P\u0159i postupn\u00e9m o\u017eivov\u00e1n\u00ed a testech jednotliv\u00fdch \u010d\u00e1st\u00ed, jsme narazili na probl\u00e9m se stabilitou lad\u00edc\u00edho nap\u011bt\u00ed OCXO. F\u00e1zov\u00fd z\u00e1v\u011bs funguje spr\u00e1vn\u011b, po uzam\u010den\u00ed PLL nap\u011bt\u00ed osciluje kolem st\u0159edn\u00ed hodnoty. To je sice podstata funkce takov\u00e9to stabilizace, ale my se sna\u017e\u00edme co nejv\u00edce tuto chybu omezit a doc\u00edlit co nejmen\u0161\u00ed zm\u011bny nap\u011bt\u00ed. V\u00fdstup XOR hradla IC3B toti\u017e d\u00e1v\u00e1 jen dv\u011b velikosti nap\u011bt\u00ed 0V nebo 5V. Kondenz\u00e1tor v integra\u010dn\u00edm \u010dl\u00e1nku C24 se tedy bu\u010f nab\u00edj\u00ed nebo vyb\u00edj\u00ed. Nen\u00ed tam nikdy stav, kdy by se nap\u011bt\u00ed nem\u011bnilo.<\/p>\n\n\n\n \u010casov\u00e1 konstanta integra\u010dn\u00edho \u010dl\u00e1nku R18+C24 mus\u00ed b\u00fdt zvolena vhodn\u011b dlouh\u00e1. Laboroval jsem s hodnotami mezi 50ms a 40s. Nakonec se jako nejlep\u0161\u00ed jevily nejdel\u0161\u00ed \u010dasy. Nap\u011b\u0165ov\u00e1 chyba v mV zhruba odpov\u00edd\u00e1 kmito\u010dtov\u00e9 chyb\u011b v mHz.<\/em><\/p>\n\n\n\n \u0394f \u2248 \u0394U [mHz,mV]<\/p>\n\n\n\n Po zah\u0159\u00e1t\u00ed a stabilizaci parametr\u016f, nap\u011bt\u00ed osciluje kolem st\u0159edn\u00ed hodnoty s chybou a\u017e \u00b13mV<\/strong>. Ka\u017ed\u00fd OCXO m\u00e1 trochu jin\u00e9 nap\u011bt\u00ed, p\u0159i kter\u00e9m se „uzamkne PLL“. V m\u00e9m p\u0159\u00edpad\u011b se je jedn\u00e1 o dva exempl\u00e1\u0159e OCXO Trimble 34310-T a Trimble 34310-T2, <\/em>to je 4,45V a 7,75V. To si vysv\u011btluji jak jinou verz\u00ed, tak odli\u0161n\u00fdm st\u00e1rnut\u00edm ka\u017ed\u00e9ho oscil\u00e1toru. V konstrukci je na tento probl\u00e9m pamatov\u00e1no. Sta\u010d\u00ed zm\u011bnou hodnot rezistor\u016f R5+R11 p\u0159enastavit hodnotu zes\u00edlen\u00ed opera\u010dn\u00edho zesilova\u010de U3.<\/p>\n\n\n\n Taktovac\u00ed „clock“ sign\u00e1l z GPS p\u0159ij\u00edma\u010de m\u00e1 skute\u010dn\u011b velmi p\u0159esn\u00fd kmito\u010det. Jeho stabilita je a\u017e 3\u00d710-13 <\/sup>Hz (p\u0159esnost Cesiov\u00fdch hodin). Probl\u00e9m je v\u0161ak v tom, \u017ee je tento sign\u00e1l zat\u00ed\u017een f\u00e1zov\u00fdm \u0161umem. To znamen\u00e1, \u017ee se f\u00e1ze hodinov\u00e9ho sign\u00e1lu \u201cn\u00e1hodn\u011b\u201d posouv\u00e1, ne o mnoho, p\u0159esto ano. Osciluje kolem \u201cnulov\u00e9ho bodu \u2013 uzlu\u201d v intervalu asi 22ns<\/strong>. Tato chyba se v telekomunikac\u00edch naz\u00fdv\u00e1 „Jitter<\/strong>„. P\u0159i d\u00e9lce periody hodinov\u00e9ho sign\u00e1lu 1MHz (T=1\u03bcs) je to vlastn\u011b 2,2% chyba<\/strong>. F\u00e1ze tedy \u201c\u0161um\u00ed\u201d kolem \u201cnuly\u201d. Naproti tomu krystalov\u00fd oscil\u00e1tor, kter\u00fd je v OCXO, m\u00e1 velmi mal\u00fd f\u00e1zov\u00fd \u0161um. Jeho kmito\u010det v\u0161ak nebude tak p\u0159esn\u00fd jako je sign\u00e1l z GPS.<\/p>\n\n\n\n F\u00e1zov\u00fd \u0161um je ne\u017e\u00e1douc\u00ed vlastnost v\u0161ech harmonick\u00fdch oscil\u00e1tor\u016f, kter\u00e1 je zp\u016fsobena hlavn\u011b kr\u00e1tkodobou frekven\u010dn\u00ed nestabilitou. Ka\u017ed\u00fd oscil\u00e1tor je frekven\u010dn\u011b nestabiln\u00ed, tato nestabilita se pak ur\u010duje v p\u0159\u00edslu\u0161n\u00e9m \u010dasov\u00e9m intervalu. Dlouhodob\u00e1 nestabilita se ur\u010duje v r\u00e1mci hodin, dn\u00ed nebo let. Kr\u00e1tkodob\u00e1 nestabilita je definov\u00e1na na intervalu od zlomk\u016f sekund a\u017e po minuty. P\u0159\u00ed\u010dinou kr\u00e1tkodob\u00e9 nestability jsou r\u016fzn\u00e9 druhy \u0161um\u016f a jejich vz\u00e1jemn\u00e9 kombinace. Jedn\u00e1 se nap\u0159\u00edklad o term\u00e1ln\u00ed \u0161um, v\u00fdst\u0159elov\u00fd \u0161um, blikav\u00fd \u0161um<\/em>, apod. Na oscila\u010dn\u00ed frekvenci se tato nestabilita projevuje jako n\u00e1hodn\u00e1 f\u00e1zov\u00e1 modulace s mal\u00fdm zdvihem. Frekvence, na kter\u00e9 se f\u00e1zov\u00fd \u0161um m\u011b\u0159\u00ed, je ozna\u010dov\u00e1na jako offsetov\u00e1 frekvence. Ob\u00e1lka f\u00e1zov\u00e9ho \u0161umu je osov\u011b symetrick\u00e1 s osou na nosn\u00e9 frekvenci.<\/p>\n\n\n\n V tomto zapojen\u00ed tedy kombinujeme vynikaj\u00edc\u00ed dlouhodobou stabilitu GPS sign\u00e1lu s vynikaj\u00edc\u00ed kr\u00e1tkodobou stabilitou OCXO oscil\u00e1toru.<\/em><\/p>\n\n\n\n Jako modifikaci jsem sn\u00ed\u017eil pracovn\u00ed kmito\u010det PLL smy\u010dky z 1MHz na 100kHz. P\u0159i kmito\u010dtu 1MHz doch\u00e1zelo k z\u00e1kmit\u016fm ve stabilizaci a \u0159\u00edd\u00edc\u00ed nap\u011bt\u00ed VCO m\u011blo chybu i \u00b1<\/strong>25mV. P\u0159i del\u0161\u00edch pulzech se tak\u00e9 kondenz\u00e1tor C24 st\u00edh\u00e1 nab\u00edjet na vy\u0161\u0161\u00ed nap\u011bt\u00ed a m\u016f\u017eeme pou\u017eit je\u0161t\u011b del\u0161\u00ed \u010dasy RC \u010dl\u00e1nku. Velikost f\u00e1zov\u00e9ho \u0161umu n\u00e1m tedy degraduje stabilitu na\u0161eho GPSDO, proto jej chceme d\u00e1le redukovat pou\u017eit\u00edm vhodn\u00e9ho GPS p\u0159ij\u00edma\u010de. V sou\u010dasn\u00e9 dob\u011b pou\u017e\u00edv\u00e1me GPS p\u0159ij\u00edma\u010d NEO-7M, kter\u00fd je relativn\u011b levn\u00fd, m\u00e1 v\u0161ak vy\u0161\u0161\u00ed f\u00e1zovou chybu.<\/p>\n\n\n\n Na n\u00e1sleduj\u00edc\u00edch obr\u00e1zc\u00edch m\u016f\u017eeme vid\u011bt velikost f\u00e1zov\u00e9 chyby. K synchronizaci \/ triggeringu byl pou\u017eit sign\u00e1l z OCXO (modr\u00fd), \u017elut\u011b je zn\u00e1zorn\u011bn clock sign\u00e1l z GPS. \nExistuj\u00ed tzv. \u201c\u010dasovac\u00ed\u201d modely GPS p\u0159ij\u00edma\u010d\u016f s ozna\u010den\u00edm \u201cT\u201d, kter\u00e9 maj\u00ed f\u00e1zov\u00fd \u0161um men\u0161\u00ed. Jinou mo\u017enost\u00ed je p\u0159epnout GPS p\u0159ij\u00edma\u010d do „survey in“, kdy je pozice zn\u00e1m\u00e1 a chceme pouze \u010dasovou synchronizaci. V tomto re\u017eimu sta\u010d\u00ed pro p\u0159\u00edjem pouze jedna dru\u017eice.\n\n<\/p>\n\n\n\n LEA-NEO-6T_ProductSummary_(UBX-13003351)<\/a><\/p>\n\n\n\n Dal\u0161\u00edm n\u00e1padem, jak z\u00edskat stabiln\u011bj\u0161\u00ed referen\u010dn\u00ed kmito\u010det, je generov\u00e1n\u00ed lad\u00edc\u00edho nap\u011bt\u00ed OCXO p\u0159\u00edmo DAC procesoru AtXmega128. Ten disponuje 12bit DAC. Kdy\u017e nap\u011bt\u00ed vhodn\u011b zes\u00edl\u00edme, dostaneme lad\u00edc\u00ed krok cca 0,61mV.<\/p>\n\n\n\n Idea je tato: Pomoc\u00ed INT vstupu GPS p\u0159ij\u00edma\u010de (vstupu p\u0159eru\u0161en\u00ed) dostaneme p\u0159esnou \u010dasovou zna\u010dku ud\u00e1losti. Toto budeme spou\u0161t\u011bt, dejme tomu ka\u017edou celou sekundu. Z rozd\u00edlu mezi zna\u010dkami dopo\u010d\u00edt\u00e1me chybu. Del\u0161\u00edmi m\u011b\u0159en\u00edmi m\u016f\u017eeme chybu kvantifikovat a m\u011bnit podle n\u00ed lad\u00edc\u00ed nap\u011bt\u00ed OCXO.<\/em><\/p>\n\n\n\n P\u0159i m\u011b\u0159en\u00ed kmito\u010dtov\u00e9 stability je t\u0159eba m\u00edt vhodn\u00fd kmito\u010dtov\u00fd norm\u00e1l pro srovn\u00e1n\u00ed, minim\u00e1ln\u011b o jeden \u0159\u00e1d p\u0159esn\u011bj\u0161\u00ed ne\u017e je m\u011b\u0159en\u00fd sign\u00e1l.<\/p>\n\n\n\n Pro n\u00e1zornost srovn\u00e1n\u00ed jednotliv\u00fdch standard\u016f v tabulce:<\/p>\n\n\n\n Pot\u0159ebn\u00e1 m\u00edra stability je v\u017edy sv\u00e1zan\u00e1 s danou aplikac\u00ed. Pro pr\u00e1ci na kr\u00e1tk\u00fdch vln\u00e1ch n\u00e1m sta\u010d\u00ed oby\u010dejn\u00fd krystalov\u00fd oscil\u00e1tor (XO). Na VKV pak teplotn\u011b kompenzovan\u00fd (TCXO). Na mikrovln\u00e1ch bude t\u0159eba OCXO nebo GPSDO, pro spr\u00e1vnou funkci s\u00edt\u011b GSM + 4G pot\u0159ebujeme Rubidiov\u00fd standard. Abychom mohli ode\u010d\u00edtat p\u0159esnost polohy z GPS na 2m, pot\u0159ebujeme Cesiov\u00e9 hodiny.<\/p>\n\n\n\n M\u00edra stability \/ koncentrace se velmi \u010dasto uv\u00e1d\u00ed pomoc\u00ed pod\u00edlu nap\u0159. milionu. 1PPM je tedy odchylka 1\/1 000 000, 1PPB pak 1\/1 000 000 000 atd. viz tabulka.<\/p>\n\n\n\n \nJako etalon kmito\u010dtu jsem pou\u017eil Rubidiov\u00fd kmito\u010dtov\u00fd norm\u00e1l „Efratom FRK-LLN-1A1B1A“. Pomoc\u00ed m\u011b\u0159i\u010de odchylky Adret 4110<\/strong> (d\u00edky Michalovi OK3ZM), jsem mohl m\u011b\u0159it odchylku t\u011bchto dvou kmito\u010dt\u016f. M\u011b\u0159i\u010d odchylky m\u00e1 dva vstupy. Na jeden se p\u0159ivede norm\u00e1lov\u00fd sign\u00e1l<\/em> (Rb) a na druh\u00fd m\u011b\u0159en\u00fd sign\u00e1l<\/em> v na\u0161em p\u0159\u00edpad\u011b GPSDO.\n\n<\/p>\n\n\n\n Rubidiov\u00fd standard m\u00e1 stabilitu \u03c3 pro \u010das 1dne cca 1\u00d710-12 <\/sup>Hz.<\/em><\/p>\n\n\n\n \nSamotn\u00e9 m\u011b\u0159en\u00ed bylo automatizov\u00e1no. Nam\u011b\u0159en\u00e9 odnoty z odchylkom\u011bru Adret 4110 se vy\u010d\u00edtaly pomoc\u00ed s\u00e9riov\u00e9 linky do po\u010d\u00edta\u010de, kde se ka\u017edou sekundu ukl\u00e1daly v TXT souboru (op\u011bt d\u00edky OK3ZM). Na takto z\u00edskan\u00fdch datech pak byla provedena matematick\u00e1 anal\u00fdza. Zde jsem pou\u017eil program Stable32<\/a>. Autorem tohoto SW je pan William Riley, kter\u00fd jej poskytl bezplatn\u011b komunit\u011b zab\u00fdvajic\u00ed se \u010dasov\u00fdmi odchylkami. Program je p\u0159\u00edmo ur\u010den pro anal\u00fdzu frekven\u010dn\u00ed stability oscil\u00e1tor\u016f a v\u00fdpo\u010det jejich chyb (Allanova variance, kmito\u010dtov\u00fd drift, stabilita, \u0161um apod.)\n\n<\/p>\n\n\n\n \nPrvotn\u00ed nah\u0159\u00edv\u00e1n\u00ed OCXO trv\u00e1 p\u0159ibli\u017en\u011b 20min, pr\u016fb\u011bh je vid\u011bt na tomto grafu:\n\n<\/p>\n\n\n\n Na n\u00e1sleduj\u00edc\u00edm obr\u00e1zku vid\u00edme ob\u00e1lku spektra f\u00e1zov\u00e9ho \u0161umu, zat\u00edm zm\u011b\u0159eno pouze spektr\u00e1ln\u00edm analyz\u00e1torem. Tato metoda m\u00e1 \u0159adu omezen\u00ed, nev\u00fdznam\u011bj\u0161\u00ed je skute\u010dnost, \u017ee m\u011b\u0159\u00edme slo\u017eenou funkci f\u00e1zov\u00e9ho a amplitudov\u00e9ho \u0161umu. D\u00e1le jsme limitov\u00e1ni vlastn\u00edm \u0161umem LO lok\u00e1ln\u00edho oscil\u00e1toru spektr\u00e1ln\u00edho analyz\u00e1toru.<\/em><\/p>\n\n\n\n Zm\u011b\u0159en\u00e1 hodnota f\u00e1zov\u00e9ho \u0161umu je ve vzd\u00e1lenosti 2Hz od nosn\u00e9 cca -70dBc\/Hz<\/strong>.<\/p>\n\n\n\n Zm\u011b\u0159en\u00e9 odchylky od kmito\u010dtov\u00e9ho norm\u00e1lu 10MHz jsou vid\u011bt na n\u00e1sleduj\u00edc\u00edm grafu:<\/p>\n\n\n\n Pro anal\u00fdzu stability byla pou\u017eita p\u0159ekr\u00fdvaj\u00edc\u00ed se Allanova variance – to je Na n\u00e1sleduj\u00edc\u00edm obr\u00e1zku vid\u00edme hodnoty stability GPSDO pro jednotliv\u00e9 \u010dasy Tau, PLL smy\u010dka b\u011b\u017eela na 100kHz.<\/p>\n\n\n\n \nSamotn\u00fd krystalov\u00fd oscil\u00e1tor OCXO vykazuje po nah\u0159\u00e1t\u00ed takovouto stabilitu: Na grafech vid\u00edme, \u017ee GPSDO m\u00e1 na \u010dasech <100s stabilitu \u2248 3\u00d710-10 <\/sup>, zde se projevuje n\u00e1\u0161 probl\u00e9m se stabilitou lad\u00edc\u00edho nap\u011bt\u00ed (chyba \u22483mV)<\/em>. To v\u0161ak pro m\u011b\u0159\u00edc\u00ed rozli\u0161en\u00ed na\u0161eho \u010d\u00edta\u010de FC-1300 s rezervou posta\u010duje. Krystalov\u00fd oscil\u00e1tor m\u00e1 naopak v \u010dasech 1\u223c100s stabilitu \u22483\u00d710-12 <\/sup>Hz. <\/strong> Na del\u0161\u00edch \u010dasech se ji\u017e nep\u0159\u00edzniv\u011b projevuje kmito\u010dtov\u00fd drift (kter\u00fd je v GPSDO kompenzov\u00e1n).<\/p>\n\n\n\n Jak jsem ji\u017e uvedl, stabilita na\u0161eho GPSDO norm\u00e1lu pln\u011b posta\u010duje pro zam\u00fd\u0161len\u00e9 vyu\u017eit\u00ed v na\u0161em \u010d\u00edta\u010di. Pokud bychom v\u0161ak pot\u0159ebovali kmito\u010dtov\u00fd norm\u00e1l p\u0159esn\u011bj\u0161\u00ed, konstrukci bych modifikoval a lad\u00edc\u00ed nap\u011bt\u00ed VCO generoval pomoc\u00ed cca 16bit DAC p\u0159evodn\u00edku.<\/p>\n\n\n\n Co nemohu opomenout zm\u00ednit je, \u017ee p\u0159i m\u011b\u0159en\u00ed jsme se ji\u017e dot\u00fdkali hranic p\u0159esnosti Rb oscil\u00e1toru coby etalonu kmito\u010dtu. Na m\u011b\u0159en\u00edch byla poznat i nepatrn\u00e1 zm\u011bna teploty v m\u00edstnosti.<\/p>\n\n\n\n Tabulka s hodnotami \u03c3 pro p\u0159\u00edslu\u0161n\u00e9 \u010dasy Tau:<\/strong><\/em><\/p>\n\n\n\n \nPokra\u010dov\u00e1n\u00ed …\n\n<\/p>\n\n\n\n <\/p>\n","protected":false},"excerpt":{"rendered":" Postupn\u00e9 o\u017eivov\u00e1n\u00ed a testy: V tomto \u010dl\u00e1nku se budu sna\u017eit popsat dal\u0161\u00ed pr\u016fb\u011bh stavby na\u0161eho \u010d\u00edta\u010de FC-1300 a pomoc\u00ed \u0159ady m\u011b\u0159en\u00ed ov\u011b\u0159\u00edm, zda na\u0161e \u0159e\u0161en\u00ed kmito\u010dtov\u00e9ho standardu GPSDO dosta\u010duje sv\u00fdmi parametry pro na\u0161e aplikace. P\u0159i postupn\u00e9m o\u017eivov\u00e1n\u00ed a testech jednotliv\u00fdch Pokra\u010duj ve \u010dten\u00ed \u2192<\/a><\/p>\n","protected":false},"author":1,"featured_media":2914,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[253,3,123],"tags":[247,259,524,255,257,241,463,243,464,260,239],"class_list":["post-2055","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-citac-s-gps-normalem","category-elektronika","category-merici-pristroje","tag-10mhz-gpsdo","tag-adret-4110","tag-allan-variance","tag-gps-jitter","tag-gps-phase-noise","tag-gpsdo","tag-gpsdo-vs-rubidium","tag-mc100ep016","tag-rubidium-standard","tag-time-standard","tag-trimble-34310"],"_links":{"self":[{"href":"https:\/\/ok2haz.ok2kld.cz\/ok2haz\/wp-json\/wp\/v2\/posts\/2055","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/ok2haz.ok2kld.cz\/ok2haz\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/ok2haz.ok2kld.cz\/ok2haz\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/ok2haz.ok2kld.cz\/ok2haz\/wp-json\/wp\/v2\/users\/1"}],"replies":[{"embeddable":true,"href":"https:\/\/ok2haz.ok2kld.cz\/ok2haz\/wp-json\/wp\/v2\/comments?post=2055"}],"version-history":[{"count":309,"href":"https:\/\/ok2haz.ok2kld.cz\/ok2haz\/wp-json\/wp\/v2\/posts\/2055\/revisions"}],"predecessor-version":[{"id":3883,"href":"https:\/\/ok2haz.ok2kld.cz\/ok2haz\/wp-json\/wp\/v2\/posts\/2055\/revisions\/3883"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/ok2haz.ok2kld.cz\/ok2haz\/wp-json\/wp\/v2\/media\/2914"}],"wp:attachment":[{"href":"https:\/\/ok2haz.ok2kld.cz\/ok2haz\/wp-json\/wp\/v2\/media?parent=2055"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/ok2haz.ok2kld.cz\/ok2haz\/wp-json\/wp\/v2\/categories?post=2055"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/ok2haz.ok2kld.cz\/ok2haz\/wp-json\/wp\/v2\/tags?post=2055"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}![\"\"](\"https:\/\/ok2haz.ok2kld.cz\/ok2haz\/wp-content\/uploads\/2018\/12\/citac_sch.png\")
<\/a><\/figure>\n\n\n\n\u00dapravy integra\u010dn\u00edho \u010dl\u00e1nku: <\/h2>\n\n\n\n
Tedy hodnoty odpov\u00eddaj\u00edc\u00ed R18=100k, C24=100\u03bcF.
<\/em>Kmito\u010dtovou stabilitu jsem pot\u00e9 ode\u010d\u00edtal z nap\u011b\u0165ov\u00e9 zm\u011bny na v\u00fdstupu opera\u010dn\u00edho zesilova\u010de U3.<\/p>\n\n\n\n
Pokud byl pou\u017eit kmito\u010det men\u0161\u00ed, tedy 100kHz<\/strong> – relativn\u00ed f\u00e1zov\u00e1 chyba klesla na 0,22%,<\/strong> tak\u00e9 rozptyl lad\u00edc\u00edho nap\u011bt\u00ed.<\/p>\n\n\n\nMo\u017en\u00e1 dal\u0161\u00ed vylep\u0161en\u00ed: <\/h2>\n\n\n\n
Na t\u011bchto obr\u00e1zc\u00edch vid\u00edme velikost f\u00e1zov\u00e9 chyby asi 22ns<\/strong>.<\/p>\n\n\n\n<\/a><\/figure><\/li><\/a><\/figure><\/li><\/ul><\/figure>\n\n\n\nVelikosti a nejv\u011bt\u0161\u00ed zdroje f\u00e1zov\u00fdch \/ \u010dasov\u00fdch chyb: <\/h2>\n\n\n\n
<\/figure>\n\n\n\nM\u011b\u0159en\u00ed stability s Rb oscil\u00e1torem: <\/h2>\n\n\n\n
<\/a><\/figure>\n\n\n\n<\/figure>\n\n\n\n<\/a><\/figure>\n\n\n\n<\/a><\/figure><\/li><\/a><\/figure><\/li><\/ul><\/figure>\n\n\n\nDosa\u017een\u00e9 v\u00fdsledky m\u011b\u0159en\u00ed: <\/h2>\n\n\n\n
<\/a><\/figure>\n\n\n\n<\/a><\/figure>\n\n\n\n<\/a><\/figure>\n\n\n\n
univerz\u00e1ln\u00ed a \u0161iroce vyu\u017e\u00edvan\u00e1 metoda. Pou\u017e\u00edv\u00e1 se pro zv\u00fd\u0161en\u00ed p\u0159esnosti norm\u00e1ln\u00ed Allanovy variance. Je zalo\u017eena na v\u00fdpo\u010dtech, kter\u00e9 vyu\u017e\u00edvaj\u00ed n\u011bkter\u00e9 stejn\u00e9 (p\u0159ekr\u00fdvaj\u00edc\u00ed se) vzorky, p\u0159i\u010dem\u017e je mo\u017eno vyu\u017e\u00edt jakkoli velk\u00fd soubor p\u0159ekr\u00fdvaj\u00edc\u00edch se dat. Zvy\u0161uje se tak p\u0159esnost v\u00fdpo\u010dtu stability, ale zv\u00fd\u0161en\u00e1 p\u0159esnost vy\u017eaduje v\u011bt\u0161\u00ed v\u00fdpo\u010detn\u00ed v\u00fdkon a v\u011bt\u0161\u00ed dobu zpracov\u00e1n\u00ed.<\/p>\n\n\n\n<\/a><\/figure>\n\n\n\n
Lad\u00edc\u00ed nap\u011bt\u00ed bylo nastaveno fixn\u011b pomoc\u00ed v\u00edceot\u00e1\u010dkov\u00e9ho trimru.<\/em><\/p>\n\n\n\n<\/a><\/figure>\n\n\n\n<\/figure>\n\n\n\n