Topic: Wittboy settings to improve haptic rain sensor accuracy?

[kinwolf]

Hi,

      Continuing my discovery of the WS90... I did a search, but didn't find any answer on this.     

Currently, the wittboy haptic sensor has a huge difference for rain detection.  Yesterday, it rained around 21mm(reported by the WH40 and 2 older Acurite 5-1 stations) while the haptic sensor reported 26mm...!  5mm in a huge difference, and it confirmed what I noticed since I installed it, it's constantly reporting alot more mm then the other 3 devices.   Before I contact support, is there some calibration that can be done?  I saw nothing in the short installation doc, and looking in WSView, I do not see anything, but it's easy to miss a setting in that app

Thanks!

[Gyvate]

1. the WS90 rain gauge can be calibrated on five levels, depending on rain rate
2. you have to use the WS View Plus app, not the old WS View app, in order to see the 5 entries (gain) to be found under rain totals
or simply use the WebUI --> rain totals

[Rover1822]

I believe there are number floated by giusCB, for the calibration.
I have not tried them, and they seem to be dependent on the firmware version.

I do appreciate him reporting these numbers, but , I would want an official recognition from F/O or  Ecowitt.

[kinwolf]

Foudnd the "gain" settings under Rain Totals, thanks!  I guess I'll set them at 0.82 for the moment and see how it goes.

[Gyvate]

I guess you need to run the WS90 for at least 1/2 if not one year on gain 1.0 for all 5 areas, luckily have enough different rain events, have a manual rain gauge in parallel as reference and document them.
Based on this data you can calculate a gain factor for each rain rate slot and then continue observing and refining the calibration, all the time having the manual rain gauge in parallel
(or have another properly calibrated automated rain gauge in place).
I doubt Ecowitt/Fine Offset are going to publish a suggested calibration.

[Mapantz]

Although I'm about to send my Wittboy back, I don't see how the "gain" can consistently work?

I'll tell you why; The first batch of dynamic rainfall to fall on mine was grossly overreading - For every 1mm of actual rain, it was recording 3mm.

A few days ago, we had a rain shower - moderate to heavy. It failed to show anything, despite 1mm of actual rain.

If that kind of inconsistency is occurring, then using a "gain" is useless.

That, and it recording dew as rainfall all of the time, made me give up on it.

[giusCB]

 with original setting it overestimate by +50/70%
There isn't an official statement from ecowitt for calibration.
I've installed a second unit of wittboy but here in Italy we are under a severe heat wave. I've not tested piezo sensor anymore

[Gyvate]

with original setting it overestimate by +50/70%

I've got meanwhile two of them - a ß-device and a regular one.
But the overreading cannot be generalized, not with my observations now over a couple of months (six to be precise - with 73 rain days out of 179).
It's sometimes in full tune with my WH40 and even lags behind sometimes ....

It's a good dew indicator and I don't mind it reacting on dew, as dew is also rain and fills the spoons of my tipping rauge gauges.
It's also a good rain indicator. It already shows rain when it still takes a while for the WH40 to empty its first spoon.
Will be a tricky task to get it calibrated for the majority of rain events - for all it won't work, not with the calibration means we have so far. Maybe with some characteristic curve also using wind readings implemented in the firmware ....

I have one GW2000 set up with WH40 and WS90 and one with WH65/WS69 and WS90 - all gain = 1.0, and depending on rain rate and wind reality moves within the borders of the three.
Of course it's not useful for most to have three devices for what could ideally be provided by one device ....

But to reduce the WS90 onto its rain readings wouldn't be fair either. It's a solid 7-in-1 one-piece weather device with no mechanical parts, which in itself is already something special.

[kinwolf]

I contacted support to see(they do answer fast, I'll give them that!)  and here is their reply(mostly a repeast of what people mentionned here )

For calibration, there are 5 different gains you can set.

1.     For < 5m/hr, please keep 1.0 calibration factor.
2.     For the rest of 4 setting, like your recorded rain , can you check what is the average rain rate it has? You can find the most long period time of rain fall ( that contributes most to the rainfall during that time) and note what is the rain rate. Once you can determine the rain rate, then you can adjust rain gain to 0.82.
3.     Basically you can set all the rest four to 0.82, and adjust it if your future rain rate is different and thus need to adjust it slightly to meet the best result.
So it need to take a little bit time and effort to adjust this value.

[Rover1822]

Why does that sound to me like more work than the routine maintenance on a tipping bucket ?

[box]

Ah, so you need to know the rain rate to work out the rain?

[Gyvate]

Ah, so you need to know the rain rate to work out the rain?

yes - when you have five calibration tiers depending on rain rate, what else would you choose to determine the gain factor for the tiers ???
  [ You are not allowed to view attachments ]

[box]

Ah, so you need to know the rain rate to work out the rain?

yes - when you have five calibration tiers depending on rain rate, what else would you choose to determine the gain factor for the tiers ???
  [ You are not allowed to view attachments ]

dice?

[Rover1822]

What I meant was, well yes, you can calibrate, but you need a comparison unit (of some kind) to do it.
I'm just thinking of a novice, which these seem to be targeted at.

Hard enough for some to calibrate pressure, unfortunately, this is not user friendly. Having one, I would not recommend it to someone wanting a "simple" device. They also really need to improve their documentation.

The concept of plug and play and it just works , really not working here. The majority of the documentation is for connecting the GW2001.
Sadly not everyone reads this forum.

[box]

What I meant was, well yes, you can calibrate, but you need a comparison unit (of some kind) to do it.
I'm just thinking of a novice, which these seem to be targeted at.

Hard enough for some to calibrate pressure, unfortunately, this is not user friendly. Having one, I would not recommend it to someone wanting a "simple" device. They also really need to improve their documentation.

The concept of plug and play and it just works , really not working here. The majority of the documentation is for connecting the GW2001.
Sadly not everyone reads this forum.

seems to me the calibration methodology is somewhat open to question, especially if you are in an area that experiences wide variations in rainfall type and intensity

[kinwolf]

I decided to return the wittboy.  Too much discrepancies with the haptic sensor, and having to rely on another device to calibrate this one is not my idea of "no maintenance".  That feature was not ready for prime-time IMHO. 

[flyfree]

I have been able to use both the 'Ecowitt' app and the 'WSView Plus' app to set the gain for the rain calibration.
For both Apps it is a case of finding 'Rain Totals' in the top right side menu.
Then swipe/scroll down to 'Piezo Rain1 Gain'... to 'Piezo Rain5 Gain'
To determine the new values I created a spreadsheet over many months that compared the Wittboy rain rates and amounts against the totals from 2 funnel gauges which I read when ever it suited me depending on rain events sometimes several times per day.
I was then able to create formulas which compared the totals and the rates to allow playing with the calibration for each of the 5 rates until they match the actual totals.
I have readjusted the gain on three occasions as I try to fine tune its accuracy from different types of rain events.
My current adjustments I now have set the gain to the following amounts:
Piezo Rain 1 Gain, rate <4mm/h          0.90
Piezo Rain 2 Gain, 4< rate <10mm/h   0.70
Piezo Rain 3 Gain, 10< rate <30mm/h 0.60
Piezo Rain 4 Gain, 30< rate <60mm/h 0.50
Piezo Rain 5 Gain, 60< rate                 0.40
The Wittboy is mounted above my roof and the funnel gauges are at eye level on the ground on different sides of the house.
The funnel gauges can be more than 10% different to each other due to wind direction over the house so I do not expect to ever get better than 10% accuracy.
And when the power to my house blacks out during storms the Wittboy does not record any of those downpours from the blackout until power is restored which makes my long term actual totals impossible to match.

[Gyvate]

And when the power to my house blacks out during storms the Wittboy does not record any of those downpours from the blackout until power is restored which makes my long term actual totals impossible to match.

to be accurate in the statements
- your Wittboy NEVER records weather data - it only transmits the read observations every 8.8 seconds - even during a mains blackout unless you have no batteries inserted.

- A GW2000 console does not record any data either. A HP25x0 console may still record data when the WS90 ("Wittboy") array is connected to it.

- Your console may not receive this data anymore (unless it is on an uninterrupted power supply of some sort - and even if so may not post this data to e.g. ecowitt.net unless your local network and internet access infrastruture is also UPS protected.

- A data logger program may still continue logging the data if your console, network and local server where the data logging program runs on are UPS protected.

[Rover1822]

Code: [Select]

I have been able to use both the 'Ecowitt' app and the 'WSView Plus' app to set the gain for the rain calibration.
For both Apps it is a case of finding 'Rain Totals' in the top right side menu.
Then swipe/scroll down to 'Piezo Rain1 Gain'... to 'Piezo Rain5 Gain'
To determine the new values I created a spreadsheet over many months that compared the Wittboy rain rates and amounts against the totals from 2 funnel gauges which I read when ever it suited me depending on rain events sometimes several times per day.
I was then able to create formulas which compared the totals and the rates to allow playing with the calibration for each of the 5 rates until they match the actual totals.
I have readjusted the gain on three occasions as I try to fine tune its accuracy from different types of rain events.
My current adjustments I now have set the gain to the following amounts:

I would love it if there was a single set of adjustments that applied to everyone, I know from my 2 units , that won't work.
But, with that said, I am glad you have wrangled the unit to work within your parameters, and version of the unit you have,

Based on the numbers you have presented, this appears to be an early version of the WS90 , is that correct?

[flyfree]

Yes I have a WS90 from its first release.
With a GW2000C without a UPS and no data logging.

My latest rain gain settings are
< 4   < 10   < 30   < 60   60 +
1   0.8   0.7   0.6   0.4

I do not know if the gain adjusts the sensitivity.
And I do not know if the adjustments are proportional.
Initially I assumed the gain does not change the sensitivity and I also assumed the gain was proportional.
And I do not know if any software upgrades might change anything that might effect the gain.
And I did not know there was a new model Wittboy WS90.
I posted in case any one else was finding improvements by adjusting their Wittboy rain gain values and I wonder what values they may be using.

[Gyvate]

Yes I have a WS90 from its first release.
......
And I did not know there was a new model Wittboy WS90.
I posted in case any one else was finding improvements by adjusting their Wittboy rain gain values and I wonder what values they may be using.

Was it a beta-test model or an official release model ?
The official release model hardware (=hardware 2.0) wasn't changed since release. However the beta-test model (hardware 1.0) was changed to 2.0
Unless you were a ß-tester, no hardware change for you ...
there were changes in sensitivity via WS90 firmware upgrades (--> our MUST READ thread - https://www.wxforum.net/index.php?topic=40730.0)
Gain is gain - just a factor by which the sensor provided readings are multiplied by the console
If you read releated threads, especially https://www.wxforum.net/index.php?topic=43139.0, you can see findings reported by myself stating that the characteristic line of the WS90 rain sensor, especially in the lower rain/rate area (1,2), is non-linear => gain factors alone won't bring a satisfactory result .
Once a proper characterictic line is established by Ecowitt - still to happen - the linear gain factors can be used for fine tuning.
You will keep on finding new gain factors, especially in the low range areas, the more different rainfall events you meet (rain rate, wind, drop size etc.)
In the higher rain rate area (gain 3-5) the existing characteristic line can be "tamed" by gain factors and there sufficient test conditions could be set already in the laboratory I guess.

[flyfree]

Wow Gyvate you have put some time in thank you very much
I am not a beta tester.
My original kickstarter official release wittboy WS90 firmware is currently displaying 117 (when I hover over the image of the Wittboy on the Sensors ID page, using the WebUI)
I could not find the Wittboy WS90 firmware version using my current versions of my Apps.
To find the WebUI I looked up the IP address for the GW2000 on my router and entered it into my browser and found the instructions mention that first time using it the password for user admin is blank.
There was so much to read through it was difficult to find other rain calibration settings and as you point out it all depends on which firmware version is applied.
I have not researched how to update my WS90 firmware and I do not have a PC if it is required to do so.
Thank you again Gyvate for your many very useful posts