Anemometer/ Frequency Counter

Moderators: grovkillen, Stuntteam, TD-er

Post Reply
Message
Author
Ton_vN
Normal user
Posts: 132
Joined: 21 Oct 2016, 15:20
Location: Hengelo (Ov)/ NL
Contact:

Anemometer/ Frequency Counter

#1 Post by Ton_vN » 20 Apr 2019, 18:54

An electronic anemometer for windspeed quite often applies a 4cup-rotator with a contact-closure/revolution between it's interface-wire_A and interface-wire_B.
Cleaning junk, as 'left-over' found such a device (without any documentation), and ;-) that invites for experimentation .....
ESP8266_Anemo.jpg
Anemo_ESP8266
ESP8266_Anemo.jpg (18.84 KiB) Viewed 638 times
Digital interface
Connected the device with wire_A to port D4 of the WEMOS_ESP8266 (with internal pullup resistor+LED), and wire_B to GND.
At port D4 now a pulse-train appears with interval/frequency as function of windspeed hopping between VCC=3V and GND:
The LED pulses, and (under setting 'Pulse Counter / Delta') the 3 counters increment => interface function is OK!

Analogue interface
At wire_A connected parallel a very basic signal-integrating circuit to feed the ADC at port A0:
- using D4 as 'source' for the anemometer, the voltage from the anemometer will be never higher than the input-range for port A0
- a small condensator C1 for galvanic separation from port D4 and to provide pulse-signals at the edges of the pulse-train
- C1 through a diode feeding positive pulses to a bigger electrolite condensator C2 which has other side at GND
- the + of C2 is connected to A0 [perhaps a parallel resistor to GND may be useful for quicker discharge of C2]
The output of the circuit and the reading by the ADC show variation as function of windspeed => basic interface function is present!

Alignment/Calibration
Have 2 'real' meteo-anemometers, and therefore reference-info should not be a problem.
Real calibration may be a challenge, but some alignment of this setup would be nice.
Using 'formula' and/or 'rules' that must be achievable in the ESP8266 itself [I am using ESPEasy R147 at this (also left-over) WEMOS-board].
The A0-interface is the least complicated:
- just comparison of the analogue readout at A0 vs. the references, and deduct a calibration-factor as multiplier for %value% from A0.
The D4-interface is a questionmark:
- first need to get from the counter-data, either Pulsewidth-data PW, or Frequency-data F.
Starts with the question: what is represented by "Count/Time/Time"? No description in the Wiki!
- once found, then also just a matter by comparison of finding the suitable calibration-factor
But what is simplest method with 'formula' and/or 'rules' to get PW or F?
Examples not found by search at this Forum.
Somebody having a proven example?
Last edited by Ton_vN on 26 May 2019, 10:40, edited 16 times in total.

User avatar
budman1758
Normal user
Posts: 232
Joined: 15 Apr 2017, 05:13
Location: Riverside CA USA

Re: Anemometer/ Frequency Counter

#2 Post by budman1758 » 20 Apr 2019, 19:15

This thread may help with an analog output type. Works quite well.

viewtopic.php?f=4&t=3211&p=16742#p16742
"The glass is twice as big as it needs to be".

Ton_vN
Normal user
Posts: 132
Joined: 21 Oct 2016, 15:20
Location: Hengelo (Ov)/ NL
Contact:

Re: Anemometer/ Frequency Counter

#3 Post by Ton_vN » 23 Apr 2019, 09:06

@Budman

Looking at the description of your speedmeter, it could well have inside more or less the integration-circuit which is now on my breadboard-setup:
feeding with 12V~24V for strong pulses from the rotator-contact, plus application of some electronics, it is easy to get an output of 0~5V [full scale].

As stated, for me this is an experiment with a 'found/recycled' windspeedmeter: 'Empirics' rule ......

Using port D4 at the same time for measuring pulses, I have (without effort) practically limited the max. voltage available at the rotator-contacts to 3.3V.
With the present configuration of components at port A0 only max. 10 out of 1024 units from the ADC ~ 30mV, which is almost equal to 'noise'.
For stronger flank-pulses C1 must be empirically increased.
At the same time a very small C over the input port between wire_A and wire_B, to suppress high-frequency noise.
Input level for port A0 best between approx. 300mV to 1000mV at 'normal' windspeeds, with 2000mV or more for fierce gales, and 3000mV as extreme full scale:
then surely a discharge-resistor required over C2, because the ADC alone probably is not able to discharge C2 quick enough.

;-) No prescriptions, just experiments ......
Last edited by Ton_vN on 26 May 2019, 10:41, edited 4 times in total.

Ton_vN
Normal user
Posts: 132
Joined: 21 Oct 2016, 15:20
Location: Hengelo (Ov)/ NL
Contact:

Re: Anemometer/ Frequency Counter

#4 Post by Ton_vN » 25 Apr 2019, 10:52

Suspicion is just a feeling, facts count ......
But in this case the data & graphs confirm my suspicions:
the various sensors have a ;) 'very different' interpretation of the actual winds.
For this graph Accuweather acts as ExternalReference, but it's speed-info is surprisingly lower than the other sources.
piwind1_20190425 [640x480].png
Windgraph_5_Speedsensors
piwind1_20190425 [640x480].png (153.19 KiB) Viewed 574 times
Thinking exercise to get everything aligned towards 'real, practical' values .......
Perhaps I have to consider inclusion of Gust-info to complete the picture:
locally only available from TFA_Nexus, and externally from Accuweather, DarkSky etc.
Last edited by Ton_vN on 02 May 2019, 10:45, edited 1 time in total.

Ton_vN
Normal user
Posts: 132
Joined: 21 Oct 2016, 15:20
Location: Hengelo (Ov)/ NL
Contact:

Re: Anemometer/ Frequency Counter

#5 Post by Ton_vN » 25 Apr 2019, 14:19

Inserting also Gust-info in the same graph is only confusing.
Therefore a separate graph, which is easy enough with RRDTool.
piwind2_20190425 [640x480].png
Windgraph2_5_Gustsensors
piwind2_20190425 [640x480].png (149.76 KiB) Viewed 557 times
The WS7000_anemometer doesn't provide windgust-info (just 'total wind'), therefore 0 in this graph.
For the Conrad-anemometer no software yet to calculate windgust, therefore same (speed)info as in Graph W41.
The info from Accuweather again is surprisingly lower than the local measured values.

Not solving the puzzle how to align those devices ..........

Ton_vN
Normal user
Posts: 132
Joined: 21 Oct 2016, 15:20
Location: Hengelo (Ov)/ NL
Contact:

Re: Anemometer/ Frequency Counter

#6 Post by Ton_vN » 08 Sep 2019, 21:07

Larger condensator C1 improves the signals in the analogue circuit feeding WEMOS' ADC-port A0, but with negative effect on the edges of the pulses for interface-port D4.

Lessons-learned:
With this simple analogue circuit the readings from the ADC are rather ambiguous, without added value relative to the digital read-out.
Therefore the analogue circuit&read-out now removed.
Experiment finished with these results.
Concluding :( that it is most difficult to align different anemometer-systems to more-or-less agree on windspeed and windgust ........

Post Reply

Who is online

Users browsing this forum: No registered users and 1 guest