Lock laser to an atomic transition without modulation

There is always possibility that you need to stabilize a laser but do not have an appropriate modulator like AOM or EOM in hand. The dichroic atomic vapor laser lock(DAVLL) technique provides a really simple and robust modulation-free way to do that.

Although there is no modulation in DAVLL, the error signal of it is still resemble to the FM spectroscopy and the prevailing Pound-Drever-Hall(PDH) locking technique with a dispersive shape, which provide a stable zero-crossing point for locking. The dispersive error signal comes from the differential measurement of the Zeeman split levels.

The apparatus lack the complex of difficult alignment of the general modulators. Since the polarization in this technique is really important, good quality polarization beam splitter(PBS) with correct coating should be used.


The photo of the apparatus.


The coil and the heating strip.


With the pump beam(brown) blocked, and only observing the signal PD, we get the typical doppler broadened absorption signal.


With the pump beam unblocked, we can clearly see the hyperfine features in the saturated absorption (doppler free) signal from signal PD.


When we turn on the balance mode, observing signal PD with reference PD subtracted, we can see the dispersive shaped DAVLL error signal. We can also easily recognize the hyperfine transitions as well as crossovers from this snapshot.


And the last step is locking. The residue noise of after locking is shown below. This apparatus achieves sub MHz frequency noise according to a WS-7 wavemeter and will not easily get out of lock(as long as the laser is not drifting too far away).


Thanks Ben and Loic for kindly helping me debugging the system.

Laser Locked!

Finally lock my homemade diode laser onto the super stable cavity!

It is not that easy to do this because the cavity is a ULE cavity with finesse as high as 300,000. The line width of this little monster is as low as 5kHz, which is far below the typical ECDL bandwidth. (So the PDH dispersive signal we get from the oscilloscope is actually representing the line width of.our unlocked laser!) We also need to fine tune the loop filter and stabilize our modulators to get a stable lock.

The laser table. See our homemade ECDL (can be hermetic but we have not do that yet).ecdl_photo

The locking table. Mode matching, cavity, vacuum, etc. You can see the strong transmission light through the cavity after lock in the small monitor on the upper right corner of the photo.cavity_photo

Our homemade high speed photo detector. Powered by a LEMO connector with very soft silica cable. The shielding is done by CNC the copper box, which works really good that no interference from the modulator can be seen at the PD output in the spectrum analyzer down to -90dBm.pd_photo

Our all-homemade electronics!elec_photo

In order to further determine the actual line width, we need to setup another set of laser and measure the beat notes between the two lasers.

Our preliminary scheme is to collect the beat notes with a high speed photodetector and down covert it to kilohertz region using a mixer. The final result will be readout by a dynamic analyzer (SR785) and a frequency counter. The signal generator, dynamic analyzer and the frequency counter will all be synchronized to a rubidium frequency standard.

P.S. Our GIANT SR785 just arrived!



TCD1304 Verilog Driver for Spectrometer Project

Spend several hours to finish some verilog HDL code driving a TCD1304DG linear CCD from Toshiba. This work is intended for the CCD spectrometer project in the future.

The project will utilize a high precision 18bit SAR ADC (AD7960) to sample the CCD output at a very fast speed up to 5MSa/s. The code is written versatilely that many timing constraints can be directly reprogrammed through registers. When I finish the connection between this IP and the Zynq 7000 PS, it would be convenient to control the CCD sampling parameters from linux side.

The simulation result is shown below.

The 8th row: SH signal

The 9th row: ICG signal

The 10th row: MCLK signal for CCD


Danmu server deployed!

Danmu is an interesting interaction form for people who are watching videos. Unlike the comment  area beneath the video playback window, Danmu shoots the comment directly on the video overlay.

For some informal presentation, Danmu can be also used to stir up the interest of audiences. I decided to superimpose Danmu on the presentation in the recruitment speech of ‘Students Association of Science and Technology, Dept. Phys., Tsinghua Univ. (SAST-Phys-THU)’.

Thanks to the development work from Bigeagle (https://github.com/bigeagle/danmaQ) and TUNA (https://github.com/bigeagle/danmaQ), we can easily deploy a such system on our server and provide live Danmu service through web as well as WeChat.