Reparing AURIOL LD3091 weather station

My better half has a weakness for weather stations, and she especially likes this old piece. It’s not particularly smart, not particularly pretty, but it has a “frog”, which makes up for everything. It’s a cheap LIDL model — even the first one we bought didn’t pick up the signal from the external sensor. They replaced it, and it worked. For a while. Then that one stopped picking up the sensor signal too. Since it’s no longer under warranty, I took it apart.

While browsing German online forums, I came across many complaints about these cheap weather gadgets not receiving the signal from the external sensor. Often, the reason for the reception issues is simply that the external sensor is too far from the device. According to the manufacturers, the wireless sensor’s reception range is 30 meters or even 100 meters, but in practice, it’s sometimes barely ten meters. Radio interference (e.g. from a nearby TV set, wireless headphones, etc.) can also cause problems, as well as the presence of multiple walls between the transmitter (sensor) and the receiver (weather station). Ideally, the receiver should have a “line of sight” to the transmitter through a window, but keep in mind that modern insulated windows often have a metallic coating, which can significantly block radio signals. The sensor itself may be the cause of the radio connection disruption. Not only might the battery be depleted, but for example, moisture (rain or condensed dew) could have gotten inside. Alternatively, it may simply have gone out of tune over time.

Regarding the transmitter, the following points should be noted:

• It operates at 433.xx MHz; channel selection is not about choosing different frequencies — all device always transmits on the same frequency (the “channel number” is merely a code in the data packet that identifies the sensor);
• The frequency is not very stable; in many cases, if a cheaper ceramic resonator is used instead of a quartz crystal;
• Both the transmitter and the receiver have quite primitive, relatively wide-band (low selectivity) designs;
• The connection is one-way: the sensor transmits a data packet on its own approximately every half minute and does not receive any response (i.e., the weather station does not request the data—it just listens and occasionally catches a transmission);
• The duration of the transmission is very short (< 2 seconds);
• If multiple devices are operating in one area, they generally don’t interfere with each other because the chance of them transmitting at the same time is relatively low (although sometimes interference occurs, and in those cases, the received data may be misinterpreted, resulting false data displayed).

These gadgets usually contain two receivers. One is for receiving the DCF77 Central European atomic time synchronization signal. This signal operates in Mainflingen, Germany, and it transmits an amplitude-modulated signal at 77.5 kHz. That’s a very long wavelength and a very low frequency (77.5 kHz is almost within the audible range!). A ferrite antenna, which is easy to recognize, is always used for its reception. The other receiver is microwave-based, operating in the UHF band; this one picks up the sensor signal. Its antenna is a piece of wire about 15 cm long, coiled somewhere inside the device housing. If there’s no reception, it’s worth visually checking to see if everything is in order and something has not damaged it.

Josef Gajdysek reported on his blog about his experiences using an RTL-SDR to reverse-engineer the radio protocol used by his home weather station. Josef’s weather station is an ISM band device, and he found that it employs differential pulse-position modulation, meaning that it sends very short pulses and the interval between pulses determines whether a bit is a 1 or a 0. This modulation scheme provides extremely low energy consumption. Josef also wrote a Python script for decoding, which is available for download in his post.

Using an SDR receiver, I also checked and found that the sensor appears to transmit properly. The transmitter cannot be tuned, but the receiver can be tuned quite a bit with a little knob.

Since the sensor transmits very rarely and only for a very short time, it is quite difficult to tune the receiver to it. I suspect that even at the factory, they don’t tune the receiver to the specific sensor, but rather to a measurement transmitter. The measurement transmitter’s frequency is only nominally identical to that of the sensor — in practice, it differs, so the tuning is not precise. However, because the receiver is relatively wideband, it usually works. But this leads to some units being defective, so don’t be surprised if there are models that don’t receive the signal even when freshly unpacked.

Tuning to the specific sensor is time-consuming, and it may be necessary later on as well. That’s why, in a rather vandal-like fashion, I drilled a tiny hole in the device’s case so that tuning can be done even when the unit is fully assembled and installed.

The tuning procedure is as follows:

1. Note the position of the tuning screw in the coil (e.g., at the 10 o’clock position).
2. Power on the sensor.
3. If the receiver doesn’t pick up the signal, turn the screw about 1/12 of a rotation clockwise (to the 11 o’clock position).
4. Wait, wait, wait…
5. If no data is received within about 10 minutes, try in the opposite direction: turn the screw 2/12 of a rotation counterclockwise (to the 9 o’clock position).
6. Wait, wait, wait…
7. If there’s still no reception, repeat the process from step 3, but now shift the tuning by 2/12 of a turn—i.e., turn the screw by 3/12 of a rotation (to the 12 o’clock position in this example).

Keep detuning the receiver little by little until it works… If you need to turn more than a full rotation, then the issue might be something else entirely, not just detuning. But that’s another story.