Monday, September 15, 2014

Dummy 14dB Pulse Amplifier

much of a theoretical concept ...
fast schmitt trigger how to :: (uses v12.c from the SN7404's 'Spice alternate)
 Dummy 14dB Pulse Amplifier in action
.
.
.
.

this is amazing the i-net search can't allocate nor sine nor trapezoidal wave energy formulas ???

as much as i comprehend it'd go like ::

A(work J) = E(nrg J) = P(power W)dt = 1/R∫U²(t)dt

so for sine ::

E = U² / R ∫ Sin²( ωt ) dt = [ 1 - Cos α = 2 Sin ² ( α / 2 ) → { 1 - Cos ( 2 α ) } / 2 = Sin ² α ] =
= U² / ( 2R ) ∫ { 1 - Cos( 2ωt ) } dt = U² / ( 2R ) [ ∫ dt - ∫ Cos( 2ωt ) dt ] =
= [ screw this . . . ------------------------------- . . . , ok] =
= predicting : [ Sin ' ( 2ωt ) / 2ω = const. 1 / 2ω · outer fn. derivate Cos(arg.-s) · inner fn. drvt. 2ω =
= Cos ( 2ωt ) . . . - so - . . . ] = U² / ( 2R ) [ t - Sin( 2ωt ) / 2ω ] = [ for the half cycle ] =
= 2 · U² / ( 2 ... = U² / R [ τ / 4 - Sin{ 2 Pi / τ · ( τ / 4 ) } / ( 2 Pi / τ ) ] =
= U² / R [ τ / 4 - 2 τ · Sin( Pi / 2 ) / ( 4 Pi ) ] = U² / R · τ / 4 [ 1 - Sin( Pi / 2 ) / ( Pi / 2 ) ] =
= U² / R · τ / 4 [ 1 - 1 / ( Pi / 2 ) ] = . . . there's long time since i last did such  - - check multiple times before you use any of it !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! . . .
. . . U C :: E = h ν ( in µ-world ) , but here the E became C / f , as for f → ∞ E = 0 it doesn't quite match what i remember about . . . however for ⌂t and P it'd be C · τ / t , IF τ = t THEN P = C for all frequencies likely applies for non relativistic world thus the E.Sine formula might after all be and what it was found here . . . might !!!

&Shit , i gess i see the error (always post error-check myself) :: for each quadrant wave we have a bit different result ::

= U² / R · τ / 4 [ 1 - 1 / ( Pi / 2 ) ] = as infact =
= U² / R · τ / 4 Σ0,3 [ 1 - Sign( 1 - 2 (( tLOWER div τ ) mod 2 )) / ( Pi / 2 ) ] = U² / R · τ / 4 =
= not exactly sure what i'm doing (a progressive exacting) // what we should like get is
average U of sine that is ∫ASin(t)dt = -ACos(t)=A at t=(0,Pi/2) for 2Pi 4A relative value for average deviation is thus 2A/Pi . . . as U²/R·τ/4 = (2U)²/R/f , hmm for P the f goes F'd and it's OK, but for NRG ...

oscillation is ± disturbance/deviation  from system balance center ??? E = P·t = P/f . . . or dE * = Pdt = ...
... = [ if f = 1/τ = 1/(t/n) → 1/f = t/n → 1/fconst = dt/n * ] = . . . ??? PMAX · f³(t) / (3f) , f → ∞ 1 / Pi² = 1/(±Arg(-1))² . . . 1/arg , i = exp(ln(i)) = exp(ln(1·e^(i · (2n ± 1)π))) = exp(i · (2n ± 1)π) , ln i = ±iπ , (±iπ)² = -π² , f(-x)=1/f(x) , . . . = this whole computation must be started by some other way (we are missing stuff here ... ) / halt // halt /// halt

& for trapezoid ::

E = 1 / R [ a0² ∫ t0 dt + U² ∫ 11 dt + a2² ∫ t2 dt ] =
= 1 / R [ a0² t0³ / 3 + U² t1 + a2² t2³ / 3 ] = [ U = UMAX = ai ti ]
= U² / R [ U / (3a0) + t1 + U / (3a2) ] =
= U² / R [ t0 / 3 + t1 + t2 / 3 ] = [t0 -- rise time ; t1 -- ON time ; t2 -- fall time ]
P = E / ⌂t = E / ( t0 + t1 + t2 )

Monday, September 8, 2014

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

Wednesday, August 13, 2014

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]