Topic: WH40 vs WH65 Rain Measurements

[galfert]

With the smaller funnel on the WH40, shouldn't you be bumping up your calibration to account for the difference in the diameters and surface areas, as opposed to backing it down to 1?. I believe this was previously discussed. Or did I miss something entirely?

Yes you are correct. We have calculated that the required adjustment offset should be 2.56. I think however in this case platokidd is looking to do their own imperical testing and setting the offset to 1.0 in order to compare with the Stratus gauge and to then determine based on that difference what the offset should be.

The thing that plottkidd should realize is that the offset based on imperical testing should be something that is determined after several rain events of different intensities to arrive at a good average. I'm interested to see what he comes up with as opposed to the mathematical approach.

[Rover1822]

Yeah I can see starting from a baseline and then working it out. This will be interesting. Downside is that is will take considerable higher amount of rain to register.

[galfert]

Yeah I can see starting from a baseline and then working it out. This will be interesting. Downside is that is will take considerable higher amount of rain to register.

Yes but it shouldn't be any different than a WH65. Meaning it seems like the only advantage of the WH40 was that it could potentially register smaller amount of rainfall sooner. The WH40 with its original funnel would most likely indicate the start of rainfall sooner too. But maybe not because of it flat section if the rain wasn't really rolling down to be collected, not being able to overcome surface tension and maybe even evaporation on a light rain if it was somewhat hot and sunny as it began to drizzle.

We can confirm that the internal tipping mechanism is identical to the WH65. If we use the funnel from the WH65 on the WH40 and use the proper offset for that funnel size then we have essentially recreated the WH65 as a standalone. There shouldn't be any negatives in comparison to a WH65 is what I'm saying....I think it will be a perfect WH65. We have already determined that the WH65 performs better. I just don't see any downside to this conversion compared to a real WH65 is what I'm saying.

[Rover1822]

I wasn't saying there was a downside, it's a good experiment and Kudos to Platokidd for doing it

[Jai Soone]

The difference in collector size is probably what causes ambient weather to publish:
Rain Resolution: 0.01 in  on the WS-2000 versus Rain Resolution: 0.004 in on the WS-5000. Out of curiosity does anyone spray the collector with a teflon coating as suggested under item "How to DIY a High Precision Rain Gauge with Teflon Coating to Collect Small Raindrops 08-24" on page:
http://www.ecowitt.com/news/

[GHammer]

The difference in collector size is probably what causes ambient weather to publish:
Rain Resolution: 0.01 in  on the WS-2000 versus Rain Resolution: 0.004 in on the WS-5000. Out of curiosity does anyone spray the collector with a teflon coating as suggested under item "How to DIY a High Precision Rain Gauge with Teflon Coating to Collect Small Raindrops 08-24" on page:
http://www.ecowitt.com/news/

I have and it does keep rain from standing in the funnel

[ramblinche81]

the funnel slope isn't material. Below is a lot of rounding as I do math in my head, just follow the methodology.

My tipper might be 2 ml. My lab syringe is only 1ml markings. With 18cm funnel that means each tip is 2cc per (3.14x81cmsq) or less than 0.01cm (0.1mm) rain level and rate per bucket tip. If I have a 10 tip rain event and I have  mostly full bucket left over, I have a 10% error. And the next rain event is over stated a like amount less evaporation. I would call a 10 tip event a misting, not a shower as it is 1mm of rain.

In deluge rain events, 5cm per hour that is 500 tips over 3600sec. That is 2ml per 7.2 sec. During the tip, how much drip through occurs per tip which is not captured as bucket is sloped. The heavier the rate, the more spillage.

The residual on the funnel is less than 1ml. I will do dry before wet after weights. In a 5cm deluge (500 tips x 2ml per tip) that is 0.1% throughput volume error. In a mild rain that residual is maybe 0.5%. Immaterial.

[Rover1822]

Your research is admirable in the context of the residual water on the funnel and the bucket. However, a large part of this discussion and the issue of  slope of the funnel have to do with rain bouncing out of the funnel at rain velocity because of a shallow slope of the WH40, and therefore is not captured by the funnel.

For tipping bucket rigs, there will always be step error on how long and how much it takes the bucket to fill and when it actually dumps. Arguably this is a minuscule error and can be ignored for the greater benefit of a system that can report rainfall and does not require visual inspection of a rain gauge and the manual emptying of a tube. Doing the application of a non stick material (as described) can reduce part of this error for the funnel.

The larger issue here is the cross comparison of one bucket system vs another. The WH65 vs the WH40, where the WH65 is closer to a true rain gauge measurement (still with observable error as expected) , and the WH40 which is off by much larger percentage although marketed as more accurate. All of this is covered in previous posts in this thread.



 

[galfert]

I agree with Rover1822, and I have some conclusions based on feedback from Ecowitt on this issue. Because Ecowitt engineers have looked at this issue and determined that a redesigned rain funnel would not correct the issue (preventing rain bouncing out), I am therefore determining that this means that the WH40 tipper is not able to properly keep up with faster rainfall. The tipper is the same size as the WH65 but because of the larger diameter it needs to be faster to maintain accuracy. Therefore there must be great loss during tips. Sure it has been observed that rain jumps out during heavy rainfall, but that is only part of the issue.

I think the solution we are pursuing with the WH65 funnel mounted to the WH40 and then calibrating for that smaller size funnel eliminates the two problems. We eliminate the flatter slope of the WH40 funnel that causes bounce out, and we eliminate the greater tipper speed required by the tipper for a larger diameter. We essentially have created the exact working dynamics of the WH65 with this WH40 modification. I think this solution has promise. It should work with no consequences compared to the WH65. At least that is what it seems unless we are missing something that has not yet been determined.

For anyone with a DIY ability I encourage someone to create a tall cylinder for the larger WH40 funnel...keeping the diameter the same as the original. I think this cylinder needs to be thin and extend from the inside of the funnel upwards. Otherwise if mounted to the exterior then the lip of the funnel will catch water. It think this DIY solution eliminates the bounce out problem. But I still think that the results will not be ideal because of the faster tipping speed required to prevent loss during tips. At least this is the test that I think Ecowitt has performed....and why they say that a new funnel will not fix the problem. It would thus be interesting to see these results performed by someone.

Ecowitt has said (several months ago not related to this issue) that they are interested in developing a haptic rain sensor (no moving parts). I believe work on this has begun. I however feel this is not the solution we all need. From looking at the extensive work that WeatherFlow has done with their various models I don't think this technology has promise...at least not at the typical price point and target market of Ecowitt customers. I would like to see a totally new redesigned replacement for the WH40 to truly be something better than the WH65. I feel that the WH65 is not bad. I live in Florida where we get some torrential rain and I'm pleased with the results that I get compared to my Stratus. I'm not saying it is perfect, only that I'm okay with the disparity that I see.
 

[Rover1822]

I would love to take on the DIY project, however not buying a WH40 just to play with (yet)

I was all ready to buy the WH40 for the obvious benefits of supposedly increased accuracy, and the possibility to mount where maintenance would be much easier than pulling down my WS-2000. Sounded like a multiple win scenario.

Now if I was going to test it , I would want something like a large rain style shower head (one of those with a large head and a straight down water flow) on a pole regulated by a gate valve for different intensities, my Stratus and the WH40 co-located with non interference under the "rain" . Would have to swap positions and do multiple runs and under different  intensities, to get an idea of modification viability without waiting for rain.

I'm trying to decide if I'm that bored.
for the shower head... something like this
https://www.amazon.com/Derpras-Stainless-Pressure-Bathroom-Showerhead/dp/B072MHQSYT/ref=sr_1_24_sspa?dchild=1&keywords=large+rain+shower+head&qid=1598114705&sr=8-24-spons&psc=1&spLa=ZW5jcnlwdGVkUXVhbGlmaWVyPUEyMUFFNUNPR1ZVMkQ5JmVuY3J5cHRlZElkPUEwODA3MTQ2MUpUNEc2MVhaWk5JNyZlbmNyeXB0ZWRBZElkPUEwOTk2ODQ4WEFDSlkxR1dVSDlYJndpZGdldE5hbWU9c3BfbXRmJmFjdGlvbj1jbGlja1JlZGlyZWN0JmRvTm90TG9nQ2xpY2s9dHJ1ZQ==

And yes... my test scenario would still be far from reality

If Platokidds experiment pans out, I'll just go that route

[ramblinche81]

I agree with Rover1822, and I have some conclusions based on feedback from Ecowitt on this issue. Because Ecowitt engineers have looked at this issue and determined that a redesigned rain funnel would not correct the issue (preventing rain bouncing out), I am therefore determining that this means that the WH40 tipper is not able to properly keep up with faster rainfall. The tipper is the same size as the WH65 but because of the larger diameter it needs to be faster to maintain accuracy. Therefore there must be great loss during tips. Sure it has been observed that rain jumps out during heavy rainfall, but that is only part of the issue.

I doubt bounce out is significant as it only applies to outer rim couple mm. And only deluge type rains.  2mm edge is 10cmcm outer perimeter area of 250cmcm total area 4% and only 10% of drop bounces off funnel means loss is only 0.4%

5cm per hour rain is 7.2 sec between  dump. 2ml flow. A .25 sec dump cycle start to finish means about 3.5% (.25 / 7.2) possible loss per dump. 2ml is about 40 drops, which means 6 drops per sec (of 7.2 sec per 2ml rate). That is no longer distinct droplet, that is flow. Couple drops lost per dump. 2-4% possible.

I am very satisfied with 96% accuracy in a 5cm rate per hour.

Light moderate rain, great accuracy likely. Deluge typr rain 5cm per hour up a 5% error rate possible.

[Rover1822]

Sorry, I get a lot of "And only deluge type rains" , so that is ultimately important to me.

But, let's go back to the original issue. The WH65 has been demonstrated to be much more accurate than the WH40. And we want to use the WH40, so we are looking for solutions.

I see you have a WH40, have you tested it against a reputable rain gauge, such as a Stratus?

[ramblinche81]

Sorry, I get a lot of "And only deluge type rains" , so that is ultimately important to me.

But, let's go back to the original issue. The WH65 has been demonstrated to be much more accurate than the WH40. And we want to use the WH40, so we are looking for solutions.

I see you have a WH40, have you tested it against a reputable rain gauge, such as a Stratus?

I have not tried to measure vs any standard. Not sure it would be statistically valid in my tree shrouded property unless the two devices were immediately adjacent.

Statistically, two devices with 5% detection errors or variations are going to display as much as 10% differences.

I use the deluge reference because high flow rates, high droplet size, high sensor tip frequency are all additive for error. Lot of test devices have useful ranges and beyond the range, errors creep in.

Time permitting Sunday I will set up a shower head spray and measure flow rate vs sensor readings. Shouldn't be that hard.

If we demonstrate system under reports by X% above flow rate Y, then should be easy to modify reported rate.

[Rover1822]

I was kind of joking about my test setup or proposed one. (the actual effort would take a lot more thought)

I was suggesting that you do measure your rig against a known valid rain device that you can co-locate or approximate. The Stratus, $37 on Amazon seems to be a good one. I just got one in and I am using it against my WH65 (osprey). With one rain I was .16"  osprey vs 0.15 Stratus (there is a 30' distance and 8' height variance  ) . Of course that was a single instance so really nothing should be inferred . Also , I hastily mounted the Stratus because of an imminent rain event that I wanted to test, and I have limited mounting options.

If after a period of time, you are getting readings that you are content with vs the rain gauge, well maybe then "you" don't have an issue, and the rain you predominately receive fits with the parameters. Others are  reporting variances in their environment that frankly I would not live with, which prevents me from purchasing a wh40.

I am curious on why you have decided to blindly accept the measurements of your WH40 after having read through the previous 5 pages of this thread, and seem to want to argue it. Especially without a comparison point?

[ramblinche81]

The Stratus, $37 on Amazon seems to be a good one. I just got one in and I am using it against my WH65 (osprey). With one rain I was .16"  osprey vs 0.15 Stratus
I am curious on why you have decided to blindly accept the measurements of your WH40 after having read through the previous 5 pages of this thread, and seem to want to argue it. Especially without a comparison point?

Not arguing at any level. Showing the sources of error that are unavoidable and their relative % impact. Your own data above reflects a 1/16 variance....6%. Not surprising is all I am narrating.

Regarding accuracy and the prior pages, I am not doing climate research or anything remotely similar. At 10 inches a month precip, a 5% error is 1/2 an inch. 5 inches rain is 1/4 inch. . Other than research , that immaterial to most moisture based decision making for most homeowners. Would not change my lawn care, garden, or other practices.

[jimi]

Isn't this design always going to be inaccurate ?



Unless I'm missing something your always going to have losses when the bucket is emptying and the losses will increase with heavier rain. Surely the seesaw type setup is more accurate ?  The MET Office use that type of gauge https://www.metoffice.gov.uk/weather/guides/observations/how-we-measure-rainfall
My ancient (but still in use and working OK) WH1081 uses a plastic version of the seesaw rain gauge

[galfert]

For what it is worth Davis just this past year changed from seesaw to tipping spoon. Davis users are noticing better results.

As discussed before we are not trying to make the WH40 do anything spectacular. We are only trying to make it at least as good as the WH65. It is unacceptable that the WH40 which is advertised as being better than the WH65 is actually worse.

[jimi]

I understand what your trying to do      just strikes me that it's a bit of a "silk purse" job.

[ramblinche81]

For what it is worth Davis just this past year changed from seesaw to tipping spoon. Davis users are noticing better results.

As discussed before we are not trying to make the WH40 do anything spectacular. We are only trying to make it at least as good as the WH65. It is unacceptable that the WH40 which is advertised as being better than the WH65 is actually worse.

Time permitting today and Monday I will run a fixed volume through mine and tell you results.

Pretty easy to send 1 liter through at different flow rates ( 2, 3, 4 hours) and reference the indicated wh40 flow rate.

I will do the decimal place math but 1l should indicate just over 4cm based on 180mm diameter funnel. Metric system 1000 times easier to calibrate than imperial and inches to oz fluid volume.

[Rover1822]

I was thinking, as this thread is many pages, and since you are just working on the accuracy of the WH40 by itself. You may want to start a whole new thread discussing the accuracy of the WH40. You would probably get more viewers and it would be easier for other viewers to search for the post down the road.

[ramblinche81]

I was thinking, as this thread is many pages, and since you are just working on the accuracy of the WH40 by itself. You may want to start a whole new thread discussing the accuracy of the WH40. You would probably get more viewers and it would be easier for other viewers to search for the post down the road.

Good suggestion. I will leave it to moderators and members to decide if test is worthy of stand alone. Today's test run was successful as proof of concept. [ You are not allowed to view attachments ]

Plus 4cm rate achieved and time consistent for a projected 500ml total flow in batch 1. Rain started so test is aborted for today. Will have to resolve some drip rate variation as water column height declines.

[Jai Soone]

If the WH40 reports out rather frequently, one could calculate a semi-instant rain rate, infer how many tip's of the bucket occurred, calculate a correction value for each tip of the bucket, and factor that in to the reported amount of rain detected. One would then have to experiment to determine how best to calculate a correction value. Maybe it would have a linear relationship to the rain rate, maybe a more complicated curve. TBD

[K7RMK]

The difference in collector size is probably what causes ambient weather to publish:
Rain Resolution: 0.01 in  on the WS-2000 versus Rain Resolution: 0.004 in on the WS-5000. Out of curiosity does anyone spray the collector with a teflon coating as suggested under item "How to DIY a High Precision Rain Gauge with Teflon Coating to Collect Small Raindrops 08-24" on page:
http://www.ecowitt.com/news/

This makes sense. 0.004 x ~2.56 = 0.01

[galfert]

The difference in collector size is probably what causes ambient weather to publish:
Rain Resolution: 0.01 in  on the WS-2000 versus Rain Resolution: 0.004 in on the WS-5000. Out of curiosity does anyone spray the collector with a teflon coating as suggested under item "How to DIY a High Precision Rain Gauge with Teflon Coating to Collect Small Raindrops 08-24" on page:
http://www.ecowitt.com/news/

This makes sense. 0.004 x ~2.56 = 0.01

And thus this elegantly represents why greater precision does not equal greater accuracy. Two totally different concepts that all too often get mixed up.
The WH40 has greater precision....but way worse accuracy than the WH65. The WH40 traded the good accuracy of the WH65 for greater precision.


 [ You are not allowed to view attachments ]


But there is more to this because the WH40 has water loss based on bounce out also. And it also has loss due to tipper not keeping up to higher rain rates. This diagram was only intended to reflect what "on paper" would be true if these other issues did not exist.

[galfert]

Here is another way to look at it from a 2 axis perspective.

 [ You are not allowed to view attachments ]