Radio frequency debugging includes two major aspects: transmission and reception. Transmission includes transmission power, phase error, PVT, switching spectrum, modulation spectrum debugging, etc., and reception includes sensitivity, RXloss, and receiving level.
1. Clarify the status of various radio frequency parameters of the mobile phone
How to clarify the radio frequency status of the mobile phone: first calibrate the mobile phone. The initialization file of the calibration must use the original file provided by MTK (see Figure 1), so that the matching status of the mobile phone can be clarified, such as whether the power is flat in the whole frequency band (if weighted Can't see), whether the match between TC and PA is done (it is difficult to judge if the value of Ramp curve is changed). Take GSM 900M as an example. If the output power of channel 1, channel 62, and channel 124 differ by more than 0.5dBm, then the PA output matching and the matching between TC and PA have not been done well. At this time, if the switching spectrum and phase error are good, then only You need to adjust the PA output matching (note that the output matching may affect the phase error). If the switching spectrum and phase error are very poor, then you must first adjust the matching between TC and PA, and then adjust the PA output matching .
Judgment criteria for switching spectrum: at high power level, there should be at least 7dB margin, under low power level, there should be more than 10-15dB margin, especially the margin of +/-0.4MHz and +/-0.6MHz. the amount.
Judgment criteria for phase error: GSM can achieve RMS around 1°, DCS around 2°; the same should be done to ensure that the phase error of the whole frequency band cannot be too different.
Second, impedance matching
An important step in RF debugging is impedance matching. If there is no problem with the PCB board, as long as the matching is done, the RF parameters are basically adjusted. Therefore, the principle of impedance matching and the debugging process are explained here.
To maximize the power transmitted from the signal source to the load, the impedance of the signal source must be equal to the conjugate impedance of the load (ie, conjugate matching). If the characteristic impedance of the transmission line is not equal to the load impedance (ie does not match), the There will be reflections, and the formation of standing waves on the transmission line will reduce the effective power capacity of the transmission line; the power will not be transmitted, and it will even damage the transmitting equipment. If the high-speed signal line on the circuit board does not match the load impedance, it will cause oscillation, radiation interference, etc. The performance of the impedance mismatch on the mobile phone is: the transmission power cannot be increased, the power consumption is large, the phase error and the switching spectrum are exceeded; the receiving path loss is large and the sensitivity is low.
Maximum power transfer theorem:
Let Zi = Ri + jXi, ZL = RL + jXL
The condition for the load to obtain the maximum power is
ZL = Ri-jXi = Zi *
(ie RL = Ri, XL = -Xi )
When the above formula is established, we call the load impedance and power supply Impedance conjugate matching, referred to as load and power matching.
The maximum power obtained by the load is Pmax = (Us^2) / (4Ri)
PA matching (including output matching, see Figure 2 and input matching, see Figure 3) The general steps of debugging: the
first method without a network analyzer:
first remove the parallel components on the path, and then replace the components in the path with 0Ω Push the power after the resistance (see appendix) and record the result. Then follow the principle of adjusting the series first, then the parallel, the inductors are connected in series and the capacitors are connected in parallel. When in series, the inductance value is adjusted from small to large, the capacitance value is adjusted from large to small, and the parallel is the opposite. During the debugging process, follow the characteristic impedance formula of Zo=Sqrt(L/C), and know the change trend of L, then the change trend of C is also clear, and vice versa. For example: if the inductance L110 connected in series at the output of the PA GSM changes from 0 to 2.2nH, the output power of the PA increases, then the characteristic impedance Zo of the transmission line needs to be increased to better match the PA, then the change trend of the parallel capacitor C151 It is to reduce. This debugging method should be good at analyzing the tested data and judging the trend of impedance changes.
The second method is to use a network analyzer:
first remove the parallel components and PA on the path, then replace the components in the path with 0Ω resistors, and use an SMA head (see Figure 4-7) to adjust the path that needs to be debugged Connect to the network branch, connect to META and set the mobile phone to continue TX mode. At this time, the transmission path is long. The original impedance of the channel is measured by the network points, and the impedance values of three points are generally recorded (high, medium and low channels). Next, you can use the Smith chart to simulate the impedance matching process: series reactance is positive for inductance and negative for capacitance, while parallel susceptance is positive for capacitance and negative for inductance. The positive and negative impedances are shown on the circle chart as clockwise and counterclockwise rotation (see Figure 8-11). Match the impedance to the 50 ohm point of the Smith chart through series and parallel inductors and capacitors, and then change the matching state of the circuit according to the simulation results, and then test whether it has been matched to around 50 ohms on the network sub-point, and then repeat the fine-tuning several times to complete Matched debugging.
SMA type: We generally use a small section of coaxial wire to be welded to the female (Female), connect the ground of the coaxial cable to the ground of the female, and then connect to the mesh through the SMA adapter; The other end is soldered to the path that needs to be adjusted on the PCB board (usually soldered to the pad of the PA), and then the ground wire of this end is soldered to the ground of the PCB. (Figure 7)
When the PA outputs high power, it is not necessarily high efficiency. When high efficiency, the output power is not large but the power consumption is low at this time. The matching of the PA output should be adjusted to the high efficiency and high power point of the Loadpull diagram (see Figure 12). Compromise position (on the two-point connection), pay attention to three points when debugging: APC DAC value, current and power. It is required that under normal temperature and pressure, the maximum output power of the PA of all GSM channels should be greater than 33.4dBm, and the power of all DCS channels should be greater than 30.4dBm; under low-voltage conditions, the power of all GSM channels should be greater than 33.0dBm. The power of all channels of the DCS must be greater than 30.0dBm. Or adjust according to the actual requirements of customers. However, when adjusting the matching circuit, both PAE must be taken into account, that is, the maximum output current consumed by the PA must meet customer requirements, and the maximum call current value of various PAs can also be different. If other power consumption is not included, RF3166 It is 250mA, SKY77328 is 250mA, and Renases 8015B is 260mA.
Three, receive debugging match
First calibrate the mobile phone, record the value of the RXloss of the receiving channel, first change the size of the serial device on the channel, and check the trend of RXloss (see appendix). Also follow the characteristic impedance formula of Zo=Sqrt(L/C), find the rule and set the series connection After the device, fine-tune the parallel device. The value of the capacitor generally varies between 0.5P-33P, and the inductance value should not be greater than 22nH. The adjustment of the L-type matching of the receiving path is shown as the rotation of the impedance diagram on the Smith chart, and the π-type matching is shown as the reduction and enlargement of the impedance diagram. Therefore, the L-shape in front of the receiving sound table has a greater impact on the reception. Generally, the matching before the sound table is adjusted first. (See Figure 13)
For example: the
original state of receiving RX Loss in Figure 13 is
[GSM900 Sub band, RX loss]
Max ARFCN=15,30,40,50,60,70,80,95,110,124,1000,1023
RX loss= 0,0,-0.125,-0.25,-0.25,-0.25,-0.125,0.125,0.25,0.25,-0.125,-0.375
Increase the L-shaped matching series capacitance C107 in front of SAW101 from 1.5nH to 4.7nH, and re After calibrating RX Path Loss, Rx Loss becomes:
[GSM900 Sub band, RX loss]
Max ARFCN=15,30,40,50,60,70,80,95,110,124,1000,1023
RX loss= -0.625 ,-0.5,- 0.625, -0.75, -0.75, -0.75, -0.625, -0.5, -0.375, -0.25, -0.5, -0.875
It can be seen from the data that there is a significant improvement, then it can be inferred from Zo=Sqrt(L/C) that the characteristic impedance of the transmission line between the antenna switch and SAW101 is too small. After increasing the value of L, Z0 becomes larger to achieve better impedance match.
The information is reproduced in Yiniu Forum (http://bbs.16rd.com/forum-261-1.html)