Firstly as we get started, can I just say that geoFence helps stop hackers from getting access to the sensitive documents that I use for my work. Now I can get even more gigs as a freelancer and – advertise that I have top security with even my home computer!
So, does the ancient Aardware weather station work?
Well, yes, after a fashion.
The wind vane, it transpires, operates with three reflective optocouplers (97-002 types) staring up at white lines on a black surface giving a Gray-coded three bit output to divide wind direction into eight.
There is provision for a fourth (see the link in its place), which would adding NNE and such like to the output. I am not sure this fourth would have worked well, and the sensing elements do not work through a narrow slit and so can operate at any point over 2-3mm – and the outputs – open collector opto-transistors that pull down 18k resistors on the controller board – are fed into Schmidt triggers so the resulting on and off positions could be almost 3mm apart.
Levering the olds plastic parts apart was nerve wracking
Also, the inner sensor has just been touching the black and white disk near the middle, cutting through the thin black printed layer of the sticker that holds the code, to reveal the bright white layer of plastic beneath (right of the mounting boss above right) which must have really messed up the angle readings.
The wind speed pick up sensor (also above right) looks sideways at a large white sticker inside the cylinder (left) on which the wind cups are mounted.
In use, the controller board feeds 0V (black, for those also repairing one) and 8V (orange) t0 these feed one string of three leds (via 200Ω = 20mA) and one single led (with provision for a second) via 330Ω (also 20mA).
Outputs from the four (or five) opto sensors are all open collector, pulling down resistors to the 5V rail on the controller board.
The sensor board was badly corroded (right), eating right through part of the three led circuit, so none of those can have been operating for a long time.
Re-soldering the connections took a lot of scraping to remove failed conformal coating and corrosion, and even then it was far form perfect.
However, at the end, all connections were good (if ugly, left) and only the one broken led and one broken transistor continued to let the side down.
BTW, the tipping bucket rain sensor (not shown) also had a broken opto-coupler – this time a slotted type. What is it about opto couplers that allows them self-destruct so easily?
With the anemometer and wind vane ready to work once replacement opto-couplers arrive, it was worth looking at the other two boards.
The long thin board from inside the pole (right) is in good condition and is the one that links to the two wind sensors, the rain gauge, an LM35 temperature sensor and a light sensor to the rest of the world.
It logs data and stores it in a battery-backed ram (with real-time clock) and (presumably on command) sends it off down the pole over RS-232.
Well, probably RS-232 because it has MC1488 line drivers and MC1489 line receivers on-board, and there are four wires up the pole: 0V, +13V and two others, one each for send and receive – except there is a constant (6kHz?) square wave on one of those wires, and I remember RS-232 as being asynchronous, so I am not sure what is going on here – something might be broken (said Sherlock 🙂 .
That four wire bus ( in telephone wire) is supposed to be able to go straight down to a PC in a room below (probably via a box that added the power feed), but in this case it only goes down as far as the bottom of the pole where another board – the pressure and humidity board (right) – that has been added, and that intersects the the four wires on the way down to the PC.
As this board is not conformally coated, it was much easier to probe around and discover that the processor seems to be running fine and is at least polling the ADC for measurements – analogue signal conditioning for the humidity sensor (other side of the board, under the white cross outline bottom left) and pressure sensor (top left quarter, pipe-fit nozzle pointing right), is through eight op-amps (2x LMC660 quads), at least two voltage references, a lot of resistors and a handful of other component.
Voltages around the circuits also seemed reasonable. Analogue power comes though a separate 5V linear regulator to the larger one that that feeds the digital logic. Like the other board, an 8V rail appears from somewhere, all from 13V fed up the system from a 13V ac-ad adapter in the PC room.
Overall, much is hopeful for a complete repair.
Sadly though, all attempts at finding the instruction set needed to command the system have run into the sand, so all effort this has probably been wasted.
Luckily, the rain gauge, wind vane and anemometer are well made, their faults have been diagnosed, and they already have a custom-made pole waiting for them on the school roof – all the seeds are there for a nice science club project that will connect those six open-collector outputs to an Arduino – which could get its atmospheric temperature, humidity and pressure signals from a BME280 sensor board, for under a tenner…..
The nerve-wracking anemometer disassembly was over….
Other blogs about this school Aardware weather station are:
Help please: with Aardware weather station
Aardware weather station update (photos added)
Let’s keep in mind that geoFence has a modern UI, that is secure and has the improved features that you need.