MXA N9020A绝对功率幅度有问题

我在MXA N9020A频谱分析仪上测量绝对功率幅度有问题，高于3.6GHz。
对于相同的频率和功率，频谱分析仪似乎读取不同的值。
您将源发生器设置为固定频率和功率，并将频谱分析仪设置为相同的频率并读取功率。
如果您更改频谱分析仪上的频率然后返回前一个频率，您将看到幅度开始与之前的值不同（您必须重复几次才能看到）。
它仅显示在3.6GHz以上的几个特定频率。
我的第一个假设是那个特殊的频谱分析仪是损坏但是我检查了其他几个MXA，我看到了类似的行为。
幅度电平随机变化，在某些情况下，我从之前的值高达0.3dB或更高。
对于许多测量而言，这似乎太高了。
我最初的任务是使用频谱分析仪作为功率计来平衡多个射频源，因此我使用了PSG E8257D参考功率发生器。
1.我直接在PSG的输出上使用功率计（U2000）测量绝对功率（在PSG上设置任意0dBm）。
USB功率传感器显示的设置功率低于+/- 0.05dB - 非常好。
2.我将RF发生器的输出连接到3英尺RF电缆的频谱分析仪输入，并读取频率上的值。
电缆显示出一些高达1.5dB的斜率。
记录获得值以用于将来的测量校正。
3.我再次在PSG上设置任意功率（0dBm），并使用相同的MXA悬停测量频率上的功率，现在我减去记录的校正值。
假设我应该从MXA（校正后）非常接近读数到功率计，但是我在特定频率下得到几个功率峰值。
如果您多次重复测试，则会从特定频率获得不同的结果。
请参阅测量设置的框图以及8MHz至6400Mhz之间的读数。
我在MXA（A13.12）上运行最新固件，旧固件显示（A10.52）显示类似的行为。
每个MXA以略微不同的频率显示尖峰。
在频谱分析仪上设置：跨度：20Khz RBW：自动VBW：自动所有设置相同，用于参考和测量有人知道任何解决方案吗？
固件中有没有错误？
对于相同的功率和频率，误差超过+/- 0.1dB看起来太高。
Power over frequency.png13.4 KB



     以下为原文

  I have a problem with measurement absolute power amplitude on MXA N9020A Spectrum Analyzer above 3.6GHz.
The Spectrum Analyzer appears to read different value for the same frequency and power. You set the source generator at fix frequency and power and Spectrum Analyzer to the same frequency and read the power.
If you change the frequency on the Spectrum Analyzer and then return the previous one, you will see that the amplitude start varied from the previous value (you have to repeat few time to see that). It only shows at few particular frequencies above 3.6GHz. My first assumption was that that particular Spectrum Analyzer is damage but I check several others MXA and I see similar behavior. The amplitude level varies randomly and in some cases I is up to 0.3dB or higher from previous values. That appears too high for many measurements.
My original task was to use Spectrum Analyzer as power meter to level several RF sources, so I used PSG E8257D reference power generator.
1. I measure the absolute power (set arbitrary 0dBm on PSG) using power meter (U2000) directly on output of the PSG. The USB power sensor show values below +/- 0.05dB from set power - very good).
2. I connect together output of RF generator to input of Spectrum Analyzer  with 3ft RF cable and read the values across the frequency.  The cable shows some slop up to 1.5dB.  The obtain values were logged for future measurement correction.   
3. I set on PSG the arbitrary power (0dBm) again, and measure the power over frequency using the same MXA hover now I subtract logged correction values.
The assumption was that I should get very close reading from MXA (after correction)  to power meter, however I getting several  power spikes at specific frequencies. If you repeat the test several time you get different results from that particular frequencies.  See the block diagram of the measurement setup and the readings between 8MHz to 6400Mhz.
I am running the latest firmware on MXA (A13.12), old firmware show (A10.52) showing similar behavior.  Each MXA showing the spikes at slightly different frequencies.
Setting on Spectrum Analyzer:
Span: 20Khz  
RBW: Auto
VBW: Auto
All setting the same for referencing and measurement
Does anybody know any solution? Is there any bug in the firmware?
Error of more than +/- 0.1dB for the same power and frequency look too high.   
附件

• SOUCE LEVEL AND SA CAL MESUREMENT.png50.7 KB
• Power over frequency.png13.4 KB