Yokozawa, Ushiba, Takano

Summary

We found a mysterious peak at 28MHz in the demodulation singals. From the measurement with varied conditions, finally we conclude that this peak comes from GrPDHY servo and is coupled to other signals via crosstalk. We will replace the servo with a spare and try the measurement again.

Detail

Continued from klog 29146.

Today we checked the signal around GrPDH in high frequencies.

1. Measurement of demodulated signal

First, we measured the demodulator output for GRX. To increase the signal to noise ratio, we used the PDHX servo's first gain stage. Before the servo an RF LPF and 10dB attenuator were connected. The output of the demodulator was connected to GrPDHX servo and the gain of the first stage was set to 31dB, then OUT2 of the servo was connected to a Moku:Lab. Then, we found that at 28 MHz there was a large peak even though the signal got through the LPF. The spectrum is shown in Figure1.

2. Tuning off the oscillators

As the source of the peak, we suspected the local oscillators, such as for GR PDH, PLL, IR sidebands, etc.. First we turned off the LO for GR PDHX, at 33MHz, but nothing changed. Next we turned off the seed LO of f1 and f2 sideband, 5.6MHz, but no effect again. Therefore, we considered that this peak is not from LOs but other circuits.

3. Changing the condition

Next, we stopped to use the servo and measured the demodulated signal without amplification, then the peak still appeared. With this condition, we touched the cable, then noticed that the peak height varied with the distance from GrPDHY servo. This implies that the peak is coulpled to the demodulated signal by radiation from GrPDHY servo.

4. Measurement of the signal from GRX RFPD

We measured the signal from GRX RFPD directly, and we found a broad peak around 33 MHz. Even though turning off the LO of PDHX, this peak existed. This peak ssems to come from the resonant peak of the RFPD. We also touched the SMA cable from GRX RFPD, and noticed that the 28 MHz peak appeared when the cable is close to the PDHY servo.

5. Turning off the power of GrPDHY servo

We turned off the power of GrPDHY servo, and checked whether the 28MHz peak changes or not. Then, we found that the peak dissapeared after turning off the servo. Therefore, we were confident that the peak is generated inside the GrPDHY servo and coupled to the signals around it.

6. Measurement of the signal from GRY RFPD

Next, we measured the signal from GRY RFPD directly, and similarly we found a broad peak around 34 MHz, which is the resonant peak of GRY RFPD. We also put the SMA cable from GRY RFPD close to the GrPDHY servo, and confirmed that the 28 MHz peak appeared as shown in Figure2.

7. Turning off the power of GrPDHX servo and GrPDHY

We checked the effect of GrPDHX. We put the SMA cable close to GrPDHY servo, and turned off the power of GrPDHX servo. However, the 28MHz peak was still there. Finally, we turned off the power of GrPDHY servo, and checked that the peak dissapeared after turning off the servo as shown in Figure3. After turning on the power, the peak appeared again.

Conclusion

From these measurement, we concluded that GrPDHY servo generates and radiates a sharp peak noise at 28MHz, and the noise couples to the signal arpund the servo by crosstalk with cables. The previous measurement klog 29424 reported that the input refferred noise of GrPHDY is larger than that of GrPDHX, and this result supports our consideration.

Fortunately we have a spare of the servo. We will replace the servo with the spare and check whether the 28MHz peak disappear or not.