Errors reading fromt he weather station

[Auchtermuchty Weather]

Noticed gaps in my graphs and found lots of this in /var/log/syslog.  This is with a new SD card in the RPi, and I'm using the latest version of WeeWX (v2.7.0) with a WH1080 weather station:

Oct 15 11:29:03 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b220
Oct 15 11:30:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b240
Oct 15 11:33:02 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b260
Oct 15 11:34:05 weatherpi weewx[20726]: fousb: station status {'unknown': 0, 'lost_connection': 64, 'rain_overflow': 0} (64)
Oct 15 11:38:38 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b2c0
Oct 15 11:45:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x000000
Oct 15 11:49:03 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b360
Oct 15 11:51:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b3a0
Oct 15 11:53:50 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b3c0
Oct 15 11:55:00 weatherpi weewx[20726]: wxengine: garbage collected 2380 objects
Oct 15 12:03:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x000000
Oct 15 12:07:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x000000
Oct 15 12:13:02 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b4e0
Oct 15 12:17:50 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b540
Oct 15 12:19:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x000000
Oct 15 12:41:03 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b6a0
Oct 15 12:42:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b6c0
Oct 15 12:45:03 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b6e0
Oct 15 12:57:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x000000
Oct 15 13:01:02 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b7e0
Oct 15 13:09:03 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b860
Oct 15 13:10:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b880
Oct 15 13:13:02 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b8a0
Oct 15 13:19:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x000000
Oct 15 13:21:03 weatherpi weewx[20726]: fousb: station status {'unknown': 0, 'lost_connection': 0, 'rain_overflow': 0} (0)
Oct 15 13:21:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x000000
Oct 15 13:29:03 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b9a0
Oct 15 13:31:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00b9e0
Oct 15 13:33:50 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00ba00
Oct 15 13:35:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x000000
Oct 15 13:47:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00bae0
Oct 15 13:51:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x000000
Oct 15 13:53:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x000000
Oct 15 13:55:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x000000
Oct 15 14:01:03 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00bba0
Oct 15 14:03:59 weatherpi weewx[20726]: fousb: unstable read: blocks differ for ptr 0x00bbe0

[mwall]

On Wednesday, October 15, 2014 10:08:46 AM UTC-4, Auchtermuchty Weather wrote:

Noticed gaps in my graphs and found lots of this in /var/log/syslog.  This is with a new SD card in the RPi, and I'm using the latest version of WeeWX (v2.7.0) with a WH1080 weather station:

unstable read is not unusual for fine offset stations.  however, you are getting a lot of them, so i suspect you are not using the optimal fineoffset settings.  be sure to read the "use correct settings" section here:

https://sourceforge.net/p/weewx/wiki/folockup/

at 11:34:05 your station lost connection with the sensors.  by 13:21:03 it had re-established the connection.

what archive interval are you using?  factory default for fineoffset is 30 minutes.  you probably want 5 minutes.

http://weewx.com/docs/usersguide.htm#dots_in_plots

http://weewx.com/docs/usersguide.htm#wee_config_fousb

m

[Andrew Milner]

at 11.34 you lost the radio connection between the transmitter and console.  This would have shown as -- on the console display.  At 13.21 the connection was restored.  The WH1080 does nto exactly have the most robust of transmitters!!  If you are inside you can sometimes speed up the process by pressing and holding the down arrow key to make the console search for the remote unit - otherwise it will usually correct itself.  If it is a major problem you may need to relocate either the console or the tranmitter and get them as close as possible and preferably in line of sight.  The console radio reception can be affected by computer monitors, tv's, other electronics and badly routed cables.

Don't worry about the unstable reads - they are par for the course and do not necessarily indicate a problem.

[ramw...@uk2.net]

So ultimately the station is the problem...  What would be a reasonable upgrade without going crazy on price?  I might be able to stretch to on.

Also, the RPi is using a wired network connection at present.  If I switched to a wireless dongle I could position it better, but could that cause other problems?

[mwall]

On Wednesday, October 15, 2014 6:33:42 PM UTC-4, ramw...@uk2.net wrote:

So ultimately the station is the problem...  What would be a reasonable upgrade without going crazy on price?  I might be able to stretch to on.

the fine offset stations are definitely low end.  however, they can produce reasonable data.  we've had some running for 3 years in marine environments - salt spray, heavy winter storms, temperatures from well below freezing to above 90F.  on the other hand, in a different location we have had 2 ws2080 consoles fail in a 4 year period (one bad pressure sensor, the other started reporting bogus values from all sensors).

if you do not mind changing batteries, the lacross ws28xx stations are not bad, and often available for less than the fine offset.  avoid the early ws281x models.

the sensors in the meade/honeywell te923 series are perhaps the most accurate of any low-end to mid-range stations.

oregon scientific stations get pretty good reviews, can easily expand to multiple sensors, and they are often available at steep discount. they now have a direct competitor to the vantage in the wmr300 station.

the davis vantage stations tend to be the most reliable of the 'hobby' weather stations, and their pricing reflects this.  they still fail, but davis support is quite good.

there are many, many different stations available for over 1000$US : weatherhawk, rainwise, vaisala, irdam, kestral, environdata, capricorn, hobo, young, columbia, texas weather instrumens, instromet, novalynx, campbell scientific

you can see some alternatives here:

http://weewx.com/hwcmp.html

the hardware currently supported by weewx is here:

http://weewx.com/hardware.html

m

[Andrew Milner]

The RPI on wired is fine, use a longer USB lead between Pi and indoor part of weather station to reposition the weather station and see if it makes a difference.  I have my weather station about 7 metres from the RPi so that I can position it on a wall well away from interference and give minimum walla between transmitter and indoor station.

[ramw...@uk2.net]

Since I want a weather stations that WeeWX supports with the instruments on one pole, as per the FineOffset it looks like I'd have to fork out for the Davies Vantage - almost £400 with the 'thing' so I can plug the console into the RPi.

I'll play around with moving the receiver and since I don't want 10m of cable trailing around my livingroom, I'll get the RPi working wirelessly - I have a dongle and another Pi that is wireless. 

[Andrew Milner]

it will work wirelessly no problem but the problem you have is between the indoor cosole and the outdoor sender.  I would recommend that you use a powered USB hub for the RPI and plug weather station console and wireless dongle into the hub.  Do keep the weather station console as clear of the Pi, TVs etc as you can though.  My long wiring between console and RPi is pretty much hidden though.

[ramw...@uk2.net]

I wonder if the new model RPi (B+) would work better?  I gather they have upgraded the USB part of it, obviously the power adapter would need to be up to the mark.

OTOH I might have a powered USB hub lurking somewhere....  Repositioning the console would be a good idea as well, it's far from ideal at present.

[ramw...@uk2.net]

I should have been clear that the point of changing the RPi to Wireless is to make it easier to get the console into a better position.

[mwall]

On Thursday, October 16, 2014 11:43:41 AM UTC-4, ramw...@uk2.net wrote:

I should have been clear that the point of changing the RPi to Wireless is to make it easier to get the console into a better position.

 another option is to extend the length of the antenna in the weather station console.

see the 17sep2013 entry 'how to increase the ws2080 range':

http://lancet.mit.edu/mwall/projects/weather

with this modification we now get a consistent connection with the console located about 100 meters from the sensors, with a stone foundation and 4 wood framed walls between them.

extend the antenna in integer multiples of the existing antenna length so that you get the wavelength/frequency right.

m

[ramw...@uk2.net]

I'm not up for taking my console apart!

After some reading I've decided to get a RPi B+ with a decent power supply.  The B+ apparently has a significant improvement to the power regulation and hence power available to USB devices, so along with the decent power supply (not that I think there is a problem with the current ones!) I should be able to use the USB wireless dongle as well as plug in the weather station and have enough power.  That will let me position it much better and see how it goes. 

A powered USB hub would be a cheaper solution, but this is a much neater one - just one mains plug, far fewer cables.

I can simply copy the image from the SD card to the micro-SD card for the B+ and then use raspi-config to expand it - the SD card is 4GB and I've ordered an 8GB class 10 card.

I've also added my station to WOW, which was very easy once I'd got over the slow mess that's the Met Office website.

[Andrew Milner]

I just have the one mains plug also - to the powered hub and then take the RPi power from the hub!!  Anyway, my suggestion of the longer cable was just to see if repositioning helped rather than it being  permanent solution.  However you do it - make sure the weather station console on the end of its USB wire is well away from electrical interference as the electrical interference will mess up the weather station's ability to receive the outdoor sensors.

--
You received this message because you are subscribed to a topic in the Google Groups "weewx-user" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/weewx-user/13F6NbkPObs/unsubscribe.
To unsubscribe from this group and all its topics, send an email to weewx-user+...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

[Auchtermuchty Weather]

The penny hadn't dropped that USB -> MicroUSB would let the hub power the RPi!

However going wireless will let me move it all to the other end of the living room where the console should be able to 'see' the weather station, away from the TV etc.

I also won't have too many long wires trailing across my living room.  A neat run from one end to the other would be 10m or so - the room itself is almost 8m long.  I'm sure the cats would love playing with cables but it doesn't do them any good.

[mike.t...@noworries.plus.com]

Just to give you some confidence that it'll work I have the same set up as you; RPi B was wired and now wireless with a fine offset (Maplin) PWS. The console and the sensors are 100m apart, been running since January. I did have a couple of data issues recently, nothing significant

[Andrew Milner]

My ordinary RPi and FO2080 work fine also.

On 21 October 2014 19:53, <mike.t...@noworries.plus.com> wrote:

Just to give you some confidence that it'll work I have the same set up as you; RPi B was wired and now wireless with a fine offset (Maplin) PWS. The console and the sensors are 100m apart, been running since January. I did have a couple of data issues recently, nothing significant

[ramw...@uk2.net]

It has occurred to me that changing the batteries in the transmitter might be a good idea as well!

The SD cards are on their way (bought 2 for free shipping), the RPi B+ along with a case & a decent power supply are also on their way so I should be able to move it a few days. 

[Andrew Milner]

that should have been the first thing to have checked!!  Use good quality ones for longer life.

On 22 October 2014 13:48, <ramw...@uk2.net> wrote:

It has occurred to me that changing the batteries in the transmitter might be a good idea as well!

The SD cards are on their way (bought 2 for free shipping), the RPi B+ along with a case & a decent power supply are also on their way so I should be able to move it a few days. 

--

[Auchtermuchty Weather]

I realised that....  I have a pack of the disposable lithium ones on the way.  They work until they suddenly stop working...

[Andrew Milner]

I take it you are in a cold part of the world then with temperatures regularly below 0 - otherwise alkaline are better I believe.

[ramw...@uk2.net]

I'm in Scotland and they far outperform alkaline batteries in my mouse, which is in the nice warm house!

[ramw...@uk2.net]

And now the console cannot connect at all to the weather station...  Starting to wonder if the weather station hardware is the problem.  :(

[Andrew Milner]

For the time being just forget about weewx.  Disconnect the computer and make sure the console and outside sensors are talking to each other - if necessary get within a few metres of the transmitter.  Follow the instructions in your station manual for getting the console to 'find' the sensors - it is either down arrow or history that you press and hold for 5 seconds, .  Once they are talking bring the console back indoors and make sure that they continue to 'see' each other.  Once you are happy then reconnect to the computer - but if you are using wifi try and keep the wifi and raspberry pi such that they are not between the console and the outdoor sensors - ideally position the console higher up and away from everything.  Then see what happens.

Also I seem to recall that the outdoor sensor does not transmit if it is trying to synchronise the time with the radio time signal - if you have one of those types of station - so you may have to be patient!!

[ramw...@uk2.net]

While you were typing this is what I was doing...  I found the transmitter was 'stuck'.  Brought it in, took the batteries out of transmitter & receiver, put them back in the transmitter then the receiver and it connected.  I'm not using Wifi, I screwed a 4-way to the bookshelf I've moved everything to (all mounted on round-head screws) and used a Homeplug to connect the RPi to the network.  The cabling looks a bit of a rat's nest - they are all rather long - but it works.  I've had to coil them out of the way to avoid the cats playing with them.  The website has started service pages again so it's mostly working now.  Phew!

However I still have a couple of problems:

Oct 24 17:28:20 weatherpi weewx[2210]: fousb: unstable read: blocks differ for ptr 0x000160
Oct 24 17:30:15 weatherpi weewx[2210]: fousb: station status {'unknown': 0, 'lost_connection': 0, 'rain_overflow': 0} (0)

[mwall]

On Friday, October 24, 2014 12:40:32 PM UTC-4, ramw...@uk2.net wrote:

However I still have a couple of problems:

Oct 24 17:28:20 weatherpi weewx[2210]: fousb: unstable read: blocks differ for ptr 0x000160
Oct 24 17:30:15 weatherpi weewx[2210]: fousb: station status {'unknown': 0, 'lost_connection': 0, 'rain_overflow': 0} (0)

'unstable read' is not unusual for fine offset stations.

the 'station status' is reported every time the station status changes.  btw, the 'rain_overflow' bit is quite possibly a 'battery status' bit, not rain overflow.  if you see that one change, please let us know.

m

[ramw...@uk2.net]

OK, so mostly I'm back in business and I need to keep an eye on the station status message.  I can also get on with rejigging the HTML to suit.  :)

[ramw...@uk2.net]

Sigh...  Connection is lost, got the rain_overflow message:

Nov 28 00:52:22 weatherpi weewx[18110]: fousb: station status {'unknown': 0, 'lost_connection': 64, 'rain_overflow': 0} (64)

I put new batteries in both transmitter & receiver a month ago, they are disposable Lithium batteries so should still be good at present.

[Andrew Milner]

You can get a lost connection for a milion and one different reasons, of which battery is just one reason!  Do not panic ... you can force the console to try and reconnect by pressing history (or down arrow) - I forget which and holding it until you see the radio reception signal on the console.  It should reconnect.  It wil probably reconnect all by itself if you do nothing at all.  I presume the console is showing ---- - if it is showing the outdoor readings then weewx should have just picked up where it left off, and issued another status message this time with lost connection : 0 which indicates the connection has been restablished between console and outdoor sensors.

--

[ramw...@uk2.net]

I tried this, it's a FineOffset WH1081, and it doesn't seem to make it reconnect.  :(

[Andrew Milner]

Does it reconnect if you just move the two units closer together and maybe with fewer walls between?  Forget weewx for now - radio reception has nothing to do with the PC. You need to have the two units 'talking' reliably before you try and read the data out onto a pc.

[ramw...@uk2.net]

Moved the two units closer, tried the down-arrow and it suggested there was a battery problem with the transmitter.  Brought them in, tested them plus a couple of new ones and identical voltages all round.  They are disposable Lithium cells - is there anything better for the transmitter?

It reconnected once I'd removed and reinstalled the batteries in both transmitter & receiver.  It's been stable in the current position for a month, was stable in a worse one (next to the TV) for much longer, and would occasionally drop and pick up again with no interaction from me.  Something was locked and needed the equivalent of the 3-finger salute. 

Not sure if this is because the WH1080 is aging - it is a cheap'n'cheerful weather station.

[ramw...@uk2.net]

PS would the WS1080 be any better?  It uses solar power for the transmitter.  Not sure where I could get it though.

[Andrew Milner]

Looking at various Australian sites the solar ones seem to have their own fair share of problems also.

Your batteries could be being affected by  extreme weather conditions (cold and/or heat) that defeated even the Lithium batteries you are using.  It could be solar flare related, cellphone usage related, neighbour's radio transmission related or just be plain old simple FineOffset quality issues!!!!

As you say - they stations are cheap and cheerful, and work 99% of the time.  For the price - they do their job!

Sorry not to be more helpful.

[Mike Thompson]

I am suspicious that the disposable Lithium batteries don't like cold weather. My hypothesis is that the voltage may drop and hence the output of the transmitter drops. We haven't had enough cold weather yet to be absolutely sure..

FYI I have a fine offset (Maplin) as well and the sensors are some 100m from the console so right on the limit, but it's a reliable connection. Initially I used Alkaline disposable batteries.Several months in I did have some connections problems but this was due to the console being moved a few inches rather than the batteries. I thought Lithium might provide more power but as I say I think they are sensitive to freezing temperatures. I'm going to swap back to the Duracell Alkaline.

[ramw...@uk2.net]

I found the data sheet for the batteries I'm using and they should be fine given it rarely gets below -5C here - indeed I don't think there is a record below 0C since I put the batteries in.

http://data.energizer.com/PDFs/l91.pdf

[Andrew Milner]

So it could be another of the million reasons - or even just a poor battery connection. 

[kpol...@kpolberg.net]

I'm also seeing issues with connection on a WH1080 and a WH3080. Though for some reason they seem to be grouped around the same time(sunrise and sunset). As I am clearly getting the pressure, I have come to the conclusion that I am not having the problem where the console locks up. But that the wireless connection between the transmitter and the console has been lost(must admit that I first thought it was the upgrade from 2.6.4 to 2.7.0 that broke it).

select *,dateTime(dateTime,'unixepoch') as dato from archive where outTemp is NULL;

1416553870|1|5||30.1358535144714|30.3014424309099|41.9||68.0||||||0.0|0.0|||||0.0|0.0|||||||||||||||||||||||||||||||2014-11-21 07:11:10
1416554171|1|5||30.1388068517425|30.3044096075231|41.9||68.0||||||0.0|0.0|||||0.0|0.0|||||||||||||||||||||||||||||||2014-11-21 07:16:11
1416554471|1|5||30.1417601890136|30.3073767838777|41.9||68.0||||||0.0|0.0|||||0.0|0.0|||||||||||||||||||||||||||||||2014-11-21 07:21:11
1416555378|1|5||30.1388068517425|30.3044096075231|41.72||68.0||||||0.0|0.0|||||3.950625|0.0|||||||||||||||||||||||||||||||2014-11-21 07:36:18
1416555660|1|5||30.1388068517425|30.3044096075231|41.9||68.0||||||0.0|0.0|||||6.2221|0.0|||||||||||||||||||||||||||||||2014-11-21 07:41:00
1417333560|1|5||29.9379799173066|30.10264100718|46.04||68.0||||||0.0|0.0|||||0.0|0.0|||||||||||||||||||||||||||||||2014-11-30 07:46:00
1417333861|1|5||29.9409332545777|30.1056082011875|46.04||68.0||||||0.0|0.0|||||0.0|0.0|||||||||||||||||||||||||||||||2014-11-30 07:51:01
1417334160|1|5||29.9409332545777|30.1056082011875|46.04||68.0||||||0.0|0.0|||||0.0|0.0|||||||||||||||||||||||||||||||2014-11-30 07:56:00
1417334460|1|5||29.9409332545777|30.1056082011875|46.04||68.0||||||0.0|0.0|||||0.0|0.0|||||||||||||||||||||||||||||||2014-11-30 08:01:00
1417334761|1|5||29.9409332545777|30.1056082011875|46.04||68.0||||||0.0|0.0|||||0.0|0.0|||||||||||||||||||||||||||||||2014-11-30 08:06:01
1417335060|1|5||29.9438865918488|30.1085753949345|46.04||68.0||||||0.0|0.0|||||0.0|0.0|||||||||||||||||||||||||||||||2014-11-30 08:11:00
1417335360|1|5||29.9409332545777|30.1056082011875|46.04||68.0||||||0.0|0.0|||||0.0|0.0|||||||||||||||||||||||||||||||2014-11-30 08:16:00
1417336560|1|5||29.9379799173066|30.10264100718|46.04||68.0||||||0.0|0.0|||||2.398325|0.0|||||||||||||||||||||||||||||||2014-11-30 08:36:00
1417336861|1|5||29.9379799173066|30.10264100718|46.22||68.0||||||0.0|0.0|||||6.489775|0.0|||||||||||||||||||||||||||||||2014-11-30 08:41:01
1417337160|1|5||29.9409332545777|30.1056082011875|46.04||68.0||||||0.0|0.0|||||2.679975|0.0|||||||||||||||||||||||||||||||2014-11-30 08:46:00
1417337461|1|5||29.9409332545777|30.1056082011875|46.04||68.0||||||0.0|0.0|||||4.09145|0.0|||||||||||||||||||||||||||||||2014-11-30 08:51:01

There is some clever cut and paste in between here, so the numbers aren't exactly what the sql query is giving out. But without using too much time, this is the pattern that I think I am seeing(from both weather stations :)

kp

You received this message because you are subscribed to the Google Groups "weewx-user" group.
To unsubscribe from this group and stop receiving emails from it, send an email to weewx-user+...@googlegroups.com.

[Andrew Milner]

If sunrise and sunset seem to be the main times when you have problems then the odds are probably stacked in the direction of solar activity being disruptive, and maybe your stations are just a little too far apart to overcome any interference.  What happens at sunrise/sunset if they are closer together - do you still have problems under those conditions??

[kpol...@kpolberg.net]

Unfortunately these weather stations are located in two separate cabins a distance from where I live. So I haven't been able to test / do anything about it. As far as distance between the transmitter and the console, I'm not even sure I can get it any closer.

But I must say, for the price of the weather station, and the reliability / accuracy of it. I'm quite happy, though I have connected a remote controlled power switch to the rPi powering the console. In case the console should freeze, I would still be able to cut power and revive the console(console is without batteries).

kp