> {quote：title = sm7ovk写道：} {quote}>我认为你的问题是相反的边带。
如果你不删除这个其他边带，你将获得3dB更高的NF。
>问候> Jens实际上，增加的边带使测得的噪声系数低3 dB。
这是因为NF测量值是NF =（输出端的噪声功率）/增益通过测量DUT输出端噪声源的噪声功率来测量增益。
在这种情况下，噪声源会在所需的和不足的边带处产生功率，因此增益测量为3 dB太高，而nf为3 dB太低（这不完全正确，但足够多）。
当然，冷噪声测量是测量混合器中的总噪声，输入仅为50欧姆（与噪声源关闭或冷噪声相同）。
输出端的任何双边带噪声都是真实的，需要包括在内。
但很可能在LO输出信号的IF BW内，本地振荡器上存在过多的噪声，这些噪声会混合到输出端。
这是测量原始混频器的一个非常常见的问题（与转换器相比，转换器通常仅因此而具有LO滤波器）。
在这种情况下，无需过滤LO。
以上来自于谷歌翻译
以下为原文
> {quote:title=sm7ovk wrote:}{quote}
> I think your problem is the opposite sideband. If you don't remove this other sideband you'll get 3dB higher NF.
> Regards
> Jens
Actually, the added sideband makes the measureed noise figure appear 3 dB lower. That is because the NF measurement is NF=(Noise power at the output)/Gain
The gain is measured by measuring the noise power of the noise source at the output of the DUT. In this case, the Noise Source produces power at both desired and undersired sidebands, and so the gain is measured 3 dB too high, and the nf is 3 dB too low (this is not exactly correct, but cloes enough).
Of course, the cold-noise measurement is the measurement of total noise out of the mixer with just 50 ohms on the input (which is the same as noise source off, or cold noise). Any double side-band noise at the output is real, and needs to be included.
But most likely there is excess noise on the local oscillator, within the IF BW of the LO output signal, that is getting mixed down to the output. This is a very common problem measuring raw mixers (in contrast to converters, which often have LO filters just for this reason). Nothing to do but filter the LO in this case.
> {quote：title = sm7ovk写道：} {quote}>我认为你的问题是相反的边带。
如果你不删除这个其他边带，你将获得3dB更高的NF。
>问候> Jens实际上，增加的边带使测得的噪声系数低3 dB。
这是因为NF测量值是NF =（输出端的噪声功率）/增益通过测量DUT输出端噪声源的噪声功率来测量增益。
在这种情况下，噪声源会在所需的和不足的边带处产生功率，因此增益测量为3 dB太高，而nf为3 dB太低（这不完全正确，但足够多）。
当然，冷噪声测量是测量混合器中的总噪声，输入仅为50欧姆（与噪声源关闭或冷噪声相同）。
输出端的任何双边带噪声都是真实的，需要包括在内。
但很可能在LO输出信号的IF BW内，本地振荡器上存在过多的噪声，这些噪声会混合到输出端。
这是测量原始混频器的一个非常常见的问题（与转换器相比，转换器通常仅因此而具有LO滤波器）。
在这种情况下，无需过滤LO。
以上来自于谷歌翻译
以下为原文
> {quote:title=sm7ovk wrote:}{quote}
> I think your problem is the opposite sideband. If you don't remove this other sideband you'll get 3dB higher NF.
> Regards
> Jens
Actually, the added sideband makes the measureed noise figure appear 3 dB lower. That is because the NF measurement is NF=(Noise power at the output)/Gain
The gain is measured by measuring the noise power of the noise source at the output of the DUT. In this case, the Noise Source produces power at both desired and undersired sidebands, and so the gain is measured 3 dB too high, and the nf is 3 dB too low (this is not exactly correct, but cloes enough).
Of course, the cold-noise measurement is the measurement of total noise out of the mixer with just 50 ohms on the input (which is the same as noise source off, or cold noise). Any double side-band noise at the output is real, and needs to be included.
But most likely there is excess noise on the local oscillator, within the IF BW of the LO output signal, that is getting mixed down to the output. This is a very common problem measuring raw mixers (in contrast to converters, which often have LO filters just for this reason). Nothing to do but filter the LO in this case.
举报
