I also think Al's math is incorrect. My frequency counters measure
intervals to nanoseconds, better ones can to picosecond resolution,
these are standard features on better frequency counters. So one can
measure rotational speeds to many digits, with a one pulse per
revolution or 1000 pulses per revolution of the cups. Since the
physical measurement is the time period between pulses, converting to
speed, which takes simple arithmetic, will be the "average wind speed
between pulses" and this can be averaged over any number of pulses.
The practical limits are more likely the jitter of the pulses being
measured and that can come from a lot of sources - electrical noise,
mechanical vibration of the transducer, inaccuracies in the angular
division scheme, etc. I also wonder if the rotational speed of the cups
is precisely linear with wind velocity and suspect a calibration factor
is probably needed. Also, as others have mentioned, the rotational
inertia of the rotor is a limiting factor re measuring very short
gusts. However, since tower structural calculations are based on "3
second gusts" that is some guidance re what makes sense to measure. I
agree with the post that suggests about 2 to 4 Hz total system bandwidth
is probably enough (ie 0.25 to 0.125 second sample period to keep
Nyquist happy and a rotor inertia to match).
Arduinos can be programmed to measure to about 0.5 microsecond
resolution, so there are big bucks NOT solutions.
Grant KZ1W
On 4/29/2014 5:08 PM, Al Kozakiewicz wrote:
I'm not going to belabor the point. Wind speed is derived by measuring the
change in rotational position divided by time. The shorter the sampling
interval (time), the lower the measurement accuracy. There is no reason for
this to controversial.
I'm not the one who insisted that measurements had to be instantaneous, only
the one who pointed out the inherent problem in attempting such with a rotating
instrument that is a slave to the time domain.
If gusts lasting less than the refr3esh rate of the instrument are really that
much of a concern, then maybe the tower shouldn't be extended. If a guyed
tower, it's not information that you can do anything useful with
anyway....except maybe run away!
Al
AB2ZY
-----Original Message-----
From: TowerTalk [mailto:towertalk-bounces@contesting.com] On Behalf Of David
Gilbert
Sent: Tuesday, April 29, 2014 1:36 PM
To: towertalk@contesting.com
Subject: Re: [TowerTalk] Wired Anemometer for Tower?
No ... Al is not correct in his assessment, at least not at a practical level. The 12.5%
"inaccuracy" he claims is purely arbitrary ... it's a 12.5% inaccuracy for one
revolution, which is meaningless when you're trying to measure speed and not rotational
position. And at speeds of interest it's only an inaccuracy for the very briefest
interval of time ... 12.5 milliseconds at 25 mph. After that you inherently get an
averaging effect that greatly improves the accuracy.
You guys need to step back and actually think about what you're talking about. It does
NOT take "big $" to measure wind speed reasonably accurately for our purposes
here. The most relevant uncertainty I can think of in the case of the Vortex anemometer
is the effect its rotational mass has on it's ability to respond to CHANGES in wind speed.
Dave AB7E
On 4/28/2014 10:58 AM, Kim Elmore wrote:
Al is exactly correct in his assessment. Meteorologists (like me) worry about
this stuff all the time. If we want to measure the turbulence spectrum down to
the eddy dissipation rate, we need extremely high resolution and accuracy. That
means big $ instrumentation.
Kim N5OP
"People that make music together cannot be enemies, at least as long
as the music lasts." -- Paul Hindemith
On Apr 28, 2014, at 12:39, Al Kozakiewicz <akozak@hourglass.com> wrote:
I don't think it's misleading, but you'd have to look at my prior post.
If you're going for a "fast" measurement technique, you can measure the time
between two pulses. If you do that, the absolute position uncertainty is +/- 12.5% of a
revolution - based on the very nature of having only 8 pulses per revolution.
On the other hand, if you count pulses over a fixed period of time, the
measurement uncertainty drops (logarithmically, I suspect) with the number of
revolutions in the sample. It is not constant, but +/-12.5% of a revolution
added to the entire sample, which (without doing any math) gets down to 4% in
just a few revolutions.
All I'm pointing out is that with a digitized anemometer you cannot have both
accuracy AND high measurement speed at the same time. For accuracy you either
add more resolution to your transducer or sample over a longer period of time.
No free lunch here!
Al
AB2ZY
________________________________________
From: TowerTalk [towertalk-bounces@contesting.com] On Behalf Of David
Gilbert [xdavid@cis-broadband.com]
Sent: Monday, April 28, 2014 12:35 PM
To: towertalk@contesting.com
Subject: Re: [TowerTalk] Wired Anemometer for Tower?
Sorry, the calibration factor is 2.5 mph per revolution per second. I
got sloppy when I typed that, although I would have thought that was
almost intuitively obvious.
But your +/- 12.5% assertion is greatly misleading. If indeed the
2.5 mph per Hz factor is (as InSpeed claims) accurate to +/- 4%, then
the absolute accuracy is degraded further only by the accuracy of
whatever is measuring the length of time between pulses (which in most cases is
negligible by comparison). The 12.5% figure would apply to the
accuracy of measuring ACCELERATION of the wind speed, but not the
wind speed itself ... at least not unless somebody was silly enough
to care about wind speed accuracy within, say, a 6 msec window (160
pulses per second at 50 mph). I'd bet that the mass of the cups
dominates within intervals of that magnitude.
73,
Dave AB7E
On 4/28/2014 7:12 AM, Al Kozakiewicz wrote:
but at 2.5 mph per revolution and 8 pulses per revolution
You're missing a time constant there. 2.5 mph per revoution per ????
A single revolution or a single pulse has a measurement uncertainty of +/- 12.5
% with 8 ppr. So if you're going to measure velocity by measuring the time
between two consecutive pulses (the smallest possible sample), your accuracy is
never going to be better than that. But response time will be fast.
Al
AB2ZY
________________________________________
From: TowerTalk [towertalk-bounces@contesting.com] On Behalf Of
David Gilbert [xdavid@cis-broadband.com]
Sent: Monday, April 28, 2014 1:26 AM
To: towertalk@contesting.com
Subject: Re: [TowerTalk] Wired Anemometer for Tower?
Not to pound the InSpeed drum, especially since I have no connection
to them at all other than as a long ago customer, but InSpeed also
sells a high resolution anemometer for $89 that puts out 8 pulses
per revolution. The response of either Vortex version is of course
limited by the time between pulses, but at 2.5 mph per revolution
and 8 pulses per revolution for the hi-res version the ability to
almost instantly track wind gusts would be limited primarily by the
mass of the rotating cups (no spec that I am aware of) ... and they are very
light weight.
73,
Dave AB7E
On 4/27/2014 9:23 PM, Kim Elmore wrote:
I can't speak to the Pro2 model, but any anemometer must average over a period.
I was incorrect about the Vue: it yields 4 s averages in each packet. Sonic
anemometers average over something like 10 pulses. Depending on the brand,
that's anywhere from 1 s to 0.1 s. So, if you really need wind speed resolution
better than 2 or 4 s, you need to obtain a sonic anemometer. Because of the
mass of our antennas and towers, I'd be surprised if they have a significant
response to frequencies much higher than 0.25 to 0.5 Hz, but to know that
requires an engineer's analysis.
Low-end sonic anemometers can be had for a few thousand dollars. They can,
however, require periodic maintenance due to bird pecks (birds peck at the
transducers) and bird droppings that contaminate the transducer surfaces.
If you must have higher resolution than this you need very deep pockets indeed
and a lidar anemometer.
As for radar: the NEXRAD (WSR-88D) requires 5 min to complete a volume scan,
but the latency for each radial of data is on the order of a second or so. So,
it's available in real time. Thus, there's no significant latency once a
volume is complete. It does take a few seconds to move the antenna back to the
0.5 deg elevation and start the next volume, but it moves pretty fast in
elevation -- several degrees per second.
Kim N5OP
"People that make music together cannot be enemies, at least as
long as the music lasts." -- Paul Hindemith
On Apr 27, 2014, at 22:29, "tomkd8deg" <kd8deg@centurytel.net> wrote:
The unit I have is the Davis Vantage Pro2 wired, and the 4 second delay is what
Davis tech support told me when I questioned the accuracy of the wind readings
that were being displayed.
Tom KD8DEG
From: Kim Elmore [mailto:cw_de_n5op@sbcglobal.net]
Sent: Sunday, April 27, 2014 8:08 PM
To: tomkd8deg
Cc: EZ Rhino; <towertalk@contesting.com>
Subject: Re: [TowerTalk] Wired Anemometer for Tower?
I'm not sure that's true about the wind. I have a Davis Vantage Vue. The remote
unit does some processing to generate a 2 s mean speed and a peak gust value
within the 2 s window. Data are then sent in packets every 2 s to the base unit.
Mine is up 10 m (a WMO standard height). During calm winters nights it will
measure an anomalously high minimum temperature. The anemometer is shielded a
bit by the tower structure when it's downwind of the tower and so the wind
speed is biased low. Direction appears unaffected.
The rain gauge is a tipping spoon with a 3" diameter catchment. It under
samples only at high rain rates. The temperature sensor and dew cell are not
aspirated in this unit and so temperature is biased a bit high due to radiation
errors on calm, clear Summer days.
I highly recommend any of the Davis units, but the Peet Bros. Ultimeter 2100 is
also a very fine unit. Both are comparable in price (~$400).
Kim N5OP
"People that make music together cannot be enemies, at least as
long as the music lasts." -- Paul Hindemith
On Apr 27, 2014, at 18:29, "tomkd8deg" <kd8deg@centurytel.net> wrote:
The Davis wired version, which I am using, has a 4 second delay
for wind, and direction display. That can mean the loss of 10 to
20 mph in a gust situation.
Tom KD8DEG
-----Original Message-----
From: TowerTalk [mailto:towertalk-bounces@contesting.com] On
Behalf Of EZ Rhino
Sent: Sunday, April 27, 2014 4:10 PM
To: towertalk@contesting.com
Subject: Re: [TowerTalk] Wired Anemometer for Tower?
I should mention one other thing about the "cheap" weather station
anemometers that drive me crazy. The first WX station I had only
updated the wind speed about every minute. The LaCrosse updates
about every 12 seconds. This is almost worthless...I want to know
the wind speed NOW! The stuff from Inspeed is real time, very nice.
Chris
KF7P
On Apr 27, 2014, at 14:03 , Roger (K8RI) on TT wrote:
On 4/27/2014 3:15 PM, EZ Rhino wrote:
I have tried a few different weather stations, so I can
unequivocally tell
you what NOT to get.
DON'T buy anything under $300. *sigh*
Mine cost 40 something, works fine, is made of cheap plastic. Uses
plain old Duracell non rechargeable. Last 2 years or more outside.
The ones indoors only last 6 mo to a year. Probably cost twice
that much now...if available
Mounted on the tower where I can just reach it with an 8' stepladder.
Does nothing except report wind speed, direction,temp, humidity,
and relative temp.
73
Roger(K8RI)
Unfortunately the cheap ones are total garbage. The recent one I
got is a
LaCrosse station, purchased from Costco. Junk! Junky cheap
plastic, the rechargeable batteries in the solar powered wireless
anenmometer/wind vane unit went bad in the first year, and the
rain bucket stopped working in the first few months after purchase.
The Davis units are good, have a buddy with one and it is quality stuff.
And about $300. You pay for what you get. I agree, I think the
wired units are more reliable and last longer.
Ok now take a look here: www.inspeed.com These guys make
anemometers
that are decent quality and not expensive. They also have a cool
computer hookup version that allows your PC to track all the data
and put it in a spreadsheet. AND it has the wind switch option,
where you can program the system to turn an appliance on or off
(read: lower your tower!) if the wind goes above a preprogrammed
speed. Pretty neat. I don't have this option yet. I think AB7E
has one of the inspeed units too, maybe he can provide some feedback.
Chris
KF7P
On Apr 27, 2014, at 11:01 , Steve Jones wrote:
Can any of you recommend a wired anemometer or weather station unit?
I want to install an anemometer to monitor the wind speed at the
top
of my tower. All I can find online are wireless units. On this
reflector I have seen a number of bad reviews of wireless weather
sensors, because their little wifi transceivers get damaged by the
RF
from a ham antenna. I have a spare CAT5 cable at the base of the
tower
available.
73,
Steve
N6SJ
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
_______________________________________________
_______________________________________________
TowerTalk mailing list
TowerTalk@contesting.com
http://lists.contesting.com/mailman/listinfo/towertalk
|