[werk_ag]

I also prefer use "logs roll over" at midnight, it seems more logic when you lock at graphics, specially "rainfall today".

When I saw the image below (from your pages) I was thinking that something was wrong ... then I realized it is because of the roll over at 9am

[uncle_bob]

Yes, it's looking better as time when on.

As for the 9am roll over it's what the Weather Bureau uses here.
I'm not sure why this is used in Australia, it might be an old Naval thing from the Colonial days.

[werk_ag]

Did not know that.
This might be a problem, because the WeatherDuino Pro2 resets Rain Today counter at midnight. At best, you see one thing on the display and another on the webpages, but could be worse, I did not anticipate this possibility.

[uncle_bob]

As for the 9am roll over it's what the Weather Bureau uses here.
I'm not sure why this is used in Australia, it might be an old Naval thing from the Colonial days.

Interestingly, I found this on Google:

9am (EST) is the midnight greenwich mean time (UTC), which was why it was selected originally as the same time of measurement world-wide. All the large met balloons with radiosondes go up world-wide at that time also for the sake of cohesion of upper level observations.(Having worked at the Bureau for many years, that is what I was taught at training school BOM many years ago.)

Lifted from here: http://forum.weatherzone.com.au/ubbthrea...her_at_9AM

---

Did not know that.
This might be a problem, because the WeatherDuino Pro2 resets Rain Today counter at midnight. At best, you see one thing on the display and another on the webpages, but could be worse, I did not anticipate this possibility.

I can always change the Station to use midnight to midnight.

[werk_ag]

Perhaps I can add an option to software config, to allow the user to select the start hour of meteorological day.

Added to my to do list!

[werk_ag]

User configurable starting hour of meteorological day was been added to last software release!

Thanks uncle_bob

[Live_Steam_Mad]

OK I've got a real big problem with my rainfall readings. I first noticed it a while back when my 4.02 RX board was first powered up and connected to Cumulus, I think (from memory) I saw a reading of like 8 METRES of rain that gradually dropped to zero after a few minutes.

At first I thought it was just a glitch that was because the 2.50 TX board was not powered. However after connecting the TX I was expecting the huge amount to drop nearly straight away, but like I say the value stayed large for quite a while after I connected the TX (to 5V only, at the time) and only died away to zero after a few more minutes after I powered up the TX. I had no rain gauge or anything connected to the rain socket of the TX.

Anyway tonight I came to do some more testing and again had nothing connected to the TX rain socket. I powered on the RX, then started Cumulus, waited a minute or more, zero for the rain reading, then powered on the TX, and I again got 8 Metres of rain (8038.8mm / hr it said). I switched everything off then restarted everything in the same order as last time, and this time with an FO anemometer plugged into the rain gauge socket but with the wind cups just lying on the floor indoors and not moving, and this time I got 8 metres of rain again, then 6 metres, then 8 metres again, without me touching or doing anything.

I am completely at a loss to know why this is happening. I am on V2.50 TX board and V4.02 RX board, all the other readings seem to be behaving themselves nicely. The wind was reading nicely when I plugged in the FO anemometer to the Wind socket on the TX and moved the wind cups around my hand. Sometimes the Rain glitch doesn't happen and I have a zero reading for it that goes up very fast when I plug the FO anemometer into the rain socket and spin the wind cups.

Has the few hours worth of powering the TX board (previously, before I knew it could damage it) on 5V caused this? Is there a faulty component somewhere on my board? Dodgy rain (wind?) socket?

My transmissions are only with the FS1000A as of yet, and no pigtail or antenna is soldered / plugged into the FS1000A yet (waiting on the pigtail to arrive and 2nd antenna). However the TX and RX units are within 2m of each other, everything is room temperature and indoors and dry.

Is the rain data more susceptible to corruption via RF problems than the other data? Why is this huge glitch only happening with the rain data? Have I bridged a solder joint? Could me testing it with the wind cups in the rain socket (but with wind cups not moving at all) be a factor instead of using a FO rain gauge? But then the glitch also happened with nothing plugged into the rain socket. Could it be power problems from the RX being powered from the Vista laptop's USB socket? But I also saw this glitch when the CH340G Nano on the RX was powered with 12V as well (via 7809 reg.). Could it be the something to do with the characteristics / operation of the Prolific Nano that I used on the TX board? Maybe the way the PL Nano accesses memory or variables or something? Or maybe my Prolific Nano is faulty (zapped by static electricity from clothing etc?) ? I can swap it to another one if needed.
Have I got one or more faulty / wrong brand / wrong type 7809's or IRLZ44N ? Bad 2N7000 (s) ?

My TX software is V13b009, my RX software is V13b004. Have I accidentally screwed something up when compiling / editing user configuration options / altered some code accidentally ? I have as far as I am aware NEVER touched the actual code, as I haven't even looked at the code yet (nor would I actually understand it anyway)! Maybe it's something to do with my using only a 12V battery with +Bat and +12V connected together and no solar regulator and no solar panel ? Maybe it's the Arduino IDE 1.5.8 Beta that's at fault? Maybe an Arduino software library that I have the wrong version of? Maybe it doesn't like my user settings of fanOn_HiTemp = 32, fanOn_LowTemp = 32, and fanOn_LowWind = 15 ? I made the first 2 items the same value as you can see since I wanted to fan to be permanently on. The low wind setting that I used was significantly different than the default. They were the only settings that I altered for the TX board AFAIK.

I have no temp. / humidity sensor wired onto the DHT22 / SHT1x socket on the TX board yet, maybe that is a problem? Bad C5 capacitor near the Rain socket?

Sorry but I now rather depressed after wiring this lot up and finding this major problem . Ironically my FO 1081 (which had 2 large wind glitches within the first 2 weeks of being online since November 2014) has been working just fine now for several weeks .

Help!?

EDIT: I just refreshed the solder on D2 on the Prolific Arduino itself and on D2 on the socket for the Nano, and on C5 both sides (on top of the board as well this time) and have shiny joints and no bridging. Now going to go and test it.

2nd EDIT: I decided to desolder C5 anyway before retesting with Cumulus. But I just tested the old C5 that I pulled out on the bench and it seemed good. Darn. Anyway I desoldered it, by pulling on it with Tamiya large pliers with serrated jaws whilst moving the bit between the 2 legs on the other side of the board, then used same pliers and a panel pin to make a cenrtre dimple in each hole, then used a 0.5mm drill bit and Swivel Head Pin Vice to drill 2 new holes, inserted one of the cheapest 104 cap's that I could find (10uF) (tested it first on the bench, simply set the multimeter to 200 Ohms range, which charges up the cap, then set to 200mV to discharge it!) then soldered it in. NOT recommended for the faint of heart.

Just tested C1,2,3,4 also and they all charge / discharge EXCEPT C2. Don't know if that is significant or not. I can change C2 if it's needed?

PS. NOPE. Just tested the RX and TX and it was OK for a while, then I rotated the wind cups at a moderate speed in the rain socket and it showed nothing, then just showed 8038.8mm of rain per hour again about 5 minutes after I had done it.

Then I restarted everything completely just to be sure (started from scratch, no RX Arduino light / power, no Cumulus, no TX light / power) and didn't touch the wind cups and it still just showed me 8838.8mm of rain. Something seems to be badly wrong.

Is this because I am not waiting 3 minutes before starting Cumulus? But I'll bet that even if others do the same as me they still don't get 8m of rain registered?! How come only the rain reading is a problem? I think I will go and change my Prolific Nano to the 2nd one and see if it solves the problem. EDIT: Now testing 2nd PL Nano. Seems OK so far.

Alistair G.

[Mikko]

Hi,
Your Nano eeprom might have some bogus values and Cumulus is showing them. You can clear eeprom by uncomentting (take slashes out) following line from setup procedure. Uncomment it, then upload to Nano, comment again and upload once again to Nano. Hope this helps,
Mikko

Code:

//  erase_eeprom();

[Live_Steam_Mad]

Hi,
Your Nano eeprom might have some bogus values and Cumulus is showing them. You can clear eeprom by uncomentting (take slashes out) following line from setup procedure. Uncomment it, then upload to Nano, comment again and upload once again to Nano. Hope this helps,
Mikko

Code:

//  erase_eeprom();

Just now getting 6851.1mm / hr of rain on my 2nd Prolific Nano. So changing Prolific Nano's on the TX didn't work.

When you say my Nano EEPROM may be a problem, which Nano are you describing, the RX (CH340G) or the TX (PL2303, 2 of them tested) ?

I'll have a look at the TX and RX software and see if I can see the line of code that you are mentioning.

Thanks for the help so far, I really appreciate it. I am so nearly there with my new PWS but this rain rate issue is driving me completely nuts.

EDIT: I did exactly what you told me Mikko, with the 2nd Prolific Nano (2.50 TX board). There is one single "erase_eeprom" in the TX setup, as stated. I uncommented it, compiled and uploaded it, then commented it out and compiled and uploaded again. Using V13b009 software.

There is also one single "erase_eeprom" about 3/4 of the way through the code for the RX board's V13b004 software. I didn't do anything with that so far. Should I do anything with this RX Nano?

EDIT: NOPE that didn't work. I just got 6851.1mm of rain about 3 to 4 minutes after a restart!

Maybe I should give up with the Prolific Nano and buy a CH340G for the TX. Maybe I got 2 bad PL Nano's (they were from different sellers though!) ?

Alistair G.

[uncle_bob]

Hi Alistair,

Don't freak out about this. I'm sure between us all we'll get to the bottom of it.
Note that you should start the RX unit and wait for three+ minutes or so before starting Cumulus.

I'd also be testing this on another Cumulus installation until you're happy it's all working correctly, this will avoid editing the Cumulus files if weird values are presented.