noise as source

Hello, If you want to do noise analysis you should be using a noise analysis simulation? What kind of analysis are you trying to achieve?

Now if you would like to do a transient simulation and see the result of superimposing noise onto your circuit of interest, the first thing to do would be to simplify things by just using the noise source and running a transient simulation. You could then post-process the data in an FFT and see if you are getting what you expect. Your source 1 looks to be quite strange with a point at 10us that has a large offset voltage vs. the previous point.

-Samir

Hello Samir,

My target is to have a power of 15dbm at each of my noise source without applying any Signal in Addition and when I give These 2 noise sources as signals to my schematic I want to achive a power of -3dbm at my load side. I am concered with power at Input and Output and the power spectral density. I donot want any noise Analysis.
Can you please explain me how could you say that there is a dc Offset from the Excel sheet I made and suggest me a method to remove it. Actually I was asked to produce 0 dc Offset voltages which have a crest factor of 5. I considered the crest factor to be 5 but i dont know how would i consider 0 dc Offset while making file in excel.
Can you please explain me what do you mean by preprocess the data in an fft.

Spanika,

Let me understand this - you want to see the output response when you apply 2 uncorrelated signal at the inputs. Each input signal will have 15 dBm of power with a crest factor of 5. Is that correct? What is the end application here that you are trying to analyze?

Please look at the attached tina file which has a transient simulation up to 10 us. On the red curve (noise source 1) you will notice that the 1st and last points have an offset. This means that if you want the signal to be repetitive you are looking at a large discontinuity every 10us. Is this what you inteneded to do? Noise source 2 looks okay.

By postprocessing the data, I mean run a transient simulation in SPICE, export that output data and use a tool like Matlab to examine the output FFT.

-SamirNoiseSource.TSC

hii Samir,

yes what you said is right. At last I want to Show that when I put in 2 noise sources with 15dbm at Input, I want to Show that I can achieve -3dbm at my Output (because of losses of power due to op amp structure I made).

As you said matlab is a good choice to make noise sources. but I have only Excel as Option to do my noise file.

I have another doubt. What ever data I put in the PWL of Tina, and if I measure power across it, the power I see is always in Terms of picoWatts. Why is it so. even if I put nothing into PWL and just select PWL for Signal then also it Shows me power in picoWatts. I estimate to see power across These sources to be in Terms of some milliWatts, so that I could see some power at my load end.

Spanika,

 It looks like all you are trying to do is find the ac frequency response of your circuit. In that case I would do an AC response/simulation to find the gain/attenuation from the output of the noise source to the load.

The AC response only allows you to check one source at a time so you would run two ac simulations. When you are testing the response of noise source 1, you can remove source 2 and ground that side. Do the same thing when you do the response characteristic of noise source 2. You will then be left with two gain numbers -

Lets assume that you put in 15dBm (approx. 4Vpp) into source 1 and its attenuation is 20 dB leaving you with 400mVpp at the outputs. Similarly lets assume you put 10dBm (2Vpp) into source 2 and its attenuation of 20 dB leaves you with 200mVpp at the output. Since we are talking about noise which is uncorrelated, the total output noise due to source 1 and 2 is the rms value (square root of sum of squares), = SQRT(0.4^2+0.2^2) = SQRT(0.16+0.04) = SQRT(0.20) = 0.447Vpp.

I believe this is what you are looking to analyze.

-Samir

Hello Samir,
Thank you I understood your expaination clearly. But why is the power measured across PWL ls always in terms of picoWatts irrespective of whatever content I put in.

Spanika,

  Can you please let me know how you set up your simulation when you see the power in pW? Are you running an AC sim, Transient sim, etc.

Also, I made an error in my last response. Because you have a high crest factor you should not be using peak-to-peak voltage but RMS voltage, so for example in a 50 Ohm system, 15dBm is 1.26Vrms and 10dBm is 0.707Vrms. You can then use Square root of sum of squares to estimate the output RMS value at the load and the crest factor to calculate the peak to peak voltage at load. Hope this makes sense.

-Samir

hello Samir,

I am running ac Analysis and checking the table of ac results.

Hello Spanika,

If you want results from a a source that you have loaded you cannot run an AC simulation, you must run a transient simulation.

Since we are not modeling experts I am moving this thread to the modeling forum. Perhaps they can explain it better than we can.

Regards,
Loren