While attempting to develop a somewhat accurate noise budget for the 40m, I reasoned that while the shape of the transfer function for the ISS is important, the degree to which we can 'tune' it to a particular experiment/application is limited.
Beyond that, we're sort of limited by the desired high and low frequency behavior as well as the general principle that more electronics = more noise so we probably don't want more than 3 or 4 filter stages, if that. Additionally, the ISS can be over-engineered so that it suppresses the laser noise to levels well below other fundamental noise sources over the important regime ~10 - 10^3 Hz without particular regard to a noise budget.
The design I propose is very similar to a previous design, with a few adjustments. It consists of 3 filter stages that easily be modified to increase gain for higher frequencies if it is known/determined that the laser being stabilized has a lot of high frequency noise.
Stage 1: Basic LP Filter + Establish UGF (each stage 'turning on' will not change the UGF), Stage 2: Integrator with zero @ 10 kHz, Stage 3: Optional extra gain if necessary
With the full TF given by,
As usual we consider the noise caused by the servo itself. Noise analysis in LISO is done with a 1 V input excitation.
This servo should function sufficiently for the 40m.
After familiarizing myself with Altium, I drew up the attached schematic for the ISS to be used in the CTN experiment. The filename includes 'abbott-switch' as I am using an Altium component (the switch, in particular), that he created. The MAX333A actually has 20 pins on a single component, but the distributed component that he created is useful for drawing uncluttered schematics. I won't be using all of the pins on this switch, but for completeness, I have included the 3rd and 4th portion of the full component in the upper right hand corner.
Currently, the schematic includes the voltage reference (AD586), a LP filter for the reference signal, the differential amplifier stage to obtain the error signal and then finally all of the filter stages. The schematic does not include the RMS detection and subsequent triggering of each filter stage. The TRIGGER 1 signal is a user input (essentially the on button) while the TRIGGER 2 signal will flip the second switch when the RMS noise has decreased sufficiently after the first filter stage has been turned on.
PCB layouts will be done once I understand that part of Altium
NOTE THAT I HAVE DELETED ELOG 8798 AS IT WAS A DUPLICATE OF THIS ONE.
I wanted this elog to be in reply to a previous one and I couldn't figure out how to change that in an elog I already submitted.
There are essentially two major portions of the ISS I am designing. One system has the voltage reference, differential amplifier and filtering servo (schematic attached) while the other has a comparator circuit and a triggering mechanism. The first system amplifies an error signal obtained from the PD output and the voltage reference, which is then fed back through the AOM. I've done a lot of work designing/prototyping this first half and now I'm starting to design the second half.
The second system's main purpose is to maintain loop stability as the ISS is engaged. Let's assume a user has decided they want noise suppression. They would first close the ISS feedback loop and an error signal would pass through three unity-gain buffers, providing minimal noise reduction. The user can then send a signal to theTRIGGER 1 port to switch the first stage from its unity-gain position to its filtering position and reduce the intensity noise further. This signal will most likely be digital in origin. Alternatively, when the user first closes the ISS loop, the first stage can already be in its filtering position rather than necessitating two commands.
A test channel (not drawn in the included schematic) will monitor the RMS level of the incoming signal from the PD. This noisy AC signal will first be amplified and then passed through an RMS-to-DC converter. The resulting DC signal is used as a part of the triggering mechanism for later stages. Once the first stage has been switched manually, and the DC signal corresponding to RMS noise of the PD output drops below a certain threshold, stages 2 and 3 will be internally triggered with a short delay between them. Toward being able to detect this threshold, I have designed a simple comparator circuit with an LT1016. The circuit has a fairly low-level output when the input voltage is larger than the threshold (about 1.6 V for my simple prototype), but when the input passes below the threshold, the comparator puts out almost 4 V, a number limited by the supply voltage. The schematic is shown below.
The component V2 and the various voltage dividers serve to establish the reference/threshold voltage. Note that although the LT1016 is not powered in the schematic, it requires ±5 V (a max of 7 V). The above circuit was also prototyped on a breadboard and I characterized it with an oscilloscope. Using a CFG253, I made a low frequency (~0.3 Hz) triangle wave with an amplitude and DC offset such that it oscillates between 0 and 5 V. This was applied to the IN node in the above schematic. The input waveform and the circuit's response (voltage at the OUT node) are shown below. As expected, R2 serves to establish hysteresis. The comparator switches to 'high' output until the input drops below 1.6 V, and then it doesn't switch back to the 'low' output until the input goes up to ~3.4 V.
This behavior is ideal for our application as we can detect when the DC signal from the RMS-to-DC converter drops below a certain level (i.e. the first stage that has been activated does some amount of filtering to lower RMS noise), and then we can trigger subsequent filter stages off of the comparators high-level output.
This circuit could easily be used to drive the MAX333a switches shown in the first schematic attached. I believe the low-level output is not sufficient to switch the MAX333a although the ~4 V high-level output is quite sufficient. Regardless, these exact values (thresholds, outputs etc) will be determined after I have a better idea of the RMS noise of the laser without any intensity stabilization as well as a solid understanding of how the AD8436 RMS-to-DC converter works. This was simply a proof of concept for lower threshold detection using basic Schmitt trigger topology.
I went out on the floor to look at the transmitted signal from the PMC to get a rough idea of the noise of the unstabilized laser. There was already a scope hooked up so I just used the measurement features to find the following:
Signal average = 875 mV. Peak-to-Peak noise = 45 mV
Assuming the noise can be approximated as Gaussian noise, the heuristic for converting to RMS noise of the signal is RMS = Peak-to-Peak / 8 (or Peak-to-Peak / 6, I've used both...)
-> RMS Noise ~ 6.5 mV
When designing my filtering stages and RMS detection/triggering, I'll use relative RMS, i.e. 6 mV / 875 mV = 0.007, as a measure of unstabilized laser noise.
It would be better to measure the power spectrum density of the fluctuation.
The RMS does not tell enough information how the servo should be.
In deed, the power spctrum density gives you how much the RMS is in the entire or a specific frequency range.
I wanted the RMS noise simply to establish a very rough estimate of thresholds on RMS detectors that will be part of my device. If you refer to elog 8830, I explain it there. Essentially, when the ISS is first engaged, only one of the 2 or 3 filter stages will be active. Internal RMS threshold detection serves to create a logic input to switch subsequent filters to their 'on' stage.
Here I have included the full schematic (so far) of the proposed ISS. There are two sheets: the first schematic details the filter stages and their accompanying circuitry while the second schematic details the RMS threshold detection and subsequent triggering.
The first schematic is fairly self explanatory as to what different portions do, and I have annotated much of the second schematic as there are some non-traditional components etc.
I have not yet included some mechanism to adjust the threshold voltage in real time or any of the power regulation, but these should follow fairly quickly.
I have made significant changes to the ISS schematic, mostly in the form of adding necessary subsystems.
Some changes I have made:
On the front page, all inputs and outputs are currently BNC ports, although this is most likely not the final design that will be used. For instance, the ports ENABLE, INPUT GND and INVERT are supposed to be logic inputs for a MAX333a switch. These will most likely be front panel switches that either connect the switch's logic pin to GND (Logic 0) or something like a +5 V supply (Logic 1).
I also have not included power regulation for my board although I have some of the actual D1000217 Chasis Power Regulator boards and I'll incorporate those in my design soon.
I realized I totally forgot to post this last week, but I prototyped the comparator and boost triggering portion of the ISS, at least in part. Below is a schematic that shows the prototype circuit I made. Note that it includes ports for the oscilloscope channels that appear in the second image included. Essentially, I was able to verify that the output from the LT1016, as it's currently constructed in the ISS schematic, would be sufficient logic to switch the MAX333a.
Below, we can first see that the comparator is switching its output as desired. When the DC level of the input drops below a certain threshold (~1.6 V) the output of the comparator switches on to ~4 V. When the DC level of the input goes back up above the upper threshold (~3.2 V), the comparator switches off to ~0.3 V. The exact values of the threshold voltages can be determined/tuned at a later date, but this is the basic behavior that the comparator circuit will have.
To detect whether or not the MAX333a was switching properly, I connected the common terminal of one of the switches to a +5 V supply, and looked at the voltage coming off both the 'open' and 'closed' terminals of said SPDT switch. We can see that with Logic 0 (comparator output ~0.3 V) Channel 4 exhibits a ~5 V signal, just as we would expect from the above schematic. With Logic 1 (comparator output ~4 V), Channel 3 exhibits the characteristic 5 V signal.
I constructed a regulator board that can take ±24 V and supply a regulated ±15 V or ±5 V. I followed the schematics from LIGO-D1000217-v1.
I was going to make 2 boards, one for ±15 V and one for ±5, but Chub just gave me a second assembled board when I asked him for the parts to construct it
More changes that I've made:
In PSL elog 1270, Evan elucidated the explicit requirements for the CTN ISS board. Essentially, the transfer function of the ISS should be something like:
TF_mag = (Unstabilized RIN) / (Calculated RIN Requirement)
I took Evan's data and did exactly this. I then designed a servo (using the general design I proposed here) to meet this requirement with a safety factor of ~10. By safety factor, I mean that if the ISS operates exactly according to theory, it should suppress the noise by a factor of 10 more than what is necessary/set out by the requirement. Below is a plot of the loop gain obtained directly from the requirement (the above expression for TF_mag) and the transfer function of the servo I am proposing.
I don't have the actual schematics attached as I was working with a LISO file and have yet to update the corresponding Altium schematic. The LISO file is attached and I will add the schematics later, although one can reference the second link to find a simple drawing.
Attached is the finalized schematic. The general circuit topology should remain the same from this point forward, although individual component values are subject to change. I will also be adding some more annotations to ensure everything on the board is clear.
In general, I have finally included all of the correct components (i.e. front panel switches are now actually switches and front panel LEDs are now included). I also added an external 'Boost' switch, which can be used to enable or disable the boosts. The motivation for including this switch is that one might want to test functionality of the ISS without using the 'fancy' RMS detection and triggering circuitry. Additionally, one can disable the boosts when all the circuitry is stuffed in order to troubleshoot, so it essentially grants the board some flexibility in its operation.
I am now working on the PCB layout and I should hopefully have that done next week.
After many, many moons of getting to know exactly how frustrating Altium can be, I have completed the PCB layout for my ISS board (final page of ISS_v3.pdf).
Before I get into detail about the PCB, there is one significant schematic change to note: the comparator circuit was changed (with significant help from Koji) so that the voltage reference for boost triggering is established in a more logical way. Instead of the somewhat convoluted topology I had before, now there are only two feedback resistors, R82 and R83. Because their resistances (500k and 50k respectively) are so much larger than the total resistance of the 1k potentiometer (used to establish a tunable threshold voltage), the current flowing through the feedback loop is negligible compared to the 5 mA current flowing through the potentiometer (the pot is rated for 2 W and with 5 mA -> 25 mW dissapation). This allows one to set the threshold voltage for my schmitt trigger, at pin 2 of both the pot and the comparator, entirely with the pot. This trigger also has hysteresis given by the relation deltaV ~ (R83/R82) * (Voh - Vol) where deltaV is the separation between threshold voltages, Voh is the high-level comparator ouput and Vol is the low-level comparator output. Koji simulated this using CircuitLab and I plan to verify the behavior by making a quick prototype circuit.
Now, on to the PCB. The board itself is of a 'standard' LIGO size (11" x 6") has 3 routing layers and 3 internal planes, one for +15 V, one for -15 V and one for GND. In the attached pdf, red is the top routing layer, blue is the bottom layer and brown is the middle routing layer (used for ±5 V exclusively). The grey circles are pads and vias (drilled through) and anything in black is silkscreen overlay. I placed each component and track by hand, attempting to minimize the signal path and following the general rules below,
Sections of the board have been partitioned and labeled with silkscreen overlay to help in both signal pathway recognition as well as eventual troubleshooting.
On the board, I have also included holes so that it can be mounted inside of an enclosure. There is a DCC number printed as well as a 'barcode' (TrueType font: IDAutomationC39S), although they both contain filler asterisks as I haven't published this to the DCC and thus do not have a number.
Rana and I connected the PMC_trans output to the BNC connector board on the west end of the PSL table (the channel is labeled). I took a few spectra off of PMC_trans and the SR785 was connected directly to the PMC_trans output for about an hour.
Data will follow.
I was working on the electronics bench and what sounded like a huge truck rolled by outside. I didn't notice anything until now, but It looks like something became misaligned when the truck passed by (~6:45-6:50 pm). I can hear a lot of noise coming out of the control room speakers and pretty much all of the IOO plots on the wall have sharp discontinuities.
I haven't been moving around much for the past 2 hours so I don't think it was me, but I thought it was worth noting.
Previously in elog 8959, I gave a very simple method for determining the noise suppression behavior of the ISS. Recently, I recalculated this requirement in a more correct fashion and again redesigned the ISS to be used in the CTN experiment.
Just as before, the data from PSL elog 1270 is necessary to infer a noise suppression requirement. The data presented there by Evan consists of two noise spectra, 1) the unstabilized RIN presently observed in the CTN experiment readout and 2) the theoretical brownian noise produced by thermal processes in the mirror coating+substrate. The statement "TF_mag = (Unstabilized RIN) / (Calculated Brownian Noise Limit)", where TF_mag refers to the required open-loop gain of the ISS, is actually a first order approximation of the 'required' noise suppression. In fact if we wanted the laser noise to be suppressed below the calculated brownian noise level, it is more correct to say
Closed-loop ISS gain = (Calculated Brownian Noise Limit) / (Unstabilized RIN)
As this essentially gives a noise suppression spectrum i.e. a closed-loop gain in linear control theory. Below is a very simple block diagram showing how the ISS fits into the CTN experiment. The F(f) block represents my full servo board.
Some of the relevant quantities involved:
So looking at the block diagram, our full closed-loop transfer function is given by,
So then to determine the required F(f), i.e. the required transfer function for my servo, we consider the expression
The plant transfer function is simply Plant = (C(f) * a * P * A) ~ 0.014 V/V, where I have ignored the cavity pole around 97 kHz as our open-loop transfer function ends up crossing unity gain around 10 kHz. In the above, I have included what I call a 'safety factor' of 10. Essentially, I want to design my servo such that it suppresses noise well beyond what is actually required so that we can be sure noise contributions to experiment readouts are not significantly influenced by the laser intensity noise.
Using the data Evan reported for the brownian noise and free-running RIN, I came up with an F(f) to the meet the requirement as shown below.
Where the blue curve includes the safety factor mentioned before. This plot just demonstrates that using my modular ISS design, I can meet the given noise suppression requirements.
To be complete, I'll say a little more about the final design. As usual, the servo consists of three stages. The first is the usual LP filter that is always 'on' when the ISS loop is closed. The boosts I have chosen to use consist of an integrator with a single zero and a filter that looks somewhat like a de-whitening filter. The simulated open-loop transfer functions are shown below.
Right near the end of summer, I had an ISS board that was nominally working, but had a few problems I couldn't really sort out. Since I've been back, I've spent a lot of time just replacing parts, trying different circuit topologies and generally attempting to make the board function as I hoped it might in all those design stages. Below is a brief list of some of the problems I've been fixing as well as the first good characterization of the board transfer function that I've been able to get.
We'll start with some of the simple problems and proceed to more complicated ones.
The above list encompasses all the issues I've had in making the ISS board function correctly. No other major problems exist to my knowledge.
I was able to measure both the open- and closed-loop transfer functions of the servo with the SR785. The results are shown below.
The transfer function with the boosts on caps at a particular value set by op-amp railing, i.e. below 100 Hz, the op-amps are already putting out their max voltage. This is the usual physical limitation when measuring the transfer function of an integrator. We can also see that the measured phase follows the simulated phase above ~300 Hz. The 'phase matching' at low frequency is again do to the op-amp railing in the servo output..
The closed-loop gain is shown below,
The measured closed-loop gain with the boosts on again matches the LISO simulation quite well except at low frequency where we are limited by op-amp railing. We compare the measured closed-loop transfer function to the desired noise suppression stipulated in my previous elog 9331,
And we might hopefully conclude that my servo functions as desired. One should note that the op-amp railing seen in these measurements is not indicative of limitations we might face in some application of the ISS for the following reason. These transfer functions were measured with a 100 mV excitation signal (it is necessary to keep this signal amplitude large enough so that the inherent signal-to-noise ratio of the excitation source is large enough for accurate measurement) which leads to somewhat prompt railing of the op-amps. When the ISS operates to actually stabilize a laser, the input error signal will be much smaller (on the order of a few 10's of mV or less) and will decrease significantly assuming correct operation of the ISS. This means we won't see the same type of gain limitations.
What now, you ask?
Aside from the problem with the AD8436 chip, the ISS board seems to be functioning correctly. The transfer functions we have measured are correct to within the component tolerances and all of the various subsystems are behaving as they were designed to. Moving toward the goal of having this system work in situ for the CTN experiment, I need to do the following things,
So close, or so I say all the time
The 40m experienced a building-wide power failure for ~30 seconds at ~7:38 pm today.
Thought that might be important...
For the purpose of testing out the temperature sensors, I stole the PEM-SEIS_MC1X,Y,Z channels.
I unplugged Guralp NS1b, Guralp Vert1b, Guralp EW1b cables from the PEM ADCU(#10,#11,#12) near 1Y7 and put temp sensors in their place (temporarily).
OK. Today we did the same type of measurement for the Y arm laser as was done for the X arm laser here: http://nodus.ligo.caltech.edu:8080/40m/3759
And attached here is a preliminary plot of the outcome - oddities with adding on the fitted equations, but they go as follows
(Red) T_yarm = 1.4435*T_PSL - 14.6222
(Blue) T_yarm = 1.4223*T_PSL - 10.9818
(Green) T_yarm = 1.3719*T_PSL - 6.3917
It's a bit of a messy plot - should tidy it up later...
I'm going to take the easy question - What are the pink data points??
And I'm going to answer the easy question - they're additional beat frequency temperature pair positions which seem to correspond to additional lines of beat frequencies other than the three highlighted, but that we didn't feel we had enough data points to make it worthwhile fitting a curve.
It's still not entirely clear where the multiple lines come from though - we think they're due to the lasers starting to run multi-mode, but still need a bit of thought on that one to be sure...
Just a quick update... the Lightwave laser has now been moved up to the end of the Y arm. It's also been mounted on the new mounting block and heatsinks attached with indium as the heat transfer medium.
A couple of nice piccies...
The good news is that we seem to be running in a linear region of the PSL laser with a degree or so of range before the PSL Innolight laser starts to run multi-mode. On the attached graph we are currently running the PSL at 32.26degrees (measured) which puts us in the lower left corner of the plot. The blue data is the Lightwave set temperature (taken from the display on the laser controller) and the red data is the Lightwave laser crystal measured temperature (taken from the 10V/degC calibrated diagnostic output on the back of the laser controller - between pins 2 and 4).
The other good news is that we can see the transition between the PSL laser running in one mode and running in the next mode along. The transition region has no data points because the PMC has trouble locking on the multi-mode laser output - you can tell when this is happening because, as we approach the transition the PMC transmitted power starts to drop off and comes back up again once we're into the next mode region (top left portion of the plot).
The fitted lines for the region we're operating in are:
Y_arm_Temp_meas = 0.95152*T_PSL + 3.8672
Y_arm_Temp_set = 0.87326*T_PSL + 6.9825
X_arm and Y_arm vs PSL comparison.
Just a quick check of the performance of the X arm and Y arm lasers in comparison to the PSL. Plotting the data from the X arm vs PSL and Y arm vs PSL on the same plot shows that the X arm vs the PSL has no observable trending of mode-hopping in the laser, while the Y arm vs the PSL does. Suspect this is due to the fact that the X arm and PSL are both Innolight lasers with essentially identical geometry and crystals and they'll tend to mode-hop at roughly the same temperatures - note that the Xarm data is rough grained resolution so it's likely that any mode-hop transitions have been skipped over. The Lightwave on the other hand is a very different beast and has a different response, so won't hop modes at the same temperatures.
Given how close the PSL is to one of the mode-transition regions where it's currently operating (32.26 degC) it might be worth considering shifting the operating temperature down one degree or so to around 31 degC? Just to give a bit more headroom. Certainly worth bearing in mind if problems are noticed in the future.
Right. I've got a whole load of info and data and assorted musings I've been saving up and cogitating upon before dumping it into these hallowed e-pages. there's so much I'll probably turn it into a threaded entry rather than put everything in one massive page.
An overview of what's coming:
I started out using http://lhocds.ligo-wa.caltech.edu:8000/40m/Advanced_Techniques/Green_Locking?action=AttachFile&do=get&target=modematch_END.png as a reference for roughly what we want to achieve... and from http://nodus.ligo.caltech.edu:8080/40m/100730_093643/efficiency_waist_edit.png we need a waist of about 50um at the green oven. Everything else up to this point is pretty much negotiable and the only defining things that matter are getting the right waist at the doubling oven with enough available power and (after that point) having enough space on the bench to separate off the green beam and match it into the Y arm.
Step 1: Measure the properties of the beam out of the laser. Really just need this for reference later because we'll be using more easily measurable points on the bench.
Step 2: Insert a lens a few cm from the laser to produce a waist of about of a few 100um around the Faraday. Note that there's really quite a lot of freedom here as to where the FI has to be - on the X arm it's around columns 29/30 on the bench, but as long as we get something that works we can get it closer to the laser if we need to.
Step 3: After inserting the FI need to measure the beam after it (there *will* be some distortion and the beam is non-circular to begin with)
Step 3b: If beam is non-circular, make it circular.
Step 4: Insert a lens to produce a 50um waist at the doubling oven position. This is around holes 7/8 on the X arm but again, we're free to change the position of the oven if we find a better solution. The optical set-up is a little bit tight near that side of the bench on the X end so we might want to try aiming for something a bit closer to the middle of the bench? Depends how the lenses work out, but if it fits on the X end it will fit on the Y end.
RIght! Overview out of the way - now comes the trivial first bit
Step 1: Beam out of the laser - this will be tricky, but we'll see what we can actually measure in this set-up. Can't get the Beamscan head any closer to the laser and using a lambda/2 plate + polariser to control power until the Faraday isolator is in place. Using 1 inch separation holes as reference points for now - need better resolution later, but this is fine for now and gives an idea of where things need to go on the bench. The beam is aligned to the 3rd row up (T) for all measurements, the Beamscan spits out diameters (measuring only the 13.5% values) so convert as required to beam radius and the beam is checked to ensure a reasonable Gaussian profile throughout.
Position A1_13.5%_width A2_13.5%_width
(bench) (um mean) (um mean)
32 2166.1 1612.5
31 2283.4 1708.3
30 2416.1 1803.2
29 2547.5 1891.4
27 2860.1 2070.3
26 2930.2 2154.4
25 3074.4 2254.0
24 3207.0 2339.4
OK. As expected, this measurement is in the linear region of the beampath - i.e. not close to the waist position and beyond the Rayleigh length) so it pretty much looks like two straight lines. There's no easy way to get into the path closer to the laser, so reckon we'll just need to infer back from the waist after we get a lens in there. Attached the plot, but about all you really need to get from this is that the beam out of the laser is very astigmatic and that the vertical axis expands faster than the horizontal.
Not terribly exciting, but have to start somewhere.
Step 2: Getting the beam through the Faraday isolator (FI).
Started out with an f=100mm lens at position 32,T on the bench which gave a decent looking waist of order 100 um in the right sort of position for the FI, but after checking the FI specs, it's limited to 500W/cm^2. In other words, if we have full power from the laser passing into it we'd need a beam width of more than 211 um. Solution? Use an f=150mm lens instead and don't put the FI at the waist. I normally don't put a FI at a waist anyway, for assorted reasons - scattering, thermal lensing, non-linear magnetic fields, the sharp changing of the field components in an area where you want as constant a beam as possible. Checked with others to make sure they don't do things differently around these parts… Koji says it doesn't matter as long as it passes cleanly through the aperture. So… next step is inserting the Faraday.
The beam profiles in vertical and horizontal around the FI position with the f=150mm lens in place are attached. Note that the FI will be going in at around 0.56m.
I fired up some old waistplotter routines, and set the input conditions as the measured waist after the lens and used that to work out what the input waist is at the laser. It may not be entirely accurate, but it /will/ be self consistent later on.
Vertical waist = 105.00 um at 6.282 cm after laser output (approx)
Horizontal waist = 144.63 um at 5.842 cm after laser output (approx)
Step 3: Inserting FI and un-eliptical-ification of the beam
The FI set up on it's mount and the beam passes through it - centrally through the apertures on each side. Need to make sure it doesn't clip and also make sure we get 93% through (datasheet specs say this is what we should achieve). We will not achieve this, but anything close should be acceptable.
Setting up for minimum power through the FI is HWP @125deg.
Max is therefore @ 80deg
Power before FI = 544 mW
Power after FI = 496 mW (after optimising input polarisation)
Power dumped at input crystal = 8.6mW
Power dumped at input crystal from internal reflections etc = 3.5mW
Power dumped at output crystal on 1st pass = approx 8mW
OK. that gives us a 90.625% transmission and a 20.1mW absorption/unexplained loss.
Well - OK. The important part about isolators isn't their transmission, it's about how well they isolate. Let's see how much power gets ejected on returning through the isolator…
Using a beam splitter to pick off light going into and returning from the FI. A 50/50 BS1-1064-50-1025-45P. And using a mirror near the waist after the FI to send the beam back through. There are better ways to test the isolation performance of FI's but this will suffice for now - really only want to know if there's any reasonable isolation at all or if all of the beam is passing backwards through the device.
Power before BS = 536 mW (hmmn - it's gone down a bit)
Power through BS = (can't access ejected on first pass)
Power through FI = 164 mW (BS at odd angle to minimise refractive effect so less power gets through)
Power lost through mirror = 8.3mW (mirror is at normal incidence so a bit transmissive)
Using earlier 90.6% measurement as reference, power into FI = 170.83 mW
So BS transmission = 170.83/536 = 0.3187
BS reflectivity therefore = 1 - 0.3187 = 0.6813
Power back into FI = Thru FI - Thru mirror = 155.7 mW
Power reflected at BS after returning through FI = 2.2mW
Baseline power at BS reflection from assorted internal reflections in FI (blocked return beam) = 1.9mW
Note - these reflections don't appear to be back along the input beam, but they *are* detectable on the power meter.
Actual power returning into FI that gets reflected by BS = 0.3 mW
(note that this is in the fluctuating noise level of measurement so treat as an upper limit)
Accounting for BS reflectivity at this angle, this gives a return power = 0.3/0.6813 = 0.4403 mW
Reduction ratio (extinction ratio) of FI = 0.4403/155.7 = 0.00282
Again - note that this upper limit measurement is as rough and ready as it gets. It's easy to optimise this sort of thing later, preferable on a nice open bench with plenty of space and a well-calibrated photodiode. It's just to give an idea that the isolator is actually isolating at all and not spewing light back into the NPRO.
Next up… checking the mode-matching again now that the FI is in place. The beam profile was scanned after the FI and the vertical and horizontal waists are different...
Step 3b: Non-circular? We can fix that...
A quick Beamscan sweep of the beam after the Faraday:
25.8 503.9 478.8
25 477.5 489.0
24 447.1 512.4
21 441.6 604.5
20 476.3 645.4
19 545.4 704.1
18 620.3 762.8
OK. It looks not too bad - doesn't look too different from what we had. Note that the x axis is in local table units - I found this useful for working out where things were relative to other things (like lenses and the FI) - but it means the beam propagates from right to left in the plot. in other words, the horizontal waist occurs first and is larger than the vertical waist. Also - they're not fitted curves - they're by-eye, best guesses and there's no solution for the vertical that doesn't involve offsets... discussion in a later part of the thread.
Anyway! The wonderful thing about this plot is that the horizontal and vertical widths cross and the horizontal focussing at this crossing point is shallower than the vertical. This means that we can put a lens in at the crossing point and rotate it such that the lens is stronger in the horizontal plane. The lens can be rotated until the effective horizontal focal length is right to fix the astigmatism.
I used a 200mm lens I had handy - a rough check sweeping the Beamscan quickly indicated should be about right though. Adjusting the angle until the beam size at a distant point is approx circular - I then move the profiler and adjust again. Repeat as required. Now… taking some data. with just that lens in:
24 371.7 366.1
21 360.3 342.7
20 447.8 427.8
19 552.4 519.0
18 656.4 599.2
17 780.1 709.9
16 885.9 831.1
Well now. That looks quite OK. Fit's a bit rubbish on vertical but looks like a slight offset on the measurement again.
The angle of the lens looks awful, but if it's stupid and it works then it isn't stupid. If necessary, the lens can be tweaked a bit more, but there's always more tweaking possible further down the line and most of the astigmatic behaviour has been removed. It's now just a case of finding a lens that works to give us a 50 um beam at the oven position...
Step 4: Matching into the oven
Now that the astigmatism is substantially reduced, we can work out a lens solution to obtain a 50um waist *anywhere* on the bench as long as there's enough room to work with the beam afterwards. The waist after the Faraday and lens is at position 22.5 on the bench. A 50 mm lens placed 18 cm after this position (position 14.92 on the bench) should give a waist of 50 um at 24.57 cm after the waist (position 12.83 on the bench). This doesn't give much room to measure the beam waist in though - the Beamscan head has a fairly large finite size… wonder if there's a slightly less strong lens I could use…
OK. With a 66 mm lens at 23 cm (position 13.45 on the bench) after the waist we get a 50 um waist at 31.37 cm after the waist (position 10.15 on the bench).
Closest lens I found was 62.9mm which will put the 50um point a bit further towards the wall, but on the X-arm the oven is at position 8.75 ish. So anything around there is fine.
Using this lens and after a bit of manual fiddling and checking with the Beamscan, I figured we needed a close in, fine-grained measurement so set the Beamscan head up on a micrometer stage Took a whoie bunch of data around position 9 on the bench:
(mm) (um mean) (um mean)
-15 226.8 221.9
-14 210.9 208.3
-13 195.5 196.7
-12 181.0 183.2
-11 166.0 168.4
-10 154.0 153.1
-9 139.5 141.0
-8 127.5 130.0
-7 118.0 121.7
-6 110.2 111.6
-5 105.0 104.8
-4 103.1 103.0
-3 105.2 104.7
-2 110.9 110.8
-1 116.8 117.0
0 125.6 125.6
0 125.6 125.1
1 134.8 135.3
2 145.1 145.6
3 155.7 157.2
4 168.0 168.1
5 180.5 180.6
6 197.7 198.6
7 211.4 209.7
8 224.0 222.7
9 238.5 233.7
10 250.9 245.8
11 261.5 256.4
12 274.0 270.4
13 291.3 283.6
14 304.2 296.5
15 317.9 309.5
And at this point the maximum power available at the oven-waist is 298mW. With 663mW available from the laser with a desired power setting of 700mW on the supply. Should make sure we understand where the power is being lost. The beam coming through the FI looks clean and unclipped, but there is some stray light around.
7 868.5 739.9
6 1324 1130
5 1765 1492
4 2214 1862
The plot looks pretty good, but again, there looks to be an offset on the 'fitted' curve. Taking a couple of additional points further on to make sure it all works out as the beam propagates. I took a few extra points at the suggestion of Kiwamu and Koji - see the zoomed out plot. The zoomed in plot has by-eye fit lines - again, because to get the right shape to fit the points there appears to be an offset. Where is that coming from? My suspicion is that the Beamscan doesn't take account of the any background zero offsets when calculating the 13.5% and we've been using low power when doing these measurements - very small focussed beams and didn't want to risk damage to the profiler head.
Decided to take a few measurements to test this theory. Trying different power settings and seeing if it gives different offset and/or a changed width size
7 984.9 824.0 very low power
7 931.9 730.3 low power
7 821.6 730.6 higher power
7 816.4 729.5 as high as I'm comfortable going
Trying this near the waist…
8.75 130.09 132.04 low power
8.75 106.58 105.46 higher power
8.75 102.44 103.20 as high as it can go without making it's saturated
So it looks like offset *is* significant and the Beamscan measurements are more accurate with more power to make the offsets less significant. Additionally, if this is the case then we can do a fit to the previous data (which was all taken with the same power setting) and simply allow the offset to be a free parameter without affecting the accuracy of the waist calculation. This fit and data coming to an e-log near you soon.
Of course, it looks from the plots above (well... the code that produces the plots above) that the waist is actually a little bit small (around 46um) so some adjustment of the last lens back along the beam by about half a cm or so might be required.
I went through the entries.
1. Give us a photo of the day. i.e. Faraday, tilted lens, etc...
2. After all, where did you put the faraday in the plot of the entry 4466?
3. Zoomed-in plot for the SHG crystal show no astigmatism. However, the zoomed out plot shows some astigmatism.
How consistent are they? ==> Interested in seeing the fit including the zoomed out measurements.
OK. Taking these completely out of order in the easiest first...
2. The FI is between positions 27.75 and 32 on the bench - i.e. this is where the input and output apertures are. (corresponds to between 0.58 and 0.46 on the scale of those two plotsand just before both the vertical and horizontal waists) At these points the beam radius is around 400um and below, and the aperture of the Faraday is 4.8mm (diameter).
Laser set up - note the odd angles of the mirrors. This is where we're losing a goodly chunk of the light. If need be we could set it up with an extra mirror and send the light round a square to provide alignment control AND reduce optical power loss...
Faraday and angled lens - note that the lens angle is close to 45 degrees. In principle this could be replaced with an appropriate cylindrical lens, but as long as there's enough light passing through to the oven I think we're OK.
3. Fitting... coming soon once I work out what it's actually telling me. Though I hasten to point out that the latter points were taken with a different laser power setting and might well be larger than the actual beam width which would lead to astigmatic behaviour.
3. Zoomed-in plot for the SHG crystal show no astigmatism. However, the zoomed out plot shows some astigmatism.
How consistent are they? ==> Interested in seeing the fit including the zoomed out measurements.
Right. Fitting to the data. Zoomed out plots first. I used the general equation f(x) = w_o.*sqrt(1 + (((x-z_o)*1064e-9)./(pi*w_o.^2)).^2)+c for each fit which is basically just the Gaussian beam width parameter calculation but with an extra offset parameter 'c'
Vertical fit for zoomed out data:
Coefficients (with 95% confidence bounds):
c = 7.542e-06 (5.161e-06, 9.923e-06)
w_o = 3.831e-05 (3.797e-05, 3.866e-05)
z_o = 1.045 (1.045, 1.046)
Goodness of fit:
c = 1.083e-05 (9.701e-06, 1.195e-05)
w_o = 4.523e-05 (4.5e-05, 4.546e-05)
z_o = 1.046 (1.046, 1.046)
Adjusted R-square: 0.9998
OK. Looking at the plots and residuals for this, the deviation of the fit around the waist position, and in fact all over, looks to be of the order 10um. A bit large but is it real? Both w_o values are a bit lower than the 50um we'd like, but… let's check using only the zoomed in data - hopefully more consistent since it was all taken with the same power setting.
Vertical data fit using only the zoomed in data:
c = 1.023e-05 (9.487e-06, 1.098e-05)
w_o = 4.313e-05 (4.252e-05, 4.374e-05)
Horizontal data fit using only the zoomed in data:
c = 1.031e-05 (9.418e-06, 1.121e-05)
w_o = 4.41e-05 (4.332e-05, 4.489e-05)
The waists are both fairly similar this time 43.13um and 44.1um and the offsets are similar too - residuals are only spread by about 4um this time.
I'm inclined to trust the zoomed in measurement more due to the fact that all the data was obtained under the same conditions, but either way, the fitted waist is a bit smaller than the 50um we'd like to see. Think it's worthwhile moving the 62.9mm lens back along the bench by about 3/4 -> 1cm to increase the waist size.
The doubling oven is now ready to go for the Y arm. The PPKTP crystal is mounted in the oven:
Note - the crystal isn't as badly misaligned as it looks in this photo. It's just an odd perspective shot. I then closed it up and checked to make sure the IR beam on the Y bench passes through the crystal. It does. Just need to tweak the waist size/position a bit and then we can actually double some frequencies!
Last bit of oven matching for now.
I moved the lens before the oven position back along the beam path by about 1cm - waist should be just above position 9 in this case. Note - due to power-findings from previous time I'm maximising the power into the head to reduce the effect of offsets.
From position 9:
-1 121.1 123.6
0 112.5 113.8
1 106.4 106.1
2 102.9 103.4
3 103.6 103.6
4 106.6 107.4
5 111.8 112.5
6 118.2 120.1
7 126.3 128.8
8 134.4 137.1
9 143.8 146.5
10 152.8 156.1
11 163.8 167.1
12 175.1 176.4
13 186.5 187.0
14 197.1 198.4
15 210.3 208.9
16 223.5 218.7
17 237.3 231.0
18 250.2 243.9
19 262.8 255.4
20 274.7 269.0
21 290.4 282.3
22 304.3 295.5
23 316.7 303.1
Note - had to reduce power due to peak saturation at 15mm - don't think scale changed, but be aware just in case. And saturated again at 11. And again at 7. A little bit of power adjustment each time to make sure the Beamscan head wasn't saturating. Running the fit gives...
OK. The fit is reasonably good. Residuals around the area of interest (with one exception) are <+/- 2um and the waists are 47.5um (vertical) and 50.0um (horizontal) at a position of 9.09 on the bench. And the details of the fitting output are given below.
cf_(x) = w_o.*sqrt(1 + (((x-z_o)*1064e-9)./(pi*w_o.^2)).^2)+c
Coefficients (with 95% confidence bounds):
c = 5.137e-06 (4.578e-06, 5.696e-06)
w_o = 4.752e-05 (4.711e-05, 4.793e-05)
z_o = 1.04 (1.039, 1.04)
c = 3.81e-06 (2.452e-06, 5.168e-06)
w_o = 5.006e-05 (4.909e-05, 5.102e-05)
z_o = 1.04 (1.04, 1.04)
We now have green light at the Y end.
The set-up (with careful instructions from Kiwamu) - setting up with 100mW of IR into the oven.
Input IR power = 100mW measured.
Output green power = 0.11mW
(after using 2 IR mirrors to dump IR light before the power meter so losing a bit of green there light too)
And it's pretty circular-looking too. Think there might be a bit more efficiency to be gained near the edges of the crystal with internal reflections and suchlike things but that gives us an UGLY looking beam. Note - the polarisation is wrong for the crystal orientation so used a lambda/2 plate to get best green power out.
Efficiency is therefore 0.11/100 = 0.0011 (0.11%) at 100mW input power.
Temperature of the oven seems to be around 35.5degC for optimal conversion.
Took a picture. Ta-dah! Green light, and lots more where that came from! Well... about 3x more IR available anyway.
Every so often things just work out. You do the calculations, you put the lenses on the bench, you manually adjust the pointing and fiddle with the lenses a bit, you get massive chunks of assistance from Kiwamu to get the alignment controls and monitors set up and after quite a bit of fiddling and tweaking the cavity mirror alignment you might get some nice TEM_00 -like shapes showing up on your Y-arm video monitors.
So. We have resonating green light in the Y-arm. The beam is horribly off-axis and the mode-matching, while close enough to give decent looking spots, has in no way been optimised yet. Things to do tomorrow - fix the off-cavity-axis problem and tweak up the mode-matching... then start looking at the locking...
The Y-arm can now be locked with green light using the universal PDH servo. Modulation frequency is now 277kHz - chosen because it seems to produce smaller offsets due to AM effects
To lock, turn on the servo, align the system to give nice circular-looking TEM_00 resonances, and wait for a good one. It'll lock on a decent mode for a few seconds and then you can turn on the local boost and watch it lock for minutes and minutes and minutes.
The suspensions are bouncing around a bit on the Y-arm and the spot is quite low on the ETMY and a little low on ITMY, but from this point it can be tweaked and optimised.
OK… the Y-arm may be locked with green light, which was the goal, and this is all good but it's not yet awesome. Awesome would be locked and aligned properly and quiet and optimised. So... in order to assist in increasing the awesome-osity, here are a few stream-of-consciousness thoughts and stuff I've noticed and haven't had time to fix/investigate or have otherwise had pointed out to me that may help...
Firstly, the beam is not aligned down the centre of the cavity. It's pretty good horizontally, but vertically it's too low by about 3/4->1cm on ETMY. The mirrors steering the beam into the cavity have no more vertical range left, so in order to get the beam higher the final two mirrors will have to be adjusted on the bench. Adding another mirror to create a square will give more range AND there will be less light lost due to off 45degree incident angles. When I tried this before I couldn't get the beam to return through the Faraday, but now the cavity is properly aligned this should not be a problem.
A side note on alignment - while setting cameras and viewports and things up, Steve noticed that one of the cables to one of the coils (UL) passes behind the ETMY. One of the biggest problems in getting the beam into the system to begin with was missing this cable. It doesn't fall directly into the beam path if the beam is well aligned to the cavity, but for initial alignment it obscures the beam - this may be a problem later for IR alignment.
Next, the final lambda/2 waveplate is not yet in the beam. This will only become a problem when it comes to beating the beams together at the vertex, but it WILL be a problem. Remember to put it in before trying to extract signals for full LSC cavity locking.
Speaking of components and suchlike things, the equipment for the green work was originally stored in 3 plastic boxes which were stored near the end of the X-arm. These boxes, minus the components now used to set up the Y-end, are now similarly stored near the end of the Y-arm.
Mechanical shutter - one needs to be installed on the Y-end just like the X-end. Wasn't necessary for initial locking, but necessary for remote control of the green light on/off.
Other control… the Universal PDH box isn't hooked up to the computers. Connections and such should be identical to the X-arm set-up, but someone who knows what they're doing should hook things up appropriately.
More control - haven't had a chance to optimise the locking and stability so the locking loop, while it appears to be fairly robust, isn't as quiet as we would like. There appears to be more AM coupling than we initially thought based on the Lightwave AM/PM measurements from before. It took a bit of fiddling with the modulation frequency to find a quiet point where the apparent AM effects don't prevent locking. 279kHz is the best point I've found so far. There is still a DC offset component in the feedback that prevents the gain being turned up - unity gain appears limited to about 1kHz maximum. Not sure whether this is due to an offset in the demod signal or from something in the electronics and haven't had time left to check it out properly yet. Again, be aware this may come back to bite you later.
Follow the bouncing spot - the Y-arm suspensions haven't been optimised for damping. I did a little bit of fiddling, but it definitely needs more work. I've roughly aligned the ETMY oplev since that seems to be the mass that's bouncing about most but a bit of work might not go amiss before trusting it to damp anything.
Think that's about all that springs to mind for now…
Thanks to everyone at the 40m lab for helping at various times and answering daft questions, like "Where do you keep your screwdrivers?" or "If I were a spectrum analyser, where would I be?" - it's been most enjoyable!