I recently tried the RMYoung heated tipping bucket precipitation gauge, but after melting snow in the funnel it turned to ice before leaving the gauge, and the ice build-up would not clear itself until ambient temps reached 0C.
Can anyone recommend another tipping bucket that will work reliably to -20 or -25C? And, if necessary, a product that can recover autonomously from an occasional freeze-up?
Hi Anton,
The RM Young Tipping Bucket Rain Gauge is a nice little sensor and does very well in the right environment. It can melt snow in its catch basin down to -20, but from freezing and beyond that you may begin to grow icicles on the bottom. This is because the heating element in on the funnel and there is no heating element in the lower portion.
Let me recommend the CS700H. It has three heating elements, but make sure you can supply around 5 amps of power.
It also has an SDI-12 interface which controls the heaters to come on only when needed.
Generally this sensor is used where preservation of life is a priority such as; mine safety, dam safety, flood control, and slope stability.
http://www.campbellsci.com/cs700h
I hope this answers your question.
* Last updated by: sonoautomated on 2/17/2011 @ 3:48 PM *
There are several manufacturers that make what we call a weighing bucket. ETI sensors NOAH IV or III or the OTT Pluvio both work well in cold environments. We have 4 of the NOAH 3's and one NOAH 4. We put antifreeze in them to melt the snow (it is contained in a bucket). Their output values are similar to a tipping bucket but they are quite a bit "smarter".
I have several Geonor T200B vibrating wire buckets that are working very well for frozen precipitation. Weighing buckets have the advantage of very low power draw compared to heated tipping buckets. The downside is that you need to service them a couple of times a year to replace antifreeze and oil. Also, there is some noise in the signal from the vibrating wire, but it is a couple of orders of magnitude smaller than a 1mm precipitation event.
http://geonor.com/live/products/weather-instruments/t-200b-weather-precipitation-rain-gauge/
Paul
* Last updated by: PaulHeinrich on 2/8/2011 @ 2:27 PM *
Paul --
I am working with a Geonor T200B and had a question on how you were making the sensor measurement. Are you using PulseCount or PeriodAverage? And are you using SW 12V or are you leaving it on?
Can you post an example of your code?
Thanks....adam
We've used the Geonor T200-B in a test/evaluation setup, and have tried both pulse counting and period measurements, and both continuous cycled power to the transducer.
Here's code we've used for reading a couple of the sensors every minute:
PeriodAvg(gnr_freq_raw(1),1,mV250,GNR_VW1_SE,0,1,1000,1000,1,0)
PeriodAvg(gnr_freq_raw(2),1,mV250,GNR_VW2_SE,0,1,2000,1000,1,0)
I don't recall what those parameters are, nor why one uses 1000 and the other 2000, but I have not been satisfied with the stability of the measurement. My queries to other investigators turned up mentions of noise in the measurements, but the source of noise is not clear.
Looking at the signals on a scope seems to show that the output period actually does vary even with no disturbance to the instrument (no movement, no wind pumping, etc.).
My tentative conclusion was that taking the period measurement over a great number of cycles would be best, in part to try to average out the apparent systematic noise. This is what the manufacturer recommends (not necessary as coded above), but I was trying to see if we could read the sensor more quickly.
I think pulse counting vs period averaging may be more a matter of what channels are available on the logger than one being intrinsically superior to the other, though it would seem that pulse counting necessarily could miss any part of the first (or last) pulse.
I tried to look critically at wind pumping, which can impose a varying pressure on the bucket as the wind blows across the orifice, but did not reach any useful conclusion there, other than (again) that reading over a long interval might be best. If the pumping frequency is known, then it might be possible to average or count over a multiple of that frequency to reduce its influence on the measurement.
Another thing I tried was using the measurement as input to a digital filter, reasoning that the bucket load should vary slowly over time. But a long filter period would obscure actual data, so this is probably only useful if freqent and very short measurements are made.
Our Geonor should soon be back in the field, but I don't know how it will be programmed.
Thanks for all the interest and responses. Clearly sonautomated has had different experiences than I with the RM Young, but I'm glad to see there are so many alternatives.
In the low-to-medium budget range, does anyone have comments on the following:
Hydrological Services TB3 with TBH heater?
MetOne 375?
NovaLynx 260-2500E?
The TB3 with TBH heater is on our price list as the CS700H. Its limiting factors are its power requirements and price. It does have 3 heaters, though, to prevent it from forming icicles. 2 are on the base of the funnel, and one is on the base of the bucket itself. You can see this in pictures in our manual sonoautomated listed.
As to the other sensors, I am not familiar with them.
I'm interested in your icicle situation. At which temperatures did you encounter this problem?
Hi Skipper,
I had problems at -8C with very little wind - but weather is a complex phenomenon so your mileage may vary. I've passed on all my observations to CS Canada and RM Young, and from your vantage point perhaps you can join their discussions.
Any comments on the MetOne products that CS USA used to sell?
Anton
* Last updated by: Anton on 2/10/2011 @ 9:39 AM *
Message deleted by Anton.
* Last updated by: Anton on 2/10/2011 @ 9:40 AM *
We're using period averaging to read the Geonor gauges currently. I'm getting a very consistent and small amount of white noise from the strain gauges. It does not seem affected by wind. I had my university shop build 6" diameter 10' tall steel pedestals for the gauges and set them in 2' deep holes in a couple of hundred pounds of concrete. They are very stable. I made provisions to add guy wires, as well, but that seems unnecessary.
Here's my current code..
'Generic Period Average measurements Geonor_HZ
PeriodAvg(Geonor_HZ,1,mV250,14,0,1,500,1000,1,0)
and to calculate the precip...
'User Entered Calculation
Geonor_precip_mm = ((1.66018E-2*(Geonor_HZ-1587.6) + 9.20647E-6*(Geonor_HZ-1587.6)^2)*10)
Precip_30min_mm = Thirty.Geonor_precip_mm(1,1) - Thirty.Geonor_precip_mm(1,2)
Precip_1hour_mm = Hourly.Geonor_precip_mm(1,1) - Hourly.Geonor_precip_mm(1,2)
Precip_1day_mm = Hourly.Geonor_precip_mm(1,1) - Hourly.Geonor_precip_mm(1,24)
cheers, Paul
Below is an exchange I had with the Geonor rep I work with. I am planning on trying pulse counting to deal with the noise as he suggests below. It might also be possible to record a running average and subtract that from the signal in order to reduce noise..
Begin quote from Geonor Rep
----------------------------------------------
Instead of using the Period Averaging instruction I use the Pulse Counting instruction. By doing this with a scan rate of 1 min and Hertz chosen as the output the logger automatically calculates the average frequency of 60 because by definition Hertz is pulses per second.
With the older CR10 loggers there was a problem with the amount of ports available but the new of loggers from CSI let you use the Control Ports to make Pulse Count measurements.
The program below shows how we calculate the precipitation. In addition I calculate the running averages for the Hz and precip. This also helps smooth out the graph. I also calculate the rate of precipitation for a few different intervals. Any interval can be calculated.
BeginProg
Scan (1,Min,0,0)
PulseCount (freq1_Hz,1,11,0,1,1.0,0)'SENSOR #1, Port C1, SENSOR IN PRECIPITATION GAUGE
Prcip1_cm=0.01677395754817*(freq1_Hz-1057.5)+0.00000917225724*(freq1_Hz-1057.5)^2
Prcip1_in=Prcip1_cm*0.3937
'CALCULATING AVERAGES FOR HZ & PRECIPITATION (cm)
AvgRun (Run_Avg_Hz_5min,1,freq1_Hz,5)
AvgRun (Run_Avg_Hz_10min,1,freq1_Hz,10)
AvgRun (Run_Avg_cm_5min,1,Prcip1_cm,5)
AvgRun (Run_Avg_cm_10min,1,Prcip1_cm,10)
'CALCULATING RATE OF PRECIPITATION (cm)
Prcip1_cm_3hr=RTMC.Prcip1_cm(1,1) - RTMC.Prcip1_cm(1,180)
Prcip1_cm_24hr=RTMC.Prcip1_cm(1,1) - RTMC.Prcip1_cm(1,1440)
Prcip1_cm_1wk=Prcip_10MIN.Prcip1_cm(1,1) - Prcip_10MIN.Prcip1_cm(1,1008)
Prcip1_cm_30days=Prcip_60MIN.Prcip1_cm(1,1) - Prcip_60MIN.Prcip1_cm(1,720)
Prcip1_cm_1yr=Prcip_60MIN.Prcip1_cm(1,1) - Prcip_60MIN.Prcip1_cm(1,8760)
-------------------------------------------------------
End quote from Geonor rep
cheers, Paul
* Last updated by: PaulHeinrich on 2/14/2011 @ 12:36 PM *
Hi Anton,
I purchased the heated MetOne 385 this Fall and had no success keeping it from icing up. It has gone back to MetOne and they say it is working as it should, which indicates to me that it is a poor design for cold environments. I would not recommend it.
Alex
Hi Alex,
Where are you located? I'd very much like to hear more details since this seemed like a very good candidate to me, with quite a large heater.
Anton
* Last updated by: Anton on 2/23/2011 @ 9:30 AM *
Anton,
I'm in California, at Mammoth Mountain. We have a climate that fluctuates between maritime and continental depending on the year. Lots of snow, but generally not very cold. It was the high wattage of the MetOne 385's heaters which attracted me to it, but so far they have not proven to be effective in our environment.
Alex