Wow! This is detailed explanation! Thanks. I was not aware there are nuances. 
Clearly, USB side in the FO is the culprit. We can not change it.

I did not know some glitches are recoverable by Weewx. For the master glitch, 
we are back to killing USB power with no batteries in FO. I will wire something 
up.

As for RPi, I have 7 of them running 24/7. RPi 3 is touchy. RPi 2 and 4 not so 
much. As I said before, my only Zero works 6 weeks perfectly against all odds 
powering FO. Unfortunately, I can kill USB power only on RPI 2 and 3, not on 
Zero and 4 would be a very wrong choice for the job anyway.

Na 10. listopada 2019. 13:26:55 CEST, Andrew Milner 
<[email protected]> wrote:
>you have misunderstood completely
>
>1. fineoffset usb connection can glitch at any time for no predictable 
>reason - maybe once a month, maybe once a year.  On many occasions it
>can 
>glitch and weewx is able to recover the connection.
>
>2. sometimes fineoffset glitches in such a way that the only solution
>is to 
>completely power off the fineoffset, remove any batteries, and restart
>it - 
>this does not happen very often.  the problem is inside the fineoffset
>and 
>is not connected to weewx, rpi or anything else.
>
>3. an alternative recover for the problem in 2 above is to have no 
>batteries in the fine offset, power it from a powered usb hub (a usb
>hub 
>with its own independant power supply) which is able to selectively
>flip 
>power on individual ports and do the power cut/restore via the usb
>supply.  
>Only certain powered usb hubs are able to do this - check weewx threads
>for 
>more details.
>
>4. the rpi (and pizero) is known for a poor susceptability to power 
>issues.  To avoid any such issues it is suggested to ensure a good
>beefy 
>power supply for the rpi and to avoid powering other devices off the
>usb 
>port (eg fineoffset).  however powering the fineoffset via powered usb
>hub 
>(either without the switcheable ports) is ok
>
>5. no solution has been found to avoid the glitches occurring.  the
>best 
>one can achieve is to try and recover when weewx is unable to recover. 
>
>
>6. it can run for months with no issues and then have 3 in a month
>
>7 the problem is a fault in the fineoffset firmware.  recovery demands
>that 
>the fineoffset hardware has no power (usb or battery) and is restarted.
> 
>Just killing the power over usb and running off battery will not stop
>the 
>problem occurring.
>
>
>
>
>
>On Thursday, 10 October 2019 14:11:53 UTC+3, Pila wrote:
>>
>> What you say, I read as: FO can have its batteries, it is enough to
>kill 
>> the power to the USB cable for a short time to fix USB connectivity? 
>> Meaning: A nutered (power cut, data only) USB cable between the Zero
>and 
>> USB Hub plus a SmartSWitch on the USB Hub power supply can restore
>USB when 
>> it fails (even with batteries in FO). I understood FO needs to be
>powered 
>> down.
>>
>> My proffessional deformation is to verify all the facts and measure 
>> everythng including things that supposedly do not work or should not
>be 
>> done. And then after having checked all facts, make decisions.
>>
>> This simple USB connection was both interesting to try and the only 
>> immediate thing I could do to connect my Weather Station permanently
>to RPi 
>> Zero without moving the station itself elsewehere. For anything else,
>I 
>> needed to get some more stuf which takes time at my present location.
>First 
>> step is always the same: hook it up, and if it works after a month or
>so, 
>> go on to the next step with it.
>>
>> I was willing to bet Zero + FO would not work at all! USB cable
>powering 
>> my RPi Zero is 3 meters (10 feet) long! I better not say measurements
>under 
>> load :) I was sure it would not work. Now, I am actually perplexed:
>it 
>> works perfectly fine over a month! On a Zaro where Node-RED and
>Mosquitto 
>> are using power needlesly, plus WeeWx and of course - Zero is
>powering the 
>> Fine Offset itself for few weeks now. But, somehow, against all odds,
>5 
>> weeks later, all is well!?! If I did not try it, I would not have
>believed 
>> it.
>>
>> Previously, I tested connection from my Fine Offset to a PC and USB 
>> stopped working. I beleived USB died. Now I decided to do a SmartHome
>thing 
>> which starts with a Weather station. So, when I plugged a Zero into
>FO, it 
>> suddenly worked years after I gave up thinking USB port is dead or
>corroded 
>> (small island, a VERY corrosive surrounding). Them eneloops in FO
>last 
>> forever, over a year.
>>
>> It was only when I started reading on WeeWX that I learned FO power
>needs 
>> to be fully cycled to restore USB connection. Reading info on Fine
>Offset, 
>> I understand it needs to be completely powered down, screen dead,
>history 
>> emptied. Only then it restores USB connection. So, when I said FO did
>not 
>> react to Zero rebooting, I meant - its screen remained unaffected by
>Zero 
>> rebooting. FO did not loose power long enought to restart. For now,
>USB 
>> connection was never lost, Zero is powering FO for the last 3 weeks
>(no 
>> batteries).
>>
>> WeeWX log upon manual reboot seems unremarkable:
>>
>> Oct 10 12:17:56 RPiZero kernel: [    2.673222] usb 1-1: New USB
>device 
>> found, idVendor=1941, idProduct=8021, bcdDevice= 1.00
>> Oct 10 12:17:56 RPiZero kernel: [    2.688210] usb 1-1: New USB
>device 
>> strings: Mfr=0, Product=0, SerialNumber=0
>> Oct 10 12:17:56 RPiZero kernel: [    2.713617] hid-generic 0003:1941:
>> 8021.0001: hiddev96,hidraw0: USB HID v1.00 Device [HID 1941:8021] on
>usb-
>> 20980000.usb-1/input0
>>
>> Oct 10 12:18:16 RPiZero weewx[345]: engine: Using configuration file
>/home
>> /weewx/weewx.conf
>> Oct 10 12:18:16 RPiZero weewx[345]: engine: Loading station type 
>> FineOffsetUSB (weewx.drivers.fousb)
>> Oct 10 12:18:16 RPiZero weewx[345]: fousb: driver version is 1.10
>> Oct 10 12:18:16 RPiZero weewx[345]: fousb: polling mode is PERIODIC
>> Oct 10 12:18:16 RPiZero weewx[345]: fousb: polling interval is 60
>> Oct 10 12:18:17 RPiZero weewx[345]: fousb: found station on USB bus= 
>> device=
>> Oct 10 12:18:17 RPiZero weewx[345]: engine: StdConvert target unit is
>0x10
>> Oct 10 12:18:20 RPiZero weewx[345]: fousb: synchronising to the
>weather 
>> station (quality=1)
>>
>>
>>
>>
>>
>
>-- 
>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 [email protected].
>To view this discussion on the web visit
>https://groups.google.com/d/msgid/weewx-user/8ec4f77d-3b79-4463-ae56-22a235a8a6eb%40googlegroups.com.

-- 
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 [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/weewx-user/5204628E-BCFB-42D7-BACA-7E32E17EF320%40gmail.com.

Reply via email to