Saturday, February 22, 2014

200mW Boost Converter

what we got::
η=69% with the ® transistors (see next fig.) the η'd B around 80%
where we started from::
η=89%
CMOS supply dev. ::
η≤50%
"supply noise" ::
Low ON Resistance N-MOS alternate dev. ::
η=71% it can be greater if the useful load can be set much above the 1.6mA of the CMOS supply

[end of this post]

Thursday, February 13, 2014

Random

it appears that tomorrow is a "friend's day" since i'm from previous CCCP (where everyone was everyone's friend -- funny enough it mostly was so at the time) . . . i don't remember having a special party for - i don't even know who's friend's party it is - i only know it's something USA and 've spotted some greetings' symbolics - so the following may be interesting co incident or random occurrence of pseudo shapes due misc. frequencies ::


-.- this is soooo nuts XP . . . as it looks -- muhahaa

ha! a while ago i spotted an old function graph that goes with the logic here 'nd the function 4 ::


General

i just captured a fragment of the radio theatre (? google spell checker suggest there's an error ? my ISBN 0-86227-835-X p.419 however gives us an opposing view ??? what it should be a play/act house ??? go figure) aired locally - however this very ?error? goes well along with the fragment captured e.g "and the feature you don't even dare to say (being afraid of loosing it) . . . luck ..." (the L-word (not the F-word) -- italic makes it difficult to tell - so:) - made me think ? what are "they" on about ? isn't it a subjective light-time-temperature-blood-pressure.functionOf() . . . <fuzzy logic processing> resulted in a theorem hypothesis : "the lucky people are those who are stupid/inattentive enough to see no problems involved with their surroundings" . . . !Hei it's perfectly normal i'm mostly pissed by most everithing :

it's like i've frequently thought that "back at the old times" i was more happy about my life , everything i was doing - but i didn't have then "goes to err.-state at ..." AND "is not developed to support feature ..." lists populated to their today's volume

so you're either pissed and do something practical or add to a general Hurray Club - as the hope dies last the HC's role is not to be underestimated here

.
.
.
.
.
.
.
.
.

as a retired "dragon" - this is what i'm appyh about :

F;X the tasks don't show out too good here XD
-.- that's better

format cleanup:
ULL.X :: 220 = 2 x 2 x 5 x 11
ULL.Y :: 279 = 3 x 3 x 31
SA.X :: 708 = 2 x 2 x 3 x 59
SA.Y :: 1064 = 2 x 2 x 2 x 7 x 19 (= 266 ≈ 279)
Math. :: resize SA.Y so that the resulting XY ratio'd be equal to the one of the ® AND Δ(ULL.Y,SA.Y)'d be as small as possible ... looks we're already there - since the either of XY-s've no lesser common multiplier to fine adjust . . . (the Japanese are smaller so it kind of goes ;X)

def.-s :: retired dragon :: . . . the dragon is basically a spirit , a sub descendant part of a time(-ship) or casual sequence (a lie) that is defined to be such - as a creator spirit you create these (i might not remember it too well ...) by going into meditation (not here -- though you may start in this world) of some ultimate feature . . . it usually involves (the spirits know everything or at least are capable of) minimizing your databases such as cutting off certain relations (well to you if/as a spirit) so you create/meditate you.notYou() that sees less or perhaps more from derived/defined new base-set (basically it's like a voice/thought commanded computer) you are in this new spirit bubble (a new instance of "time" insulated from the higger 1-s - has less "dimensions"/deg.-of freedom - more casuality) study what you there're what it feels what you see - tune the compile (voice commanded 'puter) ? ? ? then you prog your way out ??? leaving inside the new featured you

at that point it must be realized that when a spirit you're always a bunch of "people" as in here world a set of internet ID-s only in more ID-wise continuous/integral form -- your www ID-s are not superpositionable in world(s) the spoirit ID-s are you can recombine them to a new ID-s and time-shift release into same world and same time . . . and the are more self-sustaining-continuous/integral . . . down here @ 3/4D+t have birth-life-death ...

if they are of the same dragon=spirit.subOf() then there's like lesser difference of me-s in lesser casuality time-scapes - it's like changing wear - the more you believe of being of a certain coat the more it starts to realize - as for here analogy the more you use specific ID the more you believe being that=internet.ID

well the more you modify/create it's "history" the more of it'll be "you"

now as a spirit if you out of the sudden recall being of something else previously - you may shift back to your previous you -- the difference compared to here is that as your all "dragons" are set to "life" on their primary appearance (in)to time then if you leave your certain ID that ID still continues it's life down to times it currently "populates" (someone else - the me.notMe part of your higher spirit - defines it's "history" forward) -- e.g. if you didn't mess up too bad keeps the "thread" "alive" and going

there might be occurrences where it's better for your "active.you" to populate lower realms only - or your energies just don't loop time-wise-continuous through uppers -- so here it's that retired dragon condition , blah ...

life sucks - muhahaa

here's how ::
Miss Miazaki's cat

japan cats (it should be a small town north of Tokyo -- this cat is from -- (in town) there's a weighing station or smith's-shop -- the spirit's/mind *view from N(N)E -- not enough data to extend* . . . F;i . . . might've do smth. with rust and nuts -- "spirit timers" (?failed in "this" occasion) ← too much guesswork already - the img. found from webshots B4 2010CE -- 1-st "decoded" then )

booty from web's Hots ;>- yummy

Wednesday, February 12, 2014

Re: the idea crisis

well - no crude snacks - a white coffee that's actually peppermint ginger cinnamon cardamom orange(and lemon) peels' flour (or oil) ... aah yes and coffee (sometimes i forgot and then reminds from irregularly transparent outcome XD) + multiviamin + B-complex + C - letting the body to figure out what it lacks

so anyway i was already able to decide that my КР538УН3А model is not exactly what it suppose to be

"they" claim it to be replacable with LM387N ?? let's see

ParameterLM387(A)КР538УН3А
ISUPPLY(mA)105
USUPPLY(V)30(40)7.5
Fq.Responce
Schematic

if i remember right then this particular chip is developed in ЭССР RET - yes a "mysterious" place nowadays inexistent - why i have a ® factory datasheet in hand - muhahaa - made a quick electro dynamic mic preamp - and it really is a low noise 1 - have couple of these chips left so . . .

the idea crisis

as i mostly run at studying the nodes of relations induced at the moment impulse (the www appears a good tool for such ... not so good resource (the most things are times faster found from a printed encyclopedia (if i remember such exists)))
- the settled mind is a big problem! (50% of the work is usualy done by when you actually decide so and start doing it!!!)

i just found myself pondering few minutes ago what's best to get engaged next with - well - ? some half on it's way Spice experiment ? some computer programming project ? an old 'puter game (moving your fingers trains your nervous impulse logistics) ? eating a curd snack <- is very exciting Xd XD XD XD (the list preceding it was ®-ly a BIT longer)

desperate times desperate measures - i guess i have to try it out now (what ideas they buried inside) (chewing improves the brain circulation + oxygen) then i have to re-scan the BIT longer list - muhahaa

note: eating makes you feel secure=lazy + energy goes to melting - lowering the room temperature keeps you active to keep warm - so if you don't over-eat and let some fresh air in you're almost in ideal conditions of doing some work

 . . . also eating a good = multi component warm meal is best anti sleep aid (coffe works 20 minutes then it'd drained your resources coffee+chocolate can be reboosted by lemon juice after that 20 minutes gives you a total hour or so - chocolate alone just distracts your attention (although the sweets're considered a brain food - well it reduces the stress level a bit) - while a warm meal guarantees next 2+h effective ops. - there's like decades experience i can tell it from)

chng chng - crude snack XD

Saturday, February 8, 2014

1.2V Discrete Logic Lab

Prj. thread history::


at this point i am likely able to design better of any type of  these xTL stages however there are some things that came out in prev. studies::
1) the simple TTL NAND (up right) supports higer frequencies but only up to 3-4 input gates
2) the CMOS type gates (lowwer right) consist of "too many" components
3) the best power_draw : component_count : input_count : MHz_per_unit_Power is top left

Stats:


what is being marked here is exceptional to studied set designs - lets take a closer look:


is the highest frequency v. of the exceptions a TTL stage based on 2N5087 (good sensitivity at low power) as the exceptional performance is due to the performance of that transistor drawing in average 130µA/stg.


the 2-nd fast exception being tested for XOR3 against XOR3=fnOf(XOR2) , it's NAND gate:


again we have the 2N5087 - used inplace of russian МП42Б germanium 1 - based design drawing in average 9µA/stg.


and the slowest exception drawing in average 1.87µA/stg.


the 2N5087 based

the primary objective of this research was to find optimum way to utilize some old transistors as misc. LE 1.2V logic gates

we made some findings on this subj. but it still remains far from being researched

Friday, February 7, 2014

Phase Shift Std.

Unfortunately i can't tell if this feature is applicable in real since i don't have scope nor fq. counter to confirm signals over 48kHz (96ksps sound card) KYORITSU KM 2K (no good for small signal or i'm stupid to Z it up right for)

anyway i'm always interested in various oscillators so at the time i played around with LC loop back 1-s and found the next to be surprisingly effective


What i call a "Phase Shift Std." is the (L-C network) part (your) right of the amplifier stage, the "S" "o" networks at left are also important to fit the supply with desired frequency range (tough -- i have some lesser experience with that part)


as it can be seen it's not fully comprehended nor studied - i always do "heuristic" research or solving rather than analytical - you do calculus , apply result and . . . nothing happens - a complete waste of time if your concept doesn't already "tick" - this resonant network supports even very bad amplifier designs (such that have no power to push the ring oscillator formed of 'em) - the left hand side "S" of "So" may be discarded if the circuit is not to be run on low power or near it's components' transition fq.-s.

here's the ring oscillator v.


i mostly use such "Phase Shift Std." to evaluate the amplifier stage - ? the higgher the fq. it can be pushed at the better the concept for the stage



also related but not strictly under the scope of the title ::

here's a more simple way to link high fq. ring oscillator



such is just a test finds not any theory based - is just what seemed to start up best. At the time i was interested of how fast the transistors might be pushed to go

here's the above it's transition point the 2N3904 circuit



lately revisited 1.2V Quartz Oscillator circuit


the particular X-Tal Rs = 35 kΩ C0 = 2.5pF C1 = 4 fF Ls = 5907.105371 H Q = 3.47E+04 is tricky to get going with any std. circuit at that voltage - i haven't managed to get it going sharp at parallel resonance ... yet - so there's an idea to use whatever the fq. it's willing to go and PLL it up then correct the higher 1 by extracting adding pulses * at first attempts i likely attempted run (deviate) it with too much energy (so it didn't start up)

Monday, February 3, 2014

Random Simulator Stuff

• characteristic plots (when precision planning a design such might become handy)


• checking the spice credibility (notice the 15n(s) minimum timestep)


side note: i had my µ$XP-crash @ that point (memory allocation problems ? buy a new problem (ups!) Windows ? - some people never take their job seriously - some things never change) -- so here we can generalize :: Everything "WORKS" as long as you start working (if you want to avoid that phenomena - build your own computer and write your own software - and this is not a joke ... #!!/!//(#"!!"))//"&!#(/%"!~ .... it'll take 2-y to build +3y to test-tune ?? perhaps it's worth it -- someone can't constantly spoil your day)


it's interesting that FF and IE require 100% of CPU to display the above script while Opera gets it done with 74% of CPU . . . however the offline display uses some 3-rd of that (there is no data updates from server required to display this -- the question is :: what demands 3x more =300% resource for online view of basically static web page ??? humane stupidity)


i ran some statistics :: it apares that the CPU Load v. Power Consumption has near linear dependency
namely P(W) = 117.5 Q(%) +187.1 such as
for 100%CPU we get 117.5×1 + 187.1 = 304.6 W
and for 74%CPU it's 117.5×0.74 + 187.1 = 274.05 W
 the ΔP = P100% - P74% = 304.6 - 274.05 = 30.55 W
13202 Mtoe 153539.26 TWh -- the year 2013 world's total annual energy production
(20407 TWh) 52325.64103 TWh -- (energy distributed as electricity) the amount of energy to produce it
34% -- the electricity part of annual energy budget
39% -- the efficiency of thermal power plant
39% -- part of world population that use internet
7100000000 -- number of Terraneans
8000000000 -- check number = 2000³
50% -- statistical nonsense -- speculating the even distribution IE=+FF : Chrome=+Opera of users
20% -- statistical simplification !! -- internet users that read blogs (there are 5 mind types)
30.54720758 -- ΔP Watts of Power
0.367879441 -- voluntary correction coefficient for next = Exp(-1) already included
6% -- statistical simplification !! -- average part of a day for a blog reader to chk the updates
39%6%50%20% -- target group · part of a day = q
10 -- q·ΔP as part per billion units of energy used for annual electric/distributed power production
:: in other words it's not a big deal that some browsers set the PC to use more power than others with respect to world's energy budget

however ::
1 Ah -- speculative capacity of the cell phones battery
0.1 A -- speculative charge current = 100mA
80% -- speculative AC to DC to Li-ion conversion efficiency
5.25 Wh -- apx. Max. amount of energy required to load your phones battery
45 Wh -- how much energy the "blog reader" wastes/saves per day
8.562 -- how many times the cell phone can be loaded with that energy

.

.

. . . i guess it apx. has a post's length
[]

Sunday, February 2, 2014

Re - Custom Simple Linear Regulator

(extends:: Custom Simple Linear Regulator)
i may be an idiot but most voltage regulator circuits i've come across in internet appear to be too unstable or of not enough range for design target when simulated !! (usually the simulation appears to be the way batter quality than the real life build) so not enough range in simulation - is this for any use

so a bit of my research history on this subj. ::

first i thought that for successful design of a voltage regulator is required a good voltage reference ...

it may be that way but the precision voltage reference alone does not define a thing - more important is what you do with it or other words how you utilize it with respect of the rest of the circuit - so it turns out you can "define" your voltage without any specific voltage references - in the following test ::


the R4 is required for the simulation to run (for the circuit to be defined in the Solve Electric) -- i guess i added the diode for a zener - anyway this modified constant current source turns out to be very precise comparator for external voltage ::


wow -- a bad example -- anyway 1 of my early struggles -- the "comparator" is formed here by Q5 Q6 (Q9) if the TVP (test voltage point) increases the Q6 and Q9 start to open cutting off the Q11 ((?? all this for voltage reference ?? for precision constant current by J1 J-fet ... wow XD)) . . . so the Q1 has ??? high precision voltage on it's base and if PF gets too high the emitter-base current of Q1 reduces reducing the darlington's - Q2 Q10 - current - which in turn maintains PF in required limits - the circuit was never built as i found it too "expensive" or time consuming (along with tuning the real thing)

here's the latest zener ::


and what it's good for ::


???? -- running a scratch supply concept and testing various features -- everything is cool ? ? ? but to get only 20mA usable power (and who gonna build that . . . i'm out XD)

...

an alternative to a grid shown in ® Custom Simple Linear Regulator ::
the minimalist concept::






and minimal cfg for practical usage (the specs on the lower simulation do not apply to this grid - look the plots) -- doing the over current limiting here was a bit tricky


[EOF]

30mW Boost Converter

this is just another of the many v. of this circuit
however here we matched the reality to simulation and
then fine tuned the simulation to match the actual circuit


the not shown series resistance for the inductors (114 µH) is apx.  2point16 mΩ

the above circuit showed some 84% conversion efficiency at lesser loads with the ® 2N3904 (not the КТ315Г it's been "converted")

* the bigger the wire diameter the grater the power that can be converted per unit time
* the more turns and (the lesser the wire diameter (the grater the series resistance)) -
 ˙ the lesser the start up voltage but also the lower the power per unit time

so if you need energy efficient microbe hour :p converter use the lower options (powered by cell button batteries)

although you may make a way (!!) more complex circuit with higher options - that runs occasionally from (!! then required) high energy source or from low energy source loaded capacity (!! requires specific "controller") to maintain the overhead for the application power regulator (!! another addition)

the capacitor in parallel to LR6 should be lesser for greater operation frequencies such as the following 40MHz circuit (not fully explored nor built)

it may become unstable from unpredicted modifications nor can i say if it has anything to do with reality

[EOF]