Preamplifier Before CN0417(Rev C)

Hi all,

I have a setup like this;

PlutoSDR Tx (Freq: 2.45 GHz, Pout: -20 dBm) ------TSS-53LNB+ LNA (Gain: 19 dB, Pout: -1 dBm) ------> CN0417 (Gain: 21 dB, Pout: 20 dBm)

With this setup, the EVM is around %1.7, without LNA, around %4. So this is why I used LNA to reach -1 dBm Pout. The tx signal is single carrier by the way.

But yesterday I burned my two CN0417s in the test setup. I've measured the CN0417(with beeper), the USB power stage seems OK.

Then I removed the E1 ferrite to disconnect power stage and RF stage. LTM8045 seems OK. I get clear 5V. But the whole RF stage is short circuit to GND.

Then I removed the ADL5606 and measured the PCB again. The RF filter(FL1) and all capacitors are OK except CIN.

Also, ADL5606 is short circuit from RF_IN to RF_OUT between GND (measured after desoldered from PCB).

I do not know how this happens, because I did not exceed the maximum ratings of ADL5606 (maximum Pin: 18 dBm). Even if I exceed, the LTM8045 gives maximum 450mA

and the ADL5606 consumes 540mA to run around 30dBm Pout. So the LTM8045 should not supply this. And maybe should burned because of overload (but seems OK).

Maybe CIN cap burned first then it burned ADL5606 but is a 50V capacitor. I should not be burned easily.

Could you help me to realize why and how this happens?

Thanks.

LNA IC: https://www.minicircuits.com/WebStore/dashboard.html?model=TSS-53LNB%2B

Parents
  • 0
    •  Analog Employees 
    on Oct 30, 2020 8:54 AM

    Hello,

    I have some questions about your setup:

    • I noticed you have some EVM results. Does this mean the setup you had was working properly until the ADL5606 burned like you mentioned?
    • Does the ADL5606 get burned right away? Or does it happen after some time?
    • Where is the output of the CN0417 connected to?
    • On the Pout and gain values you indicated, are those theoretical values or actual?
    • Does the CN0417 get burned when your setup is just Pluto + CN0417? What about if bench RF source + CN0417?

    It would be nice if you could attach a picture of your setup as well. It could help us debug the source of the issue.

    Trisha

  • Hi Trisha,

    Thanks for your answer. Regarding your questions my comments are below;

    • Yes. I've worked 2 hours at least with the setup. There was no issue. Also, there was ho overheating.(Red LED is never turned on)
    • It happened after some time.
    • The output was connected to the spectrum analyzer(Tektronix RSA507A).
    • All values are practical. I've measured via the spectrum analyzer.
    • I have not worked too much with PlutoSDR + CN0417 setup. Just measured EVM values and disconnect it. Also, I did not use and tested with an RF signal source.

    Regarding test setup. I will send a picture of my test setup in a few hours. Just would like to mention that. The LNA board is Nuand BT-200. I've disconnected the FB1 ferrite to disable bias-tee and supply from an external source.

  • 0
    •  Analog Employees 
    on Nov 4, 2020 7:31 AM in reply to psk

    Hi psk,

    Thanks for responding to my questions and for providing the image of your setup. We haven't encountered this issue before so I'm gonna be asking more questions.

    1. Is the power always on for the CN0417? Supplying an RF input (-1 dbm @ 560 mVpp) before powering up the board at a certain period of time might be enough to turn on the chip's ESD structures and damage the part.
    2. What is your power sequence for this setup?
    3. Where is the CN0417 powered up from? Was it a stable source with no power interruptions and capable of at least 500 mA?
    4. Are we positive the input to the CN0417 from the LNA never reaches the 18 dbm limit?

    setup diagram

    I've been trying to recreate your issue today but my setup is without the LNA. My setup is just SMA100A signal generator -> CN0417 -> MXA Signal Analyzer with single tone input signal at 2.45GHz, -1 dBm. I got a gain of 18. I've tried to run the system for more than two hours, the board didn't get busted and was amplifying okay. 

    close up

    Regards

    Trisha

  • Hi Trisha,

    Thanks for trying to reproduce the issue. My answers to your questions are below;

    1. Yes, the power was always on.
    2. Power up the amplifiers first then connect to PlutoSDR to USB. Since I've replaced the firmware of the PlutoSDR, the TX attenuation is -89 dB at startup. And it will never change until I start the modulation.
    3. CN0417 is supplied from the native(not a USB HUB) USB3.0 port of my PC.
    4. I've measured the output of the LNA before connecting to the CN0417. I've tuned the LNA to -1/0 dBm output. So I think there will be no limit exceed. I will be adding the Pout plot of the LNA today.

    Muhammad.

  • Hi Trisha,

    The output of the LNA is added below. I get a clear -20 dBm output from PlutoSDR. Approximately 1 dB cable loss involved.

  • Here is the output of the PlutoSDR. -19.5 dBm. Again approximately(actually less than) 1 dB cable loss involved.

  • Hi Trisha,

    I think I've found why this happens. The PlutoSDR runs tx_quad calib when LO is changed to 100 MHz below/above from the current frequency. When doing tx_quad_calb it overrides the current attenuation. It is doing the calibration at the minimum attenuation. I've measured my PlutoSDR's(without LNA) output via the spectrum analyzer and captured the below output in Max Hold mode.

    In the above image, there was a ~0 dBm peak when doing tx_quad calib. It suddenly disapears after calibration is done. So, this is why I need to capture in Max Hold mode. Here is the screenshot with modulation. It starts after the configuration is done.

    My app which runs on PlutoSDR, configures the AD9363 like this;

    1. set_tx_attenuation(-89);
    2. set_sampling_frequency(3e6);
    3. set_rf_bandwidth(3e6);
    4. set_tx_lo_frequency(2.45e6);
    5. set_tx_lo_powerdown(disable);
    6. set_tx_attenuation(-20);
    7. set_fir_filter();
    8. start_modulation();

    According to my test setup, the LNA output should not reach the 0 dBm as it will be dangerous for the CN0417. But when doing the tx_quad calib the PlutoSDR's output is already 0 dBm, so at this time, the output of the LNA is around 20 dBm. And this signal goes to the CN0417. And this level reaches the maximum input power(Pin=18 dBm) of the ADL5606.

    Actually, the tx_quad calib process finishes very quickly. So the ADL5606 should not be burn as this process is very fast. Some amplifier ICs works under maximum levels for a few minutes. But I agree with working under the maximum levels is not guaranteed.

    Muhammad.

Reply
  • Hi Trisha,

    I think I've found why this happens. The PlutoSDR runs tx_quad calib when LO is changed to 100 MHz below/above from the current frequency. When doing tx_quad_calb it overrides the current attenuation. It is doing the calibration at the minimum attenuation. I've measured my PlutoSDR's(without LNA) output via the spectrum analyzer and captured the below output in Max Hold mode.

    In the above image, there was a ~0 dBm peak when doing tx_quad calib. It suddenly disapears after calibration is done. So, this is why I need to capture in Max Hold mode. Here is the screenshot with modulation. It starts after the configuration is done.

    My app which runs on PlutoSDR, configures the AD9363 like this;

    1. set_tx_attenuation(-89);
    2. set_sampling_frequency(3e6);
    3. set_rf_bandwidth(3e6);
    4. set_tx_lo_frequency(2.45e6);
    5. set_tx_lo_powerdown(disable);
    6. set_tx_attenuation(-20);
    7. set_fir_filter();
    8. start_modulation();

    According to my test setup, the LNA output should not reach the 0 dBm as it will be dangerous for the CN0417. But when doing the tx_quad calib the PlutoSDR's output is already 0 dBm, so at this time, the output of the LNA is around 20 dBm. And this signal goes to the CN0417. And this level reaches the maximum input power(Pin=18 dBm) of the ADL5606.

    Actually, the tx_quad calib process finishes very quickly. So the ADL5606 should not be burn as this process is very fast. Some amplifier ICs works under maximum levels for a few minutes. But I agree with working under the maximum levels is not guaranteed.

    Muhammad.

Children
  • 0
    •  Analog Employees 
    on Nov 17, 2020 2:36 AM in reply to psk

    Hi Muhammad,

    I've reached out to the expert on Pluto about this issue. It appears that what you see is an expected behavior of the AD936x. The calibration process can be disabled but will result to bad images on TX.

    He can provide more detailed explanation on this. I've already tapped him about this.

    Regards,

    Trisha