AN IN DEPTH ANALYSIS OF THE UNITS LED DRIVE CIRCUITRY . . . . .
Your having supplied the units foil side "schematic," now lets me see the related two LED lines connections into their drive circuitry.
Refer to my mark up of the foil side of the board. . . . . . as is being placed below
( DOUBLE MAG UP AVAILABLE )
View attachment 58084
Top left corner we can see the dark gray / white wire pair associated with the flasks internal LED and then just to its right is the wire pair associated with the visible external "ersatz incandescent lamp" ( LED )
Take note that the left LED lines dark gray and the next LED lines yellow wires are being hard wire connected to their common foil buss which is the 4.5 VDC battery power buss (Red Wire).
That now means that BOTH of the other white wires need a path to ground to activate them. That is being provided by the left white wire passing down the yellow dotted path to connect into a 56 ohm current limiting Surface Mount resistor . Then it continues on to the right to the top collector tab of a Y1 (SS8050) marked switching transistor .
Meanwhile, back in the jungle, the other internal LED internal flask illumination LED has its white return wire initially encountering a 39 ohm current limiting resistor . . . Hmmmmmmm ∴ that means its LED is going to be driven even brighter, psuedo obscured within the flask, that foil path then continues to make a like connection to the very same Y1 collector.
We now have the two LED's all connected up and only needing the full conduction of Y1 transistor from its collector to emitter ground buss to finally . . . . . . "Let there be light(s)" .
That means that a weaker base bias voltage is now being needed to drive Y1 into its FULL conduction.
Take note of the bottom left corner of Y1 and its base pin which then tales a foil path up a bit and then travels across the longitudinal center of that transistor case and ends up as now being the yellow B mark up dots and travels down as far as the 100 ohm surface mount resistor.
Then after traveling across the "gap" of that resistor it continues as the red dot path and goes into an output drive port of the u-Processor / custom chip, COB *** located underneath the LARGE BLACK EPOXY BLOB. *** Chip On Board.
( Chip On Board.. . . . Its being quite a complex little puppy also, as I see that the green wired squawker speaker is receiving its audio drive directly via two foil lines feeding into an analog AF amp stage portion within the COB chip, )
Then there is ALSO included within it, all of the complex speech generation and sounds / words memory to be sorted out and stored for selected playback. Another of its ports is activating electronics for the ceramic transducer water mister feature.
The units power piezo mister element is fed by the pair of totem pole arranged 2SD682 (4 RED dots) in the top left quadrant of the board, which are fed by the IC dual op amp *** ( 4 GREEN dots) beside them .
*** Since its pin 2 and particularly its pin # 3 have no connections, we know that this is not being a 555 based ultrasonic oscillator.
I see the base of the top '882 being activated by an output port of the COB while the bottom '882 receives ultrasonic drive from the 4 GREEN dot I.C.
The '882 s drive the auto transformer type ferrite based inductor which zips up the voltage drive on up to a more high Z agreement with an ceramic piezo power transducer.
NOW, YOU JEST AIN"T A GONNA BEEEE-L-IIIIIIIEEEVE THIS !
In looking at the proximity sensor ? I was noticing that units circuit boards nearby marking of SQX-033S or 9 and recognized the similarity of numbering / identifying assignment of the companies of Matsushita / Panasonic , and the SQM prefix that they use for some of their other types . . . not PIR sensors.
That number was researching, ended at a dead end, but among the findings that showed up was this Halloween - Steampunk novelties source.
Would you not think, that after seeing their one offering there, that your unit is sharing a common production / design source.
Look thru all of the pics and you can see the working insides of that unit / with your squawker speaker . ALSO there is the adjacent
vacant cavity for a piezoelectric transducer, along with its resonating cavity and your mentioned sound exiting hole.
NOTICE that they assign a QUITE similar ID # to that units PCB !
Also, doesn't that brown Bakelite PCB, and all of its plug in connectors look QUITE familiar
Scroll all the way down to 'lil Johnnys Workshop after seeing the very tops " EYE " video
Halloween and Christmas displays are filled with hackable gizmos, gadgets, and figures. Learn how to add your own custom sounds, motions, and lighting effects to these popular geegaws!
learn.adafruit.com
You said . . .
yes, i unplugged the first led (which was hard to figure out, it has like an arrow mechanism that locks it 
) & it didnt turned on.
Oh tay . . . I got that . . . .
about the mister, i just fill the water a bit before the start of the bottleneck, its hard to get it at the right place for it to work, sometimes it works & sometimes it doesn't, intervals? i only filled water in it once, it works for 5 sec at a time.
I would expect it to NOT, work if OVERFILLED with water, unless they are using some metered / absorptive wick type of feed limiting.
HOWTODOITTOITONTHELED'S . . . . . . . .
Notice on my plotted drawing that I cut a diagonal white channel in the BLACK epoxy coating of the COB.
That let me show the hidden GREEN foil path from the top BLACK star tterminal (ground) down to the point where it can cross UNDER that added 100 ohm surface mount resistor. That foil path then lets the '882's emitters get their ground path.
LED ACTION . . .
At designated LED turn on time of the unit, an output port within the COB activates and travels the RED dot path with base turn on voltage, thru the 100 ohm SM resistor and on down the YELLOW B path into the base of Y1 switching transistor .
With it now having base turn on voltage, its C to E conduction completes the ground path of your two leds and there is light until that drive voltage is turned off by the COB drive port.
There is a series circuit path resistance of the visible 100 ohm SM res plus ~ 1,000 ohms of discrete resistance within the COB.
Get yourself a 1000 ohm resistor and a clip lead or adequate length connective wire, to have one wire /clip connection
at the RED star 's top 4.5VDC battery sources connection and then touching the 1K resistors free end to YELLOW B or RED dots, should light up the LED's.
For a comparative test, touch it to a RED dot connection and then cycle the unit thru a playout.
If the LED's do not get dimmer after the end of that play cycle . . . . . . you do not need to change to a slightly lower value than the estimated 1K.
You then can use a switch to then either engage that add on resistor or disengage it to revert to factory original.
Or as I earlier mentioned, that unit is using only one section of one of its switches while 3 other sections are not even wired in.
ARE WE HAVING FUN YET !
If so lets
ALSO answer that question of that bottom facing, 1/4 inch off the table piezo "tweeter".
Go to the main speaker and unplug it and then cycle the unit through one of its audio and sound effects passages.
With the MAIN speaker now dead, are there any audible high frequents or very high and tinny sounding audio, that's coming from that piezo HF tweeter ?
73's de Edd . . . . .
Are you guilty of . . . . . Asking dumb questions . . . . . because . . . . . its . . . . . easier than correcting your dumb mistakes.
.
