14 sensors on a stick

On a sunny day (Mon, 3 Jun 2019 09:37:35 -0700 (PDT)) it happened George Herold wrote in :

Well here the DHT11 tolerances:

It is not so bad, but they only go to 90% RH, above that it is soaked.. The ebay meter than shows HH (for too high) or LL (for too low): 2019 06 03 17:32 t 31.9 / rh 0.0 \ t_min 16.8 at 07:00 t_max 31.9 at 17:32 rh_min LL at 17:32 rh_max 88.0 at 06:44

So lots of things go wrong at the end of the ranges, 0% RH at 31.9 C??

That is the time when the sun hits it directly I think, it is mounted under the garden table, but late in the afternoon it gets out of the shadow... There is a whole lot of instructions on the web how to build a correct temp and RH sensor station though.

...

I'm unsure that instructions are what's required; atmospheric pressure, temperature, and humidity are all important, the 'relative humidity' combination is just an EASY one to sense, not necessarily the right one. The nominal sensor calibration ignores pressure entirely (which is important, it changes the dew point and onset of problems with mold, corrosion, etc.).

Atmospheric pressure can change suddenly (entering a low pressure area), and ought to have some consideration. Alas, that requires a weather map rather than a sensor.

On a sunny day (Mon, 3 Jun 2019 22:36:40 -0700 (PDT)) it happened whit3rd wrote in :

Yes, true, I am not much into the exact RH thing, just more curiosity. I do display and log pressure however:

Very accurate, and you can also get altitude changes from it, it sees the difference between on the floor and on a table.

Will send you an email Win.

OK, great Mark, looking for your email. Note my yahoo email address (my Harvard email inbox is clogged-up beyond use).

I'd very much like one if you have spares. That way we can test out all of those sensors at once to see which one we like best.

Thanks!

Cheers

Phil Hobbs

OK, Phil, I've cc'd your address to my email.

Hi, Win,

Did you get a chance to look at the humidity sensors? My high-Z sensor board is going into layout next week.

Thanks

Phil hobbs

The TH-stick board has four sensors** with a humidity channel. We've had three TH-sticks running in hives for about two months, so I can take a look at the data. One hive is inactive now, the queen having absconded, but the other two are cranking along. The data is in the cloud, so I should be able to see it from home.

That's the cloud data. You have to wait, it takes a while to come up. I see they're only plotting one each of the T and RH sensors, so I'll have to make an Excel version from 8.5MB of ASCII data in three files.

But first, it's a nice warm afternoon, and I'm outside adding a Spar urethane layer to our deck railings.

First results, from one bee hive. See DropBox folder

See graph and text: TH-stick_T,RH_four-sensors_Crackle Excerpt from the .txt file:

All four sensors performed properly for the roughly 3 months of working in the beehive, near the center where the queen would be, except for the BME280, whose RH readings jumped to 83.4 to 83.5% for about 23 days, then started working again.

The temperature plots lie very much on top of each other. The RH plots also lie on top of each other, except for the BME280, whose values track about 7% below the others. The HDC1080 was about 2% higher than the two Sensiron ICs.

The BME280 results are disappointing, it's a respected sensor whose main claim to fame is its precision pressure sensor.

Thanks, that's interesting. I'm using a Sensirion SHTC3, which looks like the right answer.

Just now I'm adding an MS5607-02BA3 pressure sensor to my cathodoluminescence front end board--that way we can cut off the HV and run up to room temperature when the SEM chamber is being vented.

Cheers

Phil Hobbs

I determined that it wasn't reasonable to hand assemble any of our TH-stick sensor ICs, and the SHTC3 doubly so with its hidden connections underneath. Although I tried extra hard not to make any mistakes, it was very helpful to be able to split my machine assembly order into two portions, with a 5-piece "first article", followed a second assembly after I made some changes. This was with Advanced Assembly in Denver.

Only $2, amazing. Is your PCB in the vacuum?

There are two in the chamber, this front end board and another one on the c old plate. The front end includes a pHEMT-bootstrapped TIA, a DC-50 MHz lo w noise variable gain amp (0-36dB in steps of 0.18 dB), thermistor lineariz ation, and an LPC845 micro to run it all. It's cooled by the aluminum/coppe r frame via a 3 W/m/K gap pad. (SEMs aren't UHV, so you can do that sort of stuff.)

Cheers

Phil Hobbs

Oh, and a Lite-On ambient light sensor, just in case.

It's fun adding I2C sensors, looking for candidates.

A second behive now, with similar results. This time the BME280 RH readings were about 5% lower than the other three.

In the third behive BME280 RH readings again about 5% lower than two Sensirion sensors and the TI was about 3% higher.

In all three hives the pair of Sensirion sensors closely tracked each other. Two of the hives were able to tightly regulate their internal TH-stick T and RH values, but the 3rd hive was unable to do so, and absconded after 2 months. It was unable to significantly build up its bee population.

Did anyone note any physical difference of the abandoned hive? Or reason they couldn't regulate it? Mikek

This was the first hive started, the others a day later. The hive was in trouble from the beginning, not growing as it should, etc. They never developed a significant foraging brood. This year we tried to avoid too much opening of the hives, but this one was inspected several times, and SFAIK, no obvious problem was observed.