Entry Sensor V2 Dissection Pics

Yes…I may do it this way. I will want to buy separate magnets though…you are correct that I have a bunch of V1 sensors currently shelved, but I do want to use them eventually. Honestly I want to connect them directly to Home Assistant and use them locally as makeshift light switches that will actually run fast and work without internet. Just so many other projects to get to, but I don’t want to separate them from their magnets since I have 3D printed brackets that fit them perfectly to work as great light switches…hence, I’ll consider separate magnets. I originally hoped magnetic tape would work for this, but it was too weak, so now I’ll look at stronger magnets, and maybe try mixing it with victor’s idea for sitting a magnet on a steel strip and see if that lets me use just one or 2 magnets.

Kind of disappointed to see the Texas Instruments CC1310 in there considering that’s the chip in the V1 sensors which commonly caused permanent failure when the batteries died. I sure hope something was done to address that problem in this new design. I lost a bunch of V1 contact and motion sensors due to that design flaw.

Seriously? So they all they did was use a much bigger battery. Wow, I hope someone enterprising does some low voltage testing to see if the V2s can also get wiped out.

I’m one of the folks who decided to purchase the CC1310 development board, figured out how to use it, and attempted to reprogram some of my dead sensors. The reality was that the sensors were not actually dead, but that the region(s) of the flash memory which contained one or more copies of their MAC address became corrupted at some point due to the low battery voltage.

What I determined was that even with a restored MAC address(es), the sensors still would not pair through the Wyze app. After reprogramming, they would work perfectly well with Home Assistant, which requires a Linux device and v1 sense bridge (the type normally connected via USB to v1 and v2 Wyze cameras). Home Assistant would correctly report each sensor’s MAC address and device status again. But the same reprogrammed sensors would not pair with the Wyze app. After looking at the flash memory dumps of a few of my sensors which had not yet died, I think I figured out why. It looks like there is a 16 byte number stored in the same region of flash as the MAC. That 16 byte number is random and unique to each sensor. I believe that number is somehow used in conjunction with the MAC during the pairing process for authentication when the Wyze app is used (which entails communication with the Wyze cloud servers which probably contain matching pairs of each device’s MAC + 16-byte unique number). Without both the MAC and its matching unique 16-byte number, pairing with the Wyze app fails. And unfortunately, when a sensor dies due to low voltage, both the MAC and that 16-byte value are lost since they are stored in the same region of corrupted flash. The MAC is marked on the sticker on the back of the device, so it can easily be recovered. But the matching 16-byte value is not stored anywhere else and is lost forever when the sensor dies.

So the only way to fully recover a sensor after it dies is to reprogram that sensor’s original (pre-death) flash image in its entirety. Which means that a user who wants to ensure that they can restore their sensors if they die must go through the effort of reading every one of their sensors’ flash images individually before they die and safely store those images somewhere. Probably not something any user is going to want to do.

I sure hope that either someone at Wyze changed the hardware or firmware design somehow or there has been some change to the CC1310 to prevent the low voltage flash memory corruption. I really don’t want to start losing my sensors one by one again. I don’t use Home Assistant and I never backed up any of my sensor firmwares prior to their death. So I have quite a few useless v1 sensors now. I don’t want to start adding v2 sensors to the v1 sensor cemetery.

1 Like

Hello People and @carverofchoice

Maybe this idea may work as shown in the picture below.

The closer to the door hinge the sensor is the larger the amount of hook-latch movement can be accommodated. Probably without any extra magnets or steel bars?



I thought about this, but I was afraid that if it is too close to the hinge I might have less accuracy on when the door is cracked open enough to let people or pets inside but not open all the way and so maybe it wouldn’t alert me that it was open. That was why I was mostly keeping it toward the right side, so if it opened enough for a pet or toddler to sneak in, it would tell me.

Very good, till later.

1 Like

Hello Mr. @carverofchoice
I guess, Till later has arrived?
The pictures below depict another approach to sensing your door.

I appreciate your support, Victor.


thanks for showing the pictures of the internals of the sensor. I want to remotely trigger the contact sensor for a project by using a remotely controlled contact relay from a non-Wyze system. I’ve done it with the old reed switches, but it doesn’t appear that this will work for what I planned. Any ideas?

I say, lets go with the larger magnet. :joy:

1 Like

Hi @vtester. Have a starting point for you. The power input to the hall effect sensor is 3.66V. The output is 0V when magnet is present and 3.66V when the magnet is removed.

Shorting the “OUTput” to ground (center pin) will send a “Closed” signal to the hub.


Edit: This is not a solution, it will drain battery. Tried w/resistor down to 100 ohms but still didn’t trip and still drained battery.


So the V1 chip and the V2 chip are slightly different, but I don’t know what the differences mean:

V1 Chip:

TI 7AI (I think the last character is an I?)
AY53 G4

V2 Chip:

TI 121

I do know that the information for CC1310 chip indicated that as of several years ago there were [at the time] only 2 revisions of the chip, Revision A, and Revision B. Both of these revisions were subject to the issue titled:

“Slow Transition Across Brown-Out Detect (BOD) Threshold Might Cause the Device to Hang”

Revision A was taken out of production as of April 2017, which means that neither of these should be Revision A, and they do seem to have some slight differences still. Maybe there is a Revision C now? I can only hope that the slight difference means that the V2’s have an upgraded/fixed version that will not be subject to the brownouts when they hit low voltage. Still, it is good to know that we should probably be careful not to let the batteries get too low in the sensors just in case they are also subject to the same Brown-Out concerns. Does anyone know what the implied differences are (besides obviously RAM/Memory)? Maybe RAM/Flash are the only differences and the chip itself is the same, so the BOD issue will always be same? It looks like there are several slight variations:

I still can’t make out what the last 2 lines of characters on the chips mean. I know the first line is the Chip model, the second one is how much Flash it has (128KB), but not sure about line 3 and 4.


@vtester - Removed the hall effect component and added a 10 Meg Ohm pull up resistor between VCC and the (very tiny) Output solder pad. Used tiny solder wick to release. Use 24 guage wire (or smaller).

  • Without the pullup resistor, the output floats. If you touch the output with VCC (+3.66V) it will show open.
    If you ground the output it will show closed.
  • This state may not be stable so the 10 meg ohm resistor holds output terminal up to about 2.042V and will stay (show) open.
  • When the output terminal is held to ground, the entry sensor will send a closed signal to the base.

This is my test, there is no guarantee. Your input appreciated. YMMV


I just tried the same thing and had similar results. A 0805 resistor fits perfectly between the supply and out put pads. I didn’t measure the voltage but I did measure the current with the sensor output leads open and closed and the current averaged .01 to .02 ma in both states. I’ll have to see how it holds up long term but so far so good.

Not thinking it through I had originally just soldered wires to ground and output of the hall effect ic and, as you would expect, burned through the battery in just a few hours. If I had a brain I might be dangerous.

Anyway thanks for the input.

1 Like

shooting from the hip . . . how about just plain magnetizing the tip?

A husky magnetizer is about $4 at Home Depot.

Getting extreme, you could turn it into an electromagnet . . .

gosh, in my house, it’s me that isn’t allowed in the kitchen.

I’m also not allowed to tough her appliances once I install them. Until, of course, it’s time to prepare them . . .

(this started when I bought her a super-fancy, top of the line washing machine. And, well, it was good strategy on her part, leading me to buy more fancy appliances . . .)

hmm, another thought . . . a small spring that will open the door a bit if not latched, and then just measure the door . . .

(personally, I’m considering embedding the sensor unit and Manet in my front door and the frame)

mmm, mechanical solution: something on a string for each “recognized closed” position, each of which, by gravity, pulls the magnet into position. Or, falls “far enough” to activate, unless. another has fallen “too far” and pulls its string.

[ok, so I just proposed implementing an AND gate with one input invert with rocks and twine . . . :roll_eyes:]

1 Like




Nothing is beyond the ability of the red blooded american male!

[I actually came up with DeMorgan’s theorem on my own at about 14, though expressed differently. I put it as, “invert all the inputs, outputs, and the function of the gate.” It came about when I needed a gate I didn’t have, but had a spare quad NAND, so . . .]


I considered that, but I’m not sure if the pull will be strong enough. Magnetic tape was too weak and the hook is tiny. It does have potential and it is very tempting to try though.

Good ideas.

another thought would be to make your hook out of a more serous material that can take a strong magnetization.

hmm, for that matter, what about making a hook out of a magnet, perhaps with a drip or router?

1 Like