Solar and UV sensors going off line - fix?

[JohnF]

Haven't had to post on these forums for some time as my Watson W-8681 solar has been behaving itself. However, in recent days I've lost the UV/Lux data on the display. Changed batteries, tested them etc., reset and all that. Then I read here there may be too much light for the solar sensor, it is very sunny and hot here today, so I've partly covered the solar sensor with some amalgamating tape and data is OK now - fingers crossed.
HTH
John

[Feek]

Now this is very odd.

After two months of not working, I just noticed that mine has come back!

The console has been showing low battery at night recently which I mentioned before but I didn't get a chance to do anything about it. I've just glanced at the display and saw it had numbers rather than dashes. I'm not sure exactly when it came back but I bet it's since the weather improved.

[JohnF]

I'm not sure exactly when it came back but I bet it's since the weather improved.

That observation appears to be the exact opposite of my situation
John

[Feek]

I'm not sure exactly when it came back but I bet it's since the weather improved.

That observation appears to be the exact opposite of my situation
John

I know and that's what surprises me.

Two months ago I noticed it had stopped working. After a while I spotted the low battery warning but it seemed that was only showing at night time. During the day it cleared. It's as though the batteries were on the verge of the low warning and the daylight would give them a jag of charge to cause the low battery warning to go off. Now it's been nice and sunny for a couple of days it seems that the batteries have charged enough to kick the sensors back into life again.

I've still got the Eneloops in there.

[JohnF]

Hi Feek
The odd thing is that I never got a low battery warning. ATM the UV is reading 11 and Lux 46 with the piece of tape over the sensor and the data is still solid.
What is surprising is that we have a cloudless sky, strong breeze, steady barometer and yet it is predicting rain which is spot on for Cornwall this evening. Cumulus came up with "Unsettled, some precipitation" That's clever.
John

[AllyCat]

ATM the UV is reading 11 and Lux 46 with the piece of tape over the sensor and the data is still solid.

Hi,

As Summer approaches here in the UK, we're probably now seeing the "Australian failure mode" ("crashing" of the solar module software due to high light levels) rather than the "UK failure mode" (excessive power drain by some solar pods causing flat batteries).

Note that there are three "Solar Sensors": The sloping PV panel might be charging the batteries to too high a voltage (but the circuit appears to include a voltage regulator to prevent this), or a computational "overflow" of the UV calculation might easily occur, because most/all of the units clearly over-read very considerably (the highest UV value in the UK should be about 7 or at most 8). However, John's comment above (and I think some from Australian members) suggests that it is the Lux value (from the "domed" sensor) which is causing the problems. Rather strange; have you ever seen Lux values approaching the 400k which is supposed to be the upper limit of measurement?

Cheers, Alan.

[dpmiller]

Superb service from FO, new sensor arrived today. Before installing it, I added small squares of yellow insulating tape over the housing above both Lux and UV sensors to calm their response a bit. Let's see how we get on now...

[ejay]

So has anyone worked out which sensor is causing the crashing?
I've heard of people having success covering up either or both of the UV and Lux sensors.
Is it excessive UV that causes the crashing? Or excessive Lux? To date, I have not really seen a conclusive answer to this.

[oman181]

Add another Australian user to the list of dissatisified clients, both FO and their reseller.
Mine has been playing up since January and been back to the ebay seller who had it for about 6 weeks and found no problem. I reinstall and the Solar/UV sensor info stops 2 days later.

I have been through this thread and the other one (https://cumulus.hosiene.co.uk/viewtopic.php?f=13&t=7512) and tried a few things with no luck. (1) new batteries (2) tape on both sensors, one at a time. Haven't done the open up solder thing as I keep pushing the reseller to fix it. It has to be sold fit for purpose and this model isn't so their obligation I believe. Plus I hate not knowing!

Mine has stopped working at 4am going by the data logs so not sure about the overload of sensor issue.
Have never seen a low battery alert on the display panel.

Mmm, is the low battery alert a permanent alert until cleared? ie. if there was some weird circumstance that caused the battery to go below the threshold but then recover (sunlight ?), will the low battery alert clear? If so, how can I find out it was present - log the comms?

From reading the forums and my own experience;
- I understand that the solar panel assembly contains the UV sensor, LUX sensor and solar panels.
- The solar panels recharge the batteries
- The UV and LUX are sent on a separate 'cycle' to the other readings, but they use the same radio module.

It seems the solar panel assembly has it's own processor and then communicates 'with the radio module' to make the transmission. I don't know if the solar panel assembly communicates directly with the radio module or communicates with the transmitter-battery unit which manages communications with the radio module - maybe others can help out with this one.

Possible causes:
(1) power to the solar panel failing OR a short term voltage drop in power causing processor halt
(2) processor in a loop that it can't recover from
(3) communications failure between solar panel and radio module/transmitter-battery unit


Mine is installed on a pole attached to a shade sail on the roof and a real pain to get at for a reset!

I would like to the info received by the base unit but would like to know if I hook up the USB to a laptop, can I find a program that will log the info to disk? Is there infor continually streamed to the USB?

Since I am recovering from a hernia op, I have no plans to go up the roof more than once!

Cheers, Phil

[ejay]

I had another thought of what might be causing the 'crashing'.

It is established the temp/hygro/wind readings are transmitted every 48 seconds, and the solar readings transmitted every 60 seconds.

When all is well, the two transmissions happily occur at their specified intervals. But what happens when they coincide? If the temp/hygro is transmitting and the solar pod also wants to transmit at the same time, who wins? Or does the solar pod crash out due to the timing clash? Could that be the reason why it is failing?

[mcrossley]

They never clash, 48 and 60 do not produce the same multiples - I had to sit down and figure this out when these units first came out.

[steve]

They never clash, 48 and 60 do not produce the same multiples - I had to sit down and figure this out when these units first came out.

They clash every 4 minutes - or am I misunderstanding?

[mcrossley]

They clash every 4 minutes - or am I misunderstanding?

Well they do, but only if you start the clock for both of them at the same time. If you introduce an offset at the start of the sequence for one of them then they should never clash - that offset will be maintained every 4 minutes and you start the sequence again.

[dpmiller]

this assumes that the crystal clocks in both sections do not drift appreciably. Remember that one of FOs first "fixes" was to solder down the casing of the crystal in the solar pod.

If (say) the crystal in the pod changes by 0.001% due to temperature (or humidity on the pcb, or whatever) then you could work out how often the data streams would overlap...

[AllyCat]

Hi Phil,

is the low battery alert a permanent alert until cleared?

No, on the 308x versions the Low battery alert seems to be updated at each transmission. On the earlier 108x, etc. models it appears to only update around midnight.

It seems the solar panel assembly has it's own processor and then communicates 'with the radio module' to make the transmission. I don't know if the solar panel assembly communicates directly with the radio module or communicates with the transmitter-battery unit which manages communications with the radio module - maybe others can help out with this one.

On the current 308x models the wireless transmitter is shared by the Solar and other External sensor microcontrollers. The data is carried on one of the six wires connecting the Solar and "Transmitter" modules. I think the Transmitter micro "synchronises" itself to the Solar data packets so that they never collide (when you reset or connect the Solar Pod it always seems to "unsynchronise" the other "External" data transmissions).

Internally, the Solar Pod PCB includes tracks for a separate transmitter, but I guess FO have decided to save costs and (currently) share the transmitter. However, if the pod were operated independently (neither has a wireless receiver so cannot "know" what the other is doing) then the two transmissions will "drift" between themselves because of crystal tolerances. Eventually (it could take many months), the two transmissions would collide, which is probably why FO have chosen the 60 seconds period. If two transmissions do collide, then the next 4 from the External sensors and 3 from the Solar Pod will still get through.

Cheers, Alan.