AD9361 output waveform amplitude is different

dreamandactual

AD9361 output waveform amplitude is different [Copy link]

 

   I generate a 1MHZ signal through DDS and transmit it through AD9361, then use AGC mode, FDD, LVDS interface, and check the register to see if 17 is equal to 1A, but the waveform is always different in amplitude. I don't know why, and then the waveform will be normal during the program download. The following is the transmission waveform, the image before and after the enlargement, please analyze the reason.

GEEK-13

Pay attention,

bqgup

Just like the AM modulated wave, the high frequency signal will carry the low frequency signal on itself and transmit it. The waveform is very similar to the first two pictures you uploaded.

dreamandactual

bqgup posted on 2019-3-3 09:22 It is similar to the AM modulation wave. The high-frequency signal will carry the low-frequency signal on itself and transmit it. The waveform is very similar to the first two pictures you uploaded

Yes, it does make sense. I use the 1MHz signal generated by DDS and then transmit it through AD9361. The transmission frequency of AD9361 is 600MHz. But how should this problem be solved?

gmchen

Your phenomenon is that the signal is modulated by a certain signal. The output amplitude of DDS usually does not change. The problem lies in the later stage. You'd better draw a block diagram of the system and indicate the position where each waveform is measured. The current description is difficult for others to understand.

dreamandactual

gmchen posted on 2019-3-3 10:02 Your phenomenon is that the signal is modulated by a certain signal. The output amplitude of DDS usually does not change. The problem lies in the later stage. You'd better draw a system...

This is a simple block diagram. The problem is that the DDS entering the AD9361 is correct, but it is incorrect when it comes out of the AD9361. I just don't know why the AD9361 outputs this waveform. Is it a problem with some register configuration?

gmchen

dreamandactual published on 2019-3-3 10:32 This is a simple block diagram. The problem is that the DDS entering the AD9361 is correct, but it is incorrect when it comes out of the AD9361. I don’t know why the AD9361 comes out...

1. The waveform you see is not a 1MHz modulated signal, but a 600MHz modulated wave. From the third picture, the frequency of the 600MHz signal has not changed, so your waveform should be an AM wave, and the amplitude of the AM wave should be changing. 2. If a 1MHz signal modulates a 600MHz carrier, then the envelope of the modulated wave should be 1MHz. But from the second picture, the envelope of the modulated wave is obviously not 1MHz, but like an interference signal. So you should focus on checking the AD9361 to see if there are any errors in the hardware and software.

bqgup

dreamandactual posted on 2019-3-3 10:32 This is a simple block diagram. The problem is that the DDS going into the AD9361 is correct, but coming out of the AD9361 is incorrect. I just don't know why the AD9361 comes out...

To see if the waveform generated by the DDS requires a DA module, the waveforms sent out after the FPGA configuration are all connected to the oscilloscope using the DA output. I suggest that the host test the frequency you measured with the oscilloscope to see if the carrier frequency is the frequency sent by your DDS.

dreamandactual

gmchen published on 2019-3-4 13:36 1. The waveform you see is not a 1MHz modulated signal, but a 600MHz modulated wave. From the third picture, the frequency of the 600MHz signal has not changed, so...

Yes, yes, I configured a sine wave in FPGA, and then transmitted it through AD9361, and then I transmitted it with a 600MHZ signal configured by the official software. I also suspect that there is a problem with my AD9361 register configuration, but I haven't found it yet. As for the hardware, I used the official board.

dreamandactual

bqgup posted on 2019-3-4 14:45 To see if the waveform generated by DDS requires a DA module, the waveforms sent out after FPGA configuration are all connected to the oscilloscope using DA output. I suggest that the host...

Yes, the oscilloscope frequency is indeed correct, 600Mhz, and it is transmitted through AD9361. Because my purpose is to generate a sine wave by myself, and then generate 2.4ghz through AD9361 and transmit it. Because my oscilloscope can only test 1g, so I chose 600mhz.

gmchen

dreamandactual posted on 2019-3-4 15:21 Yes, I configured a sine wave in FPGA, and then transmitted it through AD9361, and then I configured a 600MHZ signal through the official software...

I suggest you turn off the output of DDS first, or more directly ground the corresponding port, so that the output of AD9361 should be a constant amplitude wave. Check whether this constant amplitude wave is normal?

gmchen

In addition, AD9361 has multiple input ports. Is it possible that the unused ports are left hanging there and then modulated by some interference? Can all the unused ports be grounded?

dreamandactual

gmchen posted on 2019-3-4 20:57 In addition, AD9361 has multiple input ports. Is it possible that the unused ports are left hanging there, and then some kind of interference causes modulation...

Well, I have tried not to use DDS output, but if I don't give DDS output, there will be no waveform output. As for the port, I used differential LVDS and used several fixed ports on the official website. I am a little skeptical whether the output itself should be like this, because my DDS 1mhz is used as the modulation signal, and my carrier is 600mhz.

gmchen

dreamandactual published on 2019-3-5 15:23 Well, I tried not to use DDS output, but if I don't use DDS output, there will be no waveform output. Then for the port, I used differential LVDS and the official website fixed...

Wrong. Not using DDS output does not mean that the modulation signal is 0. The port originally connected to the DDS output should be connected to a DC signal (if it is a differential signal, it should be connected to a DC with equal positive and negative), and you should be able to see a 600MHz equal amplitude oscillation output.

dreamandactual

gmchen published on 2019-3-6 14:08 Wrong. Not using DDS output does not mean that the modulation signal is 0. The port that was originally connected to the DDS output should be connected to the DC signal (if it is a differential signal, then...

Yes, it is a differential signal, and then I directly give 12'h3f to the DDS port, and then I can't see any waveform display on the oscilloscope. If it is a DC with equal positive and negative, how should I give it? The data is 12 bits, and the differential is 1 to 6 bits.

zhuyebb

In addition, AD9361 has multiple input ports. Is it possible that the unused ports are left hanging there and then suffer some interference?

dreamandactual

zhuyebb posted on 2019-3-6 18:43 In addition, AD9361 has multiple input ports. Is it possible that the unused ports are left hanging there and then suffer some kind of interference?

There are indeed a few pins missing, such as enable, txnrx, but these signals seem to be useful only in TDD mode, and then I configured the register for FDD.

gmchen

dreamandactual published on 2019-3-6 17:20 Yes, it is a differential signal, and then I directly give the DDS port 12'h3f, and then I can't see any waveform display on the oscilloscope. It is a DC with equal positive and negative...

In order to avoid giving wrong instructions due to unfamiliarity with the chip, the most direct and crudest way is to disconnect the connection between DDS and AD9361, and then directly connect the positive and negative DC regulated power supply to the corresponding port of AD9361. In order to avoid damage to the chip due to misoperation, you can connect a resistor in series to each of the two ports of AD9361 before connecting to the power supply, or use a suitable resistor to divide the voltage before connecting to the power supply.

gmchen

dreamandactual posted on 2019-3-6 19:42 There are indeed a few pins that are missing, such as enable, txnrx, but these signals seem to be useful in TDD mode, and I configured the register for FDD.

Can those unused input pins be short-circuited to ground?

dreamandactual

gmchen posted on 2019-3-7 10:34 Can all the unused input pins be short-circuited to ground?

Yes, I can try it.