MN1010 Design Guidelines

Transcription

1 1 Introduction This document contains important technical information, design notes and helpful hints to assist the designer in achieving first time success in bringing up a design using the MN1010 GPS Receiver module. It contains design examples and suggestions on a wide variety of topics, including power supply connections and bypassing, special reset circuit requirements, RF interface design, shielding and filtering requirements, antenna considerations and other important subjects. 2 Power Supply The MN1010 GPS Receiver Module is designed to operate from two supply voltages, the main voltage being 1.8 volts, and a secondary low current supply voltage of 3.0 volts for the internal TCXO. Figure 1 Suggested Power Supply Circuit The schematic shown in Figure 1 is only a suggestion. A switching power supply is shown to improve battery life by increasing the power supply efficiency. If this is not an issue, then a dual linear regulator circuit would provide a lower cost solution. The MN1010 is not sensitive to the order in which the power supplies are sequenced, however, both supplies must be within the specified tolerance for at least 10 milliseconds before the RESET is released. 2.1 Main 1.8 volt supply The main 1.8 volts supply is fed into several voltage pins of the MN1010 GPS Receiver Module. Suitable decoupling and isolating of the individual power supply pins must be provided by external decoupling circuitry. In addition, optional power control may be added to select +1.8V_LNA and +3V pins to further reduce power consumption in standby and/or sleep modes. Micro Modular Technologies Pte. Ltd.

2 Either a switching power supply or a linear supply can be used to provide the main 1.8 volt power. The reference design schematic shows the decoupling necessary for use with a switching power supply which generates 50mVp-p of switching noise. In particular, the +1.8VRF (pin 34) is sensitive to switching power supply ripple. The main 1.8 volt power can be connected directly to the +1.8V lines, which power the digital portions of the MN1010. All other power supply lines (with the exception of VRF) can also be connected to the main 1.8 volt supply provided a 27 pf ceramic capacitor to ground is provided as close as possible to each of the remaining power pins. The +1.8VRF supply is further filtered by the 18 ohm series resistor and the parallel combination of 27pf, 1000pf and 4.7 uf capacitors to further reduce power supply switching noise. The 18 ohm resistor was selected to allow for around 50mV to 100mV of DC drop while providing 18 ohms of AC isolation. Due to the low switching frequency, ferrite beads and/or inductors will be of limited value in this application since the values would need to be very large, which would require large physical size as well. The schematic in Figure 2 illustrates how the power supply bypassing is configured. Figure 2 Recommended Bypassing Micro Modular Technologies Pte. Ltd. Page 2 of 14

3 2.2 LNA 1.8-volt supply In some cases it may be desirable to put the MN1010 into the minimum power sleep mode. This can be achieved by switching off the DC power, +1.8V_LNA, to the LNA. The RFEN pin is provided to drive a FET power switch for the +1.8VLNA supply. If this mode is not needed, then leave RFEN unconnected and connect +1.8V_LNA to the 1.8 volt supply along with the 27pF decoupling capacitor. A suggested circuit for this switching is shown in Figure TCXO 3.0-volt supply Figure 3 Power Control for 1.8-volt LNA The TCXO 3.0 volts supply needs to be clean to prevent unwanted FM and phase modulation onto the TCXO. If a switching power supply is used, then a 100 ohm resistor followed by the parallel combination of a 27 pf capacitor and a 4.7 uf capacitor would be suitable. However, since the current drain is very low, a tiny linear regulator may be more suitable. If so, the 4.7 uf capacitor and 100 ohm resistor could be eliminated. The user needs to insure the +3 volts is clean going into the MN1010. Excessive noise on the +3V supply would result in reduced C/Nos reported by the software, along with a position solution with additional wander. Figure 4 Power Control for 3-volt TCXO Micro Modular Technologies Pte. Ltd. Page 3 of 14

4 If the user desires to put the MN1010 into a very low power sleep state, the +3V supply into the MN1010 can also be switched. Figure 4 shows a suitable circuit using two transistors. If the user is using a small LDO, it is important to verify the LDO can be disabled and enabled with standard 1.8 volt logic levels. If the user does desire to switch the power to the TCXO, the RTC crystal and associated capacitors are required as the state machine which controls RFEN īs clocked by the RTC oscillator. 3 Reset Circuit Requirements Upon powerup, the RESET pin must be held low until 10 milliseconds after the 1.8-volt and 3-volt power supplies have stabilized to within proper tolerance. Upon powerdown or during power supply transients, the RESET pin must be pulled low before power drops below specification to prevent corruption of flash memory contents. Normally, due to the very light loading on the +3V supply, it will usually come into tolerance before the +1.8V supply and fall out of tolerance after the +1.8V supply. If this can be verified by design, then the supervisor need only be connected to the +1.8 volt supply. A suggested circuit for generating the reset is shown in Figure 5. It is not recommended to monitor only the +3V line or the supply voltage into the +1.8V power supply as that does not adequately protect the MN1010 during a power down condition. Pulling the RESET line low does not put the MN1010 into a low power state. The best method to put the MN1010 into a low power state is to issue the NMEA Sleep command. Pulling the RESET line low and then high will cause an MN1010 to exit the Sleep state, but the internal RTC will lose time whenever RESET is pulsed low. It is suggested the wake the MN1010 using a transition on the RX0 or RX1 inputs. Figure 5 Suggested Reset Circuit It is not possible to properly control the RESET line with a GPIO pin from the host microcontroller. While suitable pulldown resistors and software can generate a proper reset during power up, the RESET line will not be properly controlled during a power down situation as the host microcontroller will not have advance knowledge of the power down event. A voltage supervisor chip is absolutely required. Micro Modular Technologies Pte. Ltd. Page 4 of 14

5 Similarly, an RC network will not generate the proper signals on the reset pin and is not supported. The need for a proper reset signal can not be overemphasized. Generating an improper reset signal is the most likely cause of improper MN1010 operation. 4 Sleep Mode The standard software does support a software based sleep mode. The receiver is placed into sleep by issuing the NMEA sleep command (please refer to the Orion NMEA manual). All navigation processes are stopped, the RF put into a low power state and the baseband processor halted. If the +3V and +1.8V_LNA are not externally switched, the current consumption of the MN1010 will drop to approximately 10 ma. If both of these supplies are externally switched off (see sections 2 and 3), then the sleep mode current consumption would drop to around 200 ua. The MN1010 will awaken from sleep mode whenever a transition occurs on the RX0 or RX1 pin. Thus it is possible to send a NMEA command to waken the MN1010, but as the first character is lost (the first transition is actually lost to the software), the receiver will awaken, but not respond to the command. For sleep mode to function properly, it is important to make sure the 32KHz RTC is present. 5 Miscellaneous signals The MN1010 has a TPI and TPQ pin for use with factory testing. These pins must be individually tied to ground through separate 10K ohm resistors. Failure to do this will result in navigation failure. The SERBOOT line must be pulled high for proper startup. A 10Kohm pullup is sufficient. To reprogram the MN1010, tying SERBOOT to ground through a 100 ohm resistor and then pulsing RESET low will the MN1010 to enter programming mode. The user is strongly encouraged to provide test pads on their design to allow reprogramming of the flash. 6 Real-Time Clock (RTC) circuit The MN1010 contains a built-in 32KHz RTC to maintain time whenever the MN1010 is in the Sleep state. This allows the MN1010 to quickly reacquire satellites and enter navigation within the hot start time period assuming the receiver had been in sleep mode for less than two hours. The RTC circuit consists of a 12.5 pf 32KHz crystal and two capacitors (one on either side) returned to the +1.8V supply (see Figure 6). Do not return the capacitors to ground as this will cause RTC startup issues. Note that other non-rtc components are still required but have been omitted from Figure 6 for clarity. During customer product design, the user should test the startup characteristics of the RTC crystal by inserting a series resistor between the crystal/capacitor and the RTC_XIN input on the MN1010. The crystal would be suitable if the series resistor is 5 times the equivalent series resistance of the RTC crystal and the RTC oscillator can be verified to start over the design voltage and temperature margins. Micro Modular Technologies Pte. Ltd. Page 5 of 14

6 Figure 6 Internal RTC Circuit The internal RTC does not maintain time when the RESET line is pulsed low and then high, nor does it maintain time when power is removed to the MN1010. If either of these conditions is a requirement to maintain time, the MN1010 also supports the additional of an external RTC circuit as shown in Figure 7. Note the software only supports the Seiko S-35390A RTC IC when connected as shown. Note that other non-rtc components are still required but have been omitted from Figure 6 for clarity. Micro Modular Technologies Pte. Ltd. Page 6 of 14

7 Figure 7 External RTC Circuit While the Seiko S-35390A can operate from 1.5 volts to 5.5 volts, it is important to realize the logic levels expected by the external RTC are based upon the supply voltage. For this reason, the external RTC +1.8VBKUP must be between +1.5 volts and +2.2 volts for the MN1010 to recognize the device and read the stored time correctly. 7 RF Interface 7.1 RF Input The MN1010 GPS Receiver Module accepts a standard L1 GPS C/A code signal (from a passive or active antenna) on the RF Input pad of the module. If a passive antenna is desired, no additional circuitry is required. However if an active antenna is required, then suitable means for powering the active antenna must be provided external to the MN1010 GPS Receiver Module. The RF input is isolated from DC levels to a maximum of ±15 VDC. The noise figure of the MN1010 module is approximately 7 db. If a suitable high gain passive antenna is used (~5 dbic gain), then the receiver would operate correctly with maximum reported C/Nos in the mid 40s. Depending upon the end users application and accessibility to the sky, this may be suitable. Micro Modular Technologies Pte. Ltd. Page 7 of 14

8 However, for optimum performance, the MN1010 requires an active antenna, with a minimum gain of 16 db and a 1.5 db noise figure. If the antenna gain of the active antenna were the same 5 dbic at zenith, then reports C/Nos would be in the high 40s and low 50s. This provides more signal margin and allows use of some of the weaker satellites in the calculated position. Most active antennas require the DC supply voltage to be superimposed onto the RF output of the antenna. MMT recommends that a quarter wave stub be used to prevent disturbing the matching of the antenna and MN1010 module. The other end of the quarter wave stub should be AC grounded with a suitable microwave quality capacitor. The reference design shows the quarter wave stub in the antenna path. The quarter wave stub transforms an RF short at one end of the stub (which is created by a high quality microwave capacitor of approximately 27 pf) into an RF open at the other end of the stub. The length and width of the stub is determined by the printed circuit board material, the board stackup and the ground plane. DC is unaffected by the quarter wave stub. This means the quarter wave stub is a good means of supplying power to an active antenna without affecting the RF performance of the receiver. Several caveats are in order however. First, the DC must be clean. If there is any coupled AC noise (particularly centered around L1), it is carried directly into the input of the MN1010 and could reduce the sensitivity. In applications where the DC may be switched on or off, MMT recommends grounding the DC input side of the quarter wave stub rather than leaving it floating. Second, a means of current limiting the DC should be provided, particularly if the antenna is going to be removable, installable or exposed to the end user. A DC short on the antenna cable or substitution of a DC grounded passive antenna would cause excessive current to flow, most likely damaging the traces of the quarter wave stub Designing the quarter wave stub The quarter wave stub is a trace on the circuit board that gives a 90 degree phase shift over its length. One end of the quarter wave stub is terminated into the center of the non-grounded pad of the RF bypass capacitor. The other end of the quarter wave stub intersects the RF input to the MN1010. See Figure 8. Micro Modular Technologies Pte. Ltd. Page 8 of 14

9 Figure 8 Quarter Wave Stub schematic The trace impedance ideally should be greater than 100 ohms, although board geometry may make this difficult. Trace impedance is affected by board stackup, dielectric constant of the PCB and trace thickness. Many easy to use calculators are available from the internet to design trace impedances. Once the impedance is set, then set the length of the trace to give a 90 degree phase shift between the RF bypass capacitor and the location where the quarter wave stub intersects the main RF trace. Again, there are several easy to use calculators available on the web to perform this calculation. One such calculator can be found at Using FR-4 and a multilayer stackup of RF on the top layer and ground on the second layer yields values of around 28mm in length. Micro Modular Technologies Pte. Ltd. Page 9 of 14

10 The quarter wave stub can be folded to save board layout space. As a rough rule of thumb, the folds should be at right angles and when the trace doubles back it should be separated by at least five times the trace width. If the RF layer also has ground fill, it must be removed from around the quarter wave stub by a clearance of approximately 5 trace widths. The RF bypass capacitor should be a good quality microwave capacitor. The value is determined by setting the self-resonant frequency to as close to GHz as practical. For 0402 ceramic capacitors, this value will range between 18pF to 27pF depending upon capacitor vendor. At self resonant, the RF bypass capacitor has no reactive component, leaving a resistive component close to 0 ohms. The quarter wave stub transforms this RF short at one end of the stub into an RF open at the other end of the stub. Thus as the GPS frequency, the stub should have no affect on the matching Using a choke inductor to supply DC If there is no room for a quarter wave stub, a 47nH to 68nH inductor may be used in place of the quarter wave stub. The inductor self resonance must be higher in frequency than the GPS signal ( GHz). The use of an inductor will alter the MN1010 matching slightly and it is recommended to check it and select the final value of the inductor by using a network analyzer to check input return loss. The caveats of limiting the current and coupling noise into the front end that were mentioned in the section on the quarter wave stub also apply to using a choke. 8 Shielding and Filtering Requirements The MN1010 is designed to use a GPS signal that can be as low as -150 dbm. Any source of interference near in frequency to the GPS signal could potentially jam the MN1010 and disrupt reception of the signal. 8.1 Digital Emissions For proper system design, the GPS antenna needs to be shielded from any potential jamming source. For that reason, in most designs not containing a transmitter, it makes more sense to shield the digital portion of the product rather than the RF portion. This keeps the digital noise from radiating into the antenna and/or antenna feed lines. Generally, it is not necessary to provide additional shielding around the MN1010 and associated circuitry. It is important to note the GPS signal level is well below any regulatory emissions requirement for EMI and EMC. Thus while a product meets FCC class B or CISPR 22, it is possible the emissions from the product will still seriously impact the MN1010 performance. Excessive interference into the MN1010 via the antenna can result is low to very low reported C/Nos of the satellite signals and subsequent excessive TTFF times. Assuming an 18mm square patch antenna with good LNA, the reported C/Nos should be in the high 40s and low 50s. If the values are below this, then interference needs to be considered as a problem and resolved. This can also be checked by substituting an external active antenna and moving it closer to and away from the device and noting the change in reported C/Nos. If any improvement in signal is noted as the external antenna is moved away from the device, then additional shielding is required. Micro Modular Technologies Pte. Ltd. Page 10 of 14

11 8.2 RF Emissions If the product contains an RF transmitter or a second heterodyne receiver, then care must be taken to prevent overloading the front end of the MN1010 if simultaneous operation is required. This overloading can come from several sources. First, the input LNA of the MN1010 does not have a preselect filter and is fairly broad band. If for example a GSM transmitter (1.8 GHz) was close by, then the GSM signal could overload the LNA. The output of the LNA is going to be proportional to its input, and if the GSM signal so dominates, the GPS signal would be attenuated and sensitivity of the receiver would be reduced. The OEM designer would need to design suitable input filtering to the MN1010 to protect in this case. A second case occurs in the collocated transmitter. The power amplifier has both a gain and a noise figure. If we take an example of a power amp noise figure of 15 db and 30 db of gain, this would mean that the power amp radiates broadband noise approximately 45 db above thermal noise. This means the power amp alone could present a noise source in the GPS band of -129 dbm. While this would easily meet any regulatory emissions requirements, it would render the GPS receiver inoperative. In this case, a suitable filter must be placed on the output of the power amplifier of the collocated transmitter, not the GPS receiver, to avoid this case. 9 GPS Antenna Selection Currently there are several types of GPS antennas available for the user to chose from. Each type of antenna has both advantages and disadvantages which need to be carefully weighed in making a selection. In addition, most antenna types are available in both an active (includes built in LNA) and passive versions. When selecting the antenna it is important both to consider the characteristics of the GPS signal itself along with the characteristics of the antenna. The GPS signal is broadcast at GHz and comes from any of the GPS satellites from the sky. The receiver needs a minimum of four signals to compute a 3D position. Ideally, the antenna should have an unrestricted view of the sky. Certain locations may limit the visibility of the sky such as being close to a building, etc, so it is important that the product in which the antenna is installed does not further limitation to satellite visibility. The GPS signal is also right hand circularly polarized (RHCP) so best results are achieved under most conditions with a right hand circularly polarized antenna. Under severed obscuration where multipath signal reflections are present, a linearly polarized antenna my give better results under the assumption that a reflected signal is better than no signal. Antennas are specified by antenna type, antenna gain, antenna pattern, polarization and axial ratio. Antenna gain is typically considered to be the ratio of the signal level received by the antenna under consideration at zenith as compared to a theoretical isotropic radiator (equal signal level in all directions). The gain is measure in dbi (for a linearly polarized antenna) or dbic (for a circularly polarized antenna). The gain of an antenna is going to very depending upon elevation and azimuth of the signal source with respect to the antenna. Graphically plotting this variation results in an antenna pattern. The axial ratio of an antenna is a measure of the quality of its polarization. An axial ratio of 1 is perfect circular polarization, an infinite axial ratio in perfectly linear polarization. Micro Modular Technologies Pte. Ltd. Page 11 of 14

12 9.1 Patch Antennas Patch antenna are typically square or round ceramic elements with metallic plating on both sides, the top being the metallic antenna element and the bottom being the ground plane. Figure 9 Typical patch antenna If a patch antenna is selected, it is important that it be oriented such that the top surface of the antenna is horizontal with respect to the surface of the earth. Tilting the antenna away from the horizontal will result in an artificial obscuration of potentially visible satellites. While patch antenna are low cost and can provide good gain, it is important that the patch antenna be used with a proper ground plane. The antenna vendor can provide assistance in this area. In addition, patch antennas are detuned by the present of anything within its near field, such as a plastic cover. The antenna vendor can tune the antenna to compensate for this detuning. Micro Modular Technologies Pte. Ltd. Page 12 of 14

13 9.2 Helix Antennas Helix antennas are usually spirally wound onto a tubular ceramic piece (see Figure 10). For best performance, the helix antenna needs to be vertical with respect to the surface of the earth. Helix antennas do not require a ground plane, but may work better with one. 9.3 Chip Antennas Figure 10 Sarantel helix antenna (cover removed) Chip antennas (Figure 11) are the smallest antenna available for GPS and are quite popular in small handhelds. However, chip antennas are linearly polarized making them more receptive to multipath signals which would degrade the computed position in some cases. Chip antennas also have very specific ground plane requirements. The antenna vendor can provide assistance in this area and can possibly tune the chip for a specific application. Figure 11 Chip Antenna Micro Modular Technologies Pte. Ltd. Page 13 of 14

14 10 Notices All reference and informational documents (including marketing information, specifications, reference designs, etc.) are provided for information only and are subject to change without notice. Reasonable efforts have been made in the preparation of these document to assure their accuracy, however Micro Modular Technologies Pte. Ltd. assumes no liability resulting from errors or omissions in these, or any document, or from the use of the information contained herein. Micro Modular Technologies Pte. Ltd. reserves the right to make changes in the product design and specifications as needed and without notification to its users. Please check our website for the most current documentation. All information contained herein is the property of Micro Modular Technologies Pte Ltd. and may not be copied or reproduced, other than for your information, without prior written consent. 11 Contact Information Corporate Headquarters Micro Modular Technologies Pte. Ltd. No. 3, Ubi Avenue 3, #05-01 Crocodile House, Singapore Tel: (65) Fax: (65) Americas and Europe Micro Modular Technologies Americas Creekside Lane Longmont, CO 80503, U.S.A. Tel: (1) Fax: (1) For a list of Regional Sales Representatives, please see our web page: Document no: MN1010-DG Micro Modular Technologies Pte. Ltd. Page 14 of 14