low cost, 10 g dual axis accelerometer with digital outputs MXD2002A/b features dual axis accelerometer fabricated on a monolithic cmos ic on-chip mixed mode signal processing resolution better than 6 milli- g 50,000 g shock survival rating 30 hz bandwidth 2.70v to 5.25v single supply operation small (5mm x 5mm x 2mm) surface mount package continuous self test independently programmable axis (factory special) applications automotive ? vehicle security/active suspension/abs hed angle control/tilt sensing security ? gas line/elevator/fatigue sensing office equipment ? computer peripherals/pda?s/mouse smart pens/cell phones gaming ? joystick/rf interface/menu selection/tilt sensing white goods ? spin/vibration control i n t e r nal os c i l l a t o r sck ( opt i ona l ) cl k h e at er c ont r o l x a x i s y ax i s f a ct o r y a d j u st o f fs e t & g a in lpf lpf tem p e r at ur e s ens o r v o l t age re f e r e n c e vr e f do u t x vd d v d a gn d 2- a x i s senso r do u t y to ut c o n tin o u s s e lf t e s t a/ d a/ d mx d2002a/ b functional block diagram the m xd2002a/ b i s a very l o w cost , dual axi s accelerom eter fabricated on a standard, subm icron cmos process. the MXD2002A/b m easures acceleration with a full-scale range of 10 g . (the memsic accelerom eter product l i n e ext e nds from 1 g to 10 g with cu sto m versi ons avai l a bl e above 10 g .) it can m easure both dynam i c acceleration (e.g., vibration) and static acceleration (e.g., gravity). the MXD2002A/b design is based on heat convect i on and requi res no sol i d proof m a ss. th is elim in ates stictio n an d p a rticle p r o b l em s asso ciated with co m p etitiv e d e v i ces an d p r o v i d e s sh o c k su rv iv al u p to 50,000 g, lead in g to sig n i fican tly lo wer failu re rates an d l o wer l o ss due t o handl i ng duri ng assem b l y . the m xd2002a/ b provi des a di gi t a l out put (ref. ot her memsic data sheets for analog or ratiom e tric analog out put s). the out put s are di gi t a l si gnal s wi t h dut y cy cl es (rat i o of pul sewi dt h t o peri od) t h at are proport i onal t o acceleration. the duty cycl e outputs can be directly interfaced to a m i cro-processor. information furnished by memsic is be liev ed to be accurate and reliable. how e v e r, no responsibility is assumed by memsic for its use, nor for any infringements of patents or other rights of third parties w h ich may result from its use. no license is gr anted by implication or ot herw i se under any patent or patent rights of memsic. the t y pi cal noi se fl oor i s .0015% dut y cy cl e / hz allo win g sig n a ls b e lo w 6 m illi- g t o be resol v ed at 1 hz bandwi d t h . the m xd2002a/ b i s avai l a bl e i n a l o w profi l e lc c surface m ount package (5m m x 5m m x 2m m height). it is herm et i cal l y seal ed and operat i onal over a -40 c t o +105 c tem p erature range. due t o t h e st andard c m os st ruct ure of t h e m xd2002a/ b , ad d itio n a l circu itry can easily b e in co rp o r ated in to cu sto m versi ons for hi gh vol um e appl i cat i ons. c ont act t h e fact ory for m o re inform ation. ? memsic, inc. 100 burtt road, a ndov er, ma 01810 tel: 978.623.8188 fax: 978.623.9945 www. m e m s i c . c o m m e m s ic m xd2002a/ b r e v.e page 1 of 8 5/ 15/ 2004
m e m s ic m xd2002a/ b r e v.e page 2 of 8 5/ 15/ 2004 MXD2002A/b specifications (measurem ents @ 25 c, acceleration = 0 unless otherwise noted, v dd , v da = 5.0v unl ess ot herwi s e speci fi ed) parameter conditions min mx d2002a/b ty p max units sensor input m eas urem ent range 1 each axis 10 g nonlinearity best fit straight line 1.0 2.0 % of fs alignm ent error 2 1.0 d e g r e e s transverse sensitivity 3 2.0 % sensitivity d outx and d outy each axis @5.0v supply 1.80 2.00 2.20 % duty cy cle/ g change over temperature (uncompensated) 4 ? from 25 c, at ?40 c +93 % ? from 25 c, at +105 c - 4 7 % change over temperature (compensated) 4 ? from 25 c, ?40 c to +105 c < 3 . 0 % zero g bias level 0 g offset 5 each axis -0.70 0.00 +0.70 g 0 g duty cy cle 5 4 8 . 6 5 0 51.4 % duty cycle 0 g offs et over tem p erature ? from 25 c ? from 25 c, based on 2%/g 2.0 004 m g / c % / c noise performance noise density , % duty cy cle .0015 0.005 % duty cy cle/ hz frequency response 3db bandwidth 3 0 h z temp e rature outp u t t out voltage 1.15 1.25 1.35 v sensitivity 4.6 5.0 5.4 mv/ k voltage reference v re f @ 2 . 7 v - 5 . 0 v supply 2 . 4 2 . 5 2.65 v change over temperature 0.1 mv/ c current drive capability source 100 a s elf tes t continuous voltage at d outx , d outy under failure @5.0v supply , output rails to supply voltage 5.0 v continuous voltage at d outx , d outy under failure @2.7v supply , output rails to supply voltage 2.7 v d outx and d outy outputs digital signal of 100 hz or 400hz normal output range @5.0v supply @2.7v supply 0.1 0.1 4 . 9 2.6 v v current source or sink, @ 2.7v-5.0v supply 100 a rise/fall time 2.7 to 5.0v supply 90 100 110 nsec powe r suppl y operating voltage range 2.7 5.25 v supply current @ 5.0v 3.0 4.2 4.9 ma supply current 6,7 @ 2 . 7 v 3 . 0 5 . 4 6 m a temperature range operating range -40 +105 c notes 1 guaranteed by m easurem ent of initial offset and sensitivity. 2 alignm ent error is specif i ed as the angle between the true and indicated axis of sensitivity. 3 transverse sensitivity is the alge braic sum of the alignm ent and the inherent sensitivity errors 4 the sensitivity change over tem p er ature f o r therm a l accelerom eters is based on var i ations in heat tr ansfer that ar e gover n ed by the laws of phy sics and it is highly consistent fr om device to device. please ref e r to the section in this data sheet title d ? c om pensation for the change of sensitivity over tem p erature? for m o re inform ation. 5 t h e device oper a tes over a 2. 7v to 5. 25v supply r a nge. please note that sensitivity and zero g bias level will be slightly different at 2.7v operation. for devices to be oper a ted at 2. 7v/3. 0 v in pr oduction, they can be tr im m e d at the factor y specifically for this lower supply voltage oper a tion, in which case the sensitivity and zero g bias level specifications on this page will be m e t. please contact the factory for specially trim m e d devices for low supply voltage oper a tion. 6 note that the accelerom eter has a cons tant heater power control circuit ther eby display i ng higher supply cu r r e nt at lower oper a ting voltage.
m e m s ic m xd2002a/ b r e v.e page 3 of 8 5/ 15/ 2004 absolute max i mum ratings* suppl y vol t a ge (v dd , v da ) ??????...-0.5 to +7.0v storage tem p erature ???.????-65 c t o +150 c acceleration ??????????????..50,000 g *stresses above those liste d under a b solute maxim u m ratings m a y cause perm anent dam a ge to the device. t h is is a stress rating only ; the functional operation of the device at these or any other conditions above those indicated in the operational sections of this specification is not im p lied. e xposure to absolute m a xim u m rating conditions for extended periods m a y affect device reliability . package characteri sti c s package ja jc device weight lcc-8 110 c/w 2 2 c/w < 1 gram 8 4 1 2 3 7 6 5 t op v i ew me m s i c x + g y + g note: the memsic logo?s arrow indicates the +x sensing direction of the device. the +y sensing direction is rotated 90 away from the +x direction. s m all circle indicates pin one(1). pin description: lcc-8 package pi n n a m e descri pt i o n 1 t out tem p erat ure (anal og vol t a ge) 2 d outy y-axis acceleration digital signal 3 g n d g r o u n d 4 v da anal og suppl y vol t a ge 5 d outx x-axis acceleration digital signal 6 v ref 2.5v r e ference 7 sck opt i onal ext e rnal c l ock 8 v dd di gi t a l suppl y vol t a ge ordering guide model p a c k a g e styl e di gi tal output d2002al lc c - 8 sm d* 100 hz d2002b l lc c - 8 sm d* 400 hz *lcc parts are shipped in tape and reel packaging. caution esd (electro static d i sch a rg e) sen s itiv e d e v i ce. theory of operation the memsic device is a com p lete dual-axis acceleration m easu r em en t system fab r icated o n a m o n o lith ic cmos ic process. the devi ce operat i on i s based on heat t r ansfer by natural convection and opera tes like other accelerom eters havi ng a proof m a ss. the st at i onary el em ent , or ?proof m a ss?, in the memsic sensor is a gas. a sin g l e h eat so u r ce, cen tered in th e silico n ch ip is suspended across a cavity. equally spaced alum inum /polysilicon therm opiles (groups of t h erm o coupl es) are l o cat ed equi di st ant l y on al l four si des of the heat source (dual axis). under zero acceleration, a t e m p erat ure gradi e nt i s sy m m e t r i cal about t h e heat source, so th at th e tem p eratu r e is th e sam e at all fo u r th erm o p iles, causi ng t h em t o out put t h e sam e vol t a ge. acceleration in any direction will disturb the tem p erature profi l e , due t o free convect i on heat t r ansfer, causi ng i t t o be asy m m e t r i cal . the t e m p erat ure, and hence vol t a ge out put o f th e fo u r th erm o p iles will th en b e d i fferen t . th e d i fferen tial v o ltag e at th e th erm o p ile o u t p u t s is d i rectly proportional to the acceleration. there are two identical acceleration signal paths on the accelerom eter, one to m easure acceleration in the x-axis and one to m easure acceleration in the y-axis. please visit the memsic website at www.m e m s ic.com for a picture/graphic descri pt i on of t h e free convect i on heat t r ansfer pri n ci pl e. pin descriptions v dd ? thi s i s t h e suppl y i nput for t h e di gi t a l ci rcui t s and the sensor heater in the accelerom eter . the dc voltage shoul d be bet w een 2.70 and 5.25 vol t s . r e fer t o t h e sect i on on pc b l a y out and fabri cat i on suggest i ons for gui dance on ext e rnal part s and connect i ons recom m e nded. v da ? thi s i s t h e power suppl y i nput for t h e anal og am p lifiers in th e accelerom eter . refer to th e sectio n o n pcb l a y out and fabri cat i on suggest i ons for gui dance on ext e rnal part s and connect i ons recom m e nded. gnd ? thi s i s t h e ground pi n for t h e accelerom eter . d outx ? thi s pi n i s t h e di gi t a l out put of t h e x-axi s acceleration sensor. it is f actory program m a ble to 100 hz or 400 hz. the user shoul d ensure t h e l o ad i m pedance i s suffi ci ent l y hi gh as t o not source/ si nk >100 a typical. w h ile th e sen s itiv ity o f th is ax is h a s b een p r o g r am m e d at th e facto r y to b e th e sam e as th e sen s itiv ity fo r th e y-ax is, th e accelerom eter can be program m e d for non-equal sen s itiv ities o n th e x - an d y-ax es. co n t act th e facto r y fo r ad d itio n a l in fo rm atio n . d outy ? thi s pi n i s t h e di gi t a l out put of t h e y - axi s acceleration sensor. it is f actory program m a ble to 100 hz
m e m s ic m xd2002a/ b r e v.e page 4 of 8 5/ 15/ 2004 or 400 hz. the user shoul d ensure t h e l o ad i m pedance i s suffi ci ent l y hi gh as t o not source/ si nk >100 a typical. w h ile th e sen s itiv ity o f th is ax is h a s b een p r o g r am m e d at th e facto r y to b e th e sam e as th e sen s itiv ity fo r th e x - ax is, th e accelerom eter can be program m e d for non-equal sen s itiv ities o n th e x - an d y-ax es. co n t act th e facto r y fo r ad d itio n a l in fo rm atio n . t out ? thi s pi n i s t h e buffered out put of t h e t e m p erat ure sensor. the anal og vol t a ge at t out i s an i ndi cat i on of t h e di e t e m p erat ure. thi s vol t a ge i s useful as a di fferent i a l m easurem ent of tem p erature from am bient and not as an absolute m easurem ent of tem p er atu r e. after co rrelatin g th e vol t a ge at t out to 2 5 c am b i en t, th e ch an g e in th is v o ltag e due to changes in the am bient tem p erature can be used to com p ensate for the change over tem p erature of the accelerom eter offset and sensitivity. please refer to the sectio n o n co m p en satio n fo r th e ch an g e in sen s itiv ity over tem p erature for m o re inform ation. sck ? the st andard product i s del i v ered wi t h an i n t e rnal cl ock opt i on (800khz). this pin should be grounded when operating with the internal clock. an external cl ock opt i on can be speci al ordered from t h e fact ory al l o wi ng t h e user t o i nput a cl ock si gnal bet w een 400khz and 1.6m hz v ref ? a referen ce v o ltag e is av ailab l e fro m th is p i n . it is set at 2.50v t y pi cal and has 100 a o f d r iv e cap ab ility. compensation for the change in sensitivity over temperature all therm a l accelerom eters di splay the sam e sensitivity ch an g e with tem p eratu r e. th e sen s itiv ity ch an g e d e p e n d s on vari at i ons i n heat t r ansfer t h at are governed by t h e l a ws of phy si cs. m a nufact uri ng vari at i ons do not i n fl uence t h e sen s itiv ity ch an g e , so th ere are n o u n it to u n it d i fferen ces in sensitivity change. the sensitivity change is governed by t h e fol l o wi ng equat i on (and shown i n fi gure 1 i n c): s i x t i 2.67 = s f x t f 2.67 where s i is th e sen s itiv ity at an y in itial tem p eratu r e t i , and s f is th e sen s itiv ity at an y o t h e r fin a l tem p eratu r e t f with the tem p erature values in k an d th e sen s itiv ity ex p r essed in % duty cycle/g. 0.0 0.5 1.0 1.5 2.0 - 4 0 - 20 0 2 0 4 0 6 0 8 0 100 t e m p er at u r e ( c ) s e n s itiv ity (n o rma lize d ) figure 1: t h ermal a ccelerometer sensitivity in gam i ng appl i cat i ons where t h e gam e or cont rol l e r i s t y pi cal l y used i n a const a nt t e m p erat ure envi ronm ent , sen s itiv ity m i g h t n o t n eed to b e co m p en sated in h a rd ware or soft ware. the com p ensat i on for t h i s effect coul d be done i n st i n ct i v el y by t h e gam e pl ay er. fo r ap p licatio n s wh ere sen s itiv ity ch an g e s o f a few p e rcen t are acceptable, the above e quation can be approxim a ted wi t h a l i n ear funct i on. usi ng a l i n ear approxi m a t i on, an ext e rnal ci rcui t t h at provi des a gai n adjust m e nt of ?0.9%/ c wo u l d k eep th e sen s itiv ity with in 1 0 % o f its ro o m tem p erature value over a 0 c to +5 0 c range. for appl i cat i ons t h at dem a nd hi gh perform ance, a l o w cost m i cro-cont rol l e r can be used t o i m pl em ent t h e above equat i on. a reference desi gn usi ng a m i crochi p m c u (p/ n 16f873/ 04-so) and m e m s ic devel oped fi rm ware i s available by contacting the factory. w ith this reference d e sig n , th e sen s itiv ity v a riatio n o v e r th e fu ll tem p eratu r e range (-40 c t o +105 c ) can be kept bel o w 3%. pl ease v i sit th e memsic web site at www.m e m s ic.com for reference desi gn i n form at i on on ci rcui t s and program s including look up tables for eas ily incorporating sensitivity com p ensation. discussion of tilt applications and minimum resolution tilt applica t io ns: one of t h e m o st popul ar appl i cat i ons of the memsic accelerom eter product line is in tilt/inclination m easurem ent. an accelerom eter uses the force of gravi t y as an i nput t o det e rm i n e t h e i n cl i n at i on angl e of an object . a memsic accelerom eter is m o st sensitive to changes in position, or tilt, when the accelerom eter?s sensitive axis is p e rp en d i cu lar to th e fo rce o f g r av ity, o r p a rallel to th e earth?s surface. sim ilarly, wh en the accelerom eter?s axis is p a rallel to th e fo rce o f g r av ity (p erp e n d i cu lar to th e earth ? s surface), it is least sensitiv e to changes in tilt. tab l e 1 an d fig u r e 2 h e lp to illu strate th e o u t p u t ch an g e s in th e x- an d y-ax es as th e u n it is tilted fro m +9 0 to 0 . not i ce t h at when one axi s has a sm al l change i n out put per d e g r ee o f tilt (in m g ), t h e second axi s has a l a rge change i n o u t p u t p e r d e g r ee o f tilt. th e co m p lem e n t ary n a tu re o f these two signals perm its low cost accurate tilt sensing to be achi e ved wi t h t h e m e m s ic devi ce (reference appl i cat i on not e an-00m x-007). t op v i ew x y +90 0 0 0 gravi t y me ms i c figure 2: accelerometer po sition relative to gravity
m e m s ic m xd2002a/ b r e v.e page 5 of 8 5/ 15/ 2004 x - a x i s y - a x i s x-axis ori e nt at i on to earth?s surface (deg.) x output ( g ) c h ange per deg. o f tilt (m g ) y output ( g ) c h ange per deg. o f tilt (m g ) 9 0 1 . 0 0 0 0 . 1 5 0 . 0 0 0 1 7 . 4 5 8 5 0 . 9 9 6 1 . 3 7 0 . 0 8 7 1 7 . 3 7 8 0 0 . 9 8 5 2 . 8 8 0 . 1 7 4 1 7 . 1 6 7 0 0 . 9 4 0 5 . 8 6 0 . 3 4 2 1 6 . 3 5 6 0 0 . 8 6 6 8 . 5 9 0 . 5 0 0 1 5 . 0 4 4 5 0 . 7 0 7 1 2 . 2 3 0 . 7 0 7 1 2 . 2 3 3 0 0 . 5 0 0 1 5 . 0 4 0 . 8 6 6 8 . 5 9 2 0 0 . 3 4 2 1 6 . 3 5 0 . 9 4 0 5 . 8 6 1 0 0 . 1 7 4 1 7 . 1 6 0 . 9 8 5 2 . 8 8 5 0 . 0 8 7 1 7 . 3 7 0 . 9 9 6 1 . 3 7 0 0 . 0 0 0 1 7 . 4 5 1 . 0 0 0 0 . 1 5 table 1: changes in tilt for x- and y - axes minimum resolution : accelerom eters can be used in a wid e v a riety o f lo w g ap p licatio n s su ch as tilt an d o r ien t atio n . th e d e v i ce n o i se flo o r will v a ry with th e m easurem ent bandwi d t h . w i t h t h e reduct i on of t h e b a n d w id th th e n o i se flo o r d r o p s . th is will im p r o v e th e si gnal t o noi se rat i o of t h e m easurem ent and resol u t i on. the out put noi se scal es di rect l y wi t h t h e square root of t h e m easurem ent bandwi d t h . the m a xi m u m am pl i t ude of t h e noi se, i t s peak- t o - peak val u e, approxi m a t e l y defi nes t h e wo rst case reso lu tio n o f th e m easurem ent. the peak-to- peak noi se i s approxi m a t e l y equal t o 6.6 t i m es as t h e rm s value (with an average uncerta inty of .1%). the m a xim u m n o i se fo r 1 . 0 h z b a n d w id th will b e 1 m g / hz . if the bandwi d t h i s i n creased t o 10 hz for exam pl e: 3.162 m g i s t h e m a xi m u m rm s noi se and 20.87m g i s t h e m a xi m u m peak -t o-peak noi se. digital interface the MXD2002A/b is easily in terfaced with low cost m i crocontrollers. for the di gital output accelerom eter, one digital input port is require d to read one accelerom eter output. for the analog output accel erom eter, m a ny low cost m i crocontrollers are available t oday that feature integrated a/ d (anal og t o di gi t a l convert ers) wi t h resol u t i ons rangi ng from 8 t o 12 bi t s . in m a ny appl i cat i ons t h e m i crocont rol l e r provi des an effective approach for the te m p erature com p ensation of the sen s itiv ity an d th e zero g offset. specific code set, reference desi gns, and appl i cat i ons not es are avai l a bl e from t h e fact ory . the fol l o wi ng param e t e rs m u st be consi d ered i n a digital interface: reso lu tio n : sm allest detectable ch ange in input acceleration bandw i d t h : detectable accelerations in a given period of tim e acqui si t i on ti me : the duration of the m easurem ent of the acceleration signal duty cycle definition the m xd2002a/ b has t w o pw m dut y cy cl e out put s (x,y ). the acceleration is proportional to the ratio t1/t2. the zero g o u t p u t is set to 5 0 % d u t y cycle an d th e sen s itiv ity scale factor is set to 2% duty cycle change per g . these n o m in al v a lu es are affected b y th e in itial to leran ce o f th e devi ce i n cl udi ng zero g o ffset erro r an d sen s itiv ity erro r. thi s devi ce i s offered from t h e fact ory program m e d t o ei t h er a 10m s peri od (100 hz) or a 2.5m s peri od (400hz). t1 lengt h of t h e ?on? port i on of t h e cy cl e. t2 (peri od) lengt h of t h e t o t a l cy cl e. du ty cycle ratio o f th e ?0 n ? tim e (t1) of the cycle to the total cycle (t2). defined as t1/t2. pul s e wi dt h ti m e peri od of t h e ?on? pul se. defi ned as t1. t2 t1 a (g)= (t1/t2 - 0.5)/2% 0g = 50% duty cy cle t2= 2.5ms or 10ms (factory programmable) figure 4: typical output duty c ycle choosing t2 and counter frequency design trade-offs the noise level is one determ inant of accelerom eter resol u t i on. the second rel a t e s t o t h e m easurem ent resol u t i on of t h e count er when decodi ng t h e dut y cy cl e output. the actual resolution of the acceleration signal is l i m i t e d by t h e t i m e resol u t i on of t h e count i ng devi ces used t o decode t h e dut y cy cl e. the fast er t h e count er cl ock, t h e hi gher t h e resol u t i on of t h e dut y cy cl e and t h e short e r t h e t2 peri od can be for a gi ven resol u t i on. tabl e 2 shows so m e o f th e trad e-o ffs. it is im p o r tan t to n o t e th at th is is th e resol u t i on due t o t h e m i croprocessors? count er. it i s probable that the accelerom ete r?s noise floor m a y set the lo wer lim it o n th e reso lu tio n . t2 (m s) memsic sa mple rate counter- clock rate (mhz) counts pe r t 2 cy cle counts per g reso- lution (m g ) 2 . 5 4 0 0 2 . 0 5000 1 0 0 1 0 2 . 5 4 0 0 1 . 0 2500 5 0 2 0 2 . 5 4 0 0 0 . 5 1250 2 5 4 0 1 0 . 0 1 0 0 2 . 0 20000 4 0 0 2 . 5 1 0 . 0 1 0 0 1 . 0 10000 2 0 0 5 . 0 1 0 . 0 1 0 0 0 . 5 5000 1 0 0 1 0 table 2: trade-offs between microcontroller counter rate and t2 period.
m e m s ic m xd2002a/ b r e v.e page 6 of 8 5/ 15/ 2004 using the accelerometer in very low power applications (battery operation) in ap p licatio n s with p o w er lim itatio n s , p o w er cyclin g can be used t o ext e nd t h e bat t e ry operat i ng l i f e. one i m port a nt consideration when power cyc ling is that the accelerom eter tu rn o n tim e lim its th e freq u e n c y b a n d w id th o f th e accelerations to be m easured . for exam ple, operating at 2.7v t h e t u rn on t i m e i s 40m s. to doubl e t h e operat i ng tim e, a particular application m a y cycle power on for 40m s, t h en off for 40m s, resul t i ng i n a m easurem ent peri od of 80m s, or a frequency of 12.5hz. w i t h a frequency of m easurem ents of 12.5hz, accelerations changes as hi gh as 6.25hz can be det ect ed. power cycling can be used eff ectiv ely in m a n y in clin o m etry appl i cat i ons, where i n cl i n at i on changes can be sl ow and i n frequent . compensation for z e ro g offset change over temperature the com p ensation of offset is p e rfo rm ed with th e fo llo win g equat i on: aoc = a + ( a + b * t + c * t * t) where aoc is the offset com p ensated acceleration, a is the uncom pensated acceleration, t is tem p erature and a, b, c are constants characteristic to each accelerom eter. com puter program s are used to determ ine these constants. th e co n s tan t s can b e read fro m an d written to th e mcu eepr o m vi a t h e r s -232. the const a nt s a,b,c are norm a l l y sto r ed in th e mcu eeprom. to d e term in e th e v a lu es o f the constants, each accelerom eter is taken to three different tem p eratures, preferably even ly spread across the desired tem p erature span. the zero g bias (a0, a1 and a2) and the tem p eratures (t0, t1 and t2) are recorded at each tem p erature. the data collected (a0, t0, a1, t1, a2, t2) is used i n a quadrat i c i n t e rpol at i on (or lagrange pol y nom i a l ) to d e term in e a, b an d c as fo llo ws: r0 = a0 / ( (t0-t1)*(t0-t2) ) r1 = a1 / ( (t1-t0)*(t1-t2) ) r2 = a2 / ( (t2-t0)*(t2-t1) ) a = r0 * t1 * t2 + r1 * t0 * t2 + r2 * t0 * t1 b = - r0 * (t1+t2) ? r1 * (t0+t2) ? r2 *(t0+t1) c = r0 + r1 + r2 in m a ny cases a com puter is used to control the tem p erature, com m unicate with the mcu, and to calculate the constants. after calculating the constants, the com puter downl oads t h e const a nt s t o eepr o m . for a m o re det a i l di scussi on of t e m p erat ure com p ensat i on reference m e m s ic appl i cat i on not e #an-00m x-002 m i croc ont r o l l e r me m s i c ac ce l i/ o i/ o a/ d ax ay t figure 5: zero g offset te mperature compensation circuit compensation for ex tending the frequency response the response of the therm a l accelerom eter is a function of t h e i n t e rnal gas phy si cal propert i e s, t h e nat u ral convect i on m echanism and the sensor electronics. since the gas properties of memsic' s m a ss produced accelerom eter are u n i fo rm , a d i g ital filter can b e u s ed to eq u a lly co m p en sate all sen s o r s. th e co m p en satin g filter d o e s n o t req u i re adjustm e nt for individual accel erom eters. the function of the com p ensating filter is to apply gain in proportion with the acceleration changes. the faster the acceleration ch an g e s o ccu r, th e h i g h e r th e g a in th at th e filter ap p lies. for analog output accelerom eter s, the com p ensating filter can be i m pl em ent e d wi t h a ci rcui t i nvol vi ng t w o op-am ps and som e resi st ors and capaci t o rs. for di gi t a l out put accelerom eters, a digital filter is necessary. in applications where high fre quency accelerations need to be m easured, a dsp (di g i t a l si gnal processor) m a y be n ecessary to im p l em en t th e d i g ital filter. dsp ic?s an d devel opm ent t ool s are readi l y available from m a jor ic m a nufact urers. ho wev e r, if th e b a n d w id th req u i rem e n t is relativ ely lo w (i .e. 100hz), i t i s possi bl e t o i m pl em ent a di gi t a l frequency co m p en satin g filter with an 8 b it m i cro c o n t ro ller. th e m i cro c o n t ro ller will lik ely h a v e to b e cap ab le o f o p e ratin g at rel a t i v el y hi gh cl ock frequenci e s (20m hz). converting the digital output to an analog output the pw m out put can be easi l y convert ed i n t o an anal og o u t p u t b y in teg r atio n . a sim p le rc filter can d o th e conversi on. not e t h at t h at t h e i m pedance of t h e ci rcui t fol l o wi ng t h e i n t e grat or m u st be m u ch hi gher t h an t h e im p e d a n ce o f th e rc filter. referen ce fig u r e 6 fo r an exam ple. 1uf dout aout 10k memsic accel. figure 6: converting the dig ital output to an analog voltage
m e m s ic m xd2002a/ b r e v.e page 7 of 8 5/ 15/ 2004 temperature output noise reduction it is reco m m e n d e d th at a sim p le rc lo w p a ss filter is u s ed when m easuring the tem p erature output. tem p erature out put i s t y pi cal l y a very sl ow changi ng si gnal , so a very lo w freq u e n c y filter elim in ates erro n e o u s read in g s th at m a y resul t from t h e presence of hi gher frequency noi se. a sim p le filter is sh o w n in fig u r e 8 . filtered tout 8.2k 0.1uf memsic accel. tout fi gure 8: temperat ure out put n o i s e reduct i o n power supply noise rejection two capacitors and a resistor are recom m e nded for best reject i on of power suppl y noi se (reference fi gure 9 bel o w). the capaci t o rs shoul d be l o cat ed as cl ose as possi bl e t o t h e devi ce suppl y pi ns (v da , v dd ). the capacitor lead length should be as short as possibl e, and surface m ount capacitors are preferred. for typical a pplications, capacitors c1 and c 2 can be ceram i c 0.1 f, and t h e resi st or r can be 10 ? . in 5v appl i cat i ons where power consum pt i on i s not a concern, m a xi m u m suppl y noi se reject i on can be obt ai ned by si gni fi cant l y i n creasi ng t h e val u es of c 1 , c 2 and r . for exam pl e, c 1 = c 2 = 0.47 f and r = 270 ? will v i rtu a lly el i m i n at e power suppl y noi se effect s. r memsic accelerometer vda c1 c2 vdd v supply figure 9: power supply noise rejection pcb layout and fabrication suggestions 1. the sck pi n shoul d be grounded t o m i ni m i ze noi se. 2. li beral use of ceram i c by pass capaci t o rs i s recom m e nded. 3. r obust l o w i nduct a nce ground wi ri ng shoul d be used. 4. c a re shoul d be t a ken t o ensure t h ere i s ?t herm al sy m m e t r y ? on t h e pc b i m m e di at el y surroundi ng t h e m e m s ic devi ce and t h at t h ere is no significant heat source nearby. 5. a m e t a l ground pl ane shoul d be added di rect l y beneat h t h e m e m s ic devi ce. the si ze of t h e ground pl ane shoul d be si m i l a r t o t h e m e m s ic devi ce?s foot pri n t and as t h i c k as possi bl e. 6. vi as can be added sy m m e t r i cal l y around t h e ground plane. vias increase therm a l isolation of the device from the rest of the pcb
m e m s ic m xd2002a/ b r e v.e page 8 of 8 5/ 15/ 2004 package d r aw ing fig 1 0 : hermetica lly s e a l ed pa cka g e ou tlin e
|
