ALS Active Feedback System Manual
Michael C. Martin
This brief manual is intended to show how to operate the ALS Active Feedback electronics on a synchrotron infrared beamline. It is not exhaustive, so please contact me should you have any questions that remain unanswered.
1. Optical setup
The ALS Active Feedback system is intended to stabilize the infrared beam on a synchrotron infrared beamline at low frequencies (typically below 500 Hz). This system locks the position of the beam in four degrees of freedom, and significantly reduces beam motion in this low frequency region. The implementation of this system followed much effort in identifying and passively correcting many noise sources at high and low frequencies observed at ALS Beamline 1.4. The remaining low frequency noise after passive fixes were installed is what the feedback system was designed to help mitigate.
Figure 1 shows a schematic layout of the active feedback system as implemented on ALS beamline 1.4.3. A second duplicate system was later installed for beamline 1.4.4. Dichroic beamplitters are used to allow a portion of the visible synchrotron beam to continue to position sensitive detectors (PSD’s) while the infrared light is reflected on down the beamline and into the FTIR systems. The x and y positions read from the PSD’s are used via control electronics to drive the tip and tilt of piezo actuated mirrors further upstream. The control electronics drive the active mirror such that the light is kept at the center of the PSD’s.
Because the piezo actuated mirror mounts have a first resonance frequency at about 2.5 KHz, the correction circuit is tuned to have a pole at 1 KHz. This ensures that no corrections are applied close to the ~2.5 KHz resonance which could drive the system unstable. The active feedback does however provide significant stabilization of the beam position, as shown in Figure 2. A noise reduction factor greater than 10 is achieved for frequencies below 100 Hz, with this factor going to 1 at about 1 KHz, as designed.
2. Electronic connections
The heart of the active feedback system is the control circuit. This electronic unit must read the beam position from the PSD and apply a correction voltage to the active optic to maintain the beam on the center of the PSD. At LBNL, Mike Chin designed a self-contained NIM module to accomplish this task for two axes (x and y) at a time. These feedback modules power, read and control one Hamamatsu S1880 PSD via its matching C4674 control card, and one Physik Instrumente S330.30 piezo actuated tip-tilt platform. A photograph of one feedback control module is shown in Figure 3, along with a description of what each connection and control is used for.
The feedback control module is connected to the PI mirror via three BNC connectors: 1 main drive voltage connection (V+), and two correction signals (Mirror Drive X and Mirror Drive Y). Since the mirror is typically inside a vacuum chamber of the infrared beamline, these connections will have to pass through an appropriate electrical feedthrough, such as a multi-BNC vacuum feedthrough flange, and then be connected to the piezo leads.
The output of the PSD electronics card is connected to the PSD circuit board connection (DB9) on the feedback control module. This powers the PSD and reads out the x and y positions of the light incident on the PSD. Voltages proportional to these x and y positions can be monitored using the X OUT and Y OUT BNC connections and a voltmeter or the display unit described below. A third voltage is also provided which is proportional to the total light intensity incident on the PSD. This can be useful for sorting out if you have aligned the main beam onto the PSD, or if the detector is instead picking up a reflected beam or other stray light.
A display module is very useful for initial alignment of the optical system and for verification of correct feedback operations. The X and Y OUT’s from the feedback control module are simply voltages proportional to the centroid x and y position read by the PSD (1 volt = 1 mm). These voltages can be read by any voltmeter. For convenience, a four voltage NIM display module was designed at LBNL providing displays for a complete four-axis feedback system which fits along size the feedback control modules in a NIM enclosure. The display module and its connectors are shown in Figure 4. These displays show mV (= microns) and go up to 2 volts full scale.
3. Feedback operations
Once all electrical connections are in place, start with the feedback electronics switched off (toggle switches on the lower part of the control module switched to the right), and power on the NIM enclosure. The first step will be to manually align the visible synchrotron beam to approximately the center of the PSD’s. This is the case when the X OUT and Y OUT voltages are close to zero (< 100 mV). It is important to make sure the direct beam is on the PSD and not a reflection or stray light. This can be done visually if there is adequate access to the beamline, or the incident light monitor can be used to see the total amount of light on the PSD (direct beam will have a higher signal level than reflected beam). And one should be sure that you can move fully and uniformly across the PSD by steering the beam in the x and y directions. If when moving continuously in one direction the readout voltage first goes up and then goes down, you might be looking at a reflection instead of the main beam.
Once the x and y positions are close to zero, turn on the x feedback by flipping the toggle switch labeled x on the control module fully to the left. If you see a large jump in the y display value, then you have x and y orientation incorrect between the piezo and the PSD. Simply power down the NIM enclosure, swap the X and Y mirror drive BNC’s, and then power on the NIM enclosure and try again. If you see the x display go off scale and the x feedback unlocked light illuminate when switching on the x feedback loop, the feedback electronics are probably set to the incorrect polarity (applying a positive correction voltage when a negative correction voltage is required, or vice versa). In this case power down the NIM enclosure, remove the feedback control module and change the polarity as described in the next section. If the x value goes towards zero volts when switching on the x feedback loop, then the feedback is engaged and is working.
The gain setting for the feedback loop can be adjusted by using a small electronics screwdriver and turning the potentiometer screw just under the Mirror Drive X output. By increasing the gain you will see the X OUT voltage get closer to zero volts until the gain is so high that the feedback loop becomes unlocked and the red feedback unlocked light illuminates. In practice I adjust the gain until I see the system become unlocked and then I back the gain down a couple of turns below where the unlocked light comes on. This way you are getting a strong feedback signal, but the system remains stably locked.
Once the x feedback is working properly, repeat the above procedures for the y feedback loop, verifying polarity and setting the feedback gain. Finally the entire procedure must be repeated for the second active optic and PSD to ensure its feedback loops are correctly functioning. The feedback system is now fully functional. A photo of a functioning feedback system at the ALS is shown in Figure 5.
The display unit and feedback unlocked lights can be monitored to ensure that the feedback system continues to function. If the synchrotron light moves farther than the feedback system can compensate for, the x and/or y readout voltages will drift away from zero volts. In this case, or whenever you want to realign the beamline, switch all four feedback loops off, and manually align the beamline optics to get the first PSD x and y position readouts to near zero voltage. Switch on these first two feedback loops to hold their x and y positions. Then adjust the further downstream beamline optics to get the second PSD x and y position readouts near zero voltage and then switch on the second feedback module’s feedback loops to return to normal operations.
The X and Y OUT monitors can also be conveniently used to connect to a spectrum analyzer to determine the frequency components of noise on the beamline with and without the feedback switched on, and also at much higher frequencies since the PSD’s will respond to significantly higher frequencies than the active optics can correct for. We’ve used the PSD’s to diagnose synchrotron beam motions up to 22 KHz.
4. Polarity Settings
The polarity of the feedback system needs to be set to match the physical setup of your beamline’s optics. The polarity of the feedback loops in the feedback control modules can be changed by removing the NIM module from the NIM crate, opening up the side panel with a screw driver, and finding the blue jumpers (labeled H5 and H3, or H11 and H21) shown in Figure 6. This figure shows the correct jumper positions for positive or negative polarity, so if the polarity requires reversing, simple change both jumpers (H5 and H3, or H11 and H21) to the opposite positions as you found them. Once the polarity setting is changed, replace the module cover and put the module back in the NIM crate. Power up the NIM crate and re-test for correct feedback functionality.