AD9910 amplitude VS temperature

Hi, everyone, I have developed a PCB board with eight AD9910s. However, I have some problems which will be listed below. Would you please give me some help? Thanks in advance.

  1. The temperature of the AD9910 can reach up to 60°C when the frequency of the output signal is 98.6MHz. Have you ever encounter similar phenomenon?
  2. The crucial problem which bothers me a lot is that the amplitude of the output signal varies when AD9910 heated up. The frequency of the output signal is same as 98.6MHz.

As to the second problem, I have searched the DDS>Q&A section and the related links will be given at last. From the links, I learned about two things. The first one is that the temperature drift of the resistor which connected to "DAC_RSET" pin should be as low as possible.  The second one is that the power supply for AVDD and DVDD should be stable. Based on those information, I have made some improvements of my design. First, I have used a very low temperature drift resistor( Temperature coefficient is 0.2 ppm/°C ) for "DAC_RSET" pin. Second, I have used six LDOs to respectively power the DVDD_3.3V,  AVDD_3.3V,  DVDD_1.8,  AVDD_1.8V(pin 6),   AVDD_1.8V(pin 3), AVDD_1.8V(pin 89, 92). The output pins 80, 81 of AD9910 are connected to ADT1-1WT's pin 4 and pin 6 as the below figure shows. And I measured the power of the signal on pin 3 of the ADT1-1WT with a frequency spectrograph. The measured power varies from -2.54dBm to -3.17dBm from 08:22am to 11:40am. However, we would ideally like an amplitude(power) drift of less than 1%. Now, I have purchased a low temperature drift(1 ppm/°C) voltage reference for DVDD_3.3V. I will post the new results when the new measured data is available. 

Lastly, I am wondering whether the amplitude(power) drift of AD9910 can meet the 1% requirement.  Please forgive me for the poor English.

 

AD9910 Temperature stability - Q&A - Direct Digital Synthesis (DDS) - EngineerZone (analog.com)

Output amplitude change with temperature (AD995x) - Q&A - Direct Digital Synthesis (DDS) - EngineerZone (analog.com)

The amplitude of sine wave in AD9854 is unstable! - Q&A - Direct Digital Synthesis (DDS) - EngineerZone (analog.com)

Ad9910 temprature problem - Q&A - Direct Digital Synthesis (DDS) - EngineerZone (analog.com)

DDS amplitude - Q&A - Direct Digital Synthesis (DDS) - EngineerZone (analog.com)

AD9951/AD9859 temperature-depended stability issue - Q&A - Direct Digital Synthesis (DDS) - EngineerZone (analog.com)

AD9834 - How to improve the accuracy and temperature stability - Q&A - Direct Digital Synthesis (DDS) - EngineerZone (analog.com)

AD9833 DDS chip - Output Amplitude Stability - Q&A - Direct Digital Synthesis (DDS) - EngineerZone (analog.com)

  • Per the 1st thread you noted in your post, the variation appears to be ~2.3%. But that is the result of variation over the full temperature range and across multiple parts. It may be possible to get a particular part to be within 1%, assuming the temperature variation is not too great. However, there is no guarantee.

    If <1% temperature stability is a requirement, you may need to incorporate a temperature sensor and a gain compensation circuit to mitigate the drift of the AD9910. Because we do not specify a thermal coefficient (which is likely different from part to part), you would need to characterize and model the thermal behavior of the part(s), and implement the appropriate thermal compensation. Unfortunately, there is no simple solution to stabilizing output power over temperature.

    You might try employing a heat sink. All things being equal, this would lower the upper temperature reached by the die, which effectively reduces the range of temperature variation. A heat sink may get you one step closer to your 1% target.