Humididty meters

I would expect they are using an off the shelf sensor "chip". Most of these measure temperature and humidity in the same device and are self calibrating. The oldest examples like the DHT11 had a resistive element for sensing RH, but the newer ones use a capacitive one.

DHT22/AM2302 are newer and more accurate and can work in a wider range of temp and RH.

Another BigClive video - including checking accuracy

The Lidl temp+humidity meter I bought recently showed 69% in my 3rd bedroom/office/computer room. MY ebac dehumidifier got that down to 50% but the recent cold spell brought it down to 43%. Now it is back up 54%.

FWIW around £20 gets you a Raspberyy Pi Pico W, a BME280 sensor and a couple of AA batteries to log temperature, humidity (+/- 3% RH accuracy) and pressure for weeks.

(The "W" so you can sync the Pico's internal clock from wifi to timestamp the data).

A fun little DIY project, especially if, like me, you've never worked with microcontrollers before.

Thanks both. I'm familiar with saturated salt solutions for checking the accuracy of these things, and have already checked my device in just the way Big Clive did, in a sealed sandwich box but using several salts in sequence.

But knowing they can use either resistive or capacitative sensors set me off on another search, and I've found a web site of a Mfr who seems to accept questions, so I'll ask there.

The one Clive was looking at there was a more basic device - basically just the sensor element itself, and it relies on external circuitry to drive it, also to read temperature, and compute the RH.

The ones I linked to ealier have their own built in microcontroller, that handle all the reading, calibration and outputs the final value. Those rely on a proprietary protocol that you need to manually "bit bang" to communicate with. There are also some that are similar but they support I2C which makes them much simpler to interface with[1]

Unnecessarily complicated. Those little circular guages you get from pet shops for reptiles are cheap as chips, use no battery and are surprisingly accurate and sensitive. I keep one in the room where I work on a lot of static sensitive test equipment. If it drops below

Normally now. Previously it could be hair whose length changes with humidity.

Wet and dry bulb is also used but that needs a water reservoir that needs to be topped up so you only see that with proper scientific instruments now.

There are also chilled mirror and gravimetric systems.

It shouldnt be hard to check it visually. It will be the sensor whose resistance changes with relative humidity.

Wrong forum if you're not DIYing a solution!

But surely it depends on what your requirements are, I was curious how the temperature and humidity changed in our garage overnight during the freeze. The Pico logs timestamped temperature, humidity and pressure at

I'd assumed the humidity part of the sensor worked via capacitance. As it also knows T and P it can work out RH.

I don't think so. The resistance changes with the absolute moisture content, and the device then goes on to calculate the relative humidity from the measured temperature and the saturation humidity for that temperature. The latter at least would require some sort of look-up table I would think.

Did you miss the fun DIY project criterion?

So how do you get that to log data then?

I bought a basic one a few years ago, not from an instrument supplier. Works pretty well; it's right beside me now, on the wall.

I don't think they are officially called that I thought it was hygrometers, but I do recall they must now be electronic ones as a clock sold by the Science Museum some time ago had one as well as other sensors and it could speak them as well. Brian

But they only work correctly when there is an airflow across the wet bulb and a specific speed. Kept indoors and away from drafts they give a higher wet bulb reading than expected. I have a Brannen greenhouse mercury wet/dry thermometer in the house and even in that heatwave the difference between dry and wet never exceeded 4C

Dunno.

That's bullshit with the specific speed claim. They were in fact the standard instrument used in standard stevenson screens used in stardard weather stations for more than a century.

That's because the humidity was so high.

I built several raspberry pi based environmental controllers over the years. Many of the cheaper humidity sensors are only accurate in the middle of the range. I wanted something that was accurate up nearer

I went for the Sensirion SHT75 which is very accurate from 10-90%RH, and outside that the accuracy only tails off by a factor of 2, so still quite good. They were expensive at around £25 each. They also sold the ones which failed to meet spec as SHT71 for half the price, which was good enough for one of my applications. They are a proprietary interface, a bit like I2C without addressing, which meant I had to bit-bang them, but that worked fine. I've had some of these operating continuously for 10 years now, including outdoors, and they've been rock solid. Sadly, Sensirion stopped producing them a few years ago because they couldn't get the cost any lower - they were very expensive to manufacture. I think I'm down to one left in my sock now.

I bought an SHT31 which is a new design from them, and is properly I2C so I don't have to write a bit-banging driver, just software to use the standard I2C API. This works fine. It's almost as accurate as the SHT75 was, but because the accuracy doesn't vary between 0-100%RH, it's actually much more accurate at the extremes. It's cheaper too, but being surface mount, is a bit more fiddly to deal with at the DIY level (not just the mounting issues, but also that if just mounted on a normal board, it's in the dead air space which wouldn't be good for accurately measuring temperature and humidity, so it needs raising off the board).

Andrew

It is a small device. I was going to say that with a small soldering iron, flux and some copper braid it's shouldn't be difficult to mount.

However, why not use something like:

That's exactly what I did do to test it:

I modified someone else's driver to make full use of its features, adding actually checking the CRC and calculating the dew point, and measuring its response time (out of curiosity):

Andrew