I would like to ask how to use INTERCONNECT for modelling a system for mm-wave generation like the one in (figure 7) to reproduce the EVM (error vector magnitude) and constellation (figure 9). srep19891.pdf (1.8 MB)
Thank you for the paper, it is a very nice application. For the setup in reference Figure 7., I think you can reproduce it by using the elements in the Element Library. We don’t have a very good EDFA model yet but we are working on it. As written in the reference, the EDFA is " low-noise, almost insensitive to polarization of the signal and can be relatively simply realized", which is a simple amplifier model and you can use the “Optical Amplifier” instead.
The most tricky part in the setup is the dual-wavelength laser. I think you can use two CW laser models and an “Optical Splitter/coupler” to achieve a simple version of the dual-wavelength laser as shown in the reference Figure 3(a).
The measurements and plots in Figure 9. can be measured by the “Vector Signal Analyzer”. However, the figures are for measurements with optical link plus a 6 m wireless link. There is no way to model the wireless link in INTERCONNECT. However if you have some information about the transfer function/s-parameter of the wireless link, you can use the s-parameter elements to implement it into the system.
I hope this could help The reference indeed is a very good application. I’d like to discuss more with you if you are going to model it.
Thanks a lot for your interest. I started building the first part of the system till the transmitter stage. The preliminary file is attached. mm-wave signal generation.icp (276.9 KB)
Here are some points to discuss:
I agree with you the EDFA shouldn’t be a problem. An OA is enough.
I modelled the Dual wavelength source as you suggested and it works pretty well. I checked that using the OSA.
I don’t know how to generate the OFDM signal
I can’t find an electrical mixer to up-convert the OFDM signal on it to the IF = 5GHz. Do we need integration with cadence to do this? If so, how to start, please?
There are 2 types of photo-detectors in the Element library. (PIN or APD?) which one is convenient for this application?
For the OFDM signal, you could add up several electrical "Sine Wave"s using the "Electrical Adder"s. According to the paper, I guess the simplest case is to generate two electrical sine waves which are out of phase as the I and Q part of the input to the I/Q mixer.
I am not so sure whether Cadence already has the build in I/Q mixer (and the arbitrary waveform generator for the OFDM signal generation as well) in the library, if it does, then probably use the Cadence interoperability is a good idea. You will need an extra interoperability license for the integration. Please contact firstname.lastname@example.org for the details. Otherwise, a simple I/Q mixer can be achieved by using some “Electrical Multiplexer” and "Electrical Adder"s. I made a sample of the OFDM plus the I/Q mixer up converter iq_mixer.icp (501.3 KB) and it works as expected with the output spectrum showing a 10 GHz channel spacing and the LSB suppressed.
In the sample file I used single frequency for the OFDM signals. The paper shows a 5 GHz bandwidth for each channel, so in this case, you could use a PRBS generator and a Gaussian pulse shaper to achieve that.
Another way of generating the signal, if you have the AWG as referenced in the paper, you could generate the OFDM signal using the AWG and save it into a file to load into INTERCONNECT as well.
They both will do the job, and the PIN photodetector is the common one to use. If you use phase modulation/coherent modulation, a coherent detector will be needed. You could refer to the online page Optical QPSK for more information on coherent detection.
I hope this could help Please keep me posted if you make any progress on this
Thanks a lot.
I tried to model the system based on your recommendationin the file attached. In this model, I assumed only a single unmodulated carrier exists at the input of the MZI. The spectrum of the final output has a peak at 60 GHz as designed. However, there is a DC component. I don’t know where does it come from. And if you zoom in, you will find lots of wiggles. What do you think?
I also tried to model an I/Q mixer. electrical_IQmixerTxRx.icp (1022.4 KB)
The file has the modulator and de-modulator. To my understanding, the IQ data should be carried on orthogonal baseband frequencies from 0 to 5 GHz. Then this whole spectrum will be mixed with the IF (=5 GHz) and drives the MZI.
The power spectrum measured at the end (after PD) should also has two peaks with 60 GHz spacing in between. It is only because that the Spectrum Analyzer put the reference frequency at 0 Hz, so it looks like a DC component.
I hope this answers your question. Let me know if you have any further questions