8 x DS18B20's for $10 i....showing a temperature of about 4 degrees celsius cooler than another bought sensor with LCD display. I now need to look into calibration testing with an ice water bath and boiling water bath to work out if they are reasonably accurate.
This is an interesting example. Is this an issue of the quality from China, or the quality of the type of part itself?
I ended up with about 16 different temperature/humidity sensors, all locally purchased. I put new batteries in all of them, and put them in a box.
Every one of them had a different reading. No two gave the same reading for temperature or for humidity.
I did the calibration thing with sticking a cup of saturated salt water in the sealed box, and letting it sit for 24 hours.
Again, every one was different. Some were closer to the expected reading, but not the same ones that were closer at the previous temperature & humidity.
Some proved to be a bit more consistent, but none of them were more accurate.
So maybe there is just a wide variance of parts, regardless, and for certain things we just have to do the extra work ourselves to get them measured and calibrated?
I guess for temperature, the only sensor to trust is the thermistor probe on the multimeter?
I suppose that the point I have tried to make is this is NOT a China specific issue, but an issue with buying small quantities of parts from dealers that are selling excess inventory and possibly inventories purchased in bulk from bankrupt businesses.
Excess inventory may often include rejects.
Digikey, Mouser, and RS Solutions are out there for people that want or need first rate quality.
I bought 8 humidity/temperature sensors from China via ebay, these are just some small (approximately) 2"x1.5"x0.5" integrated units with an LCD display, powered by an lr44 battery. I bought them in two batches, a year apart. They cost around $4 (IIRC) a piece, free shipping. What's nice is that they ended up showing exactly the same values for humidity/temperature, they needed two-three days in the same spot in the room to converge (for the units from the second shipment to agree with the ones I had already). I have checked with a laboratory humidity instrument and they're less than 5% off, which is pretty good. Closer to 2% actually (although 'accurate' humidity level is very hard to measure). Temperature is spot on.
So YMMV I guess.
-Tor
(so why did I buy all of those? It's because I have a number of solid wood guitars and other instruments, and I need to keep an eye on humidity, and, to some extent, temperature. Especially during winter. So I have these cheap sensors everywhere, in rooms I intend to play in, and in guitar cases and such. Even in my office, where I currently read 36% 25C, so only the laminated office guitar will do.)
Yet from some testing on the weekend with an ice water bath, I tested some of the unit's, the best read 1.56C, the worst was over 4C. Each read differently. Each was in the bath via a thin plastic bag for over 10 min. I need to do further testing. I'm yet to adjust the code to display multiple at that same time to see if they really are displaying differently.
Maybe the DS18B20 was a calibrated device that at default gave you +-0.5C temperature accuracy....
ice water bath,... for over 10 min.
Maybe they need a longer time to sit. I had mine in a large plastic box together, and then individually in plastic bags. But only left them overnight, not two or three days.
I guess the quality of the local testing can have issues too.
Also, I thought the water in the ice water bath needs to be circulating at a certain rate, or the water will be warm or supercooled?
Comments
This is an interesting example. Is this an issue of the quality from China, or the quality of the type of part itself?
I ended up with about 16 different temperature/humidity sensors, all locally purchased. I put new batteries in all of them, and put them in a box.
Every one of them had a different reading. No two gave the same reading for temperature or for humidity.
I did the calibration thing with sticking a cup of saturated salt water in the sealed box, and letting it sit for 24 hours.
Again, every one was different. Some were closer to the expected reading, but not the same ones that were closer at the previous temperature & humidity.
Some proved to be a bit more consistent, but none of them were more accurate.
So maybe there is just a wide variance of parts, regardless, and for certain things we just have to do the extra work ourselves to get them measured and calibrated?
I guess for temperature, the only sensor to trust is the thermistor probe on the multimeter?
Excess inventory may often include rejects.
Digikey, Mouser, and RS Solutions are out there for people that want or need first rate quality.
So YMMV I guess.
-Tor
(so why did I buy all of those? It's because I have a number of solid wood guitars and other instruments, and I need to keep an eye on humidity, and, to some extent, temperature. Especially during winter. So I have these cheap sensors everywhere, in rooms I intend to play in, and in guitar cases and such. Even in my office, where I currently read 36% 25C, so only the laminated office guitar will do.)
cause I have a HUGE L shaped desk. 8 feet long one side 6 feet other .
Attachment not found.
http://pdf1.alldatasheet.com/datasheet-pdf/view/58557/DALLAS/DS18B20.html
Yet from some testing on the weekend with an ice water bath, I tested some of the unit's, the best read 1.56C, the worst was over 4C. Each read differently. Each was in the bath via a thin plastic bag for over 10 min. I need to do further testing. I'm yet to adjust the code to display multiple at that same time to see if they really are displaying differently.
Maybe they need a longer time to sit. I had mine in a large plastic box together, and then individually in plastic bags. But only left them overnight, not two or three days.
I guess the quality of the local testing can have issues too.
Also, I thought the water in the ice water bath needs to be circulating at a certain rate, or the water will be warm or supercooled?
You automatically win the "cool lab + lab wear" contest.