Version: 0.3
Author: Bernard PH1M
Last change: 2025-04-15
The Vaisala RS41 radiosondes are reused by radio/balloon enthusiasts. The RS41 is designed to be used just once and for short flights. So, when the sonde is used in another setting, some problems may occur.
One of the issues is that the sonde doesn't restart after a voltage dip. If the voltage drops below a threshold (1.15V) the sonde will turn off and will keep off, even when the voltage goes back to 1.75V. This will happen when solar panels are used or the battery voltage goes down and up due to (extreme) temperature changes.
Another issue is the large (inrush) current that occurs briefly after the sonde is turned on. Both problems are solved when the modifications, which are written down below, are done.
Note that this is just an experiment. There are maybe better ways, and some component values are chosen because they were simply available at the workbench.
For removing the components, I used a hot air station. The use of Koptan tape is also recommended.
After the modifications are applied, the RS41 will turn off at around 1.17V and turn on again if the voltage rises to 1.75V. I think it is possible to bring the last value down to change some values of the components. This will be written in the next version.
A large part of the RS41 circuit consists of a switch network that controls a MOSFET which turns on and off the power to a boost converter. This circuit is designed to turn the RS41 on once by pressing the button or by NFC, and off by the MCU. It is not designed to turn it on again.
The modification is to get rid of this circuit and just connect the power to the boost converter
Unfortunately, there is a catch, so some components should be added.
Let's get started!
In the image below (Figure 1), you see in red the components to be removed and in green the to be added bridge.
Also in red are the two traces indicated, which should be cut.
In the picture below (Figure 2), you see that the MOSFET with some related components is removed. In green is indicated the to-be-made connection between the Source and the Drain.
Figure 3 shows the soldered wire between the drain and the source.
At this point everything should be fine because the MOSFET is gone, and the boost converter always gets power, and it should be turned on again after a power dip, right? Unfortunately, not because the boost converter is not happy.
I put power (3V) to the RS41 by a lab power supply, and the current limiter kicks in directly. Did I make a mistake? So I did check for a short between + and -, but it was not there. I checked the board with a thermal imaging camera for hot spots. Only the boost converter and the inductor were hot, but no other components. (See Figure 4) So there should be something wrong with the boost converter itself.
After some search for possible causes, there was a comment that there can be a problem when the boost converter is already turned on even when the voltage is not settled. (can be caused because of the big inrush).
The boost converter has an enable pin (EN), which is directly connected to the power. So even when the voltage is not settled, the boost converter is already turned on.
A possible solution is to add an RC delay network to the enable pin. This will turn the boost converter on after a delay, which is depending on the values of R and C. I did use for the resistor a value of 47 kOhm and 100 nF for the capacitor. The R and C are placed in series. A resistor (with a value of 47 kOhm) is placed in parallel with the capacitor, so it will empty itself when the voltage is going down. Other ways the restart of the sonde will fail after a voltage drop, because the boost converter is enabled directly when the voltage is going up because the capacitor is still filled, hence the enable pin is never turned off. I did try to replace resistor R2 with a diode, but this did not work. I think that the current situation where the resistors act as a voltage divider, which brings the voltage at the enable pin down, works fine. Furthermore, I did try different values for R1, 10 kOhm and 30 kOhm, but this didn't work out. Maybe a value around 40 kOhm is best. If you have some ideas, please contact me at bernard _at_ ph1m.nl
Below you see the trace to the enable pin is cut.
The trace from the enable pin is also cut, so it is not connected to anything.
In red is indicated the resistor to be removed, so the not-needed switch logic is disconnected. The two resistors on the left are still connected because those are used for measuring the voltage.
In figure 8, you see the resistor removed.
Figure 9 shows the bridge, capacitor and resistor placed as indicated in Figure 5. The resistor in parallel is not placed yet. As a bridge, I used a 0 ohm resistor, but it can also be a wire, if desired.
And now the resistor is also added on top of the capacitor.
Figure 11 shows the components that are part of the not-needed switch network to be removed.
And the components are gone.
The button is not needed anymore, so let loose some weight.
The button is removed.
Congratulations! The modifications are done. If everything did go well, the RS41 will turn off at around 1.17V and turn on again if the voltage rises to 1.75V. As mentioned before, this last value can possibly be less, if some resistor values are changed.
Below are some modifications that will turn down the used power.
The RS41 sonde contains resistors that are used for heating. Those can be turned on and off by the microcontroller. But the firmware should be aware of this. I saw some versions where the heaters were still on and it would use much-needed power.
All the components can be removed. You can simply break the piece off, but extra care should be taken to prevent shorts between the traces.
The components on the front are removed.
The components at the back can also be removed.
The components at the back are removed.
For amateur usage, only the MCU, radio and the GPS receiver are needed. The logic for measuring the humidity and temperature can be turned off by removing the corresponding LDO.
Figure 21 shows the Low Dropout Regulator (LDO) removed.
As a reference, I used the reverse engineering of https://github.com/bazjo/radiosonde_hardware/tree/master/Vaisala_RS41.