Temperature data spikes revisited

[Gina]

OK Mark

Started testing... 10:00 10pF I'm expecting to see spikes...

[Gina]

None so far...

One thing has come out of my research into 1-wire cabling is that screened (STP) cable has a very much increased capacitance over UTP and isn't recommended for 1-wire. So it could well be that UTP is best for the FO wind sensors too.

[nking]

None so far...

One thing has come out of my research into 1-wire cabling is that screened (STP) cable has a very much increased capacitance over UTP and isn't recommended for 1-wire. So it could well be that UTP is best for the FO wind sensors too.

That's good as it's cheaper

[Gina]

Certainly is

12:00 had a spike with 10pF and now changed to 100pF.
12:25 had a spike with 100pF and now changed to 1nF. Testing continues...

Pulse wave front on anemometer.

AWS_Waveform_Anemometer_1nF.png

[Gina]

13:30 had spikes with 1nF and now changed to 10nF. Testing continues...

Anemometer cycle time for about 10 mph measures 135ms. I make that 7.4 revolutions per second.

14:15 had spike with 10nF and now changed to 100nF. This is about the top limit for wind speed measurement.

AWS_Waveform_Anemometer_100nF.png

Rising edge of anemometer waveform reaches 1.5v (half supply) after 12ms this makes the maximum rotation rate that it will measure to be 8 rps by my calculation. This corresponds to just over 100mph.

[beteljuice]

Gina your expertise and logic are beyond question, but ...

I have been concerned that your tenacity is more like OCD

[Gina]

I'm getting near the point where I've had enough I must say! But I think I'm about done - 100nF may be the optimum - more will limit the available wind speed range and less may result in spikes. I may try using a waveform generator to check the effective wind speed at the point where the microcircuit fails. Mind you, I am giving it a very severe test - there's a coil of 250ft of CAT5e UTP cable, plus a few feet of telephone cable between the anemometer and the transmitter unit. I can't imagine anyone wanting to separate their wind sensors and tx unit by anything like that distance Maybe I should test a shorter length...

For my own use, I intend cutting off something like 25ft - enough to reach from the top of my mast to the bottom and then about 5ft to the Stevenson screen. (I'm thinking of threading the cable through the hollow mast.)

As a temporary measure I might try terminating my present STP cable just at the bottom end - I'm not taking a 20ft mast down in this freezing cold weather!

[mcrossley]

I think most people will be like you, putting the transmitter somewhere near the base of the mast, so perhaps something like 50' may be a more realistic test? Next you could try putting the rain counter on a long wire too, and start another thread? I'll get my coat!

[Gina]

I think most people will be like you, putting the transmitter somewhere near the base of the mast, so perhaps something like 50' may be a more realistic test?

yes, I agree.

Next you could try putting the rain counter on a long wire too, and start another thread? I'll get my coat!

Well, I think the rain gauge would be less likely to pick up interference if it were connected using twisted pair cable.

[Gina]

As a temporary measure I might try terminating my present STP cable just at the bottom end - I'm not taking a 20ft mast down in this freezing cold weather!

Did that this afternoon - didn't work - still have spikes. So it will have to be UTP cable properly terminated (when the weather improves).

Meanwhile, I'll cut off a length of CAT5e UTP cable and set it up for a test run overnight. I'll reduce the capacitance as well. 100nF is rather critical for high winds (I would like to maintain 100mph max). 47nF should be fine - a 2:1 safety margin. But to check for a margin regarding spikes I'll try 22nF.

[Gina]

Cut off a length of CAT5e UTP cable (something over 10m/30ft) and connected that in instead of the full 250ft. Changed capacitors to 22nF and resumed testing from about 21:15. So far so good.

test-Temp-only.png

[Gina]

Overnight test was a complete success

test-Temp-only6.png

[nking]

It looks like you've cracked it . So can we conclude that it's all down to the cable one uses when extending the link between wind sensors and the transmitter?

[Gina]

It looks like you've cracked it . So can we conclude that it's all down to the cable one uses when extending the link between wind sensors and the transmitter?

Yes indeed, so it would seem Final proof will be when I install the CAT 5e UTP cable outside. It will be down the outside of the mast initially then when it's a lot warmer and dry I may see if I can thread it down inside the mast. ATM the cable is in a coil on the table indoors. I'm going to take the wind vane unit apart and solder the CAT5e pairs directly to the circuit board together with the R & C termination (though I'm not totally sure this is necessary).

Since the weather is too cold for messing about with 20ft masts I might put the test system outside on a fence post with the cable stretched out along the fence.

Now to my next project... 1-wire

[Gina]

Wind Vane PCB with CAT5e UTP cable soldered to original cable with RC termination (100 ohm resistor plus 47nF capacitor in series) soldered across anemometer connection.

Wind-Vane.jpg