September 27, 2007 - John Hansknecht
Measurements of original DEI switch and New OC100HG based switch on the primary Pockels cell for the CEBAF machine.
TEK0005: Old DEI switch at 10Hz flip rate. Analyzer is held in a fixed rotation so the difference in birefringence is apparent. KD*P Pockels cell is flipping between +/- λ/4 voltages (5160V total). I don’t know why there is a noise riding on the upper and lower plateaus.
TEK0006: Old DEI switch. Same conditions but now at 100Hz flip rate. Note the longer term relaxation effect that seems to appear on the lower plateau.
TEK0007: Old DEI switch. Same conditions but now at 600Hz. System is starting to have “issues” and cannot run at higher flip rates. The High voltage supplies driving this switch are drawing a full 3mA. The 24V feed to these DC-DC converter high voltage supplies is running at 1 amp as opposed to the normal quiescent current of 330mA. One should not be alarmed by the apparent asymmetric pattern. The pattern appears like this because of the rotation angle of the analyzer.
TEK0008: Old DEI, now at 500 Hz. It is not evident on this image, but the high voltage supplies are unstable and don’t enjoy running at this rep rate. It is difficult to take a photo of the cell flipping with an asymmetric pseudo-random drive pattern, but I think we would clearly see birefringence differences base on the amount of time in a given state.
TEK0009. The DEI is still flipping at 500Hz, but I have rotated the analyzer to a position where the electrical birefringence only appears while the cell is in transition. This fools the scope into reading it as a 1kHz pulse train. The image shows that indeed the + and – transitions are identical.
TEK0010. DEI switch at 300Hz. AC coupled scope to zoom in on transition to show rise time. I am calling the rise time the time it takes to change from one stable birefringence to the next stable birefringence. The cursors above show this to be approx 200 us.
TEK0011. The new OC100HG switch running at 300Hz. Still AC coupled on scope. Rise time is now 50us! The DC-DC converters for the high voltage supplies are running at quiescent current. They don’t even seem to notice that they are driving a Pockels cell since they only need to charge/discharge 6pf.
TEK0012. New switch at 1kHz, DC coupled. 6.8mV pk-pk out of 300mV offset shows that at this particular analysis angle, the + and – λ/4 polarization is still 99.99%. I have not optimized the voltages because I actually want to see a difference in birefringence.
TEK0013. For comparison, new switch at 500 Hz.
TEK0014. New switch at 100 Hz.
TEK0015. New switch at 10 Hz.
TEK0016. New OC100HG switch is flipping between + and – λ/4 polarization at 410 hz. A motorized λ/2 plate is spinning against a linear polarizer for fast 360 degree analysis of the beam polarization. We have a maximum sine wave value of 930mV and minimum of 890 mV, yielding an overall circular polarization of 99.975%. We are AC coupled and the cell alignment has not been optimized. The spikes that appear are the momentary full power or reduced power spikes of light that occur when the cell has passed through conditions other than λ/4. (The 50us transition times).
TEK0020. New OC100HG switch flipping at 250 Hz. Cell was aligned and voltages were optimized to the maximum extent possible. I find that I have reached an upper limit of my analysis capabilities. Even with perfect circular light into my analyzer, I will see a sine wave during rotation due to subtle mismatches between the designed wavelength of the analyzer’s waveplate and the actual wavelength of my laser. With my present setup, I can only measure to 99.98% circular. I can only guess, but I believe I have over 99.99% circular polarization in both helicity states.