Topic: Barani MS Pro vs SmartCellino radiation shield comparison

[mauro63]

For absolute humidity, without an expensive sensor to measure it directly, it's too inaccurate above 80% because of condensation deposits on shelter materials and passive sensors well before condensation in the air (by radiation without sufficient wind), which no sensor can measure correctly and even worse because it's far from the time of its actual release, as it takes time to dry a sensor (so a sponge I'm not even talking about it! ) or a shelter, all the more so in winter, or on cloudy days with little wind, when humidity has long since changed and air condensation has long since disappeared. What's more, T measurements are also biased by water deposits. The inaccuracy of relative humidity sensors is greater in the case of condensation, even when relative humidity is stabilized in the laboratory. That's why it's only correct if it's far enough away from condensation and if the RH sensor hasn't been immersed x number of times in days with condensation (it's quite clear that they need to be regenerated when a new or used sensor is placed, but still good, not subjected to the previous days' condensation, that the RH is much lower on the new and regenerated sensor for the same after condensation and don't forget too that it takes several days for the new or regenerated sensor to align itself closer to the permanently exposed sensor


Sorry to all for OT

So, in essence, what tool would you recommend to a neophyte like me in this particular parameter, to be able to obtain a realistic absolute humidity data and not through the use of formulas and conversion algorithms?

Personally I was thinking of a traditional psychrometer in a Stevenson screen with artificial ventilation that can be activated only in case of measurement, but the problem remains that the Stevenson has a very different behavior from a SRS, I could also have the same dry bulb temperature but the amount of water vapor would not be the same and this would falsify the wet bulb data, returning a distorted data anyway.

And lastly, but forgive me please, why continue to harp on this topic and its impact on our tests if we have no way to analyze it in a simple way as we can do for air temperature unless we spend several thousand of euros?

Again sorry to all for continuos OT but I think can be useful

M.

[tobyportugal]

I'm just an amateur who's trying to get the most out of my tools, without sticking to inflexible positions.
You're familiar with this graph, which has been the subject of some heated debate.
At the height of the drift, if you apply the formula, you see that a sensor is very close to the dew point.
So when Jasper says that he doesn't need to know the humidity inside the shelter to know the dew point, I think there's a problem.
With my SHT45s, knowing that they're more accurate and that they don't use filters, I can 'attempt' a more detailed analysis because I have more parameters and can therefore eliminate hypotheses depending on the case.
I don't claim to have the only truth, but I'm exploring.

 [ You are not allowed to view attachments ]
 [ You are not allowed to view attachments ]
 [ You are not allowed to view attachments ]
 [ You are not allowed to view attachments ]
 [ You are not allowed to view attachments ]

[mauro63]

I'm just an amateur who's trying to get the most out of my tools, without sticking to inflexible positions.
You're familiar with this graph, which has been the subject of some heated debate.
At the height of the drift, if you apply the formula, you see that a sensor is very close to the dew point.
So when Jasper says that he doesn't need to know the humidity inside the shelter to know the dew point, I think there's a problem.
With my SHT45s, knowing that they're more accurate and that they don't use filters, I can 'attempt' a more detailed analysis because I have more parameters and can therefore eliminate hypotheses depending on the case.
I don't claim to have the only truth, but I'm exploring.

I agree with everything and especially the last part
After many years I have reached the awareness of not understanding anything about the correct detection of air temperature, imagine t about humidity that not even WMO has been able to define correctly yet.

We have a lot to do and a lot to understand, if we ever succeed, humility is necessary in this path.

thanks, I apologize again to everyone and let's continue

M.

[Dvalente75]

My report

[Jasper3012]

Can we try to stay on topic? All this extensive talk about humidity, time constants, shelters and what not, while interesting, is largely irrelevant to the topic of the thread.

[mauro63]

Can we try to stay on topic? All this extensive talk about humidity, time constants, shelters and what not, while interesting, is largely irrelevant to the topic of the thread.

You are absolutely right, I apologized but I do it again with you and with everyone
who is interested can open a specific post without using this one

thanks and I humbly ask forgiveness again to everyone and to moderation

M.

[Jasper3012]

It would deserve a clarification: what exactly are we talking about here ?

We all understood that we are not even looking for the one that is actually the closest with the good reactivity in light winds where the problems are more significant.
why are the instantaneous differences to be looked at only punctually even in a larger site with the raw data without removing the psychro effects and the shadows cast on the ground and on the shelters ?

Can someone tell me for example if according to the extensive tests here if the psychrometric effect of the Siap is lower than the Barani 2/3 ?

I allow myself to rephrase that the differences are abnormally low because of the mechanical smoothing on mechanical smoothing shelters + sponge probes (= ecowitt), in addition to the average and that in the end with such increasingly slow sensors, in slow shelters, we will end up having a T that no longer moves to no longer find differences.

Everything becomes opaque, it is non-science not to have all the important details.

We are gradually returning to the time when the Bible was only in Latin so that the people could not understand, contradict and choose to follow a more reliable path, rationality.

I would like to conclude with one last point : I am sure that some of the inevitable edginess can be overcome as this section in question is populated by very knowledgeable users who share experiences and viewpoints that of course may differ at times.

Have a nice day !

Think the title of the thread makes it pretty clear what we’re talking about? I also don’t really understand why you keep going on about these time constants, to be honest. Both shields have the exact same sensor with a time constant of 20 sec for 1 m/s of win, so the only difference would be caused by the shield itself and its time constant. I don’t know the time constants of the shields but I suspect they don’t differ that much from each other, considering the reactivity of both shields seems to be pretty similar.

[Jasper3012]

Yep, it's that probe. Didn't @gvdb1111 provide you with the specifications for it a few months back? The 20 sec value is from the producer and was confirmed with our own testing.

[Jasper3012]

There was clear sky both yesterday evening and this morning but the wind did not drop below 5-7 km/h and therefore, there was little difference between the shields. Previous testing with the Davis also showed that the low sun angle effect on the Barani only really kicks in with a wind speed of 0-3 km/h at 5 m height, so it wasn't surprising that there wasn't much to see on the graphs. Final min was 3.3C on the Barani and 3.4C on the SC.

 [ You are not allowed to view attachments ]

[Jasper3012]

Yep, it's that probe. Didn't @gvdb1111 provide you with the specifications for it a few months back? The 20 sec value is from the producer and was confirmed with our own testing.

No, and in any case it doesn't make enough sense to compare without harmonizing in a mathematical or material way, the reactivity of the material

You'll just have to take my word for it then, Geert is an electromechanical engineer and meteorologist so he knows what he's doing and he has personally confirmed the 20 sec for 1 m/s.

[Jasper3012]

What your Belgian network professional is telling you may be right in your mind, BUT the main problem is precisely that WMO smoothing over 1 minute is a long way from making the measurements comparable depending on the shelter and probe used.
That's what bothers me about these comparisons (and therefore about the probes that are too slow, far from the WMO standards, even without smoothing) is that objectively, what's missing is either smoothing that's adapted to each shelter + probe in order to harmonise, or using probes with different constants to compensate for the speed of some shelters compared with the slowest in the tests, or at least leaving instantaneous measurements in shelters + probes that are too slow, for want of anything better.
Your experts will confirm this: for example, the average OMM Tmoy/1mn of the 60s/1m/s probe + mecha ventilation to obtain an effective constant of 20s at all times under 9 m/s, will not give you the same result as a 20s/1 m/s probe placed in the same shelter under 9 m/s, despite the same average/1mn.
For example, smoothing the value of an Apogee's 20s/1m/s sensor by 1 minute will not often enough make it comparable with the instantaneous T of a Stevenson or Barani (smoothing the values of these passive sensors by 1 minute aggravates the problem), especially as the ground is hot and dry, the sun's rays are strong and there is a little wind, but enough to encourage very localised convection but not too much so as not to stir up the air too much.

You are taking this way too far. You're correct in saying there might be bigger differences in the results if you remove smoothing and use instantaneous values, but I don't care about those values. There's a reason smoothing over 1 min for example is applied and so I only want to see the results using that method. 1 min averaging isn't even that "bad", it's more than quick enough to detect significant differences in performance between shields. I don't understand where your problem with this method is coming from.

[mauro63]

What your Belgian network professional is telling you may be right in your mind, BUT the main problem is precisely that WMO smoothing over 1 minute is a long way from making the measurements comparable depending on the shelter and probe used.
That's what bothers me about these comparisons (and therefore about the probes that are too slow, far from the WMO standards, even without smoothing) is that objectively, what's missing is either smoothing that's adapted to each shelter + probe in order to harmonise, or using probes with different constants to compensate for the speed of some shelters compared with the slowest in the tests, or at least leaving instantaneous measurements in shelters + probes that are too slow, for want of anything better.
Your experts will confirm this: for example, the average OMM Tmoy/1mn of the 60s/1m/s probe + mecha ventilation to obtain an effective constant of 20s at all times under 9 m/s, will not give you the same result as a 20s/1 m/s probe placed in the same shelter under 9 m/s, despite the same average/1mn.
For example, smoothing the value of an Apogee's 20s/1m/s sensor by 1 minute will not often enough make it comparable with the instantaneous T of a Stevenson or Barani (smoothing the values of these passive sensors by 1 minute aggravates the problem), especially as the ground is hot and dry, the sun's rays are strong and there is a little wind, but enough to encourage very localised convection but not too much so as not to stir up the air too much.

You are taking this way too far. You're correct in saying there might be bigger differences in the results if you remove smoothing and use instantaneous values, but I don't care about those values. There's a reason smoothing over 1 min for example is applied and so I only want to see the results using that method. 1 min averaging isn't even that "bad", it's more than quick enough to detect significant differences in performance between shields. I don't understand where your problem with this method is coming from.

But excuse me Jasper, at this point you are as wrong as I am
let's stop replying to posts not strictly related to the comparison in question, otherwise I personally stop posting and reading

I repeat, if someone wants to open a technical post, with all the digressions he wants in advanced thermodynamics fine, but not here

M.

[Jasper3012]

You don't want to understand, especially as your constant does not allow the objective of recovering a WMO reference T°, as is the case with many manufacturers and weather services that do not stick to the WMO standard of 20s/1 m/s, but they have the excuse of choosing slower constants to stick, for example, to the values of the Stevensons and other slow passives of the past without needing to smooth with the average T/1mn of the samples/x s.
This discussion is getting nowhere, and in any case it's off topic.
everyone has the pleasure of talking to professionals in the sector and that's the main thing, isn't it?

off course 

The probe I use is quite literally a 20 sec for 1 m/s... I've made no attempt to harmonise the time constant between the SC+probe and Barani+probe and I don't know the time constant of either of the Barani+probe and SC+probe but that's obviously taking it way too far. This comparison isn't any less legitimate because I haven't done these things, is what I'm trying to say. We both know that we are talking about details here that might be relevant from a detailed technical pov but way too far for a "simple" radiation shield comparison like what we're doing here.

[Jasper3012]

What your Belgian network professional is telling you may be right in your mind, BUT the main problem is precisely that WMO smoothing over 1 minute is a long way from making the measurements comparable depending on the shelter and probe used.
That's what bothers me about these comparisons (and therefore about the probes that are too slow, far from the WMO standards, even without smoothing) is that objectively, what's missing is either smoothing that's adapted to each shelter + probe in order to harmonise, or using probes with different constants to compensate for the speed of some shelters compared with the slowest in the tests, or at least leaving instantaneous measurements in shelters + probes that are too slow, for want of anything better.
Your experts will confirm this: for example, the average OMM Tmoy/1mn of the 60s/1m/s probe + mecha ventilation to obtain an effective constant of 20s at all times under 9 m/s, will not give you the same result as a 20s/1 m/s probe placed in the same shelter under 9 m/s, despite the same average/1mn.
For example, smoothing the value of an Apogee's 20s/1m/s sensor by 1 minute will not often enough make it comparable with the instantaneous T of a Stevenson or Barani (smoothing the values of these passive sensors by 1 minute aggravates the problem), especially as the ground is hot and dry, the sun's rays are strong and there is a little wind, but enough to encourage very localised convection but not too much so as not to stir up the air too much.

You are taking this way too far. You're correct in saying there might be bigger differences in the results if you remove smoothing and use instantaneous values, but I don't care about those values. There's a reason smoothing over 1 min for example is applied and so I only want to see the results using that method. 1 min averaging isn't even that "bad", it's more than quick enough to detect significant differences in performance between shields. I don't understand where your problem with this method is coming from.

But excuse me Jasper, at this point you are as wrong as I am
let's stop replying to posts not strictly related to the comparison in question, otherwise I personally stop posting and reading

I repeat, if someone wants to open a technical post, with all the digressions he wants in advanced thermodynamics fine, but not here

M.

Yes you are correct, I thought it could be relevant since it's about the sensors used in this comparison but as I've mentioned in my last reply, it's simply way too advanced and far beyond the scope of this comparison.

[mauro63]

My data from yesterday, cloudy and rainy day

 [ You are not allowed to view attachments ]

M.

[hmderek]

Some night time data. Basically no wind. Comparisons are currently flawed because the Gen 3 has a different sensor, but this data does show what other charts have also suggested to me. That the SmartCellino seems slightly less responsive under these conditions.

 [ You are not allowed to view attachments ]
 [ You are not allowed to view attachments ]

[Jasper3012]

Some night time data. Basically no wind. Comparisons are currently flawed because the Gen 3 has a different sensor, but this data does show what other charts have also suggested to me. That the SmartCellino seems slightly less responsive under these conditions.

 [ You are not allowed to view attachments ]
 [ You are not allowed to view attachments ]

I don’t see this effect at all in my data, the response time is pretty much the same between the shields with my test.

[hmderek]

Some night time data. Basically no wind. Comparisons are currently flawed because the Gen 3 has a different sensor, but this data does show what other charts have also suggested to me. That the SmartCellino seems slightly less responsive under these conditions.

 [ You are not allowed to view attachments ]
 [ You are not allowed to view attachments ]

I don’t see this effect at all in my data, the response time is pretty much the same between the shields with my test.

Interesting! Is that the same with zero wind as well as other conditions?

[tobyportugal]

Logically, you don't measure with the same method!

[hmderek]

Logically, you don't measure with the same method!

You are right of course. Still interesting to see if similar patterns arise regardless of the many differences.

[hmderek]

You're right too, of course

While a passive shelter initially sees its T rise without wind due to strong radiative effects, despite the long time constant in calm winds, when the wind picks up, convection provides energy in addition to the radiative bias.

Passive shelters don't pick up on rapid variations if the wind varies too much, or if a small convective bubble rises just below them, because they smooth things out too much.
Real variations are smoothed out or contaminated by radiative bias (which is why passive shelters fail to see the numerous negative diurnal troughs captured, for example, by ventilated shelters)

Thing is though, the charts I posted are night time data, so that should at least remove solar radiation as a factor. Under these situations, the slightest breeze tends to make sensors converge. But here we had 2 instances of 0 m/s where differences start to show. They are really quite marginal though, but more than what you get when there is some kind of breeze.

Edit: By the way, it turns out that my concept of what radiative cooling does and how shield design relates to how much a sensor is affected is somewhat lacking. Need to read up on that. Winter is coming.

[mauro63]

Some night time data. Basically no wind. Comparisons are currently flawed because the Gen 3 has a different sensor, but this data does show what other charts have also suggested to me. That the SmartCellino seems slightly less responsive under these conditions.

Unfortunately, not having the box relating to the real air temperature (I'm obviously joking) have you thought about the possibility, if this happened in the absence of wind, or with very weak wind even with clear skies, that there could be accentuations in the data regarding night-time radiation?

M.

[Dvalente75]

My report...Yesterday 24/10 rainy day

[tobyportugal]

Derek, is this your shadow problem? Dank u 
 [ You are not allowed to view attachments ]

[hmderek]

Derek, is this your shadow problem? Dank u 

Yes sir. 

That is, low hanging sun came out late in the afternoon, zero wind, trees intermittently shading one shield after another.

Data in the second half of the afternoon is difficult to interpret in the summer, pretty much useless later in the year. Sun is dropping very fast this time of year. Nearing december you can't make heads or tails of the data all day long pretty much.