Uploading to PVoutput stops in the morning

Topics: 5. Support
May 7, 2015 at 9:45 AM

Last monday I installed a RPi with SBFspot V3.0.3 (according 'SBFspot V3 install on Raspberry Pi EN 1.0.pdf') to log the power production of a SMA 3300 over Bluetooth. The location of the panels is a little south of Eindhoven, the Netherlands.

Although it looks to be working fine the output to PVoutput stops round 11 o'clock in the morning, see the "summary table" below:
  • The Timezone is set to Europe/Brussels, must this be set to Europe/Amsterdam?
  • The Locale is unchanged: en-US, must this be set to nl-NL?
  • The SBFspot.sh script contains
    /usr/local/bin/sbfspot.3/SBFspot -v
  • The script is set to run between 6 and 23, in the SBFspot.cfg;
  • The cleanup section isn't applied.
Can somebody point me in the right direction on how to have the entire power production be uploaded to PVoutput?

Thank you in advance.
May 7, 2015 at 10:41 AM
Check the time of your inverter.
May 7, 2015 at 11:29 AM
SBF wrote:
Check the time of your inverter.
Checked the time of the inverter in Sunny Eplorer and this is correct.

I noticed that the timezone in PVoutput was incorrect (UTC +2), updated this to UTC +1, but still no luck.
May 7, 2015 at 8:37 PM
dear "alootje"

to get better understanding on what your system is actually doing
please take following steps and provide the following info,
note that there are two distinct parts

-a- SBFspot
executes the interrogation of your inverter and sotees info in the database
for SBFspot you need to install a cron task to execute it each 5minutes

-b- send upload to PVoutput
the crontask for this is installed according the instructions found in the documentation

step/step controls - to be aware about the functions/output from your system: controls on SBFspot
-c- allow the csv outputs from your system (see parameters in the SBFspot.cfg file)
-c.1- the spot.csv file informs you about each execution of SBFspot
-c.2- jjjjmmdd.csv file informs about the production of your inverter (with intervals of 5min)
-c.3- jjjjmm.csv file informs about the production of your inverter (for each previous day)
provide a copy of these three xxx.csv files for validation

-d- execute at least once SBFspot with the following added options (use a command window or ? ? ? )
/usr/local/bin/sbfspot.3/SBFspot -d5 -v5 1>>catch-out.txt 2>>catch-err.txt
the commands 1>>catch-out.txt 2>>catch-err.txt will send the output into a file on a directory -> directory depends on set-up of your OS
if not in the dir: /usr/local/bin/sbfspot.3/ you can search on the name "catch-"

have a look into the information that is included in the catch-xxx.txt files
it might already indicate info about abnormal settings
send a copy of: -catch-out.txt -catch-err.txt -SBFspot.cfg -xxxx.csv

from step c and d you control whether the part "gather-info-from-inverter" is correct and runs during the expected time frame

form experience: if that part is ok then the next step might be ok

kr wim
May 8, 2015 at 8:53 AM
Hello Wim,

Thank you for your detailed explanation, hopefully I can send a copy of -catch-out.txt -catch-err.txt -SBFspot.cfg -xxxx.csv tomorrow.

Kind regards,
May 9, 2015 at 12:58 PM
When running the /usr/local/bin/sbfspot.3/SBFspot -d5 -v5 command the Pi was unable to connect to the inverter, therefore I'm unable to post the requested catch files. The other files can be found here:

So it could well be that the bluetooth dongle (Logitech or Trust, can't remember) is the fault. When running hcitool scan the inverter gets found on some, but not all occasions. I will exchange this dongle with another one later today to see if I can run the /usr/local/bin/sbfspot.3/SBFspot -d5 -v5 command succesfully.
May 12, 2015 at 7:10 PM
Edited May 13, 2015 at 8:28 AM
dear alootje, allow me the following remarks

from experience: your BT signal strength (ref spot-table) is at limit,
(below58% = NOK; around 60% ~not stable above 70% is OK)
read my trics below

ANYWAY - - try to get "direct-view" = NO obstacles between dongle and inverter,
-tric-1- put an USBcable between RPi and dongle - BT is low-speed - a cable, even 5meter, might not harm
-tric-2- move your RPi to reduce "distance" to the inverter,

what i'd learn from your info about the signal quality:
-a- you'll see that in the third of your "spot" records - several values are missing (zero) in the record
-b- nevertheless you obtained the "daily" and "monthly" files (perhaps not with each execution)
-c- nevertheless you have 3 valid records (but don't know how many #trials you did execute)

your system might function better with some adaptations (using your actual dongle)
  • consider another dongle (class-1) if long-term stability of your-dongle is not good
i can't judge on your local situation, but the following "trics" can help to improve the signal strength
all these trics are related to "experienced distance" (read signal loss) on the traject between
the "BT-antenna" of your inverter and the "BT-antenna" of the RPi
** "distance" - obstacles like a wall have impact on the functional distance,
the indication "10meter" for standard BT-dongle, is given as "direct-view"
ANYWAY - - try to get "direct-view" = NO obstacles between dongle and inverter,
-tric-1- put an USBcable between RPi and dongle - BT is low-speed comm's - an USB cable, even 5meter, might not harm
-tric-2- move your RPi to reduce "distance" to the inverter,
May 15, 2015 at 12:38 PM
Dear SillWimons,

I've tried three different dongles:
  1. Started with a dongle from thrust (don't know the exact type);
  2. A dongle from modmypi.com (http://www.modmypi.com/raspberry-pi/accessories/bluetooth-dongles/nano-usb-to-bluetooth-dongle-v2.0);
  3. Logilink BT015.
All dongle are nano-sized.

With the first two dongles I had no luck, with the third one (Logilink bT015) the system seems to be working as expected: the values get updated the entire day, for the last days.

Trying to export the catch-out and catch-err is still not successful. Should I bother for this?

Anyway, the system seems to be up and running!
Marked as answer by SBF on 7/12/2015 at 11:30 AM
May 16, 2015 at 11:12 AM
Edited May 17, 2015 at 10:13 AM
Dear alootje, for the files "catch-out.txt" it is abnormal that you can not export hem

please control the instructions that you use in the commands

-b- about the catch-out.txt and catch-err.txt files
-b.1- !!!!! be aware about the meaning of the redirect indications that i did give

if you do use the below command in the operational = continue mode then MODIFY THAT AS SOON AS POSSIBLE
it is intended for TEST PURPOSES ONLY (single or few executions)
/usr/local/bin/sbfspot.3/SBFspot -d5 -v5 1>>catch-out.txt 2>>catch-err.txt

!!! replace ASAP the redirect indications such that the command becomes
/usr/local/bin/sbfspot.3/SBFspot -d5 -v5 1>catch-out.txt 2>catch-err.txt

the double arrow in the redirect instruction 1>>catch-out.txt means that after each execution the additional text is added to the file
for that reason that file might become VERY LARGE and due to its size it might be that you can not export it

-a- your trials with BT seem to confirm that the "distance" between your RPi and inverter is too high
-a.1- if during your test spot values were written then you can find their an indication for the signal
-a.2- most trust BTdongles are class2
-a.3- logilink BT0015 is class1 (from my experience it works for distances >25meter)
note: the BTmodule in the inverter is class1

kind regards wim