使用PNA X进行NFX测量怎么提高测量精度

这些误差几乎肯定是由于混频器测量中缺少LO隔离造成的。
可施加到噪声接收器的最大信号大约为-50到-60dBm。
如果您的混频器具有7级（+7 dBm）和40 dB的隔离度，那么您的应用大约为-33 dBm。
太高。
所以，你需要过滤IF。
这将导致测量的噪声系数增加，因为滤波器的损耗将增加表观噪声系数。
此外，由于LO无疑在路径中具有放大器，因此LO信号上将存在过量噪声，可能在远离LO信号的+ -1IF处显着。
这将被下转换为IF并产生过量噪声。
因此，您可能还需要对LO进行滤波以消除LO噪声的这种影响。
最后，为什么要这么麻烦？
混频器增益和噪声系数相同，因此只需测量增益并取消符号。
测量无源调音台的NF真的没什么意义，因为无论如何噪音功率只是kTB。
如果您真的想要噪声系数，请注意，如果没有RF滤波（因此您将在所有3个端口上都有滤波器）理论噪声系数大约为NFmixDSB = -SC21（dB）+3 dB。
它可以更高或更低，这取决于上边带图像和下边带图像的相应增益。
顺便说一下，本书第7章第7.8节详细介绍了这一点。
http://www.wiley.com/WileyCDA/WileyTitle/productCd-1119979552.html



     以下为原文

  The errors are almost certainly caused by lack of LO isolation in the mixer measuerment.  The maximum signal that can be applied to the noise receiver is on the order of -50 to -60 dBm.  If you have a mixer with a level 7 (+7 dBm) and 40 dB of isolation, then you are applying about -33 dBm.  Too high.

So, you need to filter the IF.  This will lead to an increase in measured noise figure as the loss of the filter will increase the apparent noise figure.

Further, since the LO undoubtably has amplifiers in the path, there will be excess noise on the LO signal, possibly significant at +- 1 IF away from the LO signal. This will be down converted to the IF and generate excess noise.  Thus, you may also need to filter the LO to remove this contribution of LO noise.

Finally, why bother?  The mixer gain and noise figure are the same, so just measure gain and negate the sign.  It really makes very little sense to measure the NF of a passive mixer, since the noise power is just kTB coming out of it anyway.  If you really want the noise figure, note that without RF filtering (thus you would have filters on all 3 ports) the theoretical noise figure is on the order of  NFmixDSB=-SC21(dB)+3 dB.  It can be higher or lower depending upon the respective gains of the upper and lower sideband image.

BTW, this is all covered in detail in  section 7.8 of chapter 7 of my book. http://www.wiley.com/WileyCDA/WileyTitle/productCd-1119979552.html