I have generated a DMT signal on AWG (tektronix 70002A) and then used it to modulate LED(the signal is DC-biased to have voltages above LED threshold).
The signal after photo detection is captured on oscilloscope along with the back to back signal from AWG (A direct coax connection between AWG and oscilloscope).
For back-to back connection, the Sampled data from oscilloscope (25000 sample points) is compared with the input DMT signal shape. It is observed that the input DMT signal can be reconstructed by using the samples between 15227 to 16384. Fig1. Thus the start and end points are identified within the above range.
But the signal after photo-detection (Fig2.) is totally unidentifiable for the corresponding start and end points. Will the aforementioned sample points are the start and end points for the detected signal? Does the delay due to optical and coaxial lengths create significant shift in the start and end points between the signals? Any better method to identify the start and end of the DMT signal?
It seems you need a commen synchronization procedure for OFDM or DMT by deploying cyclic prefix or training sequences in a Maximum likelihood method.
Thank you for the reference. I will go through it.
In my experiment, i have compared the input DMT signal and signal captured from oscilloscope (after back-to-back connection between waveform generator and oscilloscope). By identifying the shape of envelope, I could trace out the start and end of the signal captured on the oscilloscope.
Next, I capture the back-to-back and photo detected signal data simultaneously (C1,C2). Unlike the back-to-back signal, the photo detected does not look anything like input DMT signal.
Are the samples on these curves/data(C1,C2) directly correlated to each other sample by sample? If they are then, I could proceed with the same procedure as above.
If they are just different/offset by delay, then is this delay computable from oscilloscope?
DMT is very sensitive to the symbol timing errors, so a standard synchronization is essential. The distortions from LED and PD limited the feasibility of identifying the symbol head or timing delay precisely through the method you mentioned above.
Hope the reference will help you out.
- Badges
- Science topic
- Similar topics
- Engineering
- Telecommunications Engineering