Example:
Similar for the AMS HT.
Is the AMS sensor just wildly inaccurate naturally, or is it placed somewhere weird that might skew its measurements, or…?
Any way to calibrate it?
Any affordable RH sensor is generally terrible at measuring <10% RH, so anything below 10% can be considered as anything below 10%… Your sensor’s 8% could be another sensor’s 0%, your sensor’s 0% could be another sensor’s 10%.
Luckily you don’t really need the accurate <10% readings. All you need to be care about is 15%, if it’s below 15% you’re generally safe. If it’s anywhere below 10%, you’re definitely safe.
OK, is it accurate at 15%?
This is usual behaviour for any type of standard humidity sensor. The measuring range of commercially available ones usually ranges from 10 to 90 RH%, with values outside of that either being inaccurate or just assumed 0 RH% like in the AMS2. The switchbot wireless hygrometre falls directly into that usual range, and the yolink one actually has a worse accuracy (±3 %RH) within the same range.
To make matters worse, the humidity sensors used in all Bambu Lab products are sensitive to high temperatures and become highly inaccurate when exposed to heat. This isn’t a problem unique to Bambu Lab — most common RH sensors struggle under elevated temperatures. If you need accurate readings in high-heat environments, you’ll have to invest in industrial-grade sensors, and even then it’s often a waste of money for typical consumer use 
For sake of argument, assume Bambulab used a relatively not-so-great humidity sensor that’s accurate to plus or minus 5% RH at 10% RH humidity and above. Then, by @hotellonely 's logic, you need to be concerned when it shows 10% RH, because that could mean 15% RH when you account for the 5% possible measurement error. Right?
Anyway, the switchbot uses the SHT40 sensor, and according to its datasheet, it is accurate to plus or minus 3% RH at low humidity levels:
source:https://sensirion.com/media/documents/33FD6951/67EB9032/HT_DS_Datasheet_SHT4x_5.pdf
So, if Bambulab’s AMS RH accuracy really is 5%, then a value of 5%RH would be consistent with the Bambulab reading of 0% and a SHT40 reading of 8%.
I have plenty of examples where AMS2 Pro shows less than 10%:
i.e. it’s not like some, where it drops to 0% whenever it gets below 10%.
What I meant is that when it’s showing below 10%, it’s usually really below 10%, but it could be anywhere below 10%.
RH sensors would also be affected by temperature and air flows. A higher temp measured would generate lower RH.
Maybe we can agree that this applies to the cheapest of cheap sensors? As soon as you spend at least a bit of money, they become quite good quickly.
E.g. a specification of ±3% RH is quite excellent. It means that this is the absolute worst deviation the sensor can have under any circumstances. Typically it is much better than that.
Apart from the sensors themselves, the manufacturer can make a lot of errors in the rest of the device. So instead of a bad sensor, the fault often is in the bad engineering of the device.
Actually i think that the AMS does not really report relative humidity but an arbitrary “quality value”. At least that is the feeling i have for the AMS 1.
I guess we just have to wait for someone to do a teardown to find out what humidity chip it uses. Once we know that, we can check the datasheet.
Acording to the Bambulab wiki:
Source: AMS HT Part Introduction and Connection Guide | Bambu Lab Wiki
So, now we at least roughly know where it is, just not exactly what it is.
Oh well, it would have been convenient to know. For now I’ll rely on my own sensors.
You’re absolutely right that once you move beyond entry-level humidity sensors, the accuracy improves significantly. A sensor with a ±3% RH specification is certainly very good—and in many real-world scenarios, it performs even better than stated.
However, in the context of a mass-produced consumer product, even a small upgrade like this can have a noticeable impact on production cost.
The price difference between a typical ±5% RH sensor (commonly used in cost-sensitive designs) and a higher-precision ±3% RH sensor—such as the Sensirion SHT31 or Bosch BME280—is roughly $2–3 per unit at high volumes (10,000+ units). While that might not seem like much on a per-unit basis, for a product shipping 100,000 units, it translates to an additional $200,000–300,000 in component costs alone.
In consumer electronics, especially at competitive price points, such an increase often requires trade-offs elsewhere—potentially in shielding, thermal design, firmware capabilities, or mechanical tolerances.
So while higher-quality sensors are certainly worth it from a technical standpoint, the cost-benefit decision becomes more complex at scale—particularly if relative humidity is not core to the product’s primary function but rather a supplementary feature, as in the case of the AMS.
In short: yes, better accuracy does come at a higher price—but it also adds pressure on every other part of the design and manufacturing process.
The chinese have their own TH sensors that are good to plus or minus 3%RH, and I’m pretty sure they’re quite cheap:
Why am I pretty sure of that? Because as a favor to another forum member I did a teardown of a $2.19 bluetooth sensor with LCD display, with free shipping to your door if you bought 5 of them, and it was only just one small component of that package:
That’s why I agree with @Alex_vG 's assertion that spending even a little bit of money gets you pretty good accuracy (3%). I really can’t fathom why the AMS 2 Pro would be worse than that, unless Bambu simply didn’t give a damn about accuracy enough to spend a few cents on it.
Is it really such a critical component that it justifies a higher price?
I would argue it’s not that essential for anyone — but that’s just my opinion.
That’s what I’m thinking as well. I’d say it’s better to be switchable to show either levels or percentages.
If they show a lower number than reality, it would benefit Bambu. Only if it was consistantly lower, and never higher than reality.
How so?
==========================================================
Because people would think the dryer worked better than it actually did. Like when I posted my tpu drying results at 0% humidity, it may have actually been higher. 0% looks awesome though. Not saying its a conspiracy. Just saying its not exactly a bad thing for bambu
I build robots for fun using all kinds of sensors, and I have some experience with measuring relative humidity (RH) – it’s a tricky subject.
Let’s say your RH sensor shows 30% RH at 20°C, 40°C, and 0°C. In reality, the corresponding absolute humidity would be approximately:
20°C = 5.1 g/m³,
40°C = 15.3 g/m³,
0°C = 1.1 g/m³ – give or take.
Another interesting point is that if I get a reading of 30% RH at 40°C and then let that environment cool down to 20°C without exchanging any air (i.e., a sealed system), the relative humidity will rise to around 88% RH.
RH is not a measurement of how much water vapor there is in the air, but rather how close the air is to saturation (condensation) at the current temperature.
Nitpicking, but it is expected that relative humidity sensors will vary when hot. They measure relative humidity, and warm air can hold more vapor.
