[SOLVED] Total Harmonic distortion in cadence

Hi i need to calculate the THD in cadence for a multiplier circuit. I need to plot the TDH percent sweeping the two current inputs ix and iy. I searched online and found pss analysis but i couldn't figure the settings to get that done. In engine there are options of shooting and harmonic balance so don't know which to choose and set the rest of the parameters. Like do i need to change or include any particular model library? And also for the input current source should i use idc or isin??? I am working in 180nm tech. Can anyone kindly help me with the steps to plot THD percent.. Thank you!!!!

I changed the two input sources Ix and Iy to isin and since it is a multiplier i gave 70kHz to Ix and 10kHz to Iy and did the transient. Now say even if i want to find THD using calculator i need to fill a section called "fundamental frequency". So how do i determine this fundamental frequency. Is it one of the input frequencies 70k (Ix) or 10k (Iy)????

For the output transient waveform (multiplied output of Ix and Iy) i did the DFT. My plot was like two high magnitude lines at freq 90k and 150k and the magnitude of both the lines were almost same. the second line was like a point above the first line. The rest of the lines were really small magnitude and spread all over the plot. So now of these two which is the fundamental freq??? Can someone let me know how to find the fundamental frequency???

Review the definition of THD and understand that it's not suitable as performance criterion when multiplying two frequencies.

Getting 90k and 150 k output lines with 10k and 70k input makes no sense, you would expect 60k and 80k.

Thanks for the reply!! I reviewed THD and found that it is the ratio of all the harmonics to the fundamental frequency. But i dont get the part wer it is not suitable for multiplying two frequencies. I have attched a image of THD analysis of a 4 quadrant multiplier taken from a paper. And i am looking to plot a graph like that. I have two inputs Ix and Iy.

The only way I can imagine to perform a THD analysis in this case to multiply a sine and a DC signal. Apparently this setup is described in the paper.

Thank you so much for the help!!! I was able to obtain the THD plot but still my only last doubt is as far as i read and understood harmonic distortion is something say the fundamental freq is a sine wave and we require a that proper sine wave in the output. But when the output appears little off from a sine wave we say it is distorted and this is due to presence of harmonics caused by non linear load. Pls correct me if i am wrong.

But here harmonics is used for "Fig. 3.4.6 shows a sine wave (the fundamental) together with its 3rd and 5th harmonics. Adding these harmonics to the fundamental produces a resultant wave (shown in red) that is beginning to look like a square wave, in fact adding an infinite number of odd harmonics in this way would produce a perfect square wave.

So can see harmonics is used to build waves and is also sometimes not preferred as it distorts the signal. In my case the question is say two signals of different frequencies are multiplied. So resultant is a modulated wave. So i should get a particular kind of modulated output and if it distorts from the required output then the o/p modulated signal has harmonics. So its quite simple right to get THD of two signals with different frequencies. After all i think THD is nothing but representation of o/p in freq domain showing all the actual frequencies of the modulated signal and the unwanted harmonics (more like fourier transform)... And then we divide all the harmonics with the fundamental frequency. So why use a DC component for another input. Though i was able to take the output i still dont understand properly why use one DC input. Thank you!!!

If you send sine signals with frequencies f1 and f2 to an ideal 4-quadrant multiplier, you get the output signals f1+2 and f1-f2 (respectively f2-f1, which is just the same signal), but no "fundamental" wave at all. So how would you calculate THD?

In case of a non-ideal multiplier, you get all kinds of modulation products N*f1+M*f2, with N and M being integer numbers. You'll notice N,M = +/-1 stand for the first order products which are also present with an ideal multiplier, the zero order products represent offsets and carrier feedthrough and |N|,|M| > 1 higher order intermodulations products. Their presence can be used as a performance criterion similar to THD for a linear circuit, but it's a different calculation.

- - - Updated - - -

I'm not an expert in this regard, but as far as I see from the manual, Spectre RF can measure mixer intermodulation effects with QPSS analysis.