{"id":1857,"date":"2024-10-16T18:22:49","date_gmt":"2024-10-16T17:22:49","guid":{"rendered":"https:\/\/www.sp3yor.net\/?p=1857"},"modified":"2024-10-16T18:37:52","modified_gmt":"2024-10-16T17:37:52","slug":"rs-cmu200-keyboard-fix-long-and-boring-story-naprawa-klawiatury","status":"publish","type":"post","link":"https:\/\/www.sp3yor.net\/?p=1857","title":{"rendered":"R&amp;S CMU200 keyboard fix &#8211; long and boring story (naprawa klawiatury)"},"content":{"rendered":"\n<p>Scroll down for English version.<\/p>\n\n\n\n<p>W czerwcu 2023 mia\u0142em szans\u0119, razem z ekip\u0105 SN7L, odwiedzi\u0107 prawdopodobnie najwi\u0119ksze targi kr\u00f3tkofalarskie w Europie &#8211; HAM RADIO we Friedrichshafen. Spotka\u0142em wiele inspiruj\u0105cych os\u00f3b, przeprowadzia\u0142em sporo d\u0142ugich dyskusji i przywioz\u0142em do domu ton\u0119 wspania\u0142ych wspomnie\u0144, ale nie tylko. Przywioz\u0142em ze sob\u0105 tak\u017ce do\u015b\u0107 fizyczn\u0105 pami\u0105tk\u0119: analiator radiokomunikacyjny Rohde &amp; Schwartz CMU200, kt\u00f3ry kupi\u0142em na tamtejszej gie\u0142dzie. Za kwot\u0119 600\u20ac, nie by\u0142o opcji, \u017ceby to urz\u0105dzenie nie sta\u0142o si\u0119 moj\u0105 w\u0142asno\u015bci\u0105. Podobny sprz\u0119t na polskim ryneczku jest wart ponad 2x wi\u0119cej (od 5-6 tys. PLN). Analizator widma w zakresie 10 MHz &#8211; 2,7 GHz, generator od zera do 2,7 GHz, -137 do +13 dBm, podw\u00f3jny, z modulacj\u0105 AM\/FM\/DSB, szerokopasmowy pomiar mocy. Oczywi\u015bcie, to sprz\u0119t przeznaczony typowo do pracy z urz\u0105dzeniami GSM\/3G, wi\u0119c brakuje mu kilku funkcji, np. prawdziwego tracking generatora, czy mo\u017cliwo\u015bci zerowania spektrum. Tak czy siak, za 600 EUR &#8211; nic lepszego nie dostan\u0119.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><a href=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_140321.jpg\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"768\" src=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_140321-1024x768.jpg\" alt=\"\" class=\"wp-image-1866\" srcset=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_140321-1024x768.jpg 1024w, https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_140321-300x225.jpg 300w, https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_140321-768x576.jpg 768w, https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_140321.jpg 1100w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/a><\/figure>\n\n\n\n<p>Kiedy tylko CMU200 trafi\u0142 pod m\u00f3j dach (TNX SP5QWB za transport &#8211; Lufthansa mog\u0142aby nie zaakceptowa\u0107 tego 18 kg pud\u0142a jako baga\u017cu podr\u0119cznego), zmierzy\u0142em oczywi\u015bcie wszystkie moje sprz\u0119ty radiowe: spektrum wyj\u015bciowe, moc, czu\u0142o\u015b\u0107 odbiornika, itd. W ko\u0144cu mog\u0142em te\u017c pomierzy\u0107 rzeczy na 2.4 GHz, jak m\u00f3j DIY upconverter do QO-100. Po kilku miesi\u0105cach zauwa\u017cy\u0142em pewien problem &#8211; po jakim\u015b czasie, ok. 1h, urz\u0105dzenie samo si\u0119 wy\u0142\u0105cza\u0142o. OK &#8211; to nie jest nowy sprz\u0119t ze sklepu, mo\u017ce jakie\u015b kondensatory straci\u0142y ju\u017c swoj\u0105 pojemno\u015b\u0107. Po nast\u0119pnych kilku miesi\u0105cach problem zacz\u0105\u0142 si\u0119 robi\u0107 dokuczliwy &#8211; po wystartowaniu mia\u0142em tylko 5-15 minut na pomiary. Po tym jak si\u0119 wy\u0142\u0105czy\u0142, nie by\u0142em ju\u017c w stanie go w\u0142\u0105czy\u0107 &#8211; dochodzi\u0142 do ekranu biosu, a czasem nawet nie. Musia\u0142em od\u0142\u0105czy\u0107 urz\u0105dzenie od pr\u0105du, odczeka\u0107 jaki\u015b czas (na sch\u0142odzenie..?) i dopiero potem spr\u00f3bowa\u0107 ponownie. Rzecz warta odnotowania &#8211; analizator nie wy\u0142\u0105cza\u0142 si\u0119 nagle (jakby wyci\u0105gn\u0105\u0107 wtyczk\u0119), tylko w taki sam spos\u00f3b jakbym nacisn\u0105\u0142 przycisk zasilania na przednim panelu, z komunikatem &#8222;Shutting down&#8221; na wy\u015bwietlaczu.<\/p>\n\n\n\n<p>Kolejnych kilka miesi\u0119cy min\u0119\u0142o i nie by\u0142em ju\u017c w stanie wcale w\u0142\u0105czy\u0107 urz\u0105dzenia &#8211; nie dochodzi\u0142 nawet do ekranu \u0142adowania w procedurze startu. Troch\u0119 zaniepokojony tym co znajd\u0119, przyst\u0105pi\u0142em do naprawy, przeczytawszy uprzednio wszystkie w\u0105tki o CMU200 na forum EEVblog i microcontroller.net (TNX google translate). Opr\u00f3cz problemu z wy\u0142\u0105czaniem si\u0119, moje urz\u0105dzenie by\u0142o w bardzo dobrym stanie &#8211; wszystkie self-testy przechodzi\u0142y bezproblemowo. Bogatszy o internetow\u0105 wiedz\u0119, zaw\u0119\u017cy\u0142em mo\u017cliwe \u017ar\u00f3d\u0142o problemu do dw\u00f3ch przypadk\u00f3w:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>problem z zasilaczem<\/li>\n\n\n\n<li>awaria klawiatury<\/li>\n<\/ul>\n\n\n\n<p>Wsp\u00f3\u0142czesne zasilacze impulsowe nie starzej\u0105 si\u0119 godnie &#8211; przegrzewaj\u0105ce si\u0119 uk\u0142ady wysuszaj\u0105 materia\u0142y termoprzewodz\u0105ce, kondensatory elektrolityczne wysychaj\u0105 (albo rozlewaj\u0105 si\u0119), uk\u0142ady zabezpieczaj\u0105ce dzia\u0142aj\u0105 zbyt agresywnie. Zasilacze do CMU200 nie s\u0105 ani \u0142atwo dost\u0119pne, ani tanie, wi\u0119c mia\u0142em szczer\u0105 nadziej\u0119, \u017ce problem le\u017cy gdzie\u015b indziej. Na szcz\u0119\u015bcie in\u017cynierowie projektuj\u0105cy CMU200 zamontowali na p\u0142ycie g\u0142\u00f3wnej urz\u0105dzenia 9 LED\u00f3w, kt\u00f3re pokazuj\u0105 stan wszystkich linii zasilaj\u0105cych wychodz\u0105cych z wy\u015bwietlacza. Urz\u0105dzenie, w trakcie kr\u00f3tkiego momentu od w\u0142\u0105czenia do resetu, zapala\u0142o wszystkie 9 &#8211; tak wi\u0119c problem musia\u0142 le\u017ce\u0107 gdzie\u015b indziej.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><a href=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241001_144617-rotated.jpg\"><img loading=\"lazy\" decoding=\"async\" width=\"768\" height=\"1024\" src=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241001_144617-768x1024.jpg\" alt=\"\" class=\"wp-image-1867\" srcset=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241001_144617-768x1024.jpg 768w, https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241001_144617-225x300.jpg 225w, https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241001_144617-rotated.jpg 825w\" sizes=\"auto, (max-width: 768px) 100vw, 768px\" \/><\/a><\/figure>\n\n\n\n<p>Zacz\u0105\u0142em wi\u0119c czyta\u0107 o przypadku alternatywnym, czyli problemach z klawiatur\u0105 na przednim panelu. Klawiatura w CMU200 to du\u017ca p\u0142ytka drukowana wykonana z kilku warstw przezroczystego plastiku, na kt\u00f3ry naniesiono metalowe \u015bcie\u017cki, po\u0142\u0105czone przelotkami. Uk\u0142ad jest matrycowy, tzn. kiedy naciskamy guzik, przewodz\u0105ca gumka zwiera dwie linie, chip odbiera naci\u015bni\u0119cie klawisza &#8211; pro\u015bcizna. Z czytania dowiedzia\u0142em si\u0119, \u017ce najpopularniejszy problem trapi\u0105cy t\u0105 konstrukcj\u0119 to elektromigracja tudzie\u017c zjawisko podobne do ro\u015bni\u0119cia cynowych w\u0105s\u00f3w &#8211; po pewnym czasie (a te urz\u0105dzenia maj\u0105 zazwyczaj &gt;20 lat), niekt\u00f3re \u015bcie\u017cki zwieraj\u0105 si\u0119, co prowadzi do tego, \u017ce urz\u0105dzenie jest przekonane, \u017ce przycisk zasilania jest ci\u0105gle wci\u015bni\u0119ty. Na forum znalaz\u0142em informacj\u0119, \u017ce mi\u0119dzy wszystkimi liniami klawiatury, oporno\u015b\u0107 powinna wynosi\u0107 &gt; 10 MOhm.<\/p>\n\n\n\n<p>Na pocz\u0105tku nie by\u0142em przekonany, \u017ce w\u0142a\u015bnie to mi si\u0119 przytrafi\u0142o, poniewa\u017c widzia\u0142em zale\u017cno\u015b\u0107 mi\u0119dzy temperatur\u0105 urz\u0105dzenia a wyst\u0119powaniem problemu &#8211; zimny analizator dzia\u0142a\u0142 zdecydowanie d\u0142u\u017cej ni\u017c po rozgrzaniu, obstawia\u0142em wi\u0119c, \u017ce to jaki\u015b zimny lut albo kondensator. Przynajmniej do czasu, a\u017c odkry\u0142em, \u017ce wci\u015bni\u0119cie przycisku &#8222;Escape&#8221; lub kt\u00f3rego\u015b z jego s\u0105siad\u00f3w pozwala urz\u0105dzeniu si\u0119 w\u0142\u0105czy\u0107 &#8211; eureka.<\/p>\n\n\n\n<p>Zacz\u0105\u0142em wi\u0119c dok\u0142adniejsz\u0105 diagnoz\u0119. Proces dostawania si\u0119 do wn\u0119trzno\u015bci CMU200 jest opisany w manualu serwisowym, ale mo\u017ce przyda si\u0119 komu\u015b, je\u015bli kr\u00f3tko go tu opisz\u0119. Stawiamy urz\u0105dzenie w pionie, na uchwytach przedniego panelu i \u015brubokr\u0119tem Torx T20 odkr\u0119camy cztery tylne n\u00f3\u017cki, po czym \u015bci\u0105gamy ca\u0142\u0105 obudow\u0119. K\u0142adziemy urz\u0105dzenie z powrotem w normalnej pozycji, tym samym torxem odkr\u0119camy przednie uchwyty i \u015bci\u0105gamy metalow\u0105 maskownic\u0119 z przedniego panelu. G\u00f3rna blacha trzyma si\u0119 na jednej \u015brubce.<\/p>\n\n\n\n<p>W \u015brodku urz\u0105dzenia przejrza\u0142em wszystkie po\u0142\u0105czenia, ale konstrukcja raczej nie pozwala na poluzowanie si\u0119 czegokolwiek. Diody na spodzie i z ty\u0142u p\u0142yty g\u0142\u00f3wnej potwierdza\u0142y prawid\u0142owe dzia\u0142anie zasilacza, wi\u0119c skupi\u0142em si\u0119 na sercu urz\u0105dzenia, czyli panelu frontowym, b\u0119d\u0105cym zasadniczo komputerem PC. Ca\u0142y panel zamontowany jest w obudowie za pomoc\u0105 5 \u015brub na obwodzie (2x g\u00f3ra i d\u00f3\u0142, 1x lewo), dw\u00f3ch od frontu (pod enkoderem) i jednej mocuj\u0105cej z\u0142\u0105cze kart PCMCIA, dost\u0119pnej od g\u00f3ry urz\u0105dzenia. Po usuni\u0119ciu \u015brub i odpi\u0119ciu trzech ta\u015bm \u0142\u0105cz\u0105cych PCMCIA z p\u0142yt\u0105 g\u0142\u00f3wn\u0105, panel mo\u017cna ostro\u017cnie wyci\u0105gn\u0105\u0107 do przodu.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><a href=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_125809.jpg\"><img loading=\"lazy\" decoding=\"async\" width=\"1024\" height=\"768\" src=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_125809-1024x768.jpg\" alt=\"\" class=\"wp-image-1868\" srcset=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_125809-1024x768.jpg 1024w, https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_125809-300x225.jpg 300w, https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_125809-768x576.jpg 768w, https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_125809.jpg 1100w\" sizes=\"auto, (max-width: 1024px) 100vw, 1024px\" \/><\/a><\/figure>\n\n\n\n<p>Dostanie si\u0119 do \u015brodka panelu jest jeszcze prostsze ni\u017c wyj\u0119cie go &#8211; wystarczy usun\u0105\u0107 kilka \u015brub na obwodzie i urz\u0105dzenie rozk\u0142ada si\u0119 na dwie po\u0142owy. Uwaga &#8211; ostro\u017cnie z ta\u015bmami (LCD, enkoder, klawiatura).<\/p>\n\n\n\n<p>Maj\u0105c wyj\u0119ty panel odpi\u0105\u0142em ta\u015bm\u0119 od dysku twardego, \u017ceby niepotrzebnie nie nadwyr\u0119\u017ca\u0107 wiekowego dysku ci\u0105g\u0142ymi restartami, a tak\u017ce wymieni\u0142em bateri\u0119 biosu. Nie rozwi\u0105za\u0142o to problemu, zacz\u0105\u0142em wi\u0119c rozbiera\u0107 klawiatur\u0119. \u017beby zdj\u0105\u0107 plastikow\u0105 maskownic\u0119 trzeba usun\u0105\u0107 wszystkie ma\u0142e \u015brubki, a wkr\u00f3tce naszym oczom uka\u017ce si\u0119 s\u0142ynna problematyczna membrana. Moja wygl\u0105da\u0142a nieco podstarzale, ale nie by\u0142o na niej widocznych \u015blad\u00f3w korozji. Wyczy\u015bci\u0142em wszystkie klawisze alkoholem izopropylowym, ale bez skutku, tak\u017ce w nast\u0119pnym kroku odpi\u0105\u0142em ta\u015bm\u0119 klawiatury od p\u0142yty g\u0142\u00f3wnej i urz\u0105dzenie w\u0142\u0105czy\u0142o si\u0119 poprawnie, zatrzymuj\u0105c si\u0119 na b\u0142\u0119dzie o braku dysku &#8211; s\u0142usznie, bo go odpi\u0105\u0142em. Zostawi\u0142em je tak na 20 minut, nic si\u0119 przez ten okres nie sta\u0142o, tak\u017ce wiem ju\u017c na 101% gdzie siedzi problem.<\/p>\n\n\n\n<p>Zacz\u0105\u0142em szuka\u0107 zwarcia, maj\u0105c na uwadze to, co przeczyta\u0142em na jednym z for\u00f3w &#8211; je\u015bli mi\u0119dzy jakimi\u015b \u015bcie\u017ckami jest &lt; 10 MOhm, to jest to uznawane za &#8222;zwarcie&#8221;, tak wi\u0119c nastawi\u0142em miernik na zakres 200 MOhm i zacz\u0105\u0142em sprawdza\u0107 linie metod\u0105 ka\u017cdy-z-ka\u017cdym. Pomijaj\u0105c fabrycznie zwarte do siebie \u015bcie\u017cki (s\u0105 takie) i wyprowadzenia dw\u00f3ch LED\u00f3w stanu zasilania, uda\u0142o mi si\u0119 znale\u017a\u0107 op\u00f3r rz\u0119du 300 kOhm pomi\u0119dzy pinami 6 i 17 we wtyczce (z\u0142\u0105czem do g\u00f3ry, ta\u015bma w kierunku z dala ode mnie). Pin 6 to jedno z dw\u00f3ch wyprowadze\u0144 klawisza zasilania, pin 17 jest pod\u0142\u0105czony do kilku klawiszy numerycznych i schodzi dalej do dolnego rz\u0119du klawiszy pod ekranem, gdzie krzy\u017cuje si\u0119 ze \u015bcie\u017ck\u0105 z pinu 6. Wychodzi na to, \u017ce &#8222;zwarcie&#8221; musia\u0142o si\u0119 pojawi\u0107 gdzie\u015b w dolnej cz\u0119\u015bci p\u0142ytki, po\u015brednio t\u0142umaczy to, dlaczego naciskaj\u0105c klawisze w okolicy &#8222;Escape&#8221;, urz\u0105dzenie zaczyna\u0142o na chwil\u0119 startowa\u0107.<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><a href=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_125906-rotated.jpg\"><img loading=\"lazy\" decoding=\"async\" width=\"768\" height=\"1024\" src=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_125906-768x1024.jpg\" alt=\"\" class=\"wp-image-1869\" srcset=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_125906-768x1024.jpg 768w, https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_125906-225x300.jpg 225w, https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_125906-rotated.jpg 825w\" sizes=\"auto, (max-width: 768px) 100vw, 768px\" \/><\/a><\/figure>\n\n\n\n<p>Klawisze ju\u017c wyczy\u015bci\u0142em, \u017caden klawisz zreszt\u0105 nie pracowa\u0142 pomi\u0119dzy liniami 6 i 17, tak\u017ce zwarcie musia\u0142o wyst\u0105pi\u0107 pomi\u0119dzy plastikowymi warstwami. Nie da si\u0119 tam dosta\u0107 bez ca\u0142kowitego zniszczenia membrany, poszed\u0142em wi\u0119c za rad\u0105 z internetu. Nastawi\u0142em m\u00f3j zasilacz regulowany na 30 V, ograniczenie pr\u0105du na 1 mA i ostro\u017cnie przy\u0142o\u017cy\u0142em napi\u0119cie do linii mi\u0119dzy kt\u00f3rymi by\u0142o przej\u015bcie (je\u015bli mo\u017cna tak nazwa\u0107 ~300 kOhm). Nic si\u0119 nie sta\u0142o, nie by\u0142o b\u0142ysku ani huku, zasilacz nie pokaza\u0142 p\u0142yn\u0105cego pr\u0105du, ale miernik pokaza\u0142 brak przej\u015bcia pomi\u0119dzy liniami. Wpi\u0105\u0142em membran\u0119 ponownie, pod\u0142\u0105czy\u0142em do CMU200 napi\u0119cie sieciowe i urz\u0105dzenie w\u0142\u0105czy\u0142o si\u0119 ca\u0142kowicie prawid\u0142owo. Ulga!<\/p>\n\n\n\n<p>Usterka zosta\u0142a usuni\u0119ta, ale jestem silnie przekonany, \u017ce pojawi si\u0119 w przysz\u0142o\u015bci &#8211; by\u0107 mo\u017ce do nast\u0119pnego zwarcia brakuje zaledwie kilka mikron\u00f3w. Na szcz\u0119\u015bcie na ebay mo\u017cna tanio naby\u0107 (~$50) zamienniki membran wykonane na normalnej cienkiej p\u0142ytce drukowanej. Wiedz\u0105c jaki problem trapi moje urz\u0105dzenie, zam\u00f3wi\u0119 sobie takow\u0105, p\u00f3ki jeszcze kto\u015b je sprzedaje. Ciesz\u0119 si\u0119 \u017ce m\u00f3j Rohde&amp;Schwartz zmierzy dla mnie jeszcze troch\u0119 rzeczy. A przy okazji jestem pod wra\u017ceniem jak porz\u0105dnie wykonane i \u0142atwe w serwisowaniu jest to urz\u0105dzenie.<\/p>\n\n\n\n<p>73 de SQ3SWF<\/p>\n\n\n\n<figure class=\"wp-block-image size-large\"><a href=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_090513-rotated.jpg\"><img loading=\"lazy\" decoding=\"async\" width=\"768\" height=\"1024\" src=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_090513-768x1024.jpg\" alt=\"\" class=\"wp-image-1870\" srcset=\"https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_090513-768x1024.jpg 768w, https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_090513-225x300.jpg 225w, https:\/\/www.sp3yor.net\/wp-content\/uploads\/2024\/10\/20241002_090513-rotated.jpg 825w\" sizes=\"auto, (max-width: 768px) 100vw, 768px\" \/><\/a><\/figure>\n\n\n\n<hr class=\"wp-block-separator has-alpha-channel-opacity\"\/>\n\n\n\n<p>In June of 2023, I had a chance to attend HAM RADIO exhibition in Friedrichshafen, Germany with my SN7L teammates. I met a ton of inspiring people, had many long discussions and brought home a ton of wonderful memories, but, I have also brought quite a physical souvenir: Rohde&amp;Schwartz CMU200 Universal Radio Communication Tester, which I got at the flea market. For the price of 600\u20ac there was no option for me not to buy it. A similar device on the domestic market would cost me more than twice that. 0.01 &#8211; 2.7 GHZ SA, -137 to +13 double generator w\/ modulation, wideband power meter. It obviously has some drawbacks, since it&#8217;s a device primarly intended for GSM and 3G, some software features are lacking, like a proper tracking generator or ability to take reference spectrum. Still, the device rocks.<\/p>\n\n\n\n<p>After getting it home (big thanks to Bartek SQ5QWB), I obviously enjoyed it a lot &#8211; measured all of my radios for output spectrum, RX sensitivity. Finally, I had a chance to measure my homebrew QO-100 equipment. After few months I&#8217;ve spotted a problem &#8211; the device started to shutdown by itself after some time &#8211; like an hour, or so. OK, it&#8217;s old, it&#8217;s been used quite a lot, maybe some caps went wrong or so. After some more time, the problem got a bit more intense &#8211; after booting, the CMU200 worked for 5-15 minutes and then shut down. After a shutdown, I could not turn it on back again, it would not even get to the BIOS screen. I had to disconnect the power using rear switch, wait some time (let it cool down?) and then the story repeat. One thing worth noticing &#8211; the shutdown was not like a &#8222;pull-the-plug&#8221; one, but looked the same way, as if I pressed the front &#8222;Power&#8221; button, with &#8222;powering off&#8221; message on the screen.<\/p>\n\n\n\n<p>Few months later &#8211; the device won&#8217;t even get to the system loading screen. A bit afraid of what I&#8217;m about to find, I started the repair, having previously read almost all of the CMU200-related threads on EEVblog and mikrocontroller.net (thanks, google translate). Other than the shutdown issue, my device was in a good shape, all the self-tests passed flawlessly. So I have narrowed the issue down to two possible causes:<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>power supply issues<\/li>\n\n\n\n<li>keyboard failure<\/li>\n<\/ul>\n\n\n\n<p>Switched mode power supplies do not age well &#8211; chips get too hot, capacitors dry out, protection circuits kick in too early. CMU200 power supplies are neither easy to get nor cheap, so I really hoped it&#8217;s about something else. Fortunately, the engineers behind CMU200 did a very good job by placing 9 LEDs on the bottom side of the main board, indicating whether all PSU voltage lines are OK. Luckily for me, all of them went bright green when I turned on the device, so maybe, maybe the problem is somewhere else.<\/p>\n\n\n\n<p>Then I read some more about the front keyboard. CMU200&#8217;s keyboard is a single sheet of PCB, made of few transparent plastic sheets with metal traces sandwiched between them. When a button is pressed, a pair of traces is shorted by a piece of conductive rubber and a key press is registered. Easy enough. Now, the most popular mode of failure is apparently caused by electromigration or whiskers growth. I can&#8217;t really tell what&#8217;s the underlying physical principle, but the final result is some keyboard lines being unintentionally shorted &#8211; just slightly, I&#8217;ve read about &lt;10 MOhm being enough to register a press.<\/p>\n\n\n\n<p>At first I did not believe this has happened to me, since I felt like the problem was temperature-related, therefore was linked to some electronic component, possibly some cold solder joint. At least until I discovered the device starts booting fine when I press &#8222;Escape&#8221; button, or one of the neighboring ones. That&#8217;s something!<\/p>\n\n\n\n<p>So I started the diagnosis. The disassembly process is described partially in the service manual, but maybe it&#8217;ll help someone if I briefly describe it here. In order to get into CMU200, you have to place the device face-down, so it rests on the front handles. Then, remove the four rear legs using T20 torx screwdriver, remove the main case by lifting it up. Lie the device down and remove the handles (4x T20) and front metal sheet. You can remove the top part of the chassis, there&#8217;s one screw on the left and it all slides right.<\/p>\n\n\n\n<p>With the device being open I inspected all the connections, but all seemed to be fine. PSU status LEDs (bottom of the mainboard, to the rear) were indicating no power issues, so I moved to the front unit, which is basically a PC. It&#8217;s mounted to the case using few regular Phillips screws (2x top, 2x bottom, 1x left, 2x front, 1 extra one accessible from the top, attaching the PCMCIA module). After removing all the screws and disconnecting 3 flat cables between main board and PCMCIA reader, the unit can be carefully slid forward and removed.<\/p>\n\n\n\n<p>Getting inside the front unit is even simpler than removing it from the device, few of the very same screws and the housing splits in half &#8211; warning: there are some flat cables for LCD, rotary encoder and obviously, keyboard.<\/p>\n\n\n\n<p>While I was there, I disconnected the HDD cable (lots of hard restarts and power cutoffs won&#8217;t do any good to a hard disk) and replaced the CR2032 bios battery. It obviously did not help, so I started disassembling the keyboard &#8211; after removing all the screws from the front (and pulling off the encoder dial), I saw the keyboard membrane itself &#8211; it did look a bit old, but no obvious signs of corrosion. Cleaned all the key contacts with IPA to no improvement &#8211; still getting stuck at a dark screen before BIOS. So, I disconnected the keyboard flat cable from the mobo entirely. After switching the mains power using switch at the rear, the unit powered on, went through POST and promptly stopped saying no boot disk was found &#8211; which was very true. I let it run like that for 20 minutes with no issues observed, so at least I was now sure &#8211; the problem was within the keyboard membrane.<\/p>\n\n\n\n<p>I started looking for a short, keeping in mind what I read on the forums &#8211; less than 10 MOhm is considered a short here, so I kept my meter in &#8222;200M&#8221; range. And, soon enough I found a resistance of ~300 KOhm between pins 6 and 17 on the flex connector (pins up, cable going away from me). I took a photo of the membrane and traced the connections &#8211; pin 6 is one of the two power button lines; pin 17 serves some numerical buttons and the bottom row of functional keys, as far as I remember. So the &#8222;short&#8221; must have appeared somewhere in the bottom part, which would partially explain why pressing &#8222;Escape&#8221; key allowed me to boot the CMU200 for a while.<\/p>\n\n\n\n<p>Since I have already cleaned the key contacts, and there was no key connected to both line 6 &amp; 17, I assumed the short must have happened between the membrane layers, since that&#8217;s the only place the two traces meet physically. There&#8217;s no way to get there between completely destroying the membrane, so I took the advice from microcontroller forums: I took my power supply, set it to 30V and 0.001 A current limit (does not really matter, since the first capacitor-powered electricity surge exceeds it anyway) and carefully touched the leads to lines 6 and 17. Nothing visible happened, supply showed no current flowing, no sparking, no smells, but: I took the meter, and it showed more than 200 MOhm as it should. So, I quickly connected the membrane straight into the mainboard, without bothering to assemble any other part of the device, flicked the mains switch and&#8230; boom. The device booted and operated perfectly. A huge relief.<\/p>\n\n\n\n<p>The issue has been fixed, but I&#8217;m quite sure it will happen again in the future, no doubt. There are replacement membranes available on ebay, so I&#8217;ll surely get one for myself, while they&#8217;re still being sold. Glad my R&amp;S is still alive. I&#8217;m amazed how well made these devices are and how easy is to work on them, in terms of dis- and reassembly.<\/p>\n\n\n\n<p>73 \/ SQ3SWF<\/p>\n","protected":false},"excerpt":{"rendered":"<p>Scroll down for English version. W czerwcu 2023 mia\u0142em szans\u0119, razem z ekip\u0105 SN7L, odwiedzi\u0107 prawdopodobnie najwi\u0119ksze targi kr\u00f3tkofalarskie w Europie &#8211; HAM RADIO we Friedrichshafen. Spotka\u0142em wiele inspiruj\u0105cych os\u00f3b, przeprowadzia\u0142em sporo d\u0142ugich dyskusji i przywioz\u0142em do domu ton\u0119 wspania\u0142ych &hellip; <a href=\"https:\/\/www.sp3yor.net\/?p=1857\">Czytaj dalej <span class=\"meta-nav\">&rarr;<\/span><\/a><\/p>\n","protected":false},"author":3,"featured_media":0,"comment_status":"closed","ping_status":"closed","sticky":false,"template":"","format":"standard","meta":{"ngg_post_thumbnail":0,"footnotes":""},"categories":[1],"tags":[],"class_list":["post-1857","post","type-post","status-publish","format-standard","hentry","category-bez-kategorii"],"_links":{"self":[{"href":"https:\/\/www.sp3yor.net\/index.php?rest_route=\/wp\/v2\/posts\/1857","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.sp3yor.net\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.sp3yor.net\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.sp3yor.net\/index.php?rest_route=\/wp\/v2\/users\/3"}],"replies":[{"embeddable":true,"href":"https:\/\/www.sp3yor.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=1857"}],"version-history":[{"count":11,"href":"https:\/\/www.sp3yor.net\/index.php?rest_route=\/wp\/v2\/posts\/1857\/revisions"}],"predecessor-version":[{"id":1874,"href":"https:\/\/www.sp3yor.net\/index.php?rest_route=\/wp\/v2\/posts\/1857\/revisions\/1874"}],"wp:attachment":[{"href":"https:\/\/www.sp3yor.net\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=1857"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.sp3yor.net\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=1857"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.sp3yor.net\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=1857"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}