Weather station for residential PV

[sensij]

I don't want to further hijack this thread on this topic, so here is a fresh start.

I'm considering adding a weather station to my home to learn more about how the weather and environment will affect the performance of my PV system. This is primarily a hobby type endeavor, but knowledge can have value and maybe someday it might actually be worth something.

There is an outfit (or two) in FL that sells Davis equipment, at what looks like very competitive prices. I'm thinking the following system would work for me, based on J.P.M.'s recommendations and information available through other forums:

6323 - Vantage Pro2 Integrated Sensor Suite with 24 hr FARS (no display console) - $407
6450 - Solar Radiation sensor - $110
6673 - Radiation sensor mounting bracket - $21
6316 - Envoy (not 8X) receiver - $120 (substitutes for the normal display console)
6510USB - Weatherlink USB logger - $109
Meteobridge hardware - ~$35 (repurposed router to act as data gateway)
Meteobridge software - ~$71 (allows me to publish data externally w/o needing my own code)
Shipping ~$50

Total is ~$923

Probably need to add a little for a mounting system, too.

Anyone else out there doing anything like this?

[lkruper]

I don't want to further hijack this thread on this topic, so here is a fresh start.

I'm considering adding a weather station to my home to learn more about how the weather and environment will affect the performance of my PV system. This is primarily a hobby type endeavor, but knowledge can have value and maybe someday it might actually be worth something.

There is an outfit (or two) in FL that sells Davis equipment, at what looks like very competitive prices. I'm thinking the following system would work for me, based on J.P.M.'s recommendations and information available through other forums:

6323 - Vantage Pro2 Integrated Sensor Suite with 24 hr FARS (no display console) - $407
6450 - Solar Radiation sensor - $110
6673 - Radiation sensor mounting bracket - $21
6316 - Envoy (not 8X) receiver - $120 (substitutes for the normal display console)
6510USB - Weatherlink USB logger - $109
Meteobridge hardware - ~$35 (repurposed router to act as data gateway)
Meteobridge software - ~$71 (allows me to publish data externally w/o needing my own code)
Shipping ~$50

Total is ~$923

Probably need to add a little for a mounting system, too.

Anyone else out there doing anything like this?

No, but I would love to. One thing you can do is check out Weather Underground. A lot of stations that broadcast for that site are Davis and you might get an idea about the features of sites in use.

I was looking at one station that reports solar insolation but they were evidently only displaying one day but I think it is configurable to store more historically.

[J.P.M.]

No, but I would love to. One thing you can do is check out Weather Underground. A lot of stations that broadcast for that site are Davis and you might get an idea about the features of sites in use.

I was looking at one station that reports solar insolation but they were evidently only displaying one day but I think it is configurable to store more historically.

Many Weatherunderground sites do indeed use Davis equipment. As mentioned in a prior post, the site near me uses Davis equipment pretty much identical to mine. The equipment is capable of 1 min. sampling and recording. Most folks seem to use larger time increments as a choice. In close to 2 years of operation, mine is on 1 min. increments and, to my knowledge, hasn't burped yet. All data is stored/archived.

[J.P.M.]

I don't want to further hijack this thread on this topic, so here is a fresh start.

I'm considering adding a weather station to my home to learn more about how the weather and environment will affect the performance of my PV system. This is primarily a hobby type endeavor, but knowledge can have value and maybe someday it might actually be worth something.

There is an outfit (or two) in FL that sells Davis equipment, at what looks like very competitive prices. I'm thinking the following system would work for me, based on J.P.M.'s recommendations and information available through other forums:

6323 - Vantage Pro2 Integrated Sensor Suite with 24 hr FARS (no display console) - $407
6450 - Solar Radiation sensor - $110
6673 - Radiation sensor mounting bracket - $21
6316 - Envoy (not 8X) receiver - $120 (substitutes for the normal display console)
6510USB - Weatherlink USB logger - $109
Meteobridge hardware - ~$35 (repurposed router to act as data gateway)
Meteobridge software - ~$71 (allows me to publish data externally w/o needing my own code)
Shipping ~$50

Total is ~$923

Probably need to add a little for a mounting system, too.

Anyone else out there doing anything like this?

FWIW, based on my experience, I'd get the display console. I'm ignorant of an 8X receiver with a display. my 8X receiver has no display. I didn't know the suite was avail. w/out the std. console/display. It may seem like an extraneous expense, but an additional (portable) receiver w/a constant display has advantages of portability, ease of use and added flexibility. Mine sits next to the keyboard I'm using and is about as portable as an ipad. Nice to have if I want to go up on the roof and root around, or watch the inverter console or other devices (revenue meter, etc. ?) simultaneously, or keep an eye on things while engaged in other activities. On the other had, it can always be added at a later time as needs/finances dictate. Either way, it's nice to have. Just sayin'.

Seems like the way Davis prices their stuff, it's possible to buy whatever configuration is wanted piece by piece.
I may look into the meteobridge stuff. Thankx for the heads' up.

[sensij]

FWIW, based on my experience, I'd get the display console. I'm ignorant of an 8X receiver with a display. my 8X receiver has no display. I didn't know the suite was avail. w/out the std. console/display. It may seem like an extraneous expense, but an additional (portable) receiver w/a constant display has advantages of portability, ease of use and added flexibility. Mine sits next to the keyboard I'm using and is about as portable as an ipad. Nice to have if I want to go up on the roof and root around, or watch the inverter console or other devices (revenue meter, etc. ?) simultaneously, or keep an eye on things while engaged in other activities. On the other had, it can always be added at a later time as needs/finances dictate. Either way, it's nice to have. Just sayin'.

Seems like the way Davis prices their stuff, it's possible to buy whatever configuration is wanted piece by piece.
I may look into the meteobridge stuff. Thankx for the heads' up.

On further review, I realized that I had decided against the 8X when I thought I would use the IP data logger... switching the the USB logger with the Meteohub will allow the 8X to be used after all. Now I'm wondering if the soil moisture sensors would help me more efficiently run my irrigation. You know, in case the drought ever ends. Slippery slope, this is.

It may be a mistake, but I'm going to try to go without a display console initially, since it looks likely that if I can get the gateway set up right, the data is available through Davis's mobile app and my phone could effectively become the display. The display can be added back in later for another $190, only a little bit more than the $160 difference in price between the 6323 (ISS only) and 6153 (ISS with console).

[J.P.M.]

On further review, I realized that I had decided against the 8X when I thought I would use the IP data logger... switching the the USB logger with the Meteohub will allow the 8X to be used after all. Now I'm wondering if the soil moisture sensors would help me more efficiently run my irrigation. You know, in case the drought ever ends. Slippery slope, this is.

It may be a mistake, but I'm going to try to go without a display console initially, since it looks likely that if I can get the gateway set up right, the data is available through Davis's mobile app and my phone could effectively become the display. The display can be added back in later for another $190, only a little bit more than the $160 difference in price between the 6323 (ISS only) and 6153 (ISS with console).

In spite of being of being about 2/3 Luddite, I think I understand. Different configurations and tools for different needs. To me anyway (and perhaps only), the data, its quality and the insights it provides are what matters much more than the electronic tools used to acquire and manipulate it. Similar to the arrays on our roofs, and the cars we drive, to me those tools are little more than a means to an end.

If you've made the decision and that decision is to buy, congrats. Respectfully suggest you think about sensor suite placement. Depending on what your purposes, location will obviously affect the data. For example, my placement makes the ambient temp. data representative of the air temp. near the array but not the ground. Even if so inclined, my data would be misleading, misrepresentative and useless as a Weatherunground site. Shading is an obvious consideration. I clean the solar sensor before each array temp. measurement, the rain gauge needs cleaning 1X/awhile and the battery seems to last about 9 mo. or so. Those things make servicing and access also an obvious consideration. Bottom line: Choose the sensor suite location just like choosing a vendor - wisely. The sensor suite can always be moved, but that breaks the data continuity a bit.

Get ready for the data deluge.

[sensij]

I finally pulled the trigger on the station, same hardware list as above. Stayed with the Envoy instead of the 8x to save a couple bucks, with the only trade seeming to be that additional temperature stations will only report at 1 deg resolution, not 0.5 deg. Figure another month or two to get it installed and the data handling software in place. I'll be sharing what I can through my PVOutput page, although it has a 5 min update rate at the fastest it will support, so some other database might become useful for higher frequency data, which can be as quick as a minute for GHI but down to 3 s for wind gusts.

There is now a Meteobrige Pro that simplifies the communication hardware significantly (receives the weather station RF directly) for only a little bit more money than the component costs above, but I found it after I had started buying stuff so it will have to wait for the next iteration of this station, if there is one.

I'd like to install one or more temp sensors on the backside of a panel to chart measured temps, and test the correlation to panel voltage. Tangentially, I've seen the optimizers do some weird things in man-made partial shade conditions, sometimes actually doing worse than what a string inverter would do because it finds a "maximum power point" at a high voltage that does not allow any of the cells to be bypassed, forcing the whole panel to be limited by the current through the shaded cells. I'm thinking it is better in natural shade, but haven't gone out of my way to look closely at it.

Adding a Davis multi-channel temperation station is around $150 + $35 per sensor,or $110 per single channel station. So, once more than two sensors is exceeded (within the sensor wiring range), the multi-channel setup becomes the better deal. I've gotten a few leads on other less expensive wifi temperature sensors, but haven't really found something that I think will be better yet.

[Sunking]

Easiest cheapest way is with a 10 watt solar panel connected to 20 watt .54 Ohm resistor and measure eith Amp Hours or Watt Hours and keep a log of daily generation. About $100.

[J.P.M.]

I finally pulled the trigger on the station, same hardware list as above. Stayed with the Envoy instead of the 8x to save a couple bucks, with the only trade seeming to be that additional temperature stations will only report at 1 deg resolution, not 0.5 deg. Figure another month or two to get it installed and the data handling software in place. I'll be sharing what I can through my PVOutput page, although it has a 5 min update rate at the fastest it will support, so some other database might become useful for higher frequency data, which can be as quick as a minute for GHI but down to 3 s for wind gusts.

There is now a Meteobrige Pro that simplifies the communication hardware significantly (receives the weather station RF directly) for only a little bit more money than the component costs above, but I found it after I had started buying stuff so it will have to wait for the next iteration of this station, if there is one.

I'd like to install one or more temp sensors on the backside of a panel to chart measured temps, and test the correlation to panel voltage. Tangentially, I've seen the optimizers do some weird things in man-made partial shade conditions, sometimes actually doing worse than what a string inverter would do because it finds a "maximum power point" at a high voltage that does not allow any of the cells to be bypassed, forcing the whole panel to be limited by the current through the shaded cells. I'm thinking it is better in natural shade, but haven't gone out of my way to look closely at it.

Adding a Davis multi-channel temperation station is around $150 + $35 per sensor,or $110 per single channel station. So, once more than two sensors is exceeded (within the sensor wiring range), the multi-channel setup becomes the better deal. I've gotten a few leads on other less expensive wifi temperature sensors, but haven't really found something that I think will be better yet.

Congratulations.

On temp. measurements/voltage: Depending on how and where the temp. sensors attach to the panel(s), my guess is you'll need to calibrate them. They will read different than the panel temp.for a lot of reasons, probably lower. They'll read closer to the immediate panel temp. if insulated on the backside (the "side" of the sensor facing the roof), but that will change the temp. of the panel at the sensor and probably make it higher than the rest of the panel, making it of little use.

When I use an IR thermometer to measure by hand, over the backside of the array - it is accessible - by design with about 10-11" clearance (tight but workable), I take readings at 4 random places over each of the 16 panels and record (write down) the average of the 4 as the "panel temp,". Two things: first, the 4 temps. always differ from one another by up to a couple deg. C. or so, indicating to me that the cell temps are different. Second, those averaged panel temps. vary over the array, one panel temp. to the next, not surprisingly, generally similar to what one would expect a temp. profile of a flat plate exposed to a wind vector to look like - the leading panels being cooler and the change from leading panels to trailing panels changing in magnitude as the wind vector magnitude changes - more wind more temp. diff. leading to trailing edge of the array, and the closer the array average temp. is to the ambient air temp.

All the above leads me to think that one temp. sensor strapped to the back of a panel might not give readings representative of the whole array which, by my experience varies over the array and over time. Just sayin'.

More FWIW, I'm also in the latter stages of seeing how the array voltages as recorded at the inverter screen at the time of minimum incidence angle for that day correlate with the temps. I've measured the panel temps over 6 min. on either side of that time and average those two sets of temps, with 4 min. around the min. time to get off the roof, record the inverter data - including the 2 string voltages, and get back to the roof for the 2d set of temp. readings. It looks like maybe there is reasonable agreement between temps. and voltages. If so, that may allow me to forego two trips to the roof on days when I take measurements for array fouling estimates, which is the purpose of the whole exercise anyway.

BTW, I also record the roof temps. under the array in the same fashion as the panel temps. and at the same time.Those under array roof temps. will be necessary, along with some estimates of effective radiant sky temp. so that when I get unbusy enough, I can start screwing around with a heat (energy) balance on the array and get really serious about array fouling, and/or the effect of reflection from the surroundings which we all seem to either avoid or don't know about.

As I've written in the past, I've found the HDKR algorithm seems to give the most consistent estimate of P.O.A. irradiance when compared to array output vs. input over many measurements. The Irradiance sensor is, of course, horizontal.

My Davis is located about 4 ft. north of the centerline of the array and is accessible for cleaning, which is done before each set of measurements. Accessibility for cleaning is the necessary key.

Take what you want of the above. Scrap the rest.

Welcome to the world of unprocessed data.

[sensij]

All the above leads me to think that one temp. sensor strapped to the back of a panel might not give readings representative of the whole array which, by my experience varies over the array and over time. Just sayin'.

More FWIW, I'm also in the latter stages of seeing how the array voltages as recorded at the inverter screen at the time of minimum incidence angle for that day correlate with the temps. I've measured the panel temps over 6 min. on either side of that time and average those two sets of temps, with 4 min. around the min. time to get off the roof, record the inverter data - including the 2 string voltages, and get back to the roof for the 2d set of temp. readings. It looks like maybe there is reasonable agreement between temps. and voltages. If so, that may allow me to forego two trips to the roof on days when I take measurements for array fouling estimates, which is the purpose of the whole exercise anyway.

BTW, I also record the roof temps. under the array in the same fashion as the panel temps. and at the same time.Those under array roof temps. will be necessary, along with some estimates of effective radiant sky temp. so that when I get unbusy enough, I can start screwing around with a heat (energy) balance on the array and get really serious about array fouling, and/or the effect of reflection from the surroundings which we all seem to either avoid or don't know about.

As I've written in the past, I've found the HDKR algorithm seems to give the most consistent estimate of P.O.A. irradiance when compared to array output vs. input over many measurements. The Irradiance sensor is, of course, horizontal.

My Davis is located about 4 ft. north of the centerline of the array and is accessible for cleaning, which is done before each set of measurements. Accessibility for cleaning is the necessary key.

Take what you want of the above. Scrap the rest.

Welcome to the world of unprocessed data.

Thanks for the suggestions. In my search for an alternative to the Davis sensors, I've been focused on transmitters for one or more external NTC thermistors. As you probably know, those are relatively accurate, and very small... less likely to have thermal side effects and more likely to represent the surface temperature of the panel if mounted with a highly conductive paste. Since SolarEdge provides individual panel voltages, I could place a few of them around a panel and see how the voltage and temperature data correspond. I have 12 panels... and could imagine moving the sensors from one panel to the next each month so I don't have to trust the voltage data for each blindly. Not perfect, since the seasonal variation could hide some behavior, but eventually the sensors would just live on one panel (and the roof beneath it) all year round.

Anyway, your response crystallized some of what I was concerned about with the Davis temperature system, which is why it hadn't made the cut for this order. Let's see what else I can turn up.

[J.P.M.]

Thanks for the suggestions. In my search for an alternative to the Davis sensors, I've been focused on transmitters for one or more external NTC thermistors. As you probably know, those are relatively accurate, and very small... less likely to have thermal side effects and more likely to represent the surface temperature of the panel if mounted with a highly conductive paste. Since SolarEdge provides individual panel voltages, I could place a few of them around a panel and see how the voltage and temperature data correspond. I have 12 panels... and could imagine moving the sensors from one panel to the next each month so I don't have to trust the voltage data for each blindly. Not perfect, since the seasonal variation could hide some behavior, but eventually the sensors would just live on one panel (and the roof beneath it) all year round.

Anyway, your response crystallized some of what I was concerned about with the Davis temperature system, which is why it hadn't made the cut for this order. Let's see what else I can turn up.

Understood. For a lot of reasons, there's probably (?) no real good system available to homeowners measure panel temps. I, and probably most folks have neither the knowledge or resources to do a very good - that is, reliable and reasonably accurate job of it. I attempt to do the same thing many times as identically the same as possible under "cloudless" skies and hope any precision in the instruments and methods will make up for some of the inaccuracies in instrumentation or measurement of such things as irradiance, conversion from GHI to POA, temp. measurement and the biggest uncontrollable and most uncertain variable, the wind vector.

One other thing - I'd suggest using their mounting hardware. Save the time and design work. Been there, done that. Like most such things, overpriced, but a time saver.

[Mike90250]

what about a cheap IR sensor ? They can put them in all sorts of motion sensors and handheld thermometers.

[J.P.M.]

what about a cheap IR sensor ? They can put them in all sorts of motion sensors and handheld thermometers.

I use a handheld IR thermometer when I measure the surface temp. at the backside of the panels and correct for the cell temp. Long story, but the method seems to give results that make sense.

[sensij]

I've got the station up, with the first day of data in the books. I'm piping some of it to PVOutput.org... the temp field now shows the measured temp near the array (5 min average), and the voltage field is being used to display the measured GHI (5 minute average). I'll get deeper into the calculations as I get more time to program, but for now, I'm just excited to have new data coming in! A picture of the installation will follow when i can get some time to get up on the roof again.

http://pvoutput.org/intraday.jsp?id=36909&sid=33758&dt=20160124

[J.P.M.]

I've got the station up, with the first day of data in the books. I'm piping some of it to PVOutput.org... the temp field now shows the measured temp near the array (5 min average), and the voltage field is being used to display the measured GHI (5 minute average). I'll get deeper into the calculations as I get more time to program, but for now, I'm just excited to have new data coming in! A picture of the installation will follow when i can get some time to get up on the roof again.

http://pvoutput.org/intraday.jsp?id=...58&dt=20160124

Congrads. Welcome to the world of unprocessed data. Looks like you got a mostly unobstructed location for the instrument, at least for this time of year. BTW, if you don't already know, solar noon at your location was about 12:00:30 or so yesterday. From the data, looks like you got the sensor level. How did the mounting of the entire instrument go ? One comment for readers of your output screen and meant as constructive/informative: The units for GHI are W/m^2. The PVOutput screen shows Volts.

FWIW, I had high cirrus yesterday so my max GHI was 640 W/m^2 and bouncing around between 608 and 640 W/m^2 between 11:45 and 12:15 P.S.T. Roof amb. T. @ that time was about 15 deg. C. I record in 1 min. increments. Check a station called Boulder Knolls on the Weather Underground. They have a Davis Pro II+. The irradiance readings are usually reasonably close to mine and it's located about 1 1/4 miles from me. That data for yesterday looks closer to yours. This is not an exact science. They have about 6-7 years worth of data. That data has a few holes in it, but it's pretty complete otherwise.