1.2V LED driver dev.

it's quite simple to make 24% to 42% efficient 1.2(read 1.46V) "Step-Up"-s or - a less simple - up to 60% efficient (but then low current) 1-s . . . . . . . . . this one promises to be exception to prev. - K-words ::

• 100µH
• a "right" frequency & duty cycle
• special set of transistors

the img. name is misleading relict !

futher to implement ::

• sw2 & feeding the oscillator from HiVtg. after start up
• adjusting the cycle according to emptying of the battery
  from (1.46V,250mΩ) dn2 (1V,1Ω)
• optional blinking mode when battery is too low
• other

Found Bug in "a bit unusual linear voltage regulator"

New v. /!\ not final /!\

? i hope it stays stable (the real build'll likely gonna explode anyway , so ... ? why to worry)
all NPN-s 2N2369 all PNP-s 2N3906 J-Fet U309 D-s 1N4148 Qz Xtal alternate RCCL 5/15pF

PS! there's some non-trivial steps in std. encoding process - whitch i dont know about if they can be simply done making it for now ::

►the mission impossible ::

• make few BJT-s based coFM encoder/transmitter
• implement highly stable Fq. auto adjust
• implement band filtering as accurate as possible

i got some ideas but no exact ideas - (a lot of research to be done)

just fixed the HF supply bug

Spice alternate for SN7404 К155ЛН1

cross domains linking test ::

summary :: at such a high frequency the operational zone for supply voltage and also component parameters appears to be unusually narrow - i used 3 methods to derive the following 1) - const. f wave forcing through chain of 3 to 9 inverters ; 2) - (for NAND gates) 5x ring osc.'s output loaded by extra 5 inputs / (for NOT buffers) dummy 5x ring osc. (no outputs) ; 3) - (for NAND gates) 5x ring osc.'s output loaded by extra 3 inputs and forwarding these to 3NAND which on turn was loaded by dummy input of a NOT buff.

PS/!\ -- i donno what happens if you vary the supply voltage by a couple of milli volts

---

waveforms ::

(?Yo! -- i can see it , i can see it (some sh¡t works (for now x;b)))	blaah

 chips ::


summary :: [▲]

[EOF]

Diode Д9К

a while ago i tried to add it's model to LT-spice - following the specs of common Spice diode model parameters' spec. - THIS didn't quite end up as expected ???

so now i just studied the response of the I(U) graph to random LT's diode model parameter change and came to smth. that in (near) static mode resembles to a real world Д9К

some process steps in graphical representation ::

? empty DIV ?? !Wow

!!"Q%%&&¤/(/ - jälle see "Googel" ei toimi (ma ei kujuta ette neid inimesi kes iga päev siia midagi toksivad , HELP! HELP!) . . . ok , mis kuradi pilt siin on ..
↑ the lab Z-Up - a bi-way scanning of the diode chain varying the curent ↑

↑ fresh 10x + old 1x graph + static I(U) dependency ↑

↑ .model D9K D(Is=40u Rs=.3 N=3.2 Cjo=12p M=1.2 tt=60n Iave=30m Vpk=30 mfg=SpiceCLOzE type=silicon) ↑
/!\use at your own risk/!\

xls's static I(U) dependency build/check plot

---

? should i attempt to switch back to Compose view . . . (should backup this 2 note' B4) -- no switching! (not yet depressed coz the otherwise soo in-common Google malfunction) ('ll Save & FO here)

---

[EOF]

a bit unusual linear voltage regulator

some months ago i tested the FM-mic circuit that made the most sense of the ones presented in www - actually i want to use it for data comm. but there's along way there

anyways (during the test) the fq. drifted down something like a half a MHz per second ???

there might be several reasons for that
- the Cu-coil acting as a spring (slowly - expanding??)
- the battery being too weak (i had no regulator in this test)
- the BJT heating up (loosing the h21e value)
- other ...

... starting from the supply i attached a simple/efficient regulator (not tested in practice yet - all simulation) ::

 
[EOF]