Big Clive had a video on checking hygrometer accuracy using wet salt. You seal the hygrometer in a bag/container along with a small dish of salt that is wet. After a few hours, the humidity in the bag should stabilise at 75% at room temperature. The range of temperatures in which it is 75% is fairly wide. I'll carry on watching to see if you bring it up...
There's a whole range of different salts you can use to repeat that method and plot the accuracy across the RH range. E.g.www.robertharrison.org/icarus/wordpress/hardware/relative-humidity-above-saturated-salt-solutions-at-various-temperatures/
The recommended and most accurate calibration method is to place the hygrometer together with a salt solution into a small plastic box. One should take a cup or spoon filled with salt and moisten it with a few drops of water. The salt should not dissolve, but merely be damp. Afterwards, place the hygrometer and the salt in a well-sealed plastic box and wait. After about 8 hours the humidity within the box should have reached a humidity of 75%.
Face capture: I used the following process in a similar scenario. Record the screen with screen capture and face with your phone. Start both of them and clap before you start talking. When you import both videos in parallel tracks you can clearly see the clap in the audio feed. Use this clap to line up the videos, and always edit them together. It will give you a very good sync between the two videos. And as a bonus you can use the audio from your phone. It is not as difficult as it sounds. Ah, and no matter how choppy the screen capture is, you face will always be smooth.
Averaging by using multiple sensors to gain accuracy is a valid strategy and works really well (if the distribution of the value from the actual value is random) for the first few added sensors but it has diminishing returns.
I remember using a sling Hydrometer and logging them back when I was a building engineer, not bad on a palmey summer night, but a bit rougher in January.
Totally agree. These little modules for Arduino were totally inaccurate for me by default. You must do a calibration. Maybe they are calibrated during manufacture and some external factors (soldering, transportation from seller in China) make the calibration invalid? Luckily we have the formula field in ESP Easy (it's a very good project), so we may adjust the humidity/temperature if it is consistently shifted in Humidity/Temperature formula field by say: %value%-1.
Adrian you have to consider the enclosured sensors vs bare sensor too. In the past when I have used various hydrometers I used to find a lot of variance according to airflow and plug in sensors always read higher than those with the sensor behind a few slits in the case. A good excuse to take them all apart if you ask me 😊
Relative humidity is hard to pin down. If you want more accuracy then take a look at the new Senserion SHT35 - it is +/-1.5DegC. However, it will be expensive though in comparison to the ubiquitous SHT21
You can determine the minimum reading (rarely it is zero, it is often 10 or 15%) by placing them in a sealed container with a strong dessicant (something much better than silica gel), such as one of the following: calcium chloride, sodium hydroxide, sulfuric acid, or molecular sieves. Then leave it a day and take the readings. Take note of the ones that have noticable minimum offsets. Now make a saturated salt (NaCl) solution and leave it in the sealed container with them for another 24 hours. It is probably ~74%. Again record each units offset from another. You can do their "maximum" with just plain water but it can damage them so that is only recommended for mechanical type hygrometers.
I really like the facecam, please keep it up! That is much better than filming the screen with your phone, IMHO its worth the extra edit time, dont you agree?
Go with saturated aqueous sodium chloride solution in a closed chamber for checking accuracy. Beware that the water and the salt must be pure and even then it varies with temperature, but it should be around 75% at room temperature
You probably have to remove the protective cover from the Si7021. The datasheet shows the chip with a hole on top. The cover protects it during the soldering process, storage and transport.
How about making.a wet bulb hygrometer? Use two temperature sensors, keeping one of them wet with a wick. Maybe a small fan to make things equalize faster. The plug the two temperatures into a relative humidity formula and Bazinga! Relative humidity.
Very nice. others have commented about the salt calibration method which would be a great tip. I've thought of making an aspirating hygrometer with ardunio. two thermometers, one dry and one with a sock on it. small pump would wet it then small fan would blow on it, then it would calculate humidity from the wet and dry bulb temperatures. It could automatically wait until the temp is stable to do the calculation. possibly take readings periodically and log them.
Clive found these cheapos reasonably exact. Silabs makes very good devices in general, but perhaps the sensor you got became contaminated with an oil or other material. An air humidity sensor necessarily consists of absorbent material, so touching it could do that.
Greetings from the chilly plains of South Dakota, once again. Just a note, I still play with the Arduinos from time to time, and like you (probably because of watching your videos) I purchased the ones with both male and female plugs. When I hook up a couple of i2c devices, I usually just use one set of female cables, and one set that has male for hooking to the Aruino then just use the SDA etc plugs, either that or pick an aruino that has the extra plugs on the bottom of the board like many do.
facecam is somewhat unnecessary. Massive frame drops and effectively broke the video when it was included, bit like watching a slide show. cool if you're experimenting but not necessary for your vids in fact prefer without imo.
The problem with the si7021 module is it could be using SI7006, SI7020, SI7021, etc. which have varying accuracy. Worst case its +/-5% (the cheapest version). There is a laser etched code on the chip, but the datasheet does not specify what this code can be.
I have had the analogue one and the round one next to each other on my desk for a few months. They seem to drift ~10% to each other, but are generally within 2%. Neither seems to be persistently high or low, they are just sometimes higher or lower.
I can't say anything about the absolute accuracy of the two, just how they performed relative to each other. I expect your Si7021 probably is more accurate though, just by the nature of the measurement element.
I think that GY21 board already has the 5->3.3 voltage regulator and the required I2C level-shifters on the breakout board: that's what those small parts are on the opposite side of the board to the sensor IC. So you didn't need to attach it to a separate level-shifter.
Most eBay listings seem to be unaware of the built-in level shifters and regulator - although this listing does mention them! www.ebay.com/itm/New-HTU21D-GY-21-HTU-Temperature-Humidity-Sensor-Breakout-Board-Module-Hot-/271731131299
Julian Ilett Why, thank you :-) You're right, though: most of the listings do rather lazily transcribe the specifications of the sensor itself and don't mention the extras.
I meant to mention you in my latest video (for spotting the level shifters on the GY21 board) but I was having so many problems with my camera, I completely forgot. Sorry - I'll try and remember in my next hygrometer video.
If you want to do a proper determination of humidity (relative and absolute) all you need is two normal thermometers (alcohol, mercury etc), a bit of damp cotton wool, and a psychrometric chart downloaded off the internet :)
Can try and calibrate with salt & water enclosure, as others have said, but still haven't had much luck with the likes of the DHT22 , which are supposed to be factory calibrated. The variance between a few in the same room even after salt water calibration and fiddle factors is pure POOH!
DHT22 lists for about the same as BMP280 and others on ebay and elsewhere so you can't just dismiss it as cheap junk oh Night.. I think its more a case of bad manufacturing & incoherent testing. BME280 is more expensive because it does barometric pressure as well. The first few DHT22's I got were okay and relatively coherent but the last few were total rubbish, and that's after extended attempts at calibration.
Hi, thank you for the work involved, I have had limited success so far. I provided the 3.3V from the arduino, and tried taking the scl and sda through 5 to 3.3 level shifter but got nothing on the serial monitor. I took the scl and sda direct to the sensor and got results of a kind on the s monitor. the humi was looking good and responsive to just blowing on it. But the temp was a constant 128.87 which is somewhat erroneous. oh dear, back to the drawing board !
Given that the 2 digial and the analogue hygrometers all agree (within the stated error) and the Arduino's sensor is quite a bit off, I'd trust them over the 7021.
i think you can set up different obs capturing scenes and you could put chrome,the IDE and you screen in different scenes. that may speed that up or at least make it less laggy.
When I did a temperature sensor controlled fan for my motorhome, I found I had to add a small load resistor to the output of the sensor in order to improve its accuracy and stability, I wonder if the same is true for the humidity sensor?
Coolkeys2009 - yes digital to the arduino, just wasn't sure it the actual sensor is analogue and then converted to digital on the pcb, but badly implemented on the pcb, but obviously I wasn't thinking straight, as there is no conversion, with digital directly out of the sensor.
Another way to get an accurate hygrometer is to build an psychrometer (wet dry hygrometer) and use an algorithm to get the reading on yer OLED. Fundamentally the psychrometer formula, which describes the relationship between the vapor pressure e, the psychrometric difference (t-t'), the saturation vapor pressure at the temperature of the wet thermometer E', the air pressure p and a constant specific psychrometer coefficient (psychrometric constant) C: e = E' - [ C x p x ( t - t' )] The relative humidity f is the ratio of the actual water vapor pressure e to the saturation vapor pressure E' at the prevailing air temperature t: E' = 6,1078 x exp [( 17,08085 x t ) / ( 234,175 + t )] f = e / E' The accuracy of a psychrometer is app. linear to the accuracy of the temperature probes, ie. a couple of +- 0,1C temp. probes will give you close to a +- 0,1% hygrometer reading.
I am playing with the DHT11 sensor recently and I have read that you have to average at least 8 samples to get the most out of these sensors. Maybe it would be worth a go. I have not tried it myself to see if there is a difference in my ranges, but maybe you can give it a go and see what difference it makes? there are multitudes of samples available on the internet. I see that you are doing a live update of your sensor output (as am I on my project), but people suggest putting a delay on refresh to let the sensor stabilize and gather data. technically, with a sample size of 8 and a 500ms update, there would a 4 second update time on the displayed results with higher precision.
You actually don't need a level translator, since they're open drain just pull the I2C bus up to 3.3V, the Atmega on Arduino will recognize the voltage level just fine.
I also noticed that the si7021 was giving a instantanious reading where the others possibly only update every 30 seconds or so, you would need to take readings at set intervals eg every 5 mins for 30 mins then average the readings, put them all in a sealed container or do this from a distance as just breathing near some sensors can affect them.
sht21 is +-2 for RH, sht25 high +-1.8, sht35+-1.5. Caliber4r is a 1% hygrometer, but not sure which sensor that one uses. Those cheaper ebay humidity meters are garbage, and use just a pcb trace for a sensor and a cheap coating to absorb moisture. The other cheap ones use a coil for rotating a needle which is also garbage. The acurite hygrometers use the sensirion sensors with +-2 RH. Boveda sells a calibration kit for 7 bucks and is better than using salt.
When doing i2c i tend to put the pullups on the 3.3V rail, so i don't need level converters. Most 5V devices go all well with I2C at 3.3V. Also: Is it just me, or are the facecam parts choppy?
it's a protective cover intended to be left in place. From the datasheet: "The optional factory-installed cover offers a low profile, convenient means of protecting the sensor during assembly (e.g., reflow soldering) *and throughout the life of the product*, excluding liquids (hydrophobic/oleophobic) and particulates"
Is the absolute relative humidity (yes, I mean the absolute number of it) that important to you? Of course it would be nice to have an accurate number, but in the end all that matters is knowing when you have to switch on the dehumidifier. I have a few cheap digital hygrometers that pretty much show all the same numbers (bought at the same time) and put them in different rooms in the basement. From experience (you could call that "mental calibration") I know that when they climb towards 70%rh I better switch on the dehumidifier otherwise it will start to smell in a room. If it's around or below 60%rh there's no point in running them more. I don't know if these numbers reflect the real value, but the change of the value is enough to see if the dehumidifier is working well enough.
Interesting question. But can you say the same about voltage? All my DVMs disagree slightly. Until I bought that precision 5V reference I had no idea which DVM was actually right. And then we discover that all those precision reference PCBs are being shipped with the same piece of paper, with identical "calibration" values written on them. So I still can't measure voltage with absolute precision. Does it matter? Most of the time, not really :)
Try recording your face with the built-in windows webcam program, and recording the screen separately, but at the same time, then just merge the two in post :)
i think if you get finger oil on sencer it does not work right. and if you put it in a air tite bag with some salt water in a small cup water for 24 hr it shoud be 75 %
I was just going to suggest salt but there's already salt test suggestions in the comments. You had the sensor exposed, might the plastic case on the eBay versions be effecting the reading?
Interesting project! I like these videos of the design and implementation progressing. The screen capture and facebox are good concepts, but like others have pointed out, you need a better PC for it. At first I thought the choppy video was painful, but then it started freezing for about 5 minutes at a time...
Julian, versions of the Arduino IDE over 1.6.4 or so do make my computers CPU run almost maximum while only having loaded a sketch window, maybe that is why the upload takes so much longer for you too ?
Julian Ilett I use 1.6.5 therefore . There is numerous people complaining online about the cpu-hog bug of services.exe and here's one: ik1zyw.blogspot.nl/2016/04/arduino-168-ide-problems-on-windows-xp.html
hi Julian, to be honest, that new recording software is crappy. It is just too laggy and hard to see what's going on. You should perhaps try screenpresso. If that also lags, then I'm afraid you don't have sufficient processing power... I much prefer it the way it was before...
Go to: Settings --> Encoding --> Max Bitrate and change that to a lower Value after that you go to Settings --> Video --> Resolution Downscale and change that to a lower value, too. The Videoquality mith not be so good as before but it wont freeze so often. Hope I could help you. Greetings from Germany.
AFAIK the breakout board for the Si7021 already contains a 3.3V regulator and level shifting circuitry. [5:05](th-cam.com/video/C1qTztVkDa4/w-d-xo.html)
Does anyone else have issue with that SI7021 reading 128.87 degrees celsius? But it reads humidity. Of course it's way too low because that high temperature.
I've read/seen somewhere: you are not suppose to touch the sensor due to skin contamination, as you weren't wearing gloves and handling the sensor you may have ruined the accuracy of the sensor.
Yes it does, but you've still got the problem of 5V signals (SCL and SDA) on the Arduino and 3.3V signals on the Si7021. However it's academic really! The little GY21 PCB has level shifters and a 3.3V regulator onboard. AAAAArrrrgggghhhh.
I'm thinking the same thing. When buying on the cheap like that, the engineering (or lack there of) tends to be very similar if not exact. Especially spec sheets that look almost exact.
hello julian i think you answered your own question at the end of your vid, i am a refrigeration & air conditioning engineer, basically to measure temperature is cheap and easy, humidity is a different beast IE costly, it can be done electronically but if you want a cheap affordable measurement, then you need a whirling hygrometer, looks like a football rattle, works on wet & dry bulb principle, thanks andrew
Sling psychrometer is the term I was taught for the wet/dry bulb measurement. But far cheaper is using saturated salts to create a known RH in a closed system.
That can be used to power the sensor. But shoving 5V signals (SCL and SDA) into a 3.3V device isn't a good idea unless the sensor has 5V tolerant inputs.
Julian Ilett Hence I pay an annual fortune to a calibration house to maintain my ISO9001 Certification. Another quality and educational video J, entertaining too
Several websites give live humidity readings for many UK locations so you put the hygrometers outside and see how well they match. Or see, for example, www.engineeringtoolbox.com/humidity-measurement-d_561.html You have a pretty accurate "dry bulb" temperature as all four devices agree, so you just need a wet bulb thermometer and you can calculate the indoor humidity. What does 5% accuracy mean when talking about percentage reading. eg if it's reading 50% does that mean the true value is somewhere between 45% and 55%, or between 47.5% and 52.5%?
Julian Ilett the setup works great! (near 1 year) but, battery consumption issn't what i want. i build this project again but with a qC on 3,3V and 8mHz by myself.. (Atmel qC, crystal, 2x Cap's or internal Clock!!) then a deep sleep with an interrupt of 1 min, display refresh.. it issn't hard but all has to be matched.
I was just thinking the same thing. My miniature Arduino hygrometer will look great, but power consumption will be terrible - probably just a few hours on a 9V battery.
Julian Ilett jep, jep, that is why i wish a battery friendly version. i think i will also add a Button or anything (PIR?! ) that activates the Display only if required.
The Arduino UNO 3.3V pin is a supply voltage to be used for powering off board sensors / boards that require 3.3V power - etc. But all Arduino UNO data pins use 5 V on it's data signals. So - an external 5V to 3.3V level shifter board would be required for the data pins. And the shifter board would require both a 5V and a 3.3 V power feed to operate. So - you could use the 3.3 v (and 5v) feeds from the Arduino to power the shifter board - so you wouldn't need a separate external 3.3V power supply !!
Man with one hygrometer know what the humidity is. Man with two is never quite sure :p
Yes
Big Clive had a video on checking hygrometer accuracy using wet salt. You seal the hygrometer in a bag/container along with a small dish of salt that is wet. After a few hours, the humidity in the bag should stabilise at 75% at room temperature. The range of temperatures in which it is 75% is fairly wide. I'll carry on watching to see if you bring it up...
But even then, we don't know how much the error is at different values. All of them may be dead accurate at 75%, but not so much at 60%.
There's a whole range of different salts you can use to repeat that method and plot the accuracy across the RH range. E.g.www.robertharrison.org/icarus/wordpress/hardware/relative-humidity-above-saturated-salt-solutions-at-various-temperatures/
The recommended and most accurate calibration method is to place the hygrometer together with a salt solution into a small plastic box. One should take a cup or spoon filled with salt and moisten it with a few drops of water. The salt should not dissolve, but merely be damp. Afterwards, place the hygrometer and the salt in a well-sealed plastic box and wait. After about 8 hours the humidity within the box should have reached a humidity of 75%.
I like the screencapture with facecam setup! 👍🏻
Face capture: I used the following process in a similar scenario. Record the screen with screen capture and face with your phone. Start both of them and clap before you start talking. When you import both videos in parallel tracks you can clearly see the clap in the audio feed. Use this clap to line up the videos, and always edit them together. It will give you a very good sync between the two videos. And as a bonus you can use the audio from your phone. It is not as difficult as it sounds. Ah, and no matter how choppy the screen capture is, you face will always be smooth.
That Video of you on the bottom while being at the PC is nice!
Thanks :)
DHT22 is a common temp/humidity sensor that works well with Arduino by permitting 3-5V power and I/O.
Averaging by using multiple sensors to gain accuracy is a valid strategy and works really well (if the distribution of the value from the actual value is random) for the first few added sensors but it has diminishing returns.
I remember using a sling Hydrometer and logging them back when I was a building engineer, not bad on a palmey summer night, but a bit rougher in January.
Screen capture! Finally! It's a bit choppy, but only on a couple of occasions (scrolling, window switching).
Leave all the hygrometers in a same place over night to see if they "normalise", then attempt a salt calibration.
Totally agree. These little modules for Arduino were totally inaccurate for me by default. You must do a calibration.
Maybe they are calibrated during manufacture and some external factors (soldering, transportation from seller in China) make the calibration invalid? Luckily we have the formula field in ESP Easy (it's a very good project), so we may adjust the humidity/temperature if it is consistently shifted in Humidity/Temperature formula field by say: %value%-1.
Adrian you have to consider the enclosured sensors vs bare sensor too. In the past when I have used various hydrometers I used to find a lot of variance according to airflow and plug in sensors always read higher than those with the sensor behind a few slits in the case.
A good excuse to take them all apart if you ask me 😊
Relative humidity is hard to pin down. If you want more accuracy then take a look at the new Senserion SHT35 - it is +/-1.5DegC. However, it will be expensive though in comparison to the ubiquitous SHT21
You can determine the minimum reading (rarely it is zero, it is often 10 or 15%) by placing them in a sealed container with a strong dessicant (something much better than silica gel), such as one of the following: calcium chloride, sodium hydroxide, sulfuric acid, or molecular sieves. Then leave it a day and take the readings. Take note of the ones that have noticable minimum offsets. Now make a saturated salt (NaCl) solution and leave it in the sealed container with them for another 24 hours. It is probably ~74%. Again record each units offset from another. You can do their "maximum" with just plain water but it can damage them so that is only recommended for mechanical type hygrometers.
Old saying: A man with a watch always knows what time it is. A man with two watches is never quite certain.
I really like the facecam, please keep it up! That is much better than filming the screen with your phone, IMHO its worth the extra edit time, dont you agree?
Go with saturated aqueous sodium chloride solution in a closed chamber for checking accuracy.
Beware that the water and the salt must be pure and even then it varies with temperature, but it should be around 75% at room temperature
You probably have to remove the protective cover from the Si7021. The datasheet shows the chip with a hole on top. The cover protects it during the soldering process, storage and transport.
How about making.a wet bulb hygrometer? Use two temperature sensors, keeping one of them wet with a wick. Maybe a small fan to make things equalize faster. The plug the two temperatures into a relative humidity formula and Bazinga! Relative humidity.
Very nice. others have commented about the salt calibration method which would be a great tip. I've thought of making an aspirating hygrometer with ardunio. two thermometers, one dry and one with a sock on it. small pump would wet it then small fan would blow on it, then it would calculate humidity from the wet and dry bulb temperatures. It could automatically wait until the temp is stable to do the calculation. possibly take readings periodically and log them.
You got the facecam out this time. Fancy!
Yeah - Logitech c920
Same one I have! It is good, isn't it? :) True 1080p in Skype chat, too :D
Not as good as I was hoping - 1080p doesn't work in OBS and the autofocus is unreliable.
Yes. Liked having the inset view. Keep that.
I'll try :)
Clive found these cheapos reasonably exact.
Silabs makes very good devices in general, but perhaps the sensor you got became contaminated with an oil or other material. An air humidity sensor necessarily consists of absorbent material, so touching it could do that.
Greetings from the chilly plains of South Dakota, once again. Just a note, I still play with the Arduinos from time to time, and like you (probably because of watching your videos) I purchased the ones with both male and female plugs. When I hook up a couple of i2c devices, I usually just use one set of female cables, and one set that has male for hooking to the Aruino then just use the SDA etc plugs, either that or pick an aruino that has the extra plugs on the bottom of the board like many do.
I only have female-female cables, so I have to think laterally ;)
+Jerry Ericsson I thought i2c can be 'daisy-chained' ?
Screen capture and facecam! Oh my how this channel has grown up :D Nice work :)
facecam is somewhat unnecessary. Massive frame drops and effectively broke the video when it was included, bit like watching a slide show.
cool if you're experimenting but not necessary for your vids in fact prefer without imo.
I like the screen cap software. Works good.
reading of si7021 is perfect
I used it in many many projects
The problem with the si7021 module is it could be using SI7006, SI7020, SI7021, etc. which have varying accuracy. Worst case its +/-5% (the cheapest version).
There is a laser etched code on the chip, but the datasheet does not specify what this code can be.
Really like the new format ^^
Might be interesting to see what chips are in the the "cheap" devices.
I have had the analogue one and the round one next to each other on my desk for a few months. They seem to drift ~10% to each other, but are generally within 2%. Neither seems to be persistently high or low, they are just sometimes higher or lower.
I was using the mechanical one, but it's 15% lower than the Si7021 :(
I can't say anything about the absolute accuracy of the two, just how they performed relative to each other. I expect your Si7021 probably is more accurate though, just by the nature of the measurement element.
I think that GY21 board already has the 5->3.3 voltage regulator and the required I2C level-shifters on the breakout board: that's what those small parts are on the opposite side of the board to the sensor IC. So you didn't need to attach it to a separate level-shifter.
AAAAArrrrrggggghhhhhhh! I do believe you're right - your's is comment of the day :) That's almost worth a video all it's own!
Most eBay listings seem to be unaware of the built-in level shifters and regulator - although this listing does mention them! www.ebay.com/itm/New-HTU21D-GY-21-HTU-Temperature-Humidity-Sensor-Breakout-Board-Module-Hot-/271731131299
Julian Ilett Why, thank you :-) You're right, though: most of the listings do rather lazily transcribe the specifications of the sensor itself and don't mention the extras.
I meant to mention you in my latest video (for spotting the level shifters on the GY21 board) but I was having so many problems with my camera, I completely forgot. Sorry - I'll try and remember in my next hygrometer video.
If you want to do a proper determination of humidity (relative and absolute) all you need is two normal thermometers (alcohol, mercury etc), a bit of damp cotton wool, and a psychrometric chart downloaded off the internet :)
Well, actually you need to know the atmospheric pressure as well.
do you have tutorial for the automatic switch on/off using hygrometer sensor ?
Can try and calibrate with salt & water enclosure, as others have said, but still haven't had much luck with the likes of the DHT22 , which are supposed to be factory calibrated. The variance between a few in the same room even after salt water calibration and fiddle factors is pure POOH!
Kosmonooit
DHT22 lists for about the same as BMP280 and others on ebay and elsewhere so you can't just dismiss it as cheap junk oh Night.. I think its more a case of bad manufacturing & incoherent testing. BME280 is more expensive because it does barometric pressure as well. The first few DHT22's I got were okay and relatively coherent but the last few were total rubbish, and that's after extended attempts at calibration.
Hi, thank you for the work involved, I have had limited success so far. I provided the 3.3V from the arduino, and tried taking the scl and sda through 5 to 3.3 level shifter but got nothing on the serial monitor. I took the scl and sda direct to the sensor and got results of a kind on the s monitor. the humi was looking good and responsive to just blowing on it. But the temp was a constant 128.87 which is somewhat erroneous. oh dear, back to the drawing board !
I myself quite like the old mechanical sensors, although it`s much more fun fiddling with electronic sensors. :)
the 2 digital hygrometers probably have identical innards so idk if its that significant for their accuracy that they agree with each other
I think it is quite significant
Given that the 2 digial and the analogue hygrometers all agree (within the stated error) and the Arduino's sensor is quite a bit off, I'd trust them over the 7021.
I like the idea to use the face cam. but please use the audio recording device from the "desk view".
You have a 3.3V output from the Arduino - you can see it at 4:20
The problem is that SDA and SCL lines are going to be run off 5v, which still isn't good for a 3.6V max device.
Arduino have 3.3V output ;)
PS: quality of video from PC is very bad and freezing
Yeah, don't I know it :)
your camera ur using for to show ur face is laggy
With OBS capturing The c920, Chrome and Arduino IDE, my PC pretty much ground to a halt !!
GOOOD TRY I LOVE IT EVEN IF IT IS LAGGY ASFUCK!
i think you can set up different obs capturing scenes and you could put chrome,the IDE and you screen in different scenes. that may speed that up or at least make it less laggy.
bigclive uses the bag method i think it will help in tracking down the specific ones accuracy/repeat ability.
You could do it the old fashioned way, with a thermometer to measure wet and dry bulb temperatures.
When I did a temperature sensor controlled fan for my motorhome, I found I had to add a small load resistor to the output of the sensor in order to improve its accuracy and stability, I wonder if the same is true for the humidity sensor?
That sensor has a serial Digital output. He can calibrate it with a mathematical formula in his code.
Coolkeys2009 - yes digital to the arduino, just wasn't sure it the actual sensor is analogue and then converted to digital on the pcb, but badly implemented on the pcb, but obviously I wasn't thinking straight, as there is no conversion, with digital directly out of the sensor.
dont you need to remove the paper cover from the sensor?
Do the rectangle and round digital hygrometers have the same layout/component inside? That might explain them showing the same temperature!
Another way to get an accurate hygrometer is to build an psychrometer (wet dry hygrometer) and use an algorithm to get the reading on yer OLED.
Fundamentally the psychrometer formula, which describes the relationship between the vapor pressure e, the psychrometric difference (t-t'), the saturation vapor pressure at the temperature of the wet thermometer E', the air pressure p and a constant specific psychrometer coefficient (psychrometric constant) C:
e = E' - [ C x p x ( t - t' )]
The relative humidity f is the ratio of the actual water vapor pressure e to the saturation vapor pressure E' at the prevailing air temperature t:
E' = 6,1078 x exp [( 17,08085 x t ) / ( 234,175 + t )]
f = e / E'
The accuracy of a psychrometer is app. linear to the accuracy of the temperature probes, ie. a couple of +- 0,1C temp. probes will give you close to a +- 0,1% hygrometer reading.
Best cheap sensor you can get this days is BME280. Si7021 and HTU21D are outdated
the difference between readings may be caused by the plastic cases of the cheap ones
and by the fact that the sensor you used was "naked" :)
Try the HYT271 for more accuracy - 1.8% RH
I'll have a look at that one - thanks :)
I am playing with the DHT11 sensor recently and I have read that you have to average at least 8 samples to get the most out of these sensors. Maybe it would be worth a go. I have not tried it myself to see if there is a difference in my ranges, but maybe you can give it a go and see what difference it makes? there are multitudes of samples available on the internet. I see that you are doing a live update of your sensor output (as am I on my project), but people suggest putting a delay on refresh to let the sensor stabilize and gather data. technically, with a sample size of 8 and a 500ms update, there would a 4 second update time on the displayed results with higher precision.
Interesting Video Julian. Keep up the great work. Nick.
If the OBS is dropping frames during the capture, try to disable the Aero effects in Windows 7. They have an official recommendation for doing that.
If the same can be done in Windows 10, I might try that.
Julian Ilett I'm not sure about Win10, but you might try to disable some/all visual effects: www.thewindowsclub.com/disable-visual-effects-windows
You actually don't need a level translator, since they're open drain just pull the I2C bus up to 3.3V, the Atmega on Arduino will recognize the voltage level just fine.
I read that on a sparkfun tutorial, I but thought it would be too difficult to implement without checking all the schematics
Make a satuated salt solution in a sealed container should give you a 75% ref, i do have somewhere info on making a 22% ref if i can find it.
I also noticed that the si7021 was giving a instantanious reading where the others possibly only update every 30 seconds or so, you would need to take readings at set intervals eg every 5 mins for 30 mins then average the readings, put them all in a sealed container or do this from a distance as just breathing near some sensors can affect them.
Shouldn't you remove the protection sticker blocking the opening on the chip, so air can get in?
sht21 is +-2 for RH, sht25 high +-1.8, sht35+-1.5. Caliber4r is a 1% hygrometer, but not sure which sensor that one uses. Those cheaper ebay humidity meters are garbage, and use just a pcb trace for a sensor and a cheap coating to absorb moisture. The other cheap ones use a coil for rotating a needle which is also garbage. The acurite hygrometers use the sensirion sensors with +-2 RH. Boveda sells a calibration kit for 7 bucks and is better than using salt.
When doing i2c i tend to put the pullups on the 3.3V rail, so i don't need level converters. Most 5V devices go all well with I2C at 3.3V.
Also: Is it just me, or are the facecam parts choppy?
hello Julian, i have try a lot of sensor for T° and H% and the best so far for me (easy use and precision) is the DHT22.
H% : +-2%
T° : +-0.5
Did you take the paper off the Arduino Humidity sensor?
it's a protective cover intended to be left in place. From the datasheet:
"The optional factory-installed
cover offers a low profile, convenient means of protecting the sensor during
assembly (e.g., reflow soldering) *and throughout the life of the product*, excluding
liquids (hydrophobic/oleophobic) and particulates"
Is the absolute relative humidity (yes, I mean the absolute number of it) that important to you? Of course it would be nice to have an accurate number, but in the end all that matters is knowing when you have to switch on the dehumidifier.
I have a few cheap digital hygrometers that pretty much show all the same numbers (bought at the same time) and put them in different rooms in the basement. From experience (you could call that "mental calibration") I know that when they climb towards 70%rh I better switch on the dehumidifier otherwise it will start to smell in a room. If it's around or below 60%rh there's no point in running them more. I don't know if these numbers reflect the real value, but the change of the value is enough to see if the dehumidifier is working well enough.
Interesting question. But can you say the same about voltage? All my DVMs disagree slightly. Until I bought that precision 5V reference I had no idea which DVM was actually right. And then we discover that all those precision reference PCBs are being shipped with the same piece of paper, with identical "calibration" values written on them. So I still can't measure voltage with absolute precision. Does it matter? Most of the time, not really :)
What software are you using to record your desktop?
OBS, but my PC can't cope with too many video sources at once.
Try recording your face with the built-in windows webcam program, and recording the screen separately, but at the same time, then just merge the two in post :)
I can't merge the two in post because I use TH-cam editor.
Julian Ilett try using davinci resolve, it is very powerful and free!
i think if you get finger oil on sencer it does not work right. and if you put it in a air tite bag with some salt water in a small cup water for 24 hr it shoud be 75 %
I was just going to suggest salt but there's already salt test suggestions in the comments.
You had the sensor exposed, might the plastic case on the eBay versions be effecting the reading?
Interesting project! I like these videos of the design and implementation progressing. The screen capture and facebox are good concepts, but like others have pointed out, you need a better PC for it. At first I thought the choppy video was painful, but then it started freezing for about 5 minutes at a time...
what about the DHT22 sensor? do you recommend it?
I am unsure, but I am sure that it just slipped your mind that the Arduino has a 3.3v Vss pin on it!
I don't know how hydrometers work, but do they depend on barometric pressure? If yes, then maybe the cheap ones are more accurate in China.
Never heard of the Si7021 before - ordered :D
also: there are 3.3V arduinos.
Didn't Confucius say "Man with one hygrometer knows what the humidity is - Man with more than one, is not quite sure..." ?
Julian, versions of the Arduino IDE over 1.6.4 or so do make my computers CPU run almost maximum while only having loaded a sketch window, maybe that is why the upload takes so much longer for you too ?
Yeah, I worry about the direction Arduino is going - the IDE is becoming bloatware :(
Julian Ilett I use 1.6.5 therefore . There is numerous people complaining online about the cpu-hog bug of services.exe and here's one: ik1zyw.blogspot.nl/2016/04/arduino-168-ide-problems-on-windows-xp.html
Now you must do a full barometer including a simple weather prediction. :)
Or I could just tear down the cheap "weather station" I bought from Clas Ohlson
Yep liked the inset picture and a proper screen capture rather than recording the monitor is a lot easier to read. Thanks
Easier to read, but I'm having so many problems with CPU loading and consequently a low frames per second count.
hi Julian, to be honest, that new recording software is crappy. It is just too laggy and hard to see what's going on. You should perhaps try screenpresso. If that also lags, then I'm afraid you don't have sufficient processing power... I much prefer it the way it was before...
Hey Julian,
try to record your screen with a resolution of 420p sothat it doesnt freeze so often.
That's difficult to do in OBS
Go to: Settings --> Encoding --> Max Bitrate and change that to a lower Value after that you go to Settings --> Video --> Resolution Downscale and change that to a lower value, too.
The Videoquality mith not be so good as before but it wont freeze so often.
Hope I could help you.
Greetings from Germany.
I don't seem to have those settings. Are there 2 versions of OBS? I'm using OBS Studio.
I dont know if there are 2 Versions. But usually you should have those settings.
Julian, i need a temperature sensor, that is Atmospheric pressure corrected, but have zero experience with arduino..Can you help me on that subject?
Where you get those level converters from ?
I just bought 10 more - 10PCS IIC I2C Level Conversion Module Sensor 5V-3V System For Arduino 5V to 3V www.ebay.com/itm/-/311576163493
I bet you have two accurate thermometers---> if so make a reference with wet bulb and dry bulb thermometers. Easy peasy ;))
Love them OLEDS, got some coming from Amazon
What about a HDC1080 or a HTU21D-F? Both of these should work on 5 volts. And according to their datasheet they should be accurate to +-2%
I think you need a new computer
Yeah, me too!!
lol. I meant your. Im lucky here, my 2 year old laptop is still better than most peoples PC.
Btw, have you seen the new Macbook Pro?
*****
Im just going to guess he has more brains than to buy locked-down hardware and software and pay more for it.
apple are retarded.
Think how stupid the average people are and you are stupider than that. Never mind, it's not fair.
What is this thing you call "humidity", i have never heard of such a thing living in southern Arizona
The oil on your fingers can permitinly mess up most hygrometers. Never touch the sensor if possible.
AFAIK the breakout board for the Si7021 already contains a 3.3V regulator and level shifting circuitry. [5:05](th-cam.com/video/C1qTztVkDa4/w-d-xo.html)
Does anyone else have issue with that SI7021 reading 128.87 degrees celsius? But it reads humidity. Of course it's way too low because that high temperature.
Did you ever get this figured out? I have the same issue.
Nope
See my post at the top: I got the same, 128.87, although the humid looked ok, and responsive to changes.
Sounds like it's pegged out. Do you need to use a pull up/down resistor?
I'm getting negative humidity, like -47% but the value changes. Is it possible I got a sensor with broken calibration?
I've read/seen somewhere: you are not suppose to touch the sensor due to skin contamination, as you weren't wearing gloves and handling the sensor you may have ruined the accuracy of the sensor.
an arduino has a 3.3V port aswell
If you power the arduino with 5v from USB what does the 3.3v pin give you on the board?
3.3v, it has a regulator on board!
Edmund Fisher
So why did he not just use the 3.3v pin?
Because the SCL and SDA pins on the Si7021 aren't 5V tolerant
Julian Ilett
Does the 3.3v pin on the ardunio not give 3.3v?
Yes it does, but you've still got the problem of 5V signals (SCL and SDA) on the Arduino and 3.3V signals on the Si7021. However it's academic really! The little GY21 PCB has level shifters and a 3.3V regulator onboard. AAAAArrrrgggghhhh.
my guess would be that those things are one and the same on the inside. That's why they show same reading.
I'm thinking the same thing. When buying on the cheap like that, the engineering (or lack there of) tends to be very similar if not exact. Especially spec sheets that look almost exact.
with an Arduino Pro Mini (3.3V), do we need the voltage converter?
hello julian i think you answered your own question at the end of your vid, i am a refrigeration & air conditioning engineer, basically to measure temperature is cheap and easy, humidity is a different beast IE costly, it can be done electronically but if you want a cheap affordable measurement, then you need a whirling hygrometer, looks like a football rattle, works on wet & dry bulb principle, thanks andrew
Thanks Andrew - I like the sound of a whirling hygrometer!
Sling psychrometer is the term I was taught for the wet/dry bulb measurement. But far cheaper is using saturated salts to create a known RH in a closed system.
you can create 75% environment with wet salt in a sealed box
I see a 3.3v pin on multiple arduino versions. Couldn't that be used ?
That can be used to power the sensor. But shoving 5V signals (SCL and SDA) into a 3.3V device isn't a good idea unless the sensor has 5V tolerant inputs.
+Julian Ilett Oh right my fault I didn't think of the signals voltages. BTW did you check you CPU usage when the IDE is loaded ?
Thanks for another great video!
try the sensor on battery voltage instead of 3.3
simon aldridge why? And what voltage? 1.5,v 3.7v, 12v?
What kind of Arduino Uno clone is that? I can't seem to find any info about this perticular one on the Internet.
It's quite old - I can't find it either.
Julian Ilett
Hence I pay an annual fortune to a calibration house to maintain my ISO9001 Certification. Another quality and educational video J, entertaining too
Several websites give live humidity readings for many UK locations so you put the hygrometers outside and see how well they match.
Or see, for example, www.engineeringtoolbox.com/humidity-measurement-d_561.html
You have a pretty accurate "dry bulb" temperature as all four devices agree, so you just need a wet bulb thermometer and you can calculate the indoor humidity.
What does 5% accuracy mean when talking about percentage reading. eg if it's reading 50% does that mean the true value is somewhere between 45% and 55%, or between 47.5% and 52.5%?
Good point - and another good point :)
my setup is: Arduino Pro mini, DHT22, BMP 180, Nokia Display.
i get Temp, Humidity, Air Pressure in hPa and Battery level in %
I'm planning a miniaturisation of this project using a Pro Mini :)
Julian Ilett the setup works great! (near 1 year) but, battery consumption issn't what i want. i build this project again but with a qC on 3,3V and 8mHz by myself.. (Atmel qC, crystal, 2x Cap's or internal Clock!!) then a deep sleep with an interrupt of 1 min, display refresh..
it issn't hard but all has to be matched.
I was just thinking the same thing. My miniature Arduino hygrometer will look great, but power consumption will be terrible - probably just a few hours on a 9V battery.
Julian Ilett jep, jep, that is why i wish a battery friendly version. i think i will also add a Button or anything (PIR?! ) that activates the Display only if required.
why do you need a voltage translator doesnt the arduino have a 3.3 v port
The Arduino UNO 3.3V pin is a supply voltage to be used for powering off board sensors / boards that require 3.3V power - etc. But all Arduino UNO data pins use 5 V on it's data signals. So - an external 5V to 3.3V level shifter board would be required for the data pins. And the shifter board would require both a 5V and a 3.3 V power feed to operate. So - you could use the 3.3 v (and 5v) feeds from the Arduino to power the shifter board - so you wouldn't need a separate external 3.3V power supply !!
Those 2 digital hygrometers may be using the same chip - teardown time.