Best BLE sensor to buy in 2025?

[ilgrank]

Hi
I have several LYWSD03MMC , and I'm quite happy with them: cheap, reliable, low battery usage, so-so antenna.
..but since @pvvx added several new supported sensors, are LYWSD03MMC still the 'best buy'?
Is any of the new supported sensors a better buy in 2025 than the LYWSD03MMC? (Better antenna, lower power consumption)
eg, the ZY-ZTH03Pro seems to have even lower power consumption, uses more cheaply available AAA batteries (which can be had for free in the battery trash bin of most supermarkets with enough juice to run such sensors for years)

[pvvx]

The best of all remains MJWSD05MMC

Large display, accurate clock, best antenna (best of all in terms of communication range).
Battery life is more than 2 years. Further depends on the settings.
2 available buttons. The second button is used to turn on the external device.

[pvvx]

eg, the ZY-ZTH03Pro seems

The ZY-ZTH03Pro display is not very contrasty (poor viewing angles).

Better than LYWSD03MMC, but worse than some others (MJWSD06MMC).

[pvvx]

LYWSD03MMC is currently the worst in terms of display.

At the same time, LYWSD03MMC has a price more than twice that.

The last one on the right is LYWSD03MMC.
Next to LYWSD03MMC are cheaper Tuya options.
It's hard to see in the photo, but the MJWSD06MMC has slightly better display contrast and viewing angles than the ZY-ZTH03Pro. (I'm too lazy to take quality photos. The readings vary - a few seconds before this photo was taken, the thermometers were in different places in the house and in the hands...)

MJWSD06MMC / MJWSD05MMC

2xAAA


ZY-ZTH02Pro displays data from an external thermometer located outside...

[ilgrank]

Wow, thanks for the detailed comparison! I'm buying some MJWSD05MMC to check if they're worth the update in my scenario

[Bjk8kds]

@pvvx

Next to LYWSD03MMC are cheaper Tuya options.

What device is that?

[GAZ082]

Hi there. I'm looking for info of the sensor used in the MJWSD06MMC (which is the mini version of MJWSD05MMC). The model is "SHV4" which found no info about. I doubt it is a Sensirion. Do you guys have some info about it?

Appreciated!

[pvvx]

MJWSD06MMC:
Sensor SHT4x id: 70410C94, I2C address: 0x88 (I2C 7bit addr = 0x44).

Sensirion SHT4x.

SHTV3-V4.pdf

[GAZ082]

Thanks! Gonna buy one then!

[joshuakraemer]

What about the CGG1? It works great for me and I much prefer its e-ink display.

[pvvx]

What about the CGG1? It works great for me and I much prefer its e-ink display.

One battery

---

Worked on CR2032 :)

---

This one works in winter, at temperatures close to zero degrees C...

PS: But the MJWSD05MMC lasts longer on a single battery than any CGG1. About 3 years...

[and-fed]

I saw a schema (attached and hope it's accurate) how MJWSD06MMC can be improved to last longer: remove one resistor and join an LDO contacts . Won't it be the best sensor to buy if assume both use custom firmware for BLE?
What is a downside if compare with MJWSD05MMC after the improvement?

[pvvx]

Would you mind if I ask you? If you believe the battery lifespan can be significantly prolonged by a capacitor for sensor devices with "coin" battery, why don't add something to home page about it with a list of several models which fits good for most populate sensors? It would be greatly appreciated by community and others will test the approach too. I am just an amateur without deep knowledge in this area and after reading all articles I have doubts about which exactly capacitor to solder. Because the devil is in the details. I realize it should be, most likely: 33-100uF, ~10-50V with leakage <=1uA, affordable/cheap, small enough to fit into box(sensor body).

My experience with electronics dates back to 1979... But I am not a specialized programmer, because I spend most of my time developing electronic systems...
And thermometers are purely part of a hobby and I post on github what came out and used for myself. I do not expect any priorities and I do not need any community support. There is a greater desire for someone else to take up the continuation of this theme. The topic with thermometers is far from the first open topic where I took part... It's time to change to something else, because in electronics everything develops very quickly.

And there is a very big difference when you select elements for production and when it needs to be recommended for ordinary users in single copies...

[pvvx]

MJWSD06MMC: Comparison of maximum pulse current without a capacitor in the power supply and with a capacitor.

Without a capacitor (window 50 ms, Power 3.3V, equivalent internal resistance of the source is 10 Ohm):

With a capacitor:

The average current over the measurement period is the same with or without a capacitor.

[and-fed]

It can help to prolong cheap CR2032 lifespan, but I'm not sure about other cases. I can't do real tests for CR2045 to estimate 7mA overall impact on longevity. I wasn't able to find independent tests too. I can just relay on specs and recommendations from manufacturers. Found one interesting from murata. Looks like the batteries can handle IoT devices like these sensors without any noticeable degradations.

I agree the capacitor helps to "smooth" bursts, but does it worth it? If the capacitor requires extra 1uA, it will be about extra +7% (15/14 )power consumption. If so, is it an evidence the longevity will be prolonged by at least same amount? I really don't know, but so far assume it won't. Again, I keep in mind reliable brands only.

Another potential issue with the capacitor. If there is not good one around, need to order it from a reliable shop with extra expenses at least for delivery for a several cents item. It's unreal to buy along with sensor from Ali a good one, most likely a fake will be soldered. I personally tried to find and eventually gave up - too risky.

If I have a good one capacitor around, I would solder it just to try. Yours has very low leakage if average is same, 0.1uA?

[pvvx]

If the capacitor requires extra 1uA,

I haven't seen capacitors like these for years. Nanoamps are still possible...
Only aluminum electrolytic capacitors have a large leakage. Tantalum and ceramic capacitors have a leakage of nanoamps at 3 V, if a capacitor of 16..25 V is selected.

For the graph, I used a previously rejected 470 μF 16V electrolytic capacitor with a large leakage and high ESR, underestimated real capacity (obvious defect) :P
As a result, the idle current increased by 0.2 μA.
With a tantalum capacitor 22uF/16V, it is impossible to register a leakage on the simplest PowerProfiler with a sensitivity of 10 nA.

[pvvx]

Example: Connecting and disconnecting a capacitor to PPK2 from nRF.

To offset relative to zero, a 4.6 MOhm resistor is connected in parallel to the PPK2 output. Otherwise, PPK2 incorrectly measures the nanoampere current due to the noise of its own power supply - the amplitude goes below zero and spoils the calculation of the average value.

When a capacitor is connected to the measuring shunt, interference is observed in the power supply, but the average current values do not change.
If you disconnect the capacitor and measure the voltage on it after a few minutes, it will be the same - 3000 mV.
The simplest multimeters with a measurement accuracy of 0.01 μA (UT71B, UT89XE, ...) will not show leakage of normal capacitors.

---

And this is the cheapest aluminum electrolytic capacitor (470 uF 16V):

And in the end the current will drop to 0.2..0.16 µA. But such capacitors have a huge noise... And this capacitor has not worked yet. And over time, when working in a stable voltage, its leakage will decrease. This is a property of electrolytic capacitors...
But they are huge and will not fit into the thermometer case :)

---

In most cases, stories about large leakage currents are stories from the last century, or about ceramic capacitors that got wet from high humidity, or were contaminated with flux during installation...

[and-fed]

I opened catalog Standard and Low Profile Tantalum Capacitors
TAJB226*020#NJ
Capacitance: 22 µF
Rated Voltage: 20 V
DCL(DC Leakage) at 20V: 4.4 µA (max)

For lower voltage using the doc I looked at "LEAKAGE CURRENT vs. RATED VOLTAGE":
Voltage: 3 V (3/20 = 15%)
Leakage should be about 0.9µA (max), average should be about 0.4µA per chart.

If you don't agree with the numbers, what is you expected leakage for the tantalum branded capacitor at 3V?
Looks like the leakage is much more smaller than estimated. If possible, please share what exactly 22uF/16V capacitor you used.
If it doesn't impact overall power consumption (<1%) - makes sense to solder to prolong lifespan at least for cheap batteries.

For the graph, I used a previously rejected 470 μF 16V electrolytic capacitor with a large leakage and high ESR, underestimated real capacity (obvious defect) :P
As a result, the idle current increased by 0.2 μA.
With a tantalum capacitor 22uF/16V, it is impossible to register a leakage on the simplest PowerProfiler with a sensitivity of 10 nA.

[ilgrank]

Some time ago, you (@pvvx ) suggested to use 6.3V, 33uF tantalum capacitors. In the light of recent discussion, and to lessen the leakage current, would you today suggest to use a 16v 100uF (or more) tantalum capacitor instead?

[pvvx]

would you today suggest to use a 16v 100uF (or more) tantalum capacitor instead?

Only after measuring the leakage current.

[ilgrank]

..that's reasonable, but required measurement apparatuses far beyond what I have access to.
Is there any part on Aliexpress (I know, I'm talking blasphemy: no way of reasonably knowing, and even if properly measured, likely the part will change anyways), Digikey or similar that you would reccommend?

[pvvx]

Take a simple multimeter with a measurement of constant voltage in mV.
Typically, the input resistance of the multimeter on the measurement of a constant voltage is 10 MOm (indicated in the documentation for the multimeter).

1. Connect the capacitor to the CR2032 battery and wait a few minutes to fully charge.
2. Disconnect one wire going to the capacitor and turn on the multimer in this gap.
3. Wait a few minutes when the readings are stabilized and record the value.
   The value in mV will indicate the current of 0.1 nA (100 pA).

It is recommended to measure without changing the temperature of the battery. That is, at a stable temperature of all components ...

[pvvx]

Digikey or similar that you would reccommend?

An example of a terrible capacitor purchased from a certified specialist store:

1330.1*0.1 = 133 nA (3.3-1.33=1.97V)
Another precision current meter shows a similar value with a difference of +-5%.
And this is not the worst of this batch of capacitors. They all go to the trash.

In a power source, such as a charger, such a capacitor will work and you will not notice anything...

And this is a film capacitor - less than 1 pA at 3.3V:

[sofakng]

Is the MJWSD05MMC still the best (recommended) sensor to buy?

Is the MJWSD06MMC the same but just smaller?

Any recommened places to buy them? I'm only seeing "Xiaomi Mijia Smart Bluetooth Thermometer 3" for sale which I think are the correct ones ...?

[ilgrank]

@sofakng :I've just asked the same question, did you read the above answers?

[sofakng]

Yes, that's why I asked if it's "still" the best ones and I also asked two other questions I didn't see answers to.

[ilgrank]

It it was 2 weeks ago, I think you can be pretty sure it is still is today, don't you think? :p
2: detailed infos about the sensors can be read here: https://github.com/pvvx/ATC_MiThermometer, just check the hw specs of the sensors you're interested in, each sensor has its specs page linked from there
3: Aliexpress usually is a good source, Taobao even better if you speak chinese.

[420pootang69]

As far as I can see MJWSD05MMC is still the top dog, but it's far more expensive than the LYWSD03MMC.
Other side of the LYWSD03MMC is that there is a ton of 3D printed enclosures for them, so you can easily use them with AAs.

The only thing I can't find on here is just how bad the new display is for battery life? I did read PVVXs comments, and it does seem like a CR2032 will not last long with one, however for those of us on AAA batteries, maybe they're still the best option?
Like can I just turn the updates down to every five minutes or something to compensate?

I think my opinion compared to PVVX is highly skewed due to location, as the ones he mentions seems plentiful in (for example) India, but for AliExpress Europe, the only ones available for ~€5 are the LYWSD03MMC. The cheapest I can see the MJWSD05MMC is from a dubious seller at €15, with most prices at €20.

[pvvx]

This year MJWSD06MMC (CR2450) is already on sale...
The MJWSD05MMC (CR2450) has been on sale for over four years. Initially, they featured Chinese day-of-the-week symbols, then a European version with English names was released. And now they're officially available everywhere.
The LKTMZL02, ZY-ZTH02Pro, ZY-ZTH03Pro (2xAAA) is the Tuya version with Zigbee.
Similar to those listed, it has a TLSR825x chip.

[pvvx]

The lowest price I've found on Aliexpress so far:

$6.99

$6.37

The lowest price I've found on Aliexpress(ru) so far: THB1 $2.2 (when purchasing 1 piece with delivery):

I am sure that if you order a dozen thermometers, you can find prices that are half as low.

[pvvx]

I think my opinion compared to PVVX is highly skewed due to location, as the ones he mentions seems plentiful in (for example) India, but for AliExpress Europe, the only ones available for ~€5 are the LYWSD03MMC.

At $5, these are old LYWSD03MMC stocks purchased from you before the introduction of increased duties on сhinese goods. Or you have expensive delivery (remote village or hamlet).
It's probably more expensive in India...

---

Russia, with delivery to St. Petersburg, MJWSD06MMC x5 $29.5 (5 x $5.85):

Russia, with delivery to St. Petersburg, MJWSD06MMC x3 $26.65 (3 x $8.88)

[420pootang69]

Thanks for the input, just confirming my suspicions that you have a better, cheaper selection over there then I have here in Ireland!
Maybe I will take a trip over and visit you and pick up a large bag of thermometers :)

I did find the TH05F below €5, but this is not recommended according to your documents.
For the same price, I also found the Tuya one in your image here:

The Tuya should get me at least a better display, however I have really cheap AA battery cases for the LYWSD03MMC already, so I'm still tempted to give them a try.
I'll get one of the new LYWSD03MMC and test what the performance is like for myself before I buy more (I need ten).

Regarding the missing capacitors on the LYWSD03MMC, I would assume when running from 2xAA batteries, this isn't an issue as the current spikes won't cause an issue for these?

Also, has anyone compared the temperature accuracy between these? I know the older LYWSD03MMC models I have here are very good, it would be great to compare them to the Tuya version.

I am also purchasing a CH582F and a INA228 so I can try out the power profiler. I wanted something like that for a project about 10 years ago. The company I was working for couldn't afford something that could sample that fast, precise and that low. It's crazy to me that I can get the equipment to do this now for less than the price of two pints of beer :)

[pvvx]

I am also purchasing a CH582F and a INA228 so I can try out the power profiler.

https://github.com/pvvx/SimplePowerProfiler/tree/main/source/ch583/WEB_USB