Wind Speed

[wd40]

I was doing a bit of rereading the manual that came with my Fine Offset, logo'ed Ambient Weather WS-2080. This time I paid more attention to the wind section.

The manual reads. "Average Wind Speed and Wind Gust. Press the up key or down key to alternate the display between the three minute Wind Average Speed and Gust Speed (or the maximum wind speed in the update period)."

I assume the three minute average speed calculation is continuous and not performed once every three minutes.
Anybody know if that is correct or not?

Randy

[steve]

That's quite interesting. That's the first time I have ever seen an 'official' mention of the wind averaging period on a Fine Offset. My manual, and others I have seen, don't quote a period. I (and I believe most others here) have always assumed that the average was over the 48-second update period. Possibly this is a change in the firmware for these newer models?

But to attempt to answer your question, I would assume it is a 'rolling' average, and doesn't change abruptly every 3 minutes.

[wd40]

The three minute running average makes sense and makes the wind speed number a little more palatable I would think
I checked the model they label as a Ws-1080 which is the one with a touch screen instead of buttons, its manual also shows the three minute average.

The guy that owns Ambient Weather went to a lot to rewrite the manuals as I understand it from my reading.

The manuals are on line under the FAQ section for each model.

http://ambientweather.wikispaces.com/

Glad I could supply some useful information.

Randy

[steve]

I checked the model they label as a Ws-1080 which is the one with a touch screen instead of buttons, its manual also shows the three minute average.

Well, that's definitely the same as the ones most people have (including me), so if that's correct they all have a 3 minute average.

[jim-easterbrook]

Interesting. 3 minutes is not an integer multiple of 48 seconds (the data update period) so I suspect the true value might be 3.2 minutes, i.e. an average of five readings.

I wonder if this can be tested by looking at the data from the weather station? The data I get seems quite 'spiky' for something that has been averaged. If I could show that mathematically such a sequence requires negative inputs to the averager, I'd have proved it isn't being averaged. Some work to do there.

[steve]

I wonder if this can be tested by looking at the data from the weather station? The data I get seems quite 'spiky' for something that has been averaged.

A quick test I thought of doing - and would have already if it hadn't been raining so much today - would be to take the console out to the anemometer, wait for an update, then stop the anemometer (my Fine Offset anemometer isn't sited very high), and then watch the subsequent updates, seeing how long it takes to get to zero.

[steve]

It's stopped raining, so I was able to carry out my experiment. The displayed wind speed had been changing between 5.4 and 6 mph for a while, and as soon as it updated, I stopped the anemometer, and waited. After the expected 48 seconds, the display changed - to zero. I released the anemometer, and after 48 seconds, it returned to 6 mph.

So I can say that my station (Maplin N96GY, which is either a WH1080 or WH1081, I can't remember which) definitely has a 48-second averaging period.

[wd40]

I will do the same to mine and see what happens.

Interesting test Steve.

I have thought about using a hair dryer to set a speed then see how it handles a constant wind. I can measure the rpm of the anemometer to have a reference.

Maybe be able to give it a go this weekend.

Randy

[wd40]

I ran the same test as Steve and got the same results as Steve.

It could be though the station has logic added if zero then zero bypassing the average to prevent customer complaints about zero wind speed when the it stops but reports the average. On one test I did delay stopping the anemometer for a few seconds anf the MPH fell way down so the 48 sec avg might still be the answer.

[dane]

I ran the same test as Steve and got the same results as Steve.

It could be though the station has logic added if zero then zero bypassing the average to prevent customer complaints about zero wind speed when the it stops but reports the average. On one test I did delay stopping the anemometer for a few seconds anf the MPH fell way down so the 48 sec avg might still be the answer.

Why don't you write the guy at ambientweather and ask him if he has documentation for the 3-minute period?

[wd40]

I did get in touch with them and they said the manual is wrong will updated to show the avg wind speed is calculated over 48 seconds.

Sorry for the miss direction.

Good quick test Steve to get this straight.

Randy

[TideMan]

When the machine calculates the average speed, does it just average the number of turns per second over 48s (with a suitable calibration), or does it calculate the vector average (i.e. including the direction)?
And if it just averages the speed, how does it figure out the average direction, especially around the northern segment where direction goes from 348.75 deg through zero to 11.25 deg?

[steve]

It just uses the number of turns. It doesn't do an average direction, it just gives a 'current' direction.

[Gina]

I don't think it does anything "clever". I don't know but I would think it just takes the number of revolutions in the 48 sec period for speed. And I'm sure the direction is simply that at the 48s sample point. However, I believe the console microprocessor can average the speed. From the way the direction jumps about, I would say you just get the direction when the data is sampled eg. 5 mins. If the direction were to be averaged over the 5mins you'd get a much better result.

EDIT - Cross posted with Steve.

[TideMan]

If the direction were to be averaged over the 5mins you'd get a much better result.

And how could you do that average without doing the vector average?
For example, say we had a North wind whose direction fluctuated between 348.75 and 11.25 deg from true N, how would you take the average of that?
That was the point of my question.

And Steve's answer is the only simple way: number of turns and current direction.