From: kochab <kochab@gm...>  20070512 23:17:32
Attachments:
radiometer_test.tgz

Hi! I'm still learning at using qucs, but it looks great, I'm trying to simulate the frequency response of a radiometer, using an AC simulation. I have set up a thermal input source and a RF amplification stage, using the 2 port S parameters component and experimental input data. The problem is on the detection stage, it is based on a square law detector, I have experimental data which relates linearly, as a function of frequency, the output Volts to the input Power. For now I've put this data on a s2p file, but I can't figure out how I can apply correctly the behaviour of the detection stage to the circuit. Could you please help?I attach the project files Thank you very much, Andrew 
From: Stefan Jahn <stefan@gr...>  20070513 07:48:10

Am So, 13.05.2007, 01:17, schrieb kochab: > Hi! Hello there, > I'm still learning at using qucs, but it looks great, I'm trying to > simulate the frequency response of a radiometer, using an AC > simulation. > I have set up a thermal input source and a RF amplification stage, > using the 2 port S parameters component and experimental input data. > The problem is on the detection stage, it is based on a square law > detector, I have experimental data which relates linearly, as a > function of frequency, the output Volts to the input Power. > For now I've put this data on a s2p file, but I can't figure out how I > can apply correctly the behaviour of the detection stage to the > circuit. > Could you please help?I attach the project files I am not sure about the actual problem here. But I can tell two things: During AC analysis the AC power sources emit the given available power "P" which are set to "0dBm" (1mW) in your example. Probably you want to change this. The second thing: If DET.s2p is meant to be a signal amplifier then have a look at "out.v" instead of "out.vn". Hope this helps, Stefan. 
From: kochab <kochab@gm...>  20070513 08:05:11

On 5/13/07, Stefan Jahn <stefan@...> wrote: > Am So, 13.05.2007, 01:17, schrieb kochab: > > > Hi! > > Hello there, Thank you very much for your answer, > > I am not sure about the actual problem here. But I can tell two > things: During AC analysis the AC power sources emit the given > available power "P" which are set to "0dBm" (1mW) in your example. > Probably you want to change this. The radiometer has to be sensitive to the temperature of the input load, therefore I regulate the input power by changing the temperature of the input resistance, which in this case is 200 K. Then I look only to the noise and not to the signal. >The second thing: If DET.s2p > is meant to be a signal amplifier then have a look at "out.v" instead > of "out.vn". The problem with DET is that it is a nonlinear amplifier, actually a square law detector, for which I have an experimental transfer function between input power (W) and output Volts. Therefore I cannot use the 2 port Sparameters component because it works between input signal (V) and output signal (V). Therefore I would need to compute the power, V^2/50 and then apply the experimental transfer function which characterize the detector. Is there a way to do this in QUCS? > > Hope this helps, > Stefan. > Thanks, Andrew 
From: Stefan Jahn <stefan@gr...>  20070513 08:38:49

Am So, 13.05.2007, 10:05, schrieb kochab: Hi! >> I am not sure about the actual problem here. But I can tell two >> things: During AC analysis the AC power sources emit the given >> available power "P" which are set to "0dBm" (1mW) in your example. >> Probably you want to change this. > > The radiometer has to be sensitive to the temperature of the input > load, therefore I regulate the input power by changing the temperature > of the input resistance, which in this case is 200 K. > Then I look only to the noise and not to the signal. Ok. >>The second thing: If DET.s2p >> is meant to be a signal amplifier then have a look at "out.v" instead >> of "out.vn". > > The problem with DET is that it is a nonlinear amplifier, actually a > square law detector, for which I have an experimental transfer > function between input power (W) and output Volts. > Therefore I cannot use the 2 port Sparameters component because it > works between input signal (V) and output signal (V). > Therefore I would need to compute the power, V^2/50 and then apply the > experimental transfer function which characterize the detector. > Is there a way to do this in QUCS? Hm. In fact Sparameters determine ratios of power waves (not voltages!). The S21 (you seem to be interested in this) is proportional sqrt(P2/P1) at given input and output impedance. So I propose to: with given transferfunction: Vout = f(Pin) to normalize Vout to an output power and then convert it to an Sparameter. Does this help?, Stefan. 
From: kochab <kochab@gm...>  20070513 18:04:12

solved!!! > > > > The problem with DET is that it is a nonlinear amplifier, actually a > > square law detector, for which I have an experimental transfer > > function between input power (W) and output Volts. > > Therefore I cannot use the 2 port Sparameters component because it > > works between input signal (V) and output signal (V). > > Therefore I would need to compute the power, V^2/50 and then apply the > > experimental transfer function which characterize the detector. > > Is there a way to do this in QUCS? > > Hm. In fact Sparameters determine ratios of power waves (not voltages!). Working at normalized impedance should be the same > The S21 (you seem to be interested in this) is proportional sqrt(P2/P1) at > given input and output impedance. So I propose to: > > with given transferfunction: Vout = f(Pin) > > to normalize Vout to an output power and then convert it to an Sparameter. Thanks, it worked: the transfer function (everything is a function of frequency) I have is v_out=tf * Pin while S21 is sqrt(Pout)= S21 * sqrt(Pin) therefore making the square root of the first equation: sqrt(Vout) = sqrt(tf) * sqrt(Pin) therefore the S21 is just the sqrt of tf and the result I get is the sqrt of Vout > Does this help?, > Stefan. A lot, thanks, Andrew 
From: Stefan Jahn <stefan@gr...>  20070513 18:29:00

Am So, 13.05.2007, 20:04, schrieb kochab: Hi Andrew, > solved!!! Nice to hear that. :) Congrats. >> > The problem with DET is that it is a nonlinear amplifier, actually a >> > square law detector, for which I have an experimental transfer >> > function between input power (W) and output Volts. >> > Therefore I cannot use the 2 port Sparameters component because it >> > works between input signal (V) and output signal (V). >> > Therefore I would need to compute the power, V^2/50 and then apply the >> > experimental transfer function which characterize the detector. >> > Is there a way to do this in QUCS? >> >> Hm. In fact Sparameters determine ratios of power waves (not >> voltages!). > > Working at normalized impedance should be the same > >> The S21 (you seem to be interested in this) is proportional sqrt(P2/P1) >> at >> given input and output impedance. So I propose to: >> >> with given transferfunction: Vout = f(Pin) >> >> to normalize Vout to an output power and then convert it to an >> Sparameter. > > Thanks, it worked: > > the transfer function (everything is a function of frequency) I have is > v_out=tf * Pin > > while S21 is sqrt(Pout)= S21 * sqrt(Pin) > > therefore making the square root of the first equation: > sqrt(Vout) = sqrt(tf) * sqrt(Pin) > > therefore the S21 is just the sqrt of tf and the result I get is the > sqrt of Vout If you want to you could contribute to the example section at <http://qucs.sourceforge.net/download.html#example>;. For this it would be nice to have your radiometer model performing some nice application with some comments (paintings>text)... Do you want to to? Cheers, Stefan. 
From: kochab <kochab@gm...>  20070514 12:34:19

> > If you want to you could contribute to the example section at > <http://qucs.sourceforge.net/download.html#example>;. > For this it would be nice to have your radiometer model performing > some nice application with some comments (paintings>text)... > Do you want to to? I'm glad you've asked, for sure I will contribute, the model is not yet finished, I will send you something in the next weeks. Thanks, andrew 
From: Stefan Jahn <stefan@gr...>  20070515 06:35:07

Am Mo, 14.05.2007, 14:34, schrieb kochab: Hello Andrew, >> If you want to you could contribute to the example section at >> <http://qucs.sourceforge.net/download.html#example>;. >> For this it would be nice to have your radiometer model performing >> some nice application with some comments (paintings>text)... >> Do you want to to? > > I'm glad you've asked, for sure I will contribute, the model is not > yet finished, I will send you something in the next weeks. Ok. Thanks in advance, Stefan. 