wind speed

T

technopengl

Member
Good evening,

I wish to make realize to my school kids (14 years old) an anemometer...
Three solutions:
1. With an "opto switch", this solution seems to me the best, but it seems difficult to realize by school kids, the assembly must be very precise...
2. With a mini alternator, but when there is a little wind, the anemometer does not turn...
3. With an ILS, as on the anemometer of a "Oregon" wather station ...
I make tests with one 08M and one 28X1 by using the command "count" (Count 1,1000, w0 )
-> ILS on the input, with a 10k connected on the +

The tests are not good...

" w0 " indicates me wrong numbers of pulses, I was able to observe the result by making turn slowly, I got along well the "tic-tic" of the ILS and the number of "tics" do not correspond to the debug value.
I conclude a problem from it of bounces!
My questions:
1. Has anybody another solution?
2. Did anybody try this last solution with a debounce network?
The idea to place a résisitor (560 ohms) and a capacitor (0,47µf) in parallels with the ILS slows down considerably the response time....
Thank you
techopengl
 
J

Janne

Senior Member
Hi,

I assume ILS is the same as a reed switch?
The RC filter network works quite well with a reed switch. In anemometer it works best if the filter consist of a capacitor and a bleeder resistor. This is because the reed switch only has a short moment to charge up the capacitor at high wind speeds. Pick the values for the components so, that the RC time constant is less than the time it takes for one revolution at maximum wind speed.
 
T

technopengl

Member
thank you!
Ok, i'm just trying with other values of RC.... So I have to take my car and try again several tests!.... tomorrow!
technopengl
 
S

SAborn

Senior Member
I use about a 10 to 1 ratio for the filter, a 1uf cap, and 1k and 10K resistors.

Using DEBUG might not be the best, as debug can mess with some timing intervals (not sure on count) it would be far better to use sertxd and view the data in the terminal window, this way you can be sure everything is operating at the correct speed.
 
erco

erco

Senior Member
Bicycle computers/odometers operate flawlessly using reed switches for millions of closures, typically as high as 12-13 closures/sec. You could literally use a bike computer ($4 Ebay China!) and calibrate it for use as your anemometer using the supplied reed switch. Depending on your anemometer's max RPM, you may want to gear it down.
 
T

technopengl

Member
Hello,

When I put a resistor of 100, the terminal shows "2" pulses while I expect that from it the only one. When I put a resistance of 10, the terminal shows "1" and it's correct.
I indeed understand that more the constant of time is raised, fewer bounces will be taken into account.... I don't understand then why with 100 the pulse is doubled. I have one magnet and one ILS (reed switch)

main:
count 1,1000,w0
sertxd (#w0,cr,lf)
goto main

Thanks

technopengl
 
T

technopengl

Member
I made a mistake in the cabling. I had left my resistor of 10k (pull up) connected on the input.... I am crazy, I drank too much...
or in french :
-> je suis une buse
-> je suis une quiche

I put 10K and 1µF and the result seems to be correct....

Thank you
 
S

SAborn

Senior Member
That French wine can cause many problems, good to see you got it sorted ( the circuit not the wine that is)

I dont read French but it looks like bad language to me, and i speak that often when things dont work, or translated to French the F!*&%# thing, &*%$#!.

I have built several aneometers from parts salveged from old bubble jet printers as they have very good optical encoders that give around 200 pulses pre rev, and a high degree of accuarcy for low wind speed.
 
J

Janne

Senior Member
Optical, hall effect and sensors of the like are in some views superior to reed relays.. no bouncing problems, better accurary, no mechanical parts to wear out. But reed relays have one very big advantage. They don't need an external power source to work, which can be a saver for example if you're short on signal leads on your underground cabling to the wind turbine (like was on my case..) It is also better in terms of lightning protection, the reed relay is easier to isolate from the rest of the system.

The biggest trouble with reed relays ultimately in this application seems to be the life expectance of the thing. Without careful design you can except the relay to die in a few months of time. Without any protection for the relay + a long cable run(load was 500µA in the form of an optocoupler) I had 2 relays die in about 4 months. There are quite a few application notes circling around about extending the life of the relay with various snubber circuits etc, but I don't know yet if they will add to the life with any significant amount.
 
S

SAborn

Senior Member
Janne,

Have you paid any attention to the mounting of the reed switch and the position of the reed itself, its worth taking a careful look at the reed and mounting it so the flat of the reed faces the magnet, reeds will normally work in most positions but tend to fail early when the flat of the reed is not positioned towards the magnet.
 
J

Janne

Senior Member
Pete,

I've been burning up the reed relays on the inspeed vortex anemometer. (http://www.inspeed.com/anemometers/Vortex_Wind_Sensor.asp)

The reed relay is mounted vertically towards the magnet pole (one leg facing directly towards the magnet pole, the other directly away from it). The magnet passes directly over the switch. Do you think this could affect the life of the poor switch, and changing the orientation of the reed would help prolong it's life? The datasheet for the reed switch doesn't really say anything about the effects of position to the expected life.

View attachment ORD211.pdf
 
Last edited:
erco

erco

Senior Member
SAborn said:
Janne,

Have you paid any attention to the mounting of the reed switch and the position of the reed itself, its worth taking a careful look at the reed and mounting it so the flat of the reed faces the magnet, reeds will normally work in most positions but tend to fail early when the flat of the reed is not positioned towards the magnet.
Click to expand...
Good call, SAborn. Properly used, reed switches are quite long-lived. The magnet orientation and travel path past the switch is critical and not at all intuitive. Best path is usually quite a bit off-center from the switch (not relay) and is best determined by experimentation. Find the "sweet spot" that works best for you with the smallest magnet.

Switch life is related to closure frequency, so gearing your anemometer down for lower RPM at the magnet should increase its life proportionally.
 
P

papaof2

Senior Member
Some washing machines use a water flow meter that counts something on the order of 800 pulses/gallon - done with a magnet in a plastic paddle wheel and a very small reed switch in the housing. The designer obviously expected tens of thousands of operations of the switch and at fairly high rates - several gallons/minute with typical U.S. city water systems.

I've used one of those flowmeters to measure the amount of water a small pump could transfer, using a PICAXE to count pulses (via interrupt), then compute and display the gallons transferred.

John
 
S

SAborn

Senior Member
@janne

I have had a lot of people complain that reed switches fail quickly, but a lot of commercial equipment use them so it must be the way they are mounted in relation to the magnet.

The reed relays are normally sealed and you cannot see the reed, but the glass bead reeds are see through and you can see the little flat reed inside, as the reed is a flat strip of metal and sideways stress will deform of fracture the reed so it should be positioned so the magnet pulls the reed up and down on the flat as it would in any switch.

I have the Davis commercial aneometers here and their reed is mounted with the flat towards the magnet.

Pete.
 
I

inglewoodpete

Senior Member
There are literally millions of reed switches in use in telehone exchanges today, installed in the 1970s, 80s and 90s, before the technology moved to digital multiplexors. Most of these are still in use today - quite reliably.

When using reed switches, it is critical to keep within the specifications. A lot of care needs to be taken when connecting them to inductive loads. The high back-emfs and associated instantaneous currents will weld the contacts together. Also, overly strong magnetic fields can magnetise the switching elements, causing them to latch up.
 
Last edited:
S

sedeap

Senior Member
Littles devices...

Ehmmm...
Maybe is a dumb question, but... one reed relay, can be used as relay and as a reed switch in the same device (not at the same time maybe) or not at all?

Just curious though...

 
I

inglewoodpete

Senior Member
I'm not 100% sure of what you mean but you can have more than 1 source of the magnetism that operates the reed. However, you need to thoroughly test your setup to prove that it will do what you want.

Just like electricity which can flow in 1 of 2 different directions, the magnetising force can be in either of 2 directions. These fields will interact, causing the reed to release at times. You will also notice that a reed will operate, release and then operate again as it passes a magnet, due to the magnetic field reversing.
 
hippy

hippy

Ex-Staff (retired)
SAborn said:
I have had a lot of people complain that reed switches fail quickly, but a lot of commercial equipment use them so it must be the way they are mounted in relation to the magnet.
Click to expand...
Reed relays seem to have a MTBF of closures ranging from below 1 million to above 1 billion depending on make. From the link by janne, 1 closure per second = 2.5mph, so with a billion closures rated switch, consistent 25mph wind ( 10 closures per second ), that should give 3 years worth of use and probably more. A million cycle MTBF may however only give 30 hours.

Mounting orientation could put additional stress on the switch which could lead to premature failure as could heavy or inappropriate load switching. Used as a normal PICAXE input switch the load would be minimal and should be non-damaging.
 
You must log in or register to reply here.
Top