After I have established one of the most noise tolerant software decoders for the DCF77 signal. I now focus on capturing a better signal, or even more important, a signal at all. Although my clock can deal with very poor signal noise ratios there must be at least some signal for it to do its work.
In order to improve on this side I followed two lines of investigation.
1) Learn about better Antenna design.
2) Evaluate different receiver boards.
With regard to the Antenna design I worked through the following three books.
The conclusions are:
(a) Bigger Antennas will result in better performance. That is they will offer more gain. In addition they can be created with a stronger directional characteristic. This in turn allows to improve signal noise ration in case the source of the noise is in a different direction than the DCF77 transmitter.
(b) Since it is technically almost impossible to have an antenna that has at least 1/4 of the wavelength of DCF77 (3868m/4 = 967m) a “small” antenna must be used. The typical solution is a resonant “magnetic” antenna. Since tuning is not completely trivial I decided to buy the biggest manufactured DCF77 antenna within my budget.
If you want to go for some really good antennas that are slightly above my budget have a look at the homepage of BAZ Spezialantennen. These Antennas are expensive but most probably worth the money. I did not try them though since my software solution did not require this at all.
With regard to different receiver modules I ordered modules from different “amateur” sources (Conrad, Reichelt and Pollin) and one “professional source” (HKW Elektronik). Here are the result of my completely unscientific benchmark.
Initially I started my experiments with a module from Conrad.
This module comes with a tuned (resonant) ferrite antenna (notice the green capacitor glued to the antenna). It also is very nice for prototyping as the leads can be attached with screws. In addition it offers a wide supply voltage range and two open collector outputs (one inverted).
The reception quality is sufficient for my clock but it was bad enough to get me started in advanced denoising techniques. It will pick up noise if my washing machnine or dryer are running.
The next module I ordered was from Reichelt.
It also has a tuned antenna. However the capacitor is an smd capacitor soldered on the board. This makes replacing the antenna slightly trickier. I can not say very much about the performance of this board because it failed during the first test. Since I did not like the antenna design anyway I did not reorder a second module.
I also ordered a very cheap module from Pollin since some people claim it is superior in performance to the Conrad board.
As the Conrad module it comes with a tuned antenna. The overall quality impression is slightly better as for the Conrad module. However it only supports a smaller voltage range and only one output. The washing machine and dryer tests showed similar results. So I did not find any significant differences in performance.
My conclusion so far is that the Conrad module is better suited for prototyping while the Pollin module is better suited for a fixed (=soldered) board.
Then I ordered a “professional module” from HKW. HKW offers lots of different modules and several different choices of antennas. I ordered one of the dumbest module they offer.
In addition I ordered all possible antenna choices.
All antennas are tuned antennas. The smallest antenna is comparable in size to the antennas of the Conrad/Reichelt/Pollin modules. However it has a beefier feel and seems to be somewhat more precise tuned. The larger antennas are comparable in appearance to the Pollin module’s antenna. Since I did not want to try all variations I immediately started to test with the largest antenna. This is of course an unfair advantage. As expected this combination blows all of the other tested modules completely out of the water. I do not know if this is due to the antenna size alone or also due to the build quality of the module. However this combination is completely unaffected by the washing machine and/or dryer. It also is completely unsensitive to the placement in my flat. The other modules can not pick up a signal everywhere. This combination can. It is so good that I decided to not experiment with any combination any further.
The only issue is that HKW does not sell to end consumers. I assume that other “professional” modules show similar performance. My conclusion is: for superior performance try to get hold of a professional module that sports a 10 cm antenna.
I extracted the following data from the different module’s datasheets. I have no means to verify most of the values due to lack of measurement equipment. But the values all look somewhat reasonable.
|Voltage [V]||1.2 – 15||3 – 12||1.2 – 3||1.2 – 5|
|Voltage typ. [V]||3 – 12||3||1.2-5|
|Voltage max [V]||15||12||3.5||5|
|Current max [uA]||200||120|
|Demodulated signal||inverted and none inverted||none inverted||inverted||inverted|
|Output Current max||2 mA||5 uA||5 uA|
|Bandwidth (Antenna) [Hz]||700||600|
|Temperature [C]||00 – +50||-40 – +85||-10 – +70|
|Sensitivity [uV/m]||25 (FTD02011R)||80||50|
|Sensitivity [uV]||0.4 – 0.6||1|
The HKW antenna comparison is based on the HKW datasheets. Notice that they do not really state how “output” is measured. So this can be only used to compare HKW antennas with other HKW antennas. However it clearly shows how size affects antenna performance.
|Size [mm]||10 x 40||10 x 60||10 x 100|
Final conclusion: size matters – bigger is better 😉