MIXer HB noise simulation problem?

Status
Not open for further replies.

liletian

Full Member level 6

Hi Guys
I am doing a noise simulation of mixer in HB simulator, it reports following warning and the report NFssb is zero which is obvious wrong, here is my noise simulation setting,
I enable nonliear noise ,choose noise Voltage/or noise figure, since my IF frequency is 5M hz, then I choose Noise Frequency 5Mhz, I also enable "compuet noise figure" and choose input frequency RF_freq+noisefreq, choose the input and output port,
I did not define noisefreq, should I define it? Can anyone comment on my noise simulation setting
and how to change it? Is there a mixer HB simulator noise simulation tutorial? The following is the warning when I ran my simulation.
Thanks a lot
"
Warning detected by hpeesofsim during spectral noise analysis HB1_NC1'.
The global circuit temperature and the IEEE standard noise temperature
(290 Kelvin) do not agree in the spectral noise analysis HB1'. The
Temp' parameter on the Options item can be set to 16.85 Celsius
to eliminate this warning message. If the circuit temperature is
different from the IEEE standard temperature the loss and the noise
figure might not agree.

Warning detected by hpeesofsim during spectral noise analysis HB1_NC1'.
Given a set of Harmonic Balance large-signal frequencies LSfreq
(determined by Freq[], Order[] and MaxOrder) and a noise frequency
FreqForNoise, the small-signal frequencies used in a nonlinear noise
analysis are SSfreq=LSfreq+FreqForNoise and SSfreq=LSfreq-FreqForNoise.
For this circuit, more than one such small-signal frequency differs
from InputFreq by less than BandwidthForNoise. This means that the
simulator is unable to uniquely determine what small-signal tone
should be used for computing the small-signal conversion gains needed
for the evaluation of NFssb and NFdsb. The small-signal frequency
corresponding to the large-signal frequency with the lowest order will
be used. In rare cases, NFssb and NFdsb are inaccurate. To eliminate
this warning and the possibility of inaccurate results:
o Reduce Order[] or MaxOrder to eliminate a conflicting large-signal
frequency
o Offset FreqForNoise and InputFreq by a small amount, like 1-2 Hz.
o Offset one of the fundamental frequencies by a small amount, like
1-2 Hz.
o Eliminate any colliding tones in your Harmonic Balance setup,
if present."

pancho_hideboo

liletian said:
I am doing a noise simulation of mixer in HB simulator, it reports following warning and the report NFssb is zero which is obvious wrong,
Although I don't know your setting, I think your setting is wrong.
Maybe NFdsb have some value unless you do set completely wrong simulation setting.

liletian said:
I enable nonliear noise ,choose noise Voltage/or noise figure, since my IF frequency is 5M hz, then I choose Noise Frequency 5Mhz, I also enable "compuet noise figure" and choose input frequency RF_freq+noisefreq, choose the input and output port,
Let's assume LO_Freq=1000MHz, RF_Freq=995MHz(LSB), IF_Freq=LO_Freq-RF_Freq=5MHz.
If we do 1 Tone(LO_freq) HB based noise analysis.
FreqForNoise(noisefreq) must be IF_Freq=5MHz and
InputFreq must be LO_Freq-noisefreq, this is LSB setting for NFssb.

If you do 2 Tone(LO_Freq, RF_Freq) HB based noise analysis, setting will be different.
It seems you don't understand the following my comments for you at all.

https://www.designers-guide.org/Forum/YaBB.pl?num=1184111670

liletian said:
Is there a mixer HB simulator noise simulation tutorial?
HB simulator noise simulation tutorial is there.
I showed you many ADS's manual location, so you can find necessary documentation by yourself.

liletian said:
Warning detected by hpeesofsim during spectral noise analysis HB1_NC1'.
The global circuit temperature and the IEEE standard noise temperature
(290 Kelvin) do not agree in the spectral noise analysis HB1'. The
Temp' parameter on the Options item can be set to 16.85 Celsius
to eliminate this warning message. If the circuit temperature is
different from the IEEE standard temperature the loss and the noise
figure might not agree.
This is very basic issue in RF. But this is not problem in actual situation.
https://www.designers-guide.org/Forum/YaBB.pl?num=1192525133

Before you play EDA tool, you should read basic RF text book, e.g. "RF Microelectronics(Behzad Razavi, 1997)".

khouly

liletian said:
Hi Guys

"

Warning detected by hpeesofsim during spectral noise analysis HB1_NC1'.
Given a set of Harmonic Balance large-signal frequencies LSfreq
(determined by Freq[], Order[] and MaxOrder) and a noise frequency
FreqForNoise, the small-signal frequencies used in a nonlinear noise
analysis are SSfreq=LSfreq+FreqForNoise and SSfreq=LSfreq-FreqForNoise.
For this circuit, more than one such small-signal frequency differs
from InputFreq by less than BandwidthForNoise. This means that the
simulator is unable to uniquely determine what small-signal tone
should be used for computing the small-signal conversion gains needed
for the evaluation of NFssb and NFdsb. The small-signal frequency
corresponding to the large-signal frequency with the lowest order will
be used. In rare cases, NFssb and NFdsb are inaccurate. To eliminate
this warning and the possibility of inaccurate results:
o Reduce Order[] or MaxOrder to eliminate a conflicting large-signal
frequency
o Offset FreqForNoise and InputFreq by a small amount, like 1-2 Hz.
o Offset one of the fundamental frequencies by a small amount, like
1-2 Hz.
o Eliminate any colliding tones in your Harmonic Balance setup,
if present."

this happen when the orders are high , so many harmonics fall in the IF spectrum , and this raise the NF

and ur done

khouly

Status
Not open for further replies.