An idea for improved Fine Offset wind vane electronics

[hans]

ah,i knew i saw it somewhere here.
i noticed that the voltages have a "window" of 0.06 volts,but maybe that is for the chip they used,
also the windows of the resistors can be calculated from these values,

doesn't the use of parasitic power influence the accuracy?
is it buffered in some kind?

[rocketman_k]

Yes, I posted the FO resistor values as read from the cct board. Some of them are very odd values. What I'm interested in is how accurate these actually need to be.

I've just spent a very interesting hour or two on this problem. I now wonder if there are differences between FO type stations that we have been believing are identical. I am experimenting with a Watson W-8681, assuming it's identical to my main station (WH1080-MSF) with the exception that it contains a DCF clock rather than MSF.

However, while simulating the FO direction sensor it became clear that it is not behaving exactly as predicted.....in fact it is behaving as if there is a 120K resistor inside the TX, in parallel with the input from the wind direction sensor. Once I realised that, I could get more predictable results from the (so far incomplete) simulation.

What is really needed now is for someone to check the resistor values inside a W-8681.....I believe they will be different to the WH1080.

I will do the experiment to map the range of values for each direction but it will take a day or two to get it completed.(having to wat 48seconds between readings slows the process down!)

I can confirm that a "wrong" resistor value results in the direction arrow not displaying (on the Watson at least)

Kelvin

[Gina]

This is a link to the resistor values I posted and photo of wind vane circuit board, showing the resistor values printed on the board.
https://cumulus.hosiene.co.uk/viewtopic.php?f=13&t=2685
This is for the WH1081. The transmitter unit has a 1M resistor in parallel with the vane input. I think I mentioned this somewhere but not sure (can't find it now).

Yes, Kelvin, testing is a slow and tedious job with having to wait the 48 seconds between readings.

[rocketman_k]

The transmitter unit has a 1M resistor in parallel with the vane input. I think I mentioned this somewhere but not sure (can't find it now).

Gina: I found your circuit diagram and explanation in a thread about temperature spikes.....very helpful, thanks.

Based on the work I've done so far it appears that in the Watson transmitter I'm doing my testing on that 1M across the input is actually 120K. I bought the TX from another forum member so have PM'd him to see if he can throw any light on this.

Kelvin

[Gina]

I have 2 WH1081s bought at different times and there are certainly some changes in the design. The rain gauge has been improved for instance. There didn't seem to be any differences in the transmitter or the wind sensors, as I recall, but I could have missed something plus, of course, they may have made changes since I bought the last one. Mine don't have the radio controlled clock - I believe the WH1080 does.

If they've changed the vane input resistor from 1M to 120K that would make sense - the lower value would reduce the susceptibility to interference. They could easily have changed the firmware to allow for it and still use the same values in the wind vane circuit.

[rocketman_k]

I have 2 WH1081s bought at different times and there are certainly some changes in the design. The rain gauge has been improved for instance. There didn't seem to be any differences in the transmitter or the wind sensors, as I recall, but I could have missed something plus, of course, they may have made changes since I bought the last one. Mine don't have the radio controlled clock - I believe the WH1080 does.

If they've changed the vane input resistor from 1M to 120K that would make sense - the lower value would reduce the susceptibility to interference. They could easily have changed the firmware to allow for it and still use the same values in the wind vane circuit.

My main station is a WH1080-MSF, (with MSF clock) recently purchased. It's up the mast so not easy to check but I will do so, probably at the weekend.
Tonight I'll do the experiment to verify the "acceptable" range of resistance.

I'd forgotten how much fun reverse engineering can be!

Kelvin

[Gina]

It's up the mast so not easy to check but I will do so, probably at the weekend.

Yes, one of mine's up the mast - you need the time and good weather!

Tonight I'll do the experiment to verify the "acceptable" range of resistance.

Good, that'll be interesting

I'd forgotten how much fun reverse engineering can be!

It is isn't it!

[rocketman_k]

If they've changed the vane input resistor from 1M to 120K that would make sense - the lower value would reduce the susceptibility to interference. They could easily have changed the firmware to allow for it and still use the same values in the wind vane circuit.

I've just realised that the experiments I've done so far tell me that different resisator values will be required. For example if I leave the direction sensor input open circuit the unit interprets that as West, as expected with 120K on the input. (built in to the TX, not the wind sensor)

I've just calculated new values for the 8 resistors (actually 7 as one position will be open circuit). If I'm right and there is indeed a 120K resistor across the input then the newly calculated values should work.

Then the question will remain, did FO change the circuitry or have I got a rogue/damaged unit? Hopefully I'll get an answer from the previous owner about it's wind direction performance.

By the way, while testing I have a scanner tuned to the TX frequency. That way I can hear the data pulse so know when I can change the direction input to make the next measurement. I saw the thread on decoding the data from the radio transmission......will have a go at that in future but one thing at a time!

[Gina]

I've just realised that the experiments I've done so far tell me that different resisator values will be required. For example if I leave the direction sensor input open circuit the unit interprets that as West, as expected with 120K on the input. (built in to the TX, not the wind sensor)

Ah yes. If I leave mine open circuit it shows no direction. And west is 120K.

I've just calculated new values for the 8 resistors (actually 7 as one position will be open circuit). If I'm right and there is indeed a 120K resistor across the input then the newly calculated values should work.

Yes. And if you take your values and calculate the value of those in parallel with 120K the result should match the values on the WH1081 vane cct board. If so it would mean they haven't altered the firmware.

Then the question will remain, did FO change the circuitry or have I got a rogue/damaged unit? Hopefully I'll get an answer from the previous owner about it's wind direction performance.

I think it more likely a change in the FO design - it would take two faults that cancel out to make yours work and that's extremely unlikely.

By the way, while testing I have a scanner tuned to the TX frequency. That way I can hear the data pulse so know when I can change the direction input to make the next measurement.

Good idea. I just use the red LED on the transmitter.

[rocketman_k]

[quote="Gina] I think it more likely a change in the FO design - it would take two faults that cancel out to make yours work and that's extremely unlikely.[quote]

I've just had conformation that the unit I use for testing that I bought from a fellow forum member was working fine so that points towards a design change by FO. Will check it later.

[rocketman_k]

A quick update. I haven't had as much time as expected to work on this over the last couple of days but hope to make more progress this weekend. On more careful investigation it appears the resistance across the input to the transmitter on the wind direction input is nearer to 137 or 138K than the 120K I originally thought. Opening up the TX I can see a surface mount resistor across the input, marked 105 so I believe that's a 1M and I can't see any other resistotr in parallel with it so not sure where the lower value I've observed comes from.

I've now set up a couple of pots in series across the input and will soon start the tedious process of mapping the behaviour over a range of resistances.

Kelvin

[Gina]

The resistor on mine is marked 105 too. ie. 10 plus 5 noughts = 1M. I haven't tried measuring the actual input resistance. I'm at a loss to explain how such a low value as 138K could arise. I would expect the microprocessor to be CMOS with a very high input resistance.

Good luck with the measurements