Designer’s Guide Community

Hans

New Member

Offline

Posts: 2
Los Angeles, CA

Jitter in Limiting Amplifier
Jul 24^th, 2003, 4:29pm

I have a question regarding the way PNoise simulates sampled noise (tdnoise). Let's say we have a limiting amplifier that is stimulated with a sinusoid at frequency f_o. The amplifier clips the signal into somewhat of a square wave and the output will be accompanied by the noise generated within the amplifier. This signal is then presented to some logic block with logic threshold V_TH,LOGIC.
The general idea is to run a tdnoise simulation and look at the noise voltage at the output when the output signal crosses V_TH,LOGIC. When plotting the integrated output noise voltage the SpectreRF User Guide suggest using the range from zero to the PSS fundamental frequency. Does this mean that the PSD for a sampled time point is only integrated over that interval? Or is the computation done in the time domain from zero to 1/f_o? How are frequencies above the fundamental taken into account? (i.e. what exactly is being integrated?) The following quote from the Affirma RF Simulator (SpectreRF) User Guide (Product Version 4.4.6 - April 2001) is the source of my confusion:
"Type 0 as the start frequency and the PSS fundamental frequency as the stop period. For example, type 1G if the PSS period is 1ns." (p. 909).
I would really appreciate it, if someone could clarify this for me. So far the Cadence documentation has been rather vague yet I believe someone out here must have stumbled across the same problem. Thanks in advance,

Hans

Back to top

IP Logged

pancho_hideboo

Senior Fellow

Offline

Posts: 1424
Real Homeless

Re: Jitter in Limiting Amplifier
Reply #1 - Jul 28^th, 2003, 8:08am

Description about TDnoise in SpectreRF manual is wrong. Cry

TDnoise setting has to be like following.

(1) Sweep Range for Pnoise.
This should be from 0 to PSSfund/2.
Sweep Range is the one we fill in the PNoise analysis form. For regular pnoise(noisetype=source), the range could be from 0 to any frequency we want.
For time-domain pnoise (noisetype=timedomain), the range should be from 0 to PSSfund/2.
Time-domain noise PSD is periodic with PSSfund so PSD is symmetric around PSSfund/2.

(2) Frequency Integration Range of PSD.
This should be from 0 to PSSfund/2.
For time-domain noise (noisetype=timedomain), SpectreRF calculates the PSD of the discrete (or sampled) noise at t0+n*T (T=1/PSSfund, 0<=t0<=T, -infinity<n<+infinity).
So its integration is the noise power at t0+n*T. Because of sampling, there is alias in its PSD, and integral frequency range should be 0 to PSSfund/2.

Back to top

Kita━━━━━━(ﾟ∀ﾟ)━━━━━━ !!!!!
http://www7.plala.or.jp/ungeromeppa/flash/kita.html
http://www.youtube.com/watch?v=mjIxGh55bMM&feature=related

IP Logged

Hans

New Member

Offline

Posts: 2
Los Angeles, CA

Re: Jitter in Limiting Amplifier
Reply #2 - Jul 28^th, 2003, 2:55pm

Let's see if I'm looking at this the right way. Your claim is that tdnoise is a purely discrete computation with a sampling rate equal to the PSS fundamental f₀ (=1/T₀). As such, the interval of interest would be [ f, f+f₀] where f can be any value due to the periodic nature of the spectrum. Any effects due to aliasing would be included in this window and integration should yield a true picture of noise across all frequencies.

Thinking about it this way, I would have to agree with an integration range from 0 to f₀. If we were only to cover half that range, we would neglect the portion that occurs at negative frequencies (and also folded to the range f₀/2 - f₀). Although there is no "new" information in those ranges, signal power does appear in it. Am I making a wrong assumption in here?

Thanks for your reply and putting this issue in a new perspective. I cretainly would appreciate it, if the Cadence manuals would go into more detail about their simulation tools - just specifying what needs to be done for a very specific example can lead to a false sense of confidence. How about some equations with the documentation?

Anyway, I'll stop my rant about poor documentation. Thanks again for taking the time to reply to my post,

Hans

Back to top

IP Logged

pancho_hideboo

Senior Fellow

Offline

Posts: 1424
Real Homeless

Re: Jitter in Limiting Amplifier
Reply #3 - Jul 28^th, 2003, 5:19pm

> Thinking about it this way,
> I would have to agree with an integration range from 0 to f0.
> If we were only to cover half that range,
> we would neglect the portion that occurs at negative frequencies
> (and also folded to the range f0/2 - f0).

Undecided

No, integration range has to be from 0 to PSSfund/2.
Since PSD of Spectre is 1-sided view spectrum
which is twice of 2-sideded view spectrum.

For example, integration range is from 0 to fmax
for normal noise analysis. This is because PSD is 1-sided view.
Here you don't have to integrate from -fmax to +fmax.

Cadence will revise TDnoise explanation in SpectreRF manual soon.
You will see the reason why both analysis range and
integration range are from 0 to PSSfund/2.

If you integrate from 0Hz to PSSfund,
then you will include the same noise many times,
so an integral range should be limited to 0Hz~PSSfund/2.

About noise folding, see below.

(3) Folding and Alias.
When inputing noise into a linear time periodic varying (LTPV) system,
because of frequency translation, the output noise at f has contribution
from input noise at f+n*PSSfund (-infinity<n<+infinity).
This is noise folding. It is physical.

When sampling the above output noise from LTPV at t0+n*T, and do
PSD for it, the PSD repeats every PSSfund.
This is aliasing. It is not physical.

Tongue

Back to top

Kita━━━━━━(ﾟ∀ﾟ)━━━━━━ !!!!!
http://www7.plala.or.jp/ungeromeppa/flash/kita.html
http://www.youtube.com/watch?v=mjIxGh55bMM&feature=related

IP Logged

jeffyan

Community Member

Offline

Posts: 47

Re: Jitter in Limiting Amplifier
Reply #4 - Jun 25^th, 2008, 5:10am

pancho_hideboo wrote on Jul 28^th, 2003, 8:08am:

Description about TDnoise in SpectreRF manual is wrong. Cry

TDnoise setting has to be like following.

(1) Sweep Range for Pnoise.
This should be from 0 to PSSfund/2.
Sweep Range is the one we fill in the PNoise analysis form. For regular pnoise(noisetype=source), the range could be from 0 to any frequency we want.
For time-domain pnoise (noisetype=timedomain), the range should be from 0 to PSSfund/2.
Time-domain noise PSD is periodic with PSSfund so PSD is symmetric around PSSfund/2.

(2) Frequency Integration Range of PSD.
This should be from 0 to PSSfund/2.
For time-domain noise (noisetype=timedomain), SpectreRF calculates the PSD of the discrete (or sampled) noise at t0+n*T (T=1/PSSfund, 0<=t0<=T, -infinity<n<+infinity).
So its integration is the noise power at t0+n*T. Because of sampling, there is alias in its PSD, and integral frequency range should be 0 to PSSfund/2.

hi,
you mean this description on tdnoise is not right?
http://www.cadence.com/whitepapers/tdnoise.pdf
thanks

Back to top

IP Logged

pancho_hideboo

Senior Fellow

Offline

Posts: 1424
Real Homeless

Re: Jitter in Limiting Amplifier
Reply #5 - Jun 25^th, 2008, 8:15am

jeffyan wrote on Jun 25^th, 2008, 5:10am:

pancho_hideboo wrote on Jul 28^th, 2003, 8:08am:

Description about TDnoise in SpectreRF manual is wrong. Cry

TDnoise setting has to be like following.

(1) Sweep Range for Pnoise.
This should be from 0 to PSSfund/2.
Sweep Range is the one we fill in the PNoise analysis form. For regular pnoise(noisetype=source), the range could be from 0 to any frequency we want.
For time-domain pnoise (noisetype=timedomain), the range should be from 0 to PSSfund/2.
Time-domain noise PSD is periodic with PSSfund so PSD is symmetric around PSSfund/2.

(2) Frequency Integration Range of PSD.
This should be from 0 to PSSfund/2.
For time-domain noise (noisetype=timedomain), SpectreRF calculates the PSD of the discrete (or sampled) noise at t0+n*T (T=1/PSSfund, 0<=t0<=T, -infinity<n<+infinity).
So its integration is the noise power at t0+n*T. Because of sampling, there is alias in its PSD, and integral frequency range should be 0 to PSSfund/2.

hi,
you mean this description on tdnoise is not right?
http://www.cadence.com/whitepapers/tdnoise.pdf
thanks

No, content of "tdnoise.pdf" is right.

My post is matched to "tdnoise.pdf".
Is there any contradiction point in "tdnoise.pdf" ?

Back to top

Kita━━━━━━(ﾟ∀ﾟ)━━━━━━ !!!!!
http://www7.plala.or.jp/ungeromeppa/flash/kita.html
http://www.youtube.com/watch?v=mjIxGh55bMM&feature=related

IP Logged

jeffyan

Community Member

Offline

Posts: 47

Re: Jitter in Limiting Amplifier
Reply #6 - Jun 27^th, 2008, 11:14pm

pancho_hideboo wrote on Jun 25^th, 2008, 8:15am:

jeffyan wrote on Jun 25^th, 2008, 5:10am:

pancho_hideboo wrote on Jul 28^th, 2003, 8:08am:

Description about TDnoise in SpectreRF manual is wrong. Cry

TDnoise setting has to be like following.

(1) Sweep Range for Pnoise.
This should be from 0 to PSSfund/2.
Sweep Range is the one we fill in the PNoise analysis form. For regular pnoise(noisetype=source), the range could be from 0 to any frequency we want.
For time-domain pnoise (noisetype=timedomain), the range should be from 0 to PSSfund/2.
Time-domain noise PSD is periodic with PSSfund so PSD is symmetric around PSSfund/2.

(2) Frequency Integration Range of PSD.
This should be from 0 to PSSfund/2.
For time-domain noise (noisetype=timedomain), SpectreRF calculates the PSD of the discrete (or sampled) noise at t0+n*T (T=1/PSSfund, 0<=t0<=T, -infinity<n<+infinity).
So its integration is the noise power at t0+n*T. Because of sampling, there is alias in its PSD, and integral frequency range should be 0 to PSSfund/2.

hi,
you mean this description on tdnoise is not right?
http://www.cadence.com/whitepapers/tdnoise.pdf
thanks

No, content of "tdnoise.pdf" is right.

My post is matched to "tdnoise.pdf".
Is there any contradiction point in "tdnoise.pdf" ?

hi,
could you tell me why integral frequency range is 0 to PSSfund/2, instead of 0 to infinte in discrete freq domain?

Back to top

IP Logged

pancho_hideboo

Senior Fellow

Offline

Posts: 1424
Real Homeless

Re: Jitter in Limiting Amplifier
Reply #7 - Jun 28^th, 2008, 12:14am

jeffyan wrote on Jun 27^th, 2008, 11:14pm:

could you tell me why integral frequency range is 0 to PSSfund/2, instead of 0 to infinte in discrete freq domain?

Rather why do you integrate 0 to infinite ?

See the following.
http://www.designers-guide.org/Forum/YaBB.pl?num=1207830622

I've been very disgusted about explaining tdnoise.
It seems people whose name is �"****yan" has same question.

Back to top

Kita━━━━━━(ﾟ∀ﾟ)━━━━━━ !!!!!
http://www7.plala.or.jp/ungeromeppa/flash/kita.html
http://www.youtube.com/watch?v=mjIxGh55bMM&feature=related

IP Logged

jeffyan Community Member Offline Posts: 47	Re: Jitter in Limiting Amplifier Reply #8 - Jun 28^th, 2008, 1:54am hi, thanks, it seem that you stated the truth, but you didn't give the reason. the integral power is infinite if the range is from 0 to infinite, and if integrated from 0 to fsample/2, the noise power is identical to the original one.
Back to top	IP Logged

pancho_hideboo

Senior Fellow

Offline

Posts: 1424
Real Homeless

Re: Jitter in Limiting Amplifier
Reply #9 - Jun 28^th, 2008, 3:38am

jeffyan wrote on Jun 28^th, 2008, 1:54am:

it seem that you stated the truth, but you didn't give the reason.

I gave enough reason which you can understand if you have basic mathematical or signal processing theory back ground.

"Sampling Theory", "Impulse Sampling", "Energy", "Repeated Spectrum Replica"
You should read the following again with considering the above issues.

http://www.designers-guide.org/Forum/YaBB.pl?num=1207830622

jeffyan wrote on Jun 28^th, 2008, 1:54am:

the integral power is infinite if the range is from 0 to infinite, and if integrated from 0 to fsample/2, the noise power is identical to the original one.

Correct.

Back to top

Kita━━━━━━(ﾟ∀ﾟ)━━━━━━ !!!!!
http://www7.plala.or.jp/ungeromeppa/flash/kita.html
http://www.youtube.com/watch?v=mjIxGh55bMM&feature=related

IP Logged