LTC ppm Noise, Low Drift Precision References FEATURES DESCRIPTION APPLICATIONS TYPICAL APPLICATION

Transcription

1 .ppm Noise, Low Drift Precision References FEATURES n Low Noise:.ppm P-P (.Hz to Hz) nv P-P for the LTC-. n Low Drift: ppm/ C Max n High Accuracy: ±.% Max n No Humidity Sensitivity (LS8 Package) n Therma Hysteresis (LS8): 3ppm ( C to 8 C) n Long-Term Drift (LS8): ppm/ khr n % Tested at C, C and C n Load Reguation: <ppm/ma n Sinks and Sources Current: ±ma n Low Dropout: mv n Maximum Suppy Votage: 3.V n Low Power Shutdown: <µa Max n Avaiabe Output Votages:.V,.8V,.V, 3V, 3.3V,.9V, V n Avaiabe in an 8-Lead MSOP and High Stabiity Hermetic mm mm LS8 Packages APPLICATIONS n Instrumentation and Test Equipment n High Resoution Data Acquisition Systems n Weigh Scaes n Precision Battery Monitors n Precision Reguators n Medica Equipment DESCRIPTION The LTC is a compete famiy of precision bandgap votage references, offering exceptiona noise and drift performance. This ow noise and drift is ideay suited for the high resoution measurements required by instrumentation and test equipment. In addition, the LTC is fuy specified over the temperature range of C to C, ensuring its suitabiity for demanding automotive and industria appications. Advanced curvature compensation aows this bandgap reference to achieve a drift of ess than ppm/ C with a predictabe temperature characteristic and an output votage accurate to ±.%, reducing or eiminating the need for caibration. The LTC can be powered from as itte as mv above the output votage to as much as 3.V. Superior oad reguation with source and sink capabiity, couped with exceptiona ine rejection, ensures consistent performance over a wide range of operating conditions. A shutdown mode is provided for ow power appications. The LTC references are offered in an 8-ead MSOP package and an 8-ead LS8 package. The LS8 is a mm mm surface mount hermetic package that provides outstanding stabiity. L, LT, LTC, LTM, Linear Technoogy and the Linear ogo are registered trademarks of Linear Technoogy Corporation. A other trademarks are the property of their respective owners. TYPICAL APPLICATION Low Frequency.Hz to Hz Noise (LTC-.) Basic Connection LTC-. 3V < 3.V C IN.µF SHDN _F _S C OUT µf nv/div TAa s/div TAb

2 ABSOLUTE MAXIMUM RATINGS Input Votage to....3v to 3.V SHDN to....3v to (.3V) Output Votage: _F....3V to (.3V) _S....3V to V Output Short-Circuit Duration... Indefinite PIN CONFIGURATION (Note ) Operating Temperature Range (Note ).. C to C Storage Temperature Range (Note )... C to C Lead Temperature Range (Sodering, sec) (Note 3)...3 C TOP VIEW * SHDN * 3 TOP VIEW 8 * 7 _F _S * MS8 PACKAGE 8-LEAD PLASTIC MSOP T JMAX = C, θ JA = 3 C/W *CONNECT PINS TO DEVICE (PIN ) SHDN * _F _S * LS8 PACKAGE 8-PIN LEADLESS CHIP CARRIER (mm mm) T JMAX = C, θ JA = C/W *CONNECT PINS TO DEVICE (PIN ) ORDER INFORMATION LEAD FREE FINISH TAPE AND REEL PART MARKING* PACKAGE DESCRIPTION SPECIFIED TEMPERATURE RANGE LTCBHMS8-.#PBF LTCBHMS8-.#TRPBF LTFDG 8-Lead Pastic MSOP C to C LTCCHMS8-.#PBF LTCCHMS8-.#TRPBF LTFDG 8-Lead Pastic MSOP C to C LTCBHMS8-.8#PBF LTCBHMS8-.8#TRPBF LTFDH 8-Lead Pastic MSOP C to C LTCCHMS8-.8#PBF LTCCHMS8-.8#TRPBF LTFDH 8-Lead Pastic MSOP C to C LTCBHMS8-.#PBF LTCBHMS8-.#TRPBF LTFCY 8-Lead Pastic MSOP C to C LTCCHMS8-.#PBF LTCCHMS8-.#TRPBF LTFCY 8-Lead Pastic MSOP C to C LTCBHMS8-3#PBF LTCBHMS8-3#TRPBF LTFDJ 8-Lead Pastic MSOP C to C LTCCHMS8-3#PBF LTCCHMS8-3#TRPBF LTFDJ 8-Lead Pastic MSOP C to C LTCBHMS8-3.3#PBF LTCBHMS8-3.3#TRPBF LTFDK 8-Lead Pastic MSOP C to C LTCCHMS8-3.3#PBF LTCCHMS8-3.3#TRPBF LTFDK 8-Lead Pastic MSOP C to C LTCBHMS8-.9#PBF LTCBHMS8-.9#TRPBF LTFDM 8-Lead Pastic MSOP C to C LTCCHMS8-.9#PBF LTCCHMS8-.9#TRPBF LTFDM 8-Lead Pastic MSOP C to C LTCBHMS8-#PBF LTCBHMS8-#TRPBF LTFDN 8-Lead Pastic MSOP C to C LTCCHMS8-#PBF LTCCHMS8-#TRPBF LTFDN 8-Lead Pastic MSOP C to C LTCBHLS8-. #PBF N/A 8-Lead Ceramic LCC (mm mm) C to C LTCCHLS8-. #PBF N/A 8-Lead Ceramic LCC (mm mm) C to C Consut LTC Marketing for parts specified with wider operating temperature ranges.*the temperature grade is identified by a abe on the shipping container. This product is ony offered in trays. For more information refer to http// Consut LTC Marketing for information on non-standard ead based finish parts. For more information on ead free part marking, go to: For more information on tape and ree specifications, go to:

3 AVAILABLE OPTIONS OUTPUT VOLTAGE INITIAL ACCURACY TEMPERATURE COEFFICIENT PART NUMBER..%.%.8.%.%..%.%.%.% 3..%.% 3.3.%.%.9.%.%..%.% See Order Information section for compete part number isting. ppm/ C ppm/ C ppm/ C ppm/ C ppm/ C ppm/ C ppm/ C ppm/ C ppm/ C ppm/ C ppm/ C ppm/ C ppm/ C ppm/ C ppm/ C ppm/ C LTCBHMS8-. LTCCHMS8-. LTCBHMS8-.8 LTCCHMS8-.8 LTCBHMS8-. LTCCHMS8-. LTCBHLS8-. LTCCHLS8-. LTCBHMS8-3. LTCCHMS8-3. LTCBHMS8-3.3 LTCCHMS8-3.3 LTCBHMS8-.9 LTCCHMS8-.9 LTCBHMS8- LTCCHMS8- ELECTRICAL CHARACTERISTICS The denotes the specifications which appy over the fu operating temperature range, otherwise specifications are at T A = C. =.V, _S connected to _F, uness otherwise noted. PARAMETER CONDITIONS MIN TYP MAX UNITS Output Votage LTCB.. % LTCC.. % Output Votage Temperature Coefficient (Note ) LTCB LTCC Line Reguation.V 3.V, SHDN = Load Reguation (Note ) I SOURCE = ma LTCMS8 Operating Votage (Note ) Output Short-Circuit Current LTCLS8 I SINK = ma LTCMS8 LTCLS ppm/ C ppm/ C ppm/v ppm/v ppm/ma ppm/ma ppm/ma ppm/ma ppm/ma ppm/ma ppm/ma ppm/ma LTC-., LTC-.8, LTC-. I SOURCE = ma, Error.% 3 3. V LTC-3, LTC-3.3, LTC-.9, LTC- I SOURCE = ±ma, Error.% I OUT = ma, Error.% Shutdown Pin (SHDN) Logic High Input Votage Logic High Input Current, SHDN = V Logic Low Input Votage Logic Low Input Current, SHDN =.8V Suppy Current No Load.. Short to Short to V V ma ma V µa V µa ma ma 3

4 ELECTRICAL CHARACTERISTICS The denotes the specifications which appy over the fu operating temperature range, otherwise specifications are at T A = C. =.V, _S connected to _F, uness otherwise noted. PARAMETER CONDITIONS MIN TYP MAX UNITS Shutdown Current SHDN Tied to µa Output Votage Noise (Note 7).Hz f Hz Hz f khz..7 ppm P-P ppm RMS Turn-On Time.% Setting, C OUT =.7µF µs Long-Term Drift of Output Votage (Note 8) LTCMS8 ppm/ khr LTCLS8 ppm/ khr Hysteresis (Note 9) LTCMS8 T = C to 7 C T = C to 8 C T = C to C LTCLS8 T = C to 7 C T = C to 8 C T = C to C ppm ppm ppm ppm ppm ppm Note : Stresses beyond those isted under Absoute Maximum Ratings may cause permanent damage to the device. Exposure to any Absoute Maximum Rating condition for extended periods may affect device reiabiity and ifetime. Note : Precision may be affected if the parts are stored outside of the specified temperature range. Large temperature changes may cause changes in device performance due to therma hysteresis. For best performance, extreme temperatures shoud be avoided whenever possibe. Note 3: The stated temperature is typica for sodering of the eads during manua rework. For detaied IR refow recommendations, refer to the Appications Information section. Note : Temperature coefficient is measured by dividing the maximum change in output votage by the specified temperature range. Note : Load reguation is measured on a puse basis from no oad to the specified oad current. Load current does not incude the ma sense current. Output changes due to die temperature change must be taken into account separatey. Note : Excudes oad reguation errors. Minimum suppy for the LTC., LTC-.8 and LTC-. is set by interna circuitry suppy requirements, regardess of oad condition. Minimum suppy for the LTC-3, LTC-3.3, LTC-.9 and LTC- is specified by oad current. Note 7: Peak-to-peak noise is measured with a -poe highpass fiter at.hz and 3-poe owpass fiter at Hz. The unit is encosed in a sti-air environment to eiminate thermocoupe effects on the eads, and the test time is seconds. Due to the statistica nature of noise, repeating noise measurements wi yied arger and smaer peak vaues in a given measurement interva. By repeating the measurement for intervas, each seconds ong, it is shown that there are time intervas during which the noise is higher than in a typica singe interva, as predicted by statistica theory. In genera, typica vaues are considered to be those for which at east % of the units may be expected to perform simiary or better. For the interva test, a typica unit wi exhibit noise that is ess than the typica vaue isted in the Eectrica Characteristics tabe in more than % of its measurement intervas. See Appication Note for noise testing detais. RMS noise is measured with a spectrum anayzer in a shieded environment. Note 8: Long-term stabiity typicay has a ogarithmic characteristic and therefore, changes after hours tend to be much smaer than before that time. Tota drift in the second thousand hours is normay ess than one-third that of the first thousand hours with a continuing trend toward reduced drift with time. Long-term stabiity is aso affected by differentia stresses between the IC and the board materia created during board assemby. Note 9: Hysteresis in output votage is created by mechanica stress that differs depending on whether the IC was previousy at a higher or ower temperature. Output votage is aways measured at C, but the IC is cyced to the hot or cod temperature imit before successive measurements. Hysteresis is roughy proportiona to the square of the temperature change. For instruments that are stored at we controed temperatures (within or 3 degrees of operationa temperature), hysteresis is usuay not a significant error source. Typica hysteresis is the worst case of C to cod to C or C to hot to C, preconditioned by one therma cyce.

5 TYPICAL PERFORMANCE CHARACTERISTICS Characteristic curves are simiar for most votage options of the LTC. Curves from the LTC-., LTC-. and the LTC- represent the range of performance across the entire famiy of references. Characteristic curves for other output votages fa between these curves and can be estimated based on their votage output. nv/ DIV.V Low Frequency.Hz to Hz Noise OUTPUT VOLTAGE (V) V Output Votage Temperature Drift 3 TYPICAL UNITS OUTPUT VOLTAGE CHANGE (ppm) 3.V Load Reguation (Sourcing) C C C s/div G.9 7 TEMPERATURE ( C)... OUTPUT CURRENT (ma) G G3 OUTPUT VOLTAGE CHANGE (ppm) 8.V Load Reguation (Sinking) C C C... OUTPUT CURRENT (ma) NOISE VOLTAGE (nv/ Hz) 3 3.V Output Votage Noise Spectrum.7µF µf µf.. FREQUENCY (khz) ma I OUT ma mv/div.v Sinking Current with a 3.3µF Output Capacitor C OUT = 3.3µF µs/div G G G ma I OUTmA mv/div.v Sourcing Current with a 3.3µF Output Capacitor SUPPLY CURRENT (µa) 8.V Shutdown Suppy Current vs Input Votage C C C NUMBER OF PARTS.V Distribution TA = C 3 C OUT = 3.3µF µs/div G7 8 INPUT VOLTAGE (V) (V) G8 G9

6 TYPICAL PERFORMANCE CHARACTERISTICS Characteristic curves are simiar for most votage options of the LTC. Curves from the LTC-., LTC-. and the LTC- represent the range of performance across the entire famiy of references. Characteristic curves for other output votages fa between these curves and can be estimated based on their votage output..v Low Frequency.Hz to Hz Noise..V Output Votage Temperature Drift 3 TYPICAL UNITS.V Load Reguation (Sourcing) nv/ DIV s/div G OUTPUT VOLTAGE (V) TEMPERATURE ( C) OUTPUT VOLTAGE CHANGE (ppm) 3 C C C... OUTPUT CURRENT (ma) G G OUTPUT VOLTAGE CHANGE (ppm) 8.V Load Reguation (Sinking) C C C... OUTPUT CURRENT (ma) G3 SUPPLY CURRENT (ma) V Suppy Current vs Input Votage C C C 8 INPUT VOLTAGE (V) G SUPPLY CURRENT (µa) 8.V Shutdown Suppy Current vs Input Votage C C C 8 INPUT VOLTAGE (V) G.V Minimum Differentia (Sourcing).V Minimum Differentia (Sinking).V Output Votage Noise Spectrum C OUT =.7µF OUTPUT CURRENT (ma). OUTPUT CURRENT (ma). NOISE VOLTAGE (nv Hz) 8 C OUT = µf C OUT = µf C C C... INPUT OUTPUT VOLTAGE (V) G C C C..... INPUT OUTPUT VOLTAGE (V) G7.. FREQUENCY (khz) F

7 TYPICAL PERFORMANCE CHARACTERISTICS Characteristic curves are simiar for most votage options of the LTC. Curves from the LTC-., LTC-. and the LTC- represent the range of performance across the entire famiy of references. Characteristic curves for other output votages fa between these curves and can be estimated based on their votage output. 9 8.V Distribution T A = C.V Temperature Drift Distribution C TO C...V SHDN Input Votage Threshods vs NUMBER OF PARTS 7 3 NUMBER OF PARTS 8 V TRIP (V)... V TH_UP V TH_DN (V) DRIFT (ppm/c).8. 8 (V) G G9 G POWER SUPPLY REJECTION RATIO (db) 8.V Power Suppy Rejection Ratio vs Frequency. C OUT =.7µF C OUT = µf C OUT = µf.. FREQUENCY (khz) G OUTPUT IMPEDENCE (Ω)..V Output Impedance vs Frequency C OUT =.7µF C OUT = µf C OUT = µf.... FREQUENCY (khz) G3 OUTPUT VOLTAGE (V) V Line Reguation 8 INPUT VOLTAGE (V) C C C G 7

8 TYPICAL PERFORMANCE CHARACTERISTICS Characteristic curves are simiar for most votage options of the LTC. Curves from the LTC-., LTC-. and the LTC- represent the range of performance across the entire famiy of references. Characteristic curves for other output votages fa between these curves and can be estimated based on their votage output. V Low Frequency.Hz to Hz Noise. V Output Votage Temperature Drift 3 TYPICAL UNITS V Load Reguation (Sourcing) nv/ DIV s/div G OUTPUT VOLTAGE (V) TEMPERATURE ( C) OUTPUT VOLTAGE CHANGE (ppm) 3 C C C.. OUTPUT CURRENT (ma) G G7 OUTPUT VOLTAGE CHANGE (ppm) 8. V Load Reguation (Sinking) C C C. OUTPUT CURRENT (ma) SUPPLY CURRENT (ma) 3 V Suppy Current vs Input Votage C C C 8 INPUT VOLTAGE (V) NOISE VOLTAGE (nv/ Hz) 8 8 V Output Votage Noise Spectrum..7µF µf µf. FREQUENCY (khz) G8 G9 G3 V Minimum - Differentia (Sourcing) V Minimum - Differentia (Sinking) V Start-Up Response with a 3.3µF Output Capacitor OUTPUT CURRENT (ma). OUTPUT CURRENT (ma). V/DIV V/DIV.. C C C. INPUT-OUTPUT VOLTAGE (V)..3 C C C... INPUT-OUTPUT VOLTAGE (V) C OUT = 3.3µF µs/div G33 8 G3 G3

9 PIN FUNCTIONS SHDN (Pin ): Shutdown Input. This active ow input powers down the device to <µa. If eft open, an interna pu-up resistor puts the part in norma operation. It is recommended to tie this pin high externay for best performance during norma operation. (Pin ): Power Suppy. Bypass with a.µf, or arger, capacitor to. (Pin ): Device Ground. This pin is the main ground and must be connected to a noise-free ground pane. _S (Pin ): Sense Pin. Connect this pin at the oad and route with a wide meta trace to minimize oad reguation errors. This pin sinks ma. Output error is R TRACE ma, regardess of oad current. For oad currents <µa, tie directy to _F pin. _F (Pin 7): Force Pin. This pin sources and sinks current to the oad. An output capacitor of.7µf to µf is required. (Pins 3,, 8): Interna Function. Ground these pins. BLOCK DIAGRAM SHDN BANDGAP _F 7 _S 3,,8 BD 9

10 APPLICATIONS INFORMATION Bypass and Load Capacitors The LTC votage references require a.µf or arger input capacitor ocated cose to the part to improve power suppy rejection. An output capacitor with a vaue between.7µf and µf is aso required. The output capacitor has a direct effect on the stabiity, turn-on time and setting behavior. Choose a capacitor with ow ESR to insure stabiity. Resistance in series with the output capacitor (ESR) introduces a zero in the output buffer transfer function and coud cause instabiity. The.7μF to μf range incudes severa types of capacitors that are readiy avaiabe as through-hoe and surface mount components. It is recommended to keep ESR ess than or equa to.ω. Capacitance and ESR are both frequency dependent. At higher frequencies capacitance drops and ESR increases. To insure stabe operation the output capacitor shoud have the required vaues at khz. In order to achieve the best performance, caution shoud be used when choosing a capacitor. X7R ceramic capacitors are sma, come in appropriate vaues and are reativey stabe over a wide temperature range. However, for a ow noise appication X7R capacitors may not be suitabe since they may exhibit a piezoeectric effect. The mechanica vibrations cause a charge dispacement in the ceramic dieectric and the resuting perturbation can ook ike noise. If X7R capacitors are necessary, a thorough bench evauation shoud be competed to verify proper performance. For very ow noise appications where every nanovot counts, fim capacitors shoud be considered for their ow noise and ack of piezoeectric effects. Fim capacitors such as poyester, poystyrene, poycarbonate, and poypropyene have good temperature stabiity. Additiona care must be taken as poystyrene and poypropyene have an upper temperature imit of 8 C to C. Above these temperatures, the working votages need to be derated according to manufacturer s specifications. Another type of fim capacitor is poyphenyene sufide (PPS). These devices work over a wide temperature range, are stabe, and have arge capacitance vaues beyond μf. In genera, fim capacitors are found in surface mount and eaded packages. Tabe is a partia ist of capacitor companies and some of their avaiabe products. In votage reference appications, fim capacitor ifetime is affected by temperature and appied votage. When poyester capacitors are operated beyond their rated temperatures (some capacitors are not rated for operation above 8 C) they need to be derated. Votage derating is usuay accompished as a ratio of appied votage to rated votage imit. Contact specific fim capacitor manufacturers to determine exact ifetime and derating information. The ifetime of X7R capacitors is ong, especiay for reference appications. Capacitor ifetime is degraded by operating near or exceeding the rated votage, at high temperature, with AC rippe or some combination of these. Most reference appications have AC rippe ony during transient events. Tabe. Fim Capacitor Companies COMPANY DIELECTRIC AVAILABLE CAPACITANCE TEMPERATURE RANGE TYPE Corne Dubier Poyester.µF to µf C to C DME Dearborn Eectronics Poyester.µF to µf C to C 8P, 3P, 3P, P, and P Tecate Poyester.µF to 8µF C to C 9, 9, and 9D Wima Poyester µf to µf C to C MKS, MKS -XL Vishay Poyester pf to µf C to C MKT8 Vishay Poycarbonate.µF to µf C to C MKC8, 3P Dearborn Eectronics Poyphenyene Sufide (PPS).µF to µf C to C 8P, 83P, 8P, 8P, and 88P Wima Poyphenyene Sufide (PPS).µF to.8µf C to C SMD-PPS

11 APPLICATIONS INFORMATION The choice of output capacitor aso affects the bandwidth of the reference circuitry and resutant noise peaking. As shown in Figure, the bandwidth is inversey proportiona to the vaue of the output capacitor. Noise peaking is reated to the phase margin of the output buffer. Higher peaking generay indicates ower phase margin. Other factors affecting noise peaking are temperature, input votage, and output oad current. NOISE VOLTAGE (nv Hz) 8 C OUT =.7µF C OUT = µf C OUT = µf LTC Start-Up and Load Transient Response Resuts for the transient response pots (Figures 3 to 8) were produced with the test circuit shown in Figure uness otherwise indicated... FREQUENCY (khz) F Figure. Output Votage Noise Spectrum The turn-on time is sew imited and determined by the short-circuit current, the output capacitor, and output votage as shown in the equation: t ON C = VOUT I OUT SC 3V, C IN.µF LTC-. 3,,,8 7 Ω C OUT 3.3µF Figure. Transient Load Test Circuit V GEN F.V For exampe, the LTC-.V, with a 3.3µF output capacitor and a typica short-circuit current of ma, the start-up time woud be approximatey: 33.. V. A F = µs The resuting turn-on time is shown in Figure 3. Here the output capacitor is 3.3µF and the input capacitor is.µf. Figure shows the output response to a mv step on. The output response to a current step sourcing and sinking is shown in Figures and, respectivey. V/DIV V/DIV C OUT = 3.3µF µs/div Figure 3. Start-Up Response F3 Figure 7 shows the output response as the current goes from sourcing to sinking. 3.V 3V Shutdown Mode The LTC famiy of references can be shut down by tying the SHDN pin to ground. There is an interna pu-up resistor tied to this pin. If eft unconnected this pin rises to and the part is enabed. Due to the ow interna pu-up current, it is recommended that the SHDN pin be pued high externay for norma operation to prevent accidenta mv/div C OUT = 3.3µF µs/div F Figure. Output Response with a mv Step On

12 APPLICATIONS INFORMATION ma I OUT ma mv/div C OUT = 3.3µF µs/div F Figure. Output Response with a ma Load Step Sourcing ma I OUT ma shutdown due to system noise or eakage currents. The turn-on/turn-off response due to shutdown is shown in Figure 8. To contro shutdown from a ow votage source, a MOSFET can be used as a pu-down device as shown in Figure 9. Note that an externa resistor is unnecessary. A MOSFET with a ow drain-to-source eakage over the operating temperature range shoud be chosen to avoid inadvertenty puing down the SHDN pin. A resistor may be added from SHDN to to overcome excessive MOSFET eakage. The SHDN threshods have some dependency on and temperature as shown in the Typica Performance Characteristics section. Avoid eaving SHDN at a votage between the threshods as this wi cause an increase in suppy current due to shoot-through current. mv/div C OUT = 3.3µF µs/div F Figure. Output Response with ma Load Step Sinking 3V 3.V TO µc C µf N7 _F LTC-. SHDN _S C µf ma I OUT ma Figure 9. Open-Drain Shutdown Circuit F9 mv/div SHDN V/DIV V/DIV C OUT = 3.3µF µs/div Figure 7. Output Response Showing a Sinking to Sourcing Transition F7 Long-Term Drift Long-term drift cannot be extrapoated from acceerated high temperature testing. This erroneous technique gives drift numbers that are widy optimistic. The ony way ong-term drift can be determined is to measure it over the time interva of interest. The LTC ong-term drift data was coected on 8 parts that were sodered into printed circuit boards simiar to a rea word appication. The boards were then paced into a constant temperature oven with a T A = 3 C, their outputs were scanned reguary and measured with an 8. digit DVM. Typica ong-term drift is iustrated in Figure a. The hermetic LS8 package provides additiona stabiity as shown in Figure b. C OUT = 3.3µF ms/div F8 Figure 8. Shutdown Response with ma Source Load

13 APPLICATIONS INFORMATION LTC LONG-TERM DRIFT (ppm) 8 TYPICAL UNITS LTC-. NUMBER OF UNITS 3 8 HOURS Figure a. Long-Term Drift F DISTRIBUTION (ppm) F Figure. Hysteresis Pot C to C ppm HOURS Fb Figure b. LT-.V CLCC Hysteresis Therma hysteresis is a measure of change of output votage as a resut of temperature cycing. Figure iustrates the typica hysteresis based on data taken from the LTC-.. A proprietary design technique minimizes therma hysteresis. Humidity Sensitivity Pastic moud compounds absorb water. With changes in reative humidity, pastic packaging materias change the amount of pressure they appy to the die inside, which can cause sight changes in the output of a votage reference, usuay on the order of ppm. The LS8 package is hermetic, so it is not affected by humidity, and is therefore more stabe in environments where humidity may be a concern. However, PC board materia may absorb water and appy mechanica stress to the LTCLS8. Proper board materias and ayout are essentia. Power Dissipation Power dissipation for the LTC depends on and oad current. Figure iustrates the power consumption versus under a no-oad and ma oad condition at room temperature for the LTC-.. Other votage options dispay simiar behavior. The MSOP8 package has a therma resistance (θ JA ) equa to 3 C/W. Under the maximum oaded condition, the increase in die temperature is over 3 C. If operated at these conditions with an ambient temperature of C, the absoute maximum junction temperature rating of the device woud be exceeded. Athough the maximum junction temperature is C, for best performance it is recommended to not exceed a junction temperature of C. The pot in Figure 3 shows the recommended maximum ambient temperature imits for differing and oad conditions using a maximum junction temperature of C. 3

14 APPLICATIONS INFORMATION POWER (W).. ma LOAD..8 NO LOAD.. on two sides of the device to reduce mechanica stress. A thicker and smaer board is stiffer and ess prone to bend. Finay, use stress reief, such as fexibe standoffs, when mounting the board. Additiona precautions incude making sure the soder joints are cean and the board is fux free to avoid eakage paths. A sampe PCB ayout is shown in Figure.. (V) F Figure. LTC-. Power Consumption F MAXIMUM AMBIENT OPERATING TEMPERATURE ( C) NO LOAD ma LOAD 3 9 (V) F3 Figure 3. LTC-. Maximum Ambient Operating Temperature PC Board Layout The LTC reference is a precision device that is factory trimmed to an initia accuracy of ±.%, as shown in the Typica Performance Characteristic section. The mechanica stress caused by sodering parts to a printed circuit board may cause the output votage to shift and the temperature coefficient to change. To reduce the effects of stress-reated shifts, mount the reference near the short edge of a printed circuit board or in a corner. In addition, sots can be cut into the board Figure. Sampe PCB Layout Load Reguation To take advantage of the Kevin force/sense pins, the _S pin shoud be connected separatey from the _F pin as shown in Figure. The _S pin sinks ma, which is unusua for a Kevin connection. However, this is required to achieve the exceptiona ow noise performance. The I R drop on the _S ine directy affects oad reguation. The _S trace shoud be as short and wide as practica to minimize series resistance The _S trace adds error as R TRACE ma, so a.ω trace adds µv error. The _F pin is not as important as the _S pin in this regard. An I R drop on the _F pin increases the minimum suppy LTC-. 7 ma MINIMIZE RESISTANCE OF METAL LOAD STAR F Figure. Kevin Connection for Good Load Reguation

15 APPLICATIONS INFORMATION votage when sourcing current, but does not directy affect oad reguation. For ight oading of the output (maximum output current <µa), _S shoud be tied to _F by the shortest possibe path to reduce errors caused by resistance in the sense trace. Carefu attention to grounding is aso important, especiay when sourcing current. The return oad current can produce an I R drop causing poor oad reguation. Use a star ground connection and minimize the ground to oad meta resistance. Athough there are severa pins that are required to be connected to ground, Pin is the actua ground for return current. Optima Noise Performance The LTC offers extraordinariy ow noise for a bandgap reference ony.ppm in.hz to Hz. As a resut, system noise performance may be dominated by system design and physica ayout. Some care is required to achieve the best possibe noise performance. The use of dissimiar metas in component eads and PC board traces creates thermocoupes. Variations in therma resistance, caused by uneven air fow, create differentia ead temperatures, thereby causing thermoeectric votage noise at the output of the reference. Minimizing the number of thermocoupes, as we as imiting airfow, can substantiay reduce these errors. Additiona information can be found in Linear Technoogy Appication Note 8. Position the input and oad capacitors cose to the part. Athough the LTC has a DC PSRR of over db, the power suppy shoud be as stabe as possibe to guarantee optima performance. A pot of the.hz to Hz ow frequency noise is shown in the Typica Performance Characteristic section. Noise performance can be further improved by wiring severa LTCs in parae as shown in the Typica Appications section. With this technique the noise is reduced by N, where N is the number of LTCs in parae. Noise Specification Noise in any frequency band is a random function based on physica properties such as therma noise, shot noise, and ficker noise. The most precise way to specify a random error such as noise is in terms of its statistics, for exampe as an RMS vaue. This aows for reativey simpe maximum error estimation, generay invoving assumptions about noise bandwidth and crest factor. Unike wideband noise, ow frequency noise, typicay specified in a.hz to Hz band, has traditionay been specified in terms of expected error, iustrated as peak-to-peak error. Low frequency noise is generay measured with an oscioscope over a second time frame. This is a pragmatic approach, given that it can be difficut to measure noise accuratey at ow frequencies, and that it can aso be difficut to agree on the statistica characteristics of the noise, since ficker noise dominates the spectra density. Whie practica, a random samping of second intervas is an inadequate method for representation of ow frequency noise, especiay for systems where this noise is a dominant imit of system performance. Given the random nature of noise, the output votage may be observed over many time intervas, each giving different resuts. Noise specifications that were determined using this method are prone to subjectivity, and wi tend toward a mean statistica vaue, rather than the maximum noise that is ikey to be produced by the device in question.

16 APPLICATIONS INFORMATION Because the majority of votage reference data sheets express ow frequency noise as a typica number, and as it tends to be iustrated with a repeatabe pot near the mean of a distribution of peak-to-peak vaues, the LTC data sheet provides a simiary defined typica specification in order to aow a reasonabe direct comparison against simiar products. Data produced with this method generay suggests that in a series of second output votage measurements, at east haf the observations shoud have a peak-to-peak vaue that is beow this number. For exampe, the LTC-. measures ess than.ppm P-P in at east % of the second observations. As mentioned above, the statistica distribution of noise is such that if observed for ong periods of time, the peak error in output votage due to noise may be much arger than that observed in a smaer interva. The ikey maximum error due to noise is often estimated using the RMS vaue, mutipied by an estimated crest factor, assumed to be in the range of to 8.. This maximum possibe vaue wi ony be observed if the output votage is measured for very ong periods of time. Therefore, in addition to the common method, a more thorough approach to measuring noise has been used for the LTC (described in detai in Linear Technoogy s AN) that aows more information to be obtained from the resut. In particuar, this method characterizes the noise over a significanty greater ength of time, resuting in a more compete description of ow frequency noise. The peak-to-peak votage is measured for second intervas over hundreds of intervas. In addition, an eectronic peak-detect circuit stores an objective vaue for each interva. The resuts are then summarized in terms of the fraction of measurement intervas for which observed noise is beow a specified eve. For exampe, the LTC-. measures ess than.7ppm P-P in 8% of the measurement intervas, and ess than.9ppm P-P in 9% of observation intervas. This statistica variation in noise is iustrated in Tabe and Figure 7. The test circuit is shown in Figure. Tabe Low Frequency Noise (ppm P-P ) %. %. 7%. 8%.8 9%.9 This method of testing ow frequency noise is superior to more common methods. The resuts yied a comprehensive statistica description, rather than a singe observation. In addition, the direct measurement of output votage over time gives an actua representation of peak noise, rather than an estimate based on statistica assumptions such as crest factor. Additiona information can be derived from a measurement of ow frequency noise spectra density, as shown in Figure 8. It shoud be noted from Figure 8 that the LTC has not ony a ow wideband noise, but an exceptionay ow ficker noise corner of Hz! This substantiay reduces ow frequency noise, as we as ong-term variation in peak noise.

17 APPLICATIONS INFORMATION IN 9V SD LTC.V REFERENCE UNDER TEST S F µf V µf A LT A = LOW NOISE PRE-AMP V k k 3µF SHIELD **.k.µf N97 V.µF µf k Q3 N97 V k* Ω* V Ω* 9Ω* Q A LT97 Q.µF k* k µf A = AND.Hz TO Hz FILTER A3 LT 33Ω* k* k* M* k* Ω*.µF.µF A LT 33µF V 33µF V AC LINE GROUND 7Ω* V INPUT SHIELDED CAN Ω* IN OUT 33µF V 33µF V ROOT-SUM-SQUARE CORRECTION SEE TEXT k * = % METAL FILM ** = % WIREWOUND, ULTRONIXA = N8 Q, Q = THERMALLY MATED SK39 (MATCH VGS %) OR LSK389 DUAL THERMALLY LAG SEE TEXT = TANTALUM,WET SLUG ILEAK < na SEE TEXT/APPENDIX B = N393 = POLYPROPELENE = / LTC A 33µF OUTPUT CAPACITORS = <na LEAKAGE AT VDC AT C SEE APPENDIX C FOR POWER, SHIELDING AND GROUNDING SCHEME F RST RST k P µf P µf T A / LT8 k k A / LT8.7k.7k k.µf.µf PEAK TO PEAK NOISE DETECTOR PEAK A7 / LT8.µF k DVM O TO V = O TO µv TO OSCILLOSCOPE INPUT VIA ISOLATED PROBE, V/DIV = µv/div, REFERRED TO INPUT, SWEEP = s/div.µf C RC RST = Q V CLR 7C A k RESET PULSE GENERATOR k B BAT-8 BAT-8 FROM OSCILLOSCOPE SWEEP GATE OUTPUT VIA ISOLATION PULSE TRANSFORMER PEAK A8 / LT8 k T P Figure. Detaied Noise Test Circuitry. See Appication Note. 7

18 APPLICATIONS INFORMATION NUMBER OF OBSERVATIONS PEAK-TO-PEAK NOISE (nv) F7 Figure 7. Low Frequency Noise Histogram of the LTC-. NOISE VOLTAGE (nv/ Hz) 8 shown in Figure 9, the output votage shifts. After the device expands, due to the heat, and then contracts, the stresses on the die have changed position. This shift is simiar, but more extreme than therma hysteresis. Experimenta resuts of IR refow shift are shown beow in Figure. These resuts show ony shift due to refow and not mechanica stress. TEMPERATURE ( C) 3 7 T S = 9 C T = C RAMP TO C 38s T P = C T L = 7 C T S(MAX) = C s 8 MINUTES t P 3s t L 3s s RAMP DOWN Figure 9. Lead-Free Refow Profie F9. FREQUENCY (Hz) 8 7 F8 Figure 8. LTC-. Low Frequency Noise Spectrum IR Refow Shift The mechanica stress of sodering a part to a board can cause the output votage to shift. Moreover, the heat of an IR refow or convection sodering oven can aso cause the output votage to shift. The materias that make up a semiconductor device and its package have different rates of expansion and contraction. After a part undergoes the extreme heat of a ead-free IR refow profie, ike the one NUMBER OF UNITS OUTPUT VOLTAGE SHIFT DUE TO IR REFLOW (%) F Figure. Output Votage Shift Due to IR Refow 8

19 LTC TYPICAL APPLICATIONS Extended Suppy Range Reference Extended Suppy Range Reference V TO 3V R BZX8C C.µF SHDN LTC-. _F _S Boosted Output Current C µf TA R k R.7k BZX8C C.µF ON SEMI MMBT SHDN _F LTC-. _S V TO 8V.µF C µf TA3 V TO 3.V TA C.µF LTC-. SHDN _F _S Q N C.7µF I MAX SET BY NPN.8V TO 3.V C3.µF Boosted Output Current C µf R Ω SHDN R k C µf N9 3mA MAX _F LTC-. _S C µf TA Output Votage Boost.V TO 3.V C µf _F LTC-. SHDN _S R R = k to k TA7.V TO.V C µf = VOLTAGE OPTION. R THIS EXAMPLE USES.V AS THE VOLTAGE OPTION FOR R USE A POTENTIOMETER THAT CAN HANDLE ma, IS LOW NOISE AND HAS A LOW TEMPERATURE COEFFICIENT Low Noise Precision Votage Boost Circuit.V TO 3.V C µf _F LTC-. SHDN _S = VOLTAGE OPTION ( R/R) THIS EXAMPLE USES.V AS THE VOLTAGE OPTION LT77 R k R k R LOAD TA8 R3 k FOR R AND R USE VISHAY TRIMMED RESISTOR ARRAY (VSR OR MPM). WITH A PRECISION ARRAY THE MATCHING AND LOW TC WILL HELP PRESERVE LOW DRIFT. R3 = R R V C µf 9

20 TYPICAL APPLICATIONS Low Noise Statistica Averaging Reference e N = e N / N; Where N is the Number of LTCs in Parae 3V TO 3.V C.µF LTC-. SHDN _F _S R 3.Ω C.7µF C9.7µF C3.µF LTC-. SHDN _F _S R 3.Ω C.7µF C.µF LTC-. SHDN _F _S R3 3.Ω C.7µF TAa C7.µF LTC-. SHDN _F _S R 3.Ω C8.7µF Low Frequency Noise (.Hz to Hz) with Four LTC-. in Parae nv/ DIV 3nV P-P.Hz to Hz s/div TAb

21 PACKAGE DESCRIPTION Pease refer to for the most recent package drawings. LTC MS8 Package 8-Lead Pastic MSOP (Reference LTC DWG # -8- Rev F).889 ±.7 (.3 ±.).3 (.) MIN (..3). ±.38 (. ±.) TYP. (.) BSC 3. ±. (.8 ±.) (NOTE 3) 8 7. (.) REF RECOMMENDED SOLDER PAD LAYOUT GAUGE PLANE.8 (.7). (.) DETAIL A NOTE:. DIMENSIONS IN MILLIMETER/(INCH). DRAWING NOT TO SCALE TYP.3 ±. (. ±.) DETAIL A SEATING PLANE.9 ±. (.93 ±.). (.3) MAX..38 (.9.) TYP. (.) BSC 3 3. DIMENSION DOES NOT INCLUDE MOLD FLASH, PROTRUSIONS OR GATE BURRS. MOLD FLASH, PROTRUSIONS OR GATE BURRS SHALL NOT EXCEED.mm (.") PER SIDE. DIMENSION DOES NOT INCLUDE INTERLEAD FLASH OR PROTRUSIONS. INTERLEAD FLASH OR PROTRUSIONS SHALL NOT EXCEED.mm (.") PER SIDE. LEAD COPLANARITY (BOTTOM OF LEADS AFTER FORMING) SHALL BE.mm (.") MAX 3. ±. (.8 ±.) (NOTE ).8 (.3) REF. ±.8 (. ±.) MSOP (MS8) 37 REV F

22 PACKAGE DESCRIPTION Pease refer to for the most recent package drawings. LS8 Package 8-Pin Leadess Chip Carrier (mm mm) (Reference LTC DWG # -8-8 Rev Ø) 8. ±. PACKAGE OUTLINE 7. ±. 3. ±.. SQ ±..8 SQ ±..7 ±. APPLY SOLDER MASK TO AREAS THAT ARE NOT SOLDERED. SQ ±.. SQ ±. 8. ±..9 ±. 8 R. REF PIN TOP MARK (SEE NOTE ) 7. REF 7. ±.. ±. 3 R. REF 3. TYP NOTE:.7 TYP. ALL DIMENSIONS ARE IN MILLIMETERS. DRAWING NOT TO SCALE 3. DIMENSIONS PACKAGE DO NOT INCLUDE PLATING BURRS PLATING BURRS, IF PRESENT, SHALL NOT EXCEED.3mm ON ANY SIDE. PLATING ELECTO NICKEL MIN.UM, ELECTRO GOLD MIN.3UM. SHADED AREA IS ONLY A REFERENCE FOR PIN LOCATION ON THE TOP AND BOTTOM OF PACKAGE. TYP. TYP LS8 9 REV Ø

23 REVISION HISTORY REV DATE DESCRIPTION PAGE NUMBER A / Votage Options Added (.,.8, 3., 3.3,.9,.), Refected Throughout the Data Sheet to B / Addition of mm x mm hermetic LS8 package Update to Eectrica Characteristics to incude LS8 package Addition of ong-term drift and hysteresis pots for LS8 package Addition of Humidity Sensitivity information Addition of Reated Parts,, 3,, 3, 3 3 Information furnished by Linear Technoogy Corporation is beieved to be accurate and reiabe. However, no responsibiity is assumed for its use. Linear Technoogy Corporation makes no representation that the interconnection of its circuits as described herein wi not infringe on existing patent rights. 3

24 TYPICAL APPLICATION Low Noise Precision -Bit Anaog-to-Digita Converter Appication.k R TD k V REF R REF Ω THERMOCOUPLE CH CH CH CH3 CH CH CH CH7 CH8 CH9 CH CH CH CH3 CH CH COM V V CC LTC9 MUXOUTN ADCINN MUXOUTP ADCINP SDI SCK SDO CS.µF SPI INTERFACE.µF nf Ω.k nf Ω / LTC.V / LTC 7.V.µF LTC _F SHDN _S 7 V REF REF REF BUSY EXT f O 3,,8 µf TA9 RELATED PARTS PART NUMBER DESCRIPTION COMMENTS LT 3 Precision Low Drift Low Noise Reference.% Max, ppm/ C Max, ppm (Peak-to-Peak) Noise LT3LS8 Precision Low Noise, Low Profie Hermetic Votage Reference.% Max, ppm/ C Max,.3µV P-P Noise, mm mm Hermetic Package LT Micropower Series References.7% Max, ppm/ C Max, ma Output Current LT Micropower Series Low Dropout.% Max, 3ppm/ C Max, ma Output Current LT79 Micropower Precision Series References.% Max, ppm/ C Max, ma Suppy, SOT3 Package LT Micropower Reference with Buffer Ampifier.% Max,.µA Suppy, SOT3 Package LTC Precision Low Drift Low Noise Reference.% Max, ppm/ C Max, C to C, MSOP8 LT Tiny Micropower Series Reference.% Max, ppm/ C Max, ma Output Current, mm mm DFN LTCLS8 High Precision, Buffered Votage Reference Famiy in mm mm Hermetic QFN Package.% Max Initia Error, ppm/ C Max Drift, Shutdown Current <µa, C to C Operation LTCLS8 Precision, Low Noise, High Output Drive Votage Reference Famiy in mm mm Hermetic QFN Package.ppm Peak-to-Peak Noise (.Hz to Hz, Sink/Source ±ma, ppm/ C Max Drift, C to C Operation LT REV B PRINTED IN USA Linear Technoogy Corporation 3 McCarthy Bvd., Mipitas, CA (8) 3-9 FAX: (8) LINEAR TECHNOLOGY CORPORATION 9