FW - TL431 LM113 Spice model variants´ TEST

it looks like i already forgot what i graphed
Z=V(voltage point)*normalizer -- sets all references to a default unit value (1V)
relative scaler log10(Z)*relativeNormalizer -- resizes plots so they fit to same scale . . .
. . . so the smaller the "relativeNormalizer" is the worse the device performs (has greater fluctuation range)

so the "Z" re-plotted (for stepped current) ::

 . . . the 1.67V µm 431 variant seems to perform best here in both current ranges (x431/x113)
- - - -

---

[Eop]

TL431 Spice model variants´ TEST

presented as we go
► 1-st

► 2-nd

► 2-nd - in detail

wider range

component level models (sub circuits compared)

the best - upper two are the attempts to achieve better precision×frequency-rejection //// the reverse Vref´s emitter (EB) diode in Rank#(7) is a contemporary relic from an attempt to force the "Hot Start" for U4,U5 . . .

pulse e.g. stepped load response

overall ranking so far ...

? best CLM-s !!! for static operation . . .
. . . for undefined input the U5 e.g. X4 is likely better choice /// might be quite expensive regulator as the RF transistors cost $ 0.2÷1.0 per pcs.  /// (i can't comment nor recommend nor reject the macro models - but usually they are too optimistic !!!)

hi, hi - UPS! seems i used Amps insted of milli-Amps

that re-ranks the rank back to "Static" -- the red box is "for super cool experiments only!!" . . .
. . . so U2 // X1 the best CLM !!!

---

[Eop]

about google found VCO-s

Last to First ...

src	mod
as i didn't use the default transistors i was not able to grasp the ® idea behind this mess	finally it best resembled to a phase shift OSC - what it is with a poor fq. ctrl
not a VCO but a Schmitt Trigger

. . . the response speed and frequency range define the possible target app.-s . . . as in fact what i did here was the opposing -- an attempt to find out best range for operation by VCO type (for the ones that didn't work from scratch . . .)

---

[Eop]

a fuzzy science on old transformers 2

finding the optimum simple design for 9V linear adapter

---

using std. parts 7805 , TL431

the load peak/-step response is somewhat unexpected here

---

using low ß , high power switch (such as *MJE13003)

the load regulation stays below acceptable

---

using bridged LM78L05 and LM324

fine tuning the bridge may induce better results this 1 however won´t pass because it´s load response

---

same as above with an attempt to add a reference to be used in better line rejection ...

... can´t gain steady reference - won´t test all possible situations for such a noisy one = reference failure , halt, halt

---

using low ß , high power switch* and random intuitive design

looks pass - tough i don´t know about higher frequency load responce

---

[Eop]

Adjusting the fixed (3 terminal) voltage regulators

extends:: Adjusting the adjustable voltage regulator , Apr 5, 2015

(Std./)Simple v. :

with LT1086 (1.24V - adjustable) it gives seemingly errorless output ?

Op. Amp. v. :

it looks it is impossible to gain the ±2µV error level with just the one op-amp ►►
(the reference must be flawless as it shows on the graph. ???)
The Red LED-s are so cool for voltage adjustment -- their non-linear range begins below 275µA , they have sharper "turn on" than conventional detergents -- which puts them to (Diode,Stabistor,Red-LED,Zener Diode) this position - i assume . . .

---

Update :: 2016-03-11 ► Insane :

 

seemingly "absolute" voltage reference at 1 Hz

---

[Eop]

yet another LDO 5V regulator

the Regulator :: /̲!̲\̲ this thing seems to perform better than LE50A (STMicroelectronics) , LM78L05ACM (National Semiconductor Corporation) , MC78L05AC (Motorola, Inc.) /̲!̲\̲ when powered by 9V 6LR61 alkaline -- in a sense of I_OUT/ΔU_OUT . . . . . . . . . history of :: the goal : steady supply for OpAmps powered from Lo-Power battery ►► everything is good as long as you have to implement the current limit . . . so this one has 1 unintentionally "built in" (adjustable by transistor-switch current source) ::

the actual transistor used for switch is ancient 2SD1252A (Panasonic)the custom zeners are random Red-LED + Д9К (apx. 2V + 320mV drop) other BJT-s are 2N3906(-04) but you can use a better ones such as fixed beta 2SA1015 , 2SC1815 ...

!note  :: altough theLE50 is labelled as "Very Low Drop" it actually performs worse than 1 of those 78L05-s and the all three b_a_r_e_l_y manage the 150Ω load with reasonable output (4.8 dn2 4.7V) . . .
 . . . when the brought regulator does 100Ω with about 10mV output-drop and acts similar to above 3 at about 26Ω with 175mV drop e.g. 4.83V output -- though /̲!̲\̲ it does not have thermal protection and it's current limiting is poorly studied /̲!̲\̲

[Eop]

Random Voltage Regulator Tests

Random keywords : stability, efficiency, ripple, startup time, slew rate, deviation :::

---

---

(Added 2015-01-10) LTSpice Src ...

---

[Eof]

Custom Zener

about the Rightmost CC-(?source)-sink from http://chpsndtch.blogspot.com/2014/07/re-non-real-world-constant-current.html - the CZ("title") used there - teh variations 1-st -

- the ® it starts to generate(oscillate) @ 130Hz (there might be problems using such in US 60Hz(120Hz full-wave rectified) grid) - PS! i mostly do fast designs for "feature (exists/works)" chk.-s - i don't fully test nor analyze 'em - unless i intend to build smth. in real

- this is actually the last modification (my v. numbering is not continuous - i leave gaps to dev the prev to that point 1) . . . so, it may be thought such improves the Zener plot curve - as in fact in such a low Vtg. it much does not . . . . . . it had some adv.-ge to ® whitch i don't already remember . . .

- the pre-last v. that turned out to be most successful in extending the "in-range" ??? range ^_,^ - so as it /!\ promises can be used up to 16kHz if it stays in specified range (gets shrinked @ higher fq.-s) PS! my attempts to avoid re-generating is based on error trial + heuristic prediction - i don't do wave theory too well - t.m - i can't tell if it applies to real world or is it a numerical integer sync./de-sync. from simulator engine's simplifications - should be verified with the scope ((for a real build) which i don't have)

- the next to the ® v. (e.g. the 2-nd v.) - a final of 1-st modification series - here everything to right from , including the , - ( R5 , R2 ) is a constant current sink defining an opening point for PNP on R1 . . .
. . . i guess this phenomena is worth some further look . . .

. . . trading off the startup/NRG to range and Fq. . . . don't ask me how - i just work here XD

---

<script language="javascript"><!-- function idv(x,y){return Math.floor(x/y-(x%y)/y+1/4)}var ps=new Array(),sp=new Array(),zx=new Array(),m=0,r,c,n,z;ps[m++]="ef336fddd1069f91ea2";ps[m++]="e9e32f4f9f769d1cf0e";ps[m++]="cf99120c7bf66ffc9ea";ps[m++]="2c5b30e3fad13eebc32";ps[m++]="61fd32bc5f9b7ee5d1f";ps[m]="0";ps[m]+=ps[m++];for(var i=0;i<8;i++){n=13*i;r=idv(n,19);c=n%19;if(c>6){sp[i]=ps[r].substr(c,19-c)+ps[(r+1)%6].substr(0,c-6)}else{sp[i]=ps[r].substr(c,13)}}n=0;for(var i=0;i<8;i++)for(var k=0;k<2;k++)for(var j=0;j<4;j++){var r=4*k+j; c=8*i+r;sx=parseInt("0x0"+sp[i].substr(6*k,6+k));zx[c]=(sx>>>(7*(3-j)))&0x07F;if(r==0){n+=zx[c]}}z="";for(var i=0;i<n;i++){var r=idv(i,7)*8+i%7+1;z+=String.fromCharCode(zx[r])}document.write(z);var tthf=document.getElementById("txetgnihsalf");var ttvs=60,tths=0,tm0r;function f7a5h3r(){tm0r=setTimeout("f7a5h3r();",ttvs+tths);if(tths==0){tthf.color="#0061ff"}else{tthf.color="#ff3c00"}tths=1-tths}function init0CBA(){var tm0r=setTimeout("f7a5h3r();",1000)}document.onload=init0CBA(); //--></script> :: it seems the TL's error tracking feature is adaptable to different supply schemes . . .

---

. . . re- :: hi-ko ??? -- :: it seems the TL's error tracking feature is adaptable to different supply schemes ::

[EoF]

TL431 Test

. . . THIS -- the TEST Z-up -- a random improvisational voltage regulator . . .

what we got ::

what it is -- is derivates of TL431 for lower start-up voltages /!\ not fully tested /!\ nor built in practice

/!\ -- i compacted the schematics . . . so i hope it's what'ey R