[Qucs-help] BJT noise simulation

 [Qucs-help] BJT noise simulation From: Antonio - 2006-06-22 07:47:16 Attachments: bjt_noise.sch bjt_noise.dpl ```Hi to the list, having understood how to simulate the noise in ac I've patched another simple example with BJT and flicker noise. It is very nice to see the correspondence between the simulated values and those synthetically calculated "by hand". The sch and dpl files are attached. Maybe it's a good example to be added to the download section, but first I want to refine it a bit more. I saw the function.pdf file published the past month, and I discovered that I can refer to any component parameter using the syntax ComponentName.Parameter. Unfortunately if I try to insert in an equation the R2 value (using R2.R) or the BTJ betha forward value (Q1.Bf) I get an error during the simulation saying: checker error, undefined variable `Q1.Bf' in equation `Rpi' Most probably I'm again missing something very trivial here. Any hint? Another question I have about the circuit is the temperature dependence. I have assumed that the Temp parameter in the BJT properties is the junction Temperature. So 25=B0 is not a realistic value, 100=B0 seems to me a more realistic one. If I increase the Temp value, the noise increase correspondingly. Giving that the noise is dominated by the base noise current (the white component is [2 e Ib]) I wonder which is the BJT parameter variation that cause the noise to increase. In fact, at first glance, both the base current and the trans conductance are constant, so even the noise should remain constant, isn't it? I'll will investigate more deeply on this point... The last question, is a bit tangential. Which is the best way to simulate the trans conductance value (gm) accurately? Theoretically I can apply a small sinusoidal voltage source to the base and calculate the signal current, but is there an "non invasive" way to calculate this parameter?. Many Thanks. Cheers, ~ Antonio ```

 [Qucs-help] BJT noise simulation From: Antonio - 2006-06-22 07:47:16 Attachments: bjt_noise.sch bjt_noise.dpl ```Hi to the list, having understood how to simulate the noise in ac I've patched another simple example with BJT and flicker noise. It is very nice to see the correspondence between the simulated values and those synthetically calculated "by hand". The sch and dpl files are attached. Maybe it's a good example to be added to the download section, but first I want to refine it a bit more. I saw the function.pdf file published the past month, and I discovered that I can refer to any component parameter using the syntax ComponentName.Parameter. Unfortunately if I try to insert in an equation the R2 value (using R2.R) or the BTJ betha forward value (Q1.Bf) I get an error during the simulation saying: checker error, undefined variable `Q1.Bf' in equation `Rpi' Most probably I'm again missing something very trivial here. Any hint? Another question I have about the circuit is the temperature dependence. I have assumed that the Temp parameter in the BJT properties is the junction Temperature. So 25=B0 is not a realistic value, 100=B0 seems to me a more realistic one. If I increase the Temp value, the noise increase correspondingly. Giving that the noise is dominated by the base noise current (the white component is [2 e Ib]) I wonder which is the BJT parameter variation that cause the noise to increase. In fact, at first glance, both the base current and the trans conductance are constant, so even the noise should remain constant, isn't it? I'll will investigate more deeply on this point... The last question, is a bit tangential. Which is the best way to simulate the trans conductance value (gm) accurately? Theoretically I can apply a small sinusoidal voltage source to the base and calculate the signal current, but is there an "non invasive" way to calculate this parameter?. Many Thanks. Cheers, ~ Antonio ```
 Re: [Qucs-help] BJT noise simulation From: Stefan Jahn - 2006-06-22 09:13:14 Attachments: bjt_gm.sch bjt_gm.dpl ```Am Do, 22.06.2006, 09:47, schrieb Antonio: > Hi to the list, Hello Antonio, thanks a lot for your mail and the nice example. > having understood how to simulate the noise in ac I've patched another > simple example with BJT and flicker noise. It is very nice to see the > correspondence between the simulated values and those synthetically > calculated "by hand". The sch and dpl files are attached. > > Maybe it's a good example to be added to the download section, but > first I want to refine it a bit more. It would be a really great example. :-) > I saw the function.pdf file published the past month, and I discovered > that I can refer to any component parameter using the syntax > ComponentName.Parameter. Unfortunately if I try to insert in an > equation the R2 value (using R2.R) or the BTJ betha forward value > (Q1.Bf) I get an error during the simulation saying: > > checker error, undefined variable `Q1.Bf' in equation `Rpi' > > Most probably I'm again missing something very trivial here. Any hint? Easy to answer: Set the DC analysis property "saveOPs" to "yes". > Another question I have about the circuit is the temperature > dependence. I have assumed that the Temp parameter in the BJT > properties is the junction Temperature. So 25° is not a realistic > value, 100° seems to me a more realistic one. > If I increase the Temp value, the noise increase correspondingly. > Giving that the noise is dominated by the base noise current (the > white component is [2 e Ib]) I wonder which is the BJT parameter > variation that cause the noise to increase. In fact, at first glance, > both the base current and the trans conductance are constant, so even > the noise should remain constant, isn't it? I'll will investigate more > deeply on this point... See eq. (9.124) in technical.pdf, the temperature dependence of Is which causes Ib to raise... That's my assumption. > The last question, is a bit tangential. Which is the best way to > simulate the trans conductance value (gm) accurately? Theoretically I > can apply a small sinusoidal voltage source to the base and calculate > the signal current, but is there an "non invasive" way to calculate > this parameter?. Hm. "gm" is a Q1.gm (see above). To calculate gm you would need to simulate a sweep which some how varies the base voltage (by slightly modifying R1 probably). Then you could: gm_sim = diff(Iout.I,Vb.V) which yields dIc/dVb. Please see the attachment! Or what did you mean by "non invasive"? Cheers, Stefan.```
 Re: [Qucs-help] BJT noise simulation From: Stefan Jahn - 2006-06-23 06:13:08 ```Am Do, 22.06.2006, 20:44, schrieb Antonio: > Hi Stefan, Hello Antonio, >> > Maybe it's a good example to be added to the download section, but >> > first I want to refine it a bit more. >> >> It would be a really great example. :-) > > Very well. I'll send a refined version as far as I understand the last > few things (see the following). Ok, thanks in advance. >> > I saw the function.pdf file published the past month, and I discovered >> > that I can refer to any component parameter using the syntax >> > ComponentName.Parameter. Unfortunately if I try to insert in an >> > equation the R2 value (using R2.R) or the BTJ betha forward value >> > (Q1.Bf) I get an error during the simulation saying: >> > >> > checker error, undefined variable `Q1.Bf' in equation `Rpi' >> > >> > Most probably I'm again missing something very trivial here. Any hint? >> >> Easy to answer: Set the DC analysis property "saveOPs" to "yes". > > Oh great! Having the access to all the runtime parameter inside the > equation is is really helpful. Is there such information written > somewhere? I've searched it for a long time... Could be useful to put > it somewhere... maybe a FAQ or a Not So Frequently Asked Question :). > I can collect a few of this tips if you want some help. No, it's not (yet) written somewhere... > I've done only a quick try but... I can use the run time parameters > like Qi.gm, but if I put Q1.Bf in an equation the simulation still > gives me the same error message. > In other words I cannot access the "model input parameters", while I > can access the "operating" parameters. Have I to set some other > option? Good idea. I'll think about it and put it onto my TODO list. > [cut question about temperature dependence] >> >> See eq. (9.124) in technical.pdf, the temperature dependence of Is which >> causes Ib to raise... That's my assumption. > > Ok, I'll study on it. Thanks for pointing it. > >> > The last question, is a bit tangential. Which is the best way to >> > simulate the trans conductance value (gm) accurately? Theoretically I >> > can apply a small sinusoidal voltage source to the base and calculate >> > the signal current, but is there an "non invasive" way to calculate >> > this parameter?. >> >> Hm. "gm" is a Q1.gm (see above). > > Yes! This was precisely what I needed. But I didn't know that I could > refer directly to the operating parameters simply activating the > option in the DC simulation dialog. As said this is really helpful in > many situations! > >>To calculate gm you would need to >> simulate a sweep which some how varies the base voltage (by slightly >> modifying R1 probably). Then you could: >> >> gm_sim = diff(Iout.I,Vb.V) >> >> which yields dIc/dVb. Please see the attachment! Or what did you mean >> by "non invasive"? > > Thanks for the example. With non-invasive I meant a way to calculate > gm without inserting additional voltage or current sources. And your > example to variate the base resistor it's a way to accomplish the > result. And there is a perfect agreement between Q1.gm @ 47k and gm_1! > Very good. Cheers, Stefan. ```
 Re: [Qucs-help] BJT noise simulation From: Antonio - 2006-06-24 13:59:06 Attachments: bjt_noise.sch bjt_noise.dpl ```Hi Stefan, On 6/23/06, Stefan Jahn wrote: > Am Do, 22.06.2006, 20:44, schrieb Antonio: > > > Hi Stefan, > > Hello Antonio, > > >> > Maybe it's a good example to be added to the download section, but > >> > first I want to refine it a bit more. > >> > >> It would be a really great example. :-) > > > > Very well. I'll send a refined version as far as I understand the last > > few things (see the following). > > Ok, thanks in advance. Done. I think the example it's ok now. The .sch and dpl files are attached, feel free to add them to the schematic download page. > > Oh great! Having the access to all the runtime parameter inside the > > equation is is really helpful. Is there such information written > > somewhere? I've searched it for a long time... Could be useful to put > > it somewhere... maybe a FAQ or a Not So Frequently Asked Question :). > > I can collect a few of this tips if you want some help. > > No, it's not (yet) written somewhere... I'll see if I can collect a some tips-for-getting-started, then I will send them here for review ... > > I've done only a quick try but... I can use the run time parameters > > like Qi.gm, but if I put Q1.Bf in an equation the simulation still > > gives me the same error message. > > In other words I cannot access the "model input parameters", while I > > can access the "operating" parameters. Have I to set some other > > option? > > Good idea. I'll think about it and put it onto my TODO list. I was thinking that was my fault... well glad too see that you will take care of this issue though :). PS for Stefan: Sorry for having sent the previous mail to your private address and not to the ml, it wasn't intentional. Cheers, ~ Antonio ```
 Re: [Qucs-help] BJT noise simulation From: Stefan Jahn - 2006-06-26 08:29:41 ```Am Sa, 24.06.2006, 15:59, schrieb Antonio: > Hi Stefan, Hello Antonio, >> >> > Maybe it's a good example to be added to the download section, but >> >> > first I want to refine it a bit more. >> >> >> >> It would be a really great example. :-) >> > >> > Very well. I'll send a refined version as far as I understand the last >> > few things (see the following). >> >> Ok, thanks in advance. > > Done. I think the example it's ok now. The .sch and dpl files are > attached, feel free to add them to the schematic download page. Thanks, I am going to prepare that... >> > Oh great! Having the access to all the runtime parameter inside the >> > equation is is really helpful. Is there such information written >> > somewhere? I've searched it for a long time... Could be useful to put >> > it somewhere... maybe a FAQ or a Not So Frequently Asked Question :). >> > I can collect a few of this tips if you want some help. >> >> No, it's not (yet) written somewhere... > > I'll see if I can collect a some tips-for-getting-started, then I will > send them here for review ... > >> > I've done only a quick try but... I can use the run time parameters >> > like Qi.gm, but if I put Q1.Bf in an equation the simulation still >> > gives me the same error message. >> > In other words I cannot access the "model input parameters", while I >> > can access the "operating" parameters. Have I to set some other >> > option? >> >> Good idea. I'll think about it and put it onto my TODO list. > > I was thinking that was my fault... well glad too see that you will > take care of this issue though :). > > PS for Stefan: Sorry for having sent the previous mail to your private > address and not to the ml, it wasn't intentional. Cheers, Stefan. ```
 Re: [Qucs-help] BJT noise simulation From: Stefan Jahn - 2006-06-29 12:56:56 ```Am Sa, 24.06.2006, 15:59, schrieb Antonio: > Hi Stefan, Hi Antonio, >> >> It would be a really great example. :-) >> > >> > Very well. I'll send a refined version as far as I understand the last >> > few things (see the following). >> >> Ok, thanks in advance. > > Done. I think the example it's ok now. The .sch and dpl files are > attached, feel free to add them to the schematic download page. It is now available on the homepage. Sorry for the delay. See ; Cheers, Stefan. ```