is depicted.

What clearly can be seen is that the ‘overlap integral’ is maximal when two identical bit patterns are in sync. In order to
determine this overlap integral properly the local clock source has to be as accurate as possible. Using more bits and
reward the outcome of a convolution process with a quality indicator gives a measure concerning the prediction skill.
In my own words, a sort of pattern recognition which is a PLL, based on the data itself!

After a long story … the result is I now have a VERY accurate DCF77 UTC clock with truly remarkable DCF signal acquisition performance!

At my location the clock needs around 5 minutes to reach the ‘synced’ state:

Decoded time: 15-03-16 1 13:06:48 CET ..
Quality (p,s,m,h,wd,d,m,y,st,tz,ls,pm): 2 (9746-0:255)(71-8:10)(56-42:2)(49-35:2)(21-14:1)(42-35:1)(30-24:1)(24-18:1)14,7,7,255
Clock state: synced
Tick: 14
confirmed_precision ?? adjustment, deviation, elapsed
0 Hz @+ , -335 Hz, 0 ticks, 0 min, 4303 cs mod 60000

Here is my preliminary Arduino sketch. It’s a somewhat patched ‘Superfilter’ sketch.

I did some simple tests. It is important to obtain a lock or sync. Once locked or synced it’s almost
impossible here to let the clock run in ‘free’ mode, except when disconnecting the DCF receiving module.