Found at least two (minor) problems:
it is not a lot of difference in slope but it is surprisingly linear over a range of 40MHz!
So, from the first plot, is the calibration factor for cable delay technique ~ 8MHz/Volt?
yes, for yesterdays measurements 7.435 MHz/V.
But that number changes from measurement to measurement as i'm changing the setup every time to improve SNR (or bring back stuff i've stolen from 40m)
Today we compared beat noise measured from phase lock loop(PLL) and cable delay technique (CD.)
The results do not agree well. Eventhough we can get the calibration factor correctly, CD is not working yet.
We want to use CD instead of PLL because CD has no phase noise added to the signal which is the limiting
source for PLL at high frequency.
One drawback is that to get the calibration factor for CD. It has to be done every time with each measurement
as the calibration factor varies with the carrier frequency of the beat, cables used for the measurement, but it's not that hard.
We use CD to measure the phase noise of IFR2023b used for PLL, its phase noise was measured and added in the noise budget.
The result at 100kHz input range from CD agrees well with the previous measurement, we can see the feature of the phase noise nicely.
This ensures that we got the CD calibration [Hz/V] correctly.
Then we switch to a 500 ft cable for better resolution and measure the phase noise of the LO again.
The results from 100 kHz input range are similar.
However, for phase noise at 10kHz input range (with lower phase noise,) CD cannot measure it properly by either short or long cables.
there is no phase noise's feature at high frequency like the previous result measured by SR785, note that
CD's noise floor is still lower than LO phase noise.
With that in mind, we measured beat noise by PLL(10kHz input range) and CD (500ft +2w amplifier).
We also injected a 1.092 kHz peak, generated by a func generator through a speaker on the table, to compare for the calibration.
1)If we use our regular calibrations for PLL and CD, the noise looks like this
2)If we match the 1.092 kHz peaks in both plot (I use a regular calibration for PLL). The result looks like this
And the coherence between PLL and CD is quite low (we have to think about it if the two signal should be coherence or not)
We will figure out what happen with the cable delay tech.
To sum up, we
1) make sure that we get the calibration factor correct for 500 ft cable setup with 2W amplifier,
by measuring ifr2023B phase noise at 100kHz range and compare it to the previous result.
2) are not certain why the LO phase noise and beat results are not similar at 10 kHz range
3) trust signal from PLL more than CD, since the beat noise from CD changed
(while PLL gave the similar result)when we adjusted the equipment, and we could not find out why.
Today, we found out that at high frequency, the limiting noise source is probably electronics noise might be from the Universal PDH box.
We tweaking the setup parameters, FSS gain, UPDH gain, sideband power, laser power to see what will change the beat noise.
We learned that
1) The current setup is not FSS gain limit, because we adjust the gain, but the beat noise remains the same,
and the range of the gain we can adjust is quite high, between 5 - 30 dB.
2) By adjusting side band power, the noise level changes.
However, when we added gain to compensate for the lower side band power, the noise does not go back to original level.
This is weird. We adjusted decrease 35.5 MHz sideband power, and noise level goes up. Increasing gain in FSS and ACAV loop does not change it back.
3) The peaks around acoustic frequency (mechanical peaks)can be suppressed by increasing PDH gain, but the flat level changes.
The noise level calculated from input referred noise sitll does not match with the beat noise, but they are close.
Now the beat noise is not limited by LO phase noise.From the flat shape of the noise, we think that it might be the electronic noise
from the PDH box.
To do next: increase the error signal slope for ACAV path. Now it is very small ~23 mV.
you have to be a bit more precisely fro points 2 and 3, e.g. no one knows what you mean as "PDH" gain as you have two loops using the PDH technique.
I had a closer look into effects caused by power fluctuations. To summarize : It is easy to measure TFs from power fluctuations to any other point in the system. SNR is good to very good.
I mainly focused on the effect of changes in shape of the TF from power fluctuations into changes of the beat frequency when changing the power levels anywhere in the system, especially in the cavity paths.
After playing a couple of hours and getting a feeling for what's going on i finally realized that only if changing the power in the ACAV-path strange things happen. This includes changing the total power going to both cavities as well. As taking TF's is taking too long at low frequencies i decided to switch from swept sine measurements to a simple digital modulation of the laser power with 100ms period. This is slow enough to see thermal effects in the cavity and other strange things which i will show below.
For the first set of measurements i reduced the power to the REFCAV path to about 460uW. The total modulation is very large, about 15.6%. The max power to ACAV was about 5.76mW, which i reduced later.
The following graphs shows the response of the beat signal to the digital modulation, beat signal calibrated to Hz, modulation signal a.u.., but the amplitude didn't change between measurements.
on the left graph (measured at very low power) one can see that the power modulation causes a sudden changes in frequency of the beat signal.
This can be caused by everything producing an offset in the EP which depends on power, e.g. RF-AM, higher-order modes in reflection etc. Nothing unusual so far.
If the optical power going to the ACAV is increased one can find a point where the change in frequency is almost gone, showed in the center graph.
When the power in increased further the sign of the frequency shift changes! Now some thermal effect becomes dominant, most likely the cavity as the period is 10s and after 5s equilibrium isn't reached.
Now this explains everything i saw before: When measuring the TF from power fluctuations to frequency shift and varying the power the shape of the TF changed because we have two different effects which can even cancel at the right power level send to the cavity!
As we know already that the size of the error signals is too tiny this effect will probably be almost gone once we fixed the demodulation for ACAV. I tried to reduce it a little bit by re-aligning the cavity but no luck - it's already aligned very well. I didn't try the EOM asthis changes the alignment to the other cavity as well and i was to lazy to re-align everything
I lock both cavities, setpoints are.
ACAV 37.0 C
RCAV 34.85 C
great, so we can check how sensitive the beat signal scatter bump is to cavity motion if we excite individual eigenmodes of both cavities before we change the suspension. This (hopefully) gives us information about which motion we are most sensitive at the moment so we can optimize the "new" suspension for that.
Beat noise measurement is measured while the table is being shaken by PZT at different frequencies.
Only 3.73 Hz (RCAV vertical) significantly affects the beat signal, even with good beam alignment.
Today I measured beat noise again to see where we are after adjusting many parameters.
The beat measurements are plotted below for 100 kHz and 10 kHz input range.
For 100kHz input range (grey), we are limited by LO phase noise. The shape of the phase noise has changed
from previous data in pink.
For 10 kHz input range (red), something else is the limiting source. I'm checking if we are limited by RFPD noise or not, but
the result is not conclusive yet.
I also measured the noise at the error point for ACAV loop (loop closed) in green, then projected it on the nb
by multiplying by the slope of the error signal (0.252 MHz/V = cavity's FWHM / error signal pkpk value = 108 kHz / 428 mVpk-pk.)
The level of the noise match the beat around 30Hz to 1kHz, but diverges at higher frequency.
Not quite sure if it makes sense or not (PLL loop is valid upto ~ 3kHz.)
The input referred noise of RFPD and the PDH box is measured, by blocking the beam on the RFPD
and measuring the singal on PDH out. The data is divided by the TF of the PDH box ( I measured this while the loop is closed,)
and multiplied by the error signal's slope, then plotted in blue.
The result is not quite right. It's higher than the beat. I'll have to check this.
*ACAV gain is set to 1.5 for all measurements.
RCAV common gain is set to 22.0 dB
power to each cavity = 1 mW
I noticed that when the power to both cavities increases (6 mW to 12 mW), the beat noise get smaller at high frequency, broadband (200Hz and above.)
The problem is when I increase the power, I change several things:
1) + 2) error signals from both cavities
3) shot noise on RFPD
4) SNR on the beat
,so I haven't show the improved beat signal before I can pinpoint what actually limits the signal.
I checked the noise from both RFPDs, from 1MHz to 50MHz, ACAV's RPFD looks very different from RCAV's RFPD.
ACAV's RFPD noise also changes between morning and evening measurements, I'm not sure what is happening here.
The first set of data (in solid lines) is taken with the amplifier to increase the noise level, then divided by 30 which is
the increasing factor of the amplifier, see below. The second set in dotted lines are taken without the amplifier.
The signal without the amplifier is quite low ~ 5 - 10 nV which is close to the noise floor of the analyzer already.
This might cause the factor of 3 in the signal of RCAV's RFPD noise.
About the changing noise shape of ACVA's RFPD, I stil don't have a clue. I'll check again tomorrow.
so I tried to switch the RFPDs to see if the beat signal can be improved.
Since the error signal on RCAV's loop is better, I thought that in can compensate with the effect from
bad RFPD. However, I could not get it locked, I increased the signal with the low noise amplifier before the mixer.
Still could not lock it. I returned the setup back to original and still could not lock it. No error signal coming out during the scan.
I worked around and then it's back on lock. So I guess there might be loosen connectors in ACAV's RFPD.
Note about the low noise amplifier I used:
The amplifier increases the signal by a factor of 30 (29.5 dB) flat from 1MHz to 100MHz, I used this to amplify the RFPD noise so that it's higher than the noise floor of the analyzer.
I checked electronic noise, of the UPDH box while tuning its gain, to understand why the beat signal gets higher when the gain on UPDH increases.
I still don't understand ACAV loop's behavior, but it is not the current limiting noise source, either RCAV loop or PLL is.
Right now, the beat signal at high frequency is limited by an unknown source(s). But by changing the gain of the UPDH which is used
for locking the beam to ACAV, the beat signal gets higher, before the oscillation in the loop occur.
I would expect the loop to become unstable first, so the beat gets worse, but this is not the case. So I measured
error signal, Vout to VCO, TF of the UPDH, noise of the UPDH, and varied the gain.
This UPDH has knob to adjust its gain. The dial reads from 0 to 10.
I check that the beat signal does not change when the gain of the UPDH is changed from 0 to 1.5, then it starts to get higher, broadband at high frequency, 200Hz and above.
So I varied the gain and recording the following(I made sure that the loop is still stable, no oscillation in ACAV_RCTRANSPD)
1)Above, beat measurement
2)Above, error signal, during closed loop
3) actuator signal to VCO, during closed loop
4)noise of the UPDH box, 50ohm terminated input,
5) TF between input and output(UPDH TF)
6) so from, 4 and 5 I can compute the input referred noise of the box. The result is unexpected. The input referred noise is getting lower
as the gain increases. With the error signal slope of 0.11 MHz/V the input referred noise is still lower than the beat (~ 10^-1 Hz/rtHz.)
So the input referred noise of the UPDH that changes with gain might not be the limiting noise.
However, when I project the noise from error point to beat noise with the slope of the error signal being 0.11 MHz/V.
The noise when the gain is 1 and 1.5 are lower than the beat. However, when the gain is 3.2 and 4 where the beat is getting higher,
the noise projection match the beat signal.
This means that
1) the noise projection factor I get this time is correct,
2) the limiting noise at gain 1.5 (usual setup) is not from ACAV loop.
3) UPDH will be the limiting noise source very soon. Since at gain 1.5, it is only a factor of 2 lower than the current beat.
So the next thing is to check if the limiting noise source comes from RCAV loop or from PLL loop.
* note: the phase shift dial on UPDH box is probably broken. I could not change the error signal with it.
I'll have to think over about what causes this.
I check the noise projection from RCAV's error signal. It is still lower than the beat frequency. So the limiting noise source is not from RCAV loop.
From previous entry, I measured the slope of the error signal for RCAV loop. The value is 0.275 MHz/V. So I can calculate the frequency noise level from RCAV loop by measureing the noise at the error point, then multiplying the noise by the slope of the error signal.
The noise level is about 2*10^-2 Hz/rtHz which is lower than the current beat noise (~0.1 Hz/rtHz).
Since the beat signal contains all noise from RCAV ACAV, and PLL loop, and I verified that the noise contributions from RCAV and ACAV loop are smaller than the measured noise. The limiting noise must come from the readout part.
So the next thing to look into is the noise in PLL loop.
I made sure that the gain setup for ACAV's servo was not too high and caused instability in the loop. The beat was not improved.
Previously from this entry, we see that noise at the error point, ACAV loop, increases with the ACAV servo gain.In general, it should decrease, since larger signal from actuator can suppress more noise. Frank told me that it might start oscillating somewhere and should check that the gain setting we use still make sense (increase gain, lower noise at error point,) so I added more data from gain 0(minimum) and 0.5 to show that the noise is actually going down as gain increase at these gain setup, and the beat signal still remain the same.
Above, noise at error point during close loop operation.
Above, noise at error point projection to frequency noise. From gain 0 to 1.5 noises at error point still lie below the measured beat.
Since the error signal measured above might not be very clear, I repeated the experiment by adding a 20 dB attenuator at the UPDH input to make sure that the signal from the mixer is not too large and cause any osscilation. However, there is no improvement in beat measurement. I think the PLL readout part is the most suspicious one right now.
I searched for the limiting noise source at high frequency of the beat measurement. I have not founded anything conclusive yet, but the key
seems to be the power on ACAV path.
From previous entries, the evidence suggested that I should look into the PLL readout technique for beat measurement, so I tried:
1) changing the beam spot size on the beat RFPD, to see if it was the result from scattering on the RFPD. the beat still looks the same even the beam is clipped.
2) adjusting + removed the 1/4, 1/2 wave plates that turn the circularly polarized beams from to cavities to linearly polarized beam,
3) using ND filters to reduce power from each of the beam, and both of them,
4) checkig the noise level of the RFPD and the mixer out, gain 10. They are lower than the beat.
So It is probably not the PLL problem. Then I revisited the idea of the noise in ACAV/ RCAV loop.
I used a 0.3 ND filter to reduce the power in RCAV path, after I adjusted the gain, the beat still remained the same.
However, when I used the filter to reduce the power into ACAV, beat was getting higher, and could not be brought down with the higher gain.
So I increased the power to ACAV from 1 mW to 2 mW, and maintained the power to RCAV at 1mW and measured the beat.
The flat high frequency noise goes down by about a factor of 2. This is surprising, I checked the error noise in ACAV loop and made sure that it was not the limiting part. I will think about it.
Note that peaks around 170 Hz are from the periscopes, and peaks around 600Hz are from the mirror mount on the periscopes.
I plan to replace them with better one after RCAV's chamber is fixed tomorrow.
The optics behind RCAV for beat measurement are re-installed, with the new periscope. I'm waiting for the temperature to settle.
The periscope behind RCAV was removed when we opened the chamber and took out the cavity. Now everything is back in place
The new periscope is installed. The bottom mirror is still on the same mount we have used before(fig1), but the top part is removed, so it should be less sensitive to seismic compared to what it was before (fig2).
I removed the hose between the RCAV chamber and the turbo pump since the valve and the turbo were turned off. Then I closed the insulation box.
The yellow foam insulation on RCAV was fixed. I melted it a bit to make sure that no part of the beam is blocked by the insulation.
fig1: new periscope setup
fig2) previous periscope
I realigned the beam on the beat path and measured the beat. With new eddy current damping, the peaks at 3.73 Hz now becomes smaller.
New beat layout, both paths travel with the same distance before combining at the beam splitter. I'll try if this layout work tomorrow.
I'm a bit worried that the CCD cameras might not fit on the table.
Beat path was setup. Beat RFPD sees the signal around 125 MHz. Beat measurement will be done tomorrow after ACAV is locked.
Beat signal from double cavities in the same chamber was measured. At DC to 100Hz, it seems to be dominated by seismic. Above 100 Hz, the frequency's laser noise is the limiting source. The setup has yet to be optimized. This is for a quick check to see how beat signal changes with the new seismic isolation setup.
I measured beat signal after Paul and Frank locked ACAV. At frequency above 100 Hz, beat signal changes with RCAV's gain setup. Error signal from RCAV's mixer out matches the shape of the beat signal ( I did not do the calibration with error signal slope, just observed the displays on SR785). However, at low frequency (DC - 100 Hz), the beat signal does not change with anything, RCAV gain, ACAV gain, PLL gain, so I'm quite certain that it's the real signal we see here.
Tuning range on PLL loop was 10kHz, the calibration is 7kHz/V. A peak close to 7 Hz might come from the beam line transverse mode of the stack.
At DC - 10 Hz, the noise is lower than before. This might be the result from common mode suppression. However, the two spacers are not identical. They have significantly different holes' sizes. This may explain why the seismic cancellation is not that good at higher frequency.
We might have to suspend the whole vacuum chamber to win against seismic.
11:47pm : the current VCO frequency @26.5degC is ~71.294MHz still slightly drifting towards higher frequencies - will wait a few more minutes before i step it up.
11:56pm : now 71.318MHz @26.5degC
12:27am : now 71.373MHz
12:42am : now 71.394MHz
1:54am : now already 71.497MHz
2:39pm : now steady at 71.858MHz
changed set point from 26.5degC to 26.6degC for a first test at 2:42PM local time on fb2. Tuned the VCO feedback to be ~zero. Tuning coeff same as mentioned in post 682
By taking the passive transfer function between a vertical seismometer on the table and the individual cavity signals, we should be able to see which direction to move the cavity supports so as to minimize the seismic coupling.
Our first iteration probably will have a sign error, but by making a few iterations we ought to be able to home in on a better support. Also, we should compare the theoretical estimate with the measured coupling in units of strain/(m/s^2).
Tara showed me a quick plot which showed the spectrum taken with the new (current) setup and the one taken when we removed the springs from the wire suspension. They look pretty identical between 10Hz and 100Hz or so. So it is likely that we see a lot of vertical seismic. I've measured a little bit at low frequencies to see where we are and we are better than before now, i would guess an order of magnitude or so, without any optimized stabilization, alignment and loops. Tara will take a nice set of measurements tomorrow and make a nice plot. The peak at 6.7Hz or so is actually the horizontal motion of the two top stack plates (not only the top plate). I used our pzt-shaker to shake the table and even with a small signal i could increase it until i got scatter noise bumps around 100Hz. So we have to damp this somehow. Is there an easy way to get some more vertical isolation? what about putting the top plate on a few springs instead of rubber? How much do you typically get when floating the table? I don't have realistic numbers for that...
We also don't have a seismometer. Jenne took her's back and Jan shipped the others back to where they came from i think. i think we should get one which we can share between labs in bridge which we can keep for longer. We needed one quite often in the past couple of months and i guess once we start with the cryo cavity we will even more frequently. Any idea where to get a cheap one? We don't need an STS-2 or so... Seismic sucks anyway in bridge...
The problem with moving supports is that the spacer has pretty wide groves. But we have to think about a clever support anyways. Currently it's sitting on viton in the groves of the spacer, which, according to comsol, is very close to the optimum position but who knows in reality.
The beat signal I measured today(orange) has noise level close to what we had before (blue,purple), so I'll try to check it again tomorrow. Meanwhile, this is the plot with beat signal when we modified the suspension (removed the spring), and the cavities were in two separate chambers (blue). The noise level around 20-100 Hz is very similar to what we have right now. Since the spring filters mostly vertical seismic, these peaks should be mostly vertical seismic coupling into the system.
as we still don't have a good model for the horizontal isolation of the stack we will try to measure the TF from horiz. accel. of the table to the beat using the PZT actuator between steel frame and table. With the table resting on the legs we only get a nice TF around the resonance of the stack at ~6.5Hz, the rest is too noisy. We will fix the broken legs to float the table tomorrow and then try again.
After switching to 14.75 MHz sideband setup, aside from the RFPDs, we did not change anything yet. We use TTFSS setup on the table. The seismic stack is still the same. The table is not floated. So we measure the beat noise as a reference.
we use 14.75 resonant EOM for adding sideband, and the broadband EOM for feedback. The resonant EOM is driven by an LO at 14.75 MHz @ 20dBm. The broadband EOM is supplied with 25V power supply, it is operated at low voltage because of the limited current source.
RCAV is locked by TTFSS, ACAV is locked by UPDH, the phase shift for both loop is done by cable length adjustment.
Note that we removed the Faraday isolator from the setup due to the limited space. It will be installed back later once we change the EOM base.
The power input is 0.2 mW on both cavities. The visibility is more than 80%.
==Problem and Plan==
We saw that beam alignment caused the bump around 1kHz to change significantly, so we will re-align everything to make sure that we have good mode matching.
The power input was chosen to be around 0.2 mW because at 1mW the error signal is irregular. It might be the RFPD problem, we will look into it.
We added a ccd camera to monitor the reflected beam from the locked cavity. The shape of the beam(turns out to be LG10) tells us that we have to do a better mode matching.
We found that bad alignment can cause extra noise in the beat noise, see (psl:796).So we monitored the reflected beam from the locked cavity to check what kind of misalignment we have to fix. Currently, the shape of LG10 is the dominant one which tells us that we do not have a good spotsize/spot position match which can be fixed by moving the lenses in front of the cavity.
above, top right panel, the shape of the beam reflected from the locked RefCav.
So we will move the lenses in front of the cavity to optimize the mode matching. The current value is ~ 96.5%, we want to improve it without spending too much time, not more than a day. We will mount the lenses on translational stages for better position adjustment.
We investigated the beat noise, and found out a few issues we have to fix to improve the sensitivity.
Resonant peaks from periscopes around 800 Hz are high. We use an aluminum bar to tighten the periscopes together. This bring down the peaks. However, the peaks are still high. We certainly have to work on the design for the periscope to get rid of the mechanical peaks.
upper beam for dual periscope. This help reducing resonant peak around 800 Hz a lot.
The local oscillator we use for driving the AOM in ACAV loop is Marconi. We are sitting on its phase noise at 10kHz tuning range .When the tuning range is set to 1kHz, noise at frequency above 1kHz drops. However, with 1kHz tuning range setup, we don't have enough gain to suppress noise at lower frequency and noise around 100Hz goes up.
We need to characterize ACAV loop to project LO's phase noise on the noise budget (this has not been done yet). We might have to add another EOM for feedback on ACAV loop.
We also noticed that talking noise can easily couple to the beat signal, but we are not certain where it happens. All mounts and posts will be checked if they are loosen or can be improved.
The noise budget is updated and plotted with today's beat measurement.
(*The electronic noise plotted in the graph is not correctly calibrated, see psl:816 for the complete calculation)
After we replaced the table's broken leg, we floated the table and measured the beat signal as a reference before modifying the seismic stack. The calculation agree with the measured data quite well.
We also measured the electronic noise from PLL. This was the signal which was fed back to the LO (with SR560 gain = 20). Apparently, we are sitting on it at 1kHz and above. We definitely need to work on PLL readout system to measure at lower sensitivity.
New traces in the noise budget:
I removed noises from room temperature fluctuation/ heater noise and spacer thermal noise because they are way smaller than coating noise and crowd the plot.
==Details about some traces in the noise budget==
The vertical seismic noise coupling is calculated by applying the seismic measurement times stack transfer function times cavity bending coupling
[ Frequency noise from seismic ] = [measured data] x [stacks TF] x [bending factor].
The peak around 6 Hz might be coupled from horizontal direction. This will be added soon.
RIN induced length noise is still an estimated. We have not been able to measured the real coupling yet, as the SNR is so low.
LO phase noise: This is from measurement. I'm not quite sure if I miss some calibration factors. The phase noise does not show up in the beat yet even though it is very close together right now.
Noise calculation from PD in PLL: ( I actually asked Koji once and did this already, see psl:730 . The results are similar)
1) determining which setup gives the best performance:
fig1: OLG TF of PLL with different gain setup.
2) Measure electronic noise from readout system with the chosen setup. This noise will show up (after some correction) in the beat and determine what is the limitation of PLL readout technique.
The PD was blocked, the feedback signal (Vfb) to the actuator (LO) was removed and measured.
3) Block diagram
[add block diagram and calculation]
4) After Koji explained on how to calculated noise budget from electronic noise in PLL to us, here the nb with PLL noise. (note: the LO phase noise has updated to 1kHz input range)
With the electronic noise from PLL, the sensitivity of this technique will prevent us from observing coating noise above 1kHz.
I'll calculate the noise from cable delay technique later and compare which one will give us better sensitivity.
mixers (which we currently have (and use) in the lab):
cables calculator for cable loss, which has a huge amount of different cable types in it's database. : http://vk1od.net/calc/tl/tllc.php
velocity ~0.66 to 7*c
RG-58C/U: loss 32 dB for 500ft @160MHz
RG-142: loss 25 dB for 500ft @160MHz
4-way splitter ZBSC-413 (datasheet)
additional insertion loss 0.5dB
power Input: 1W max
- gain 16dB
- output power: 28dBm min. (1dB compression)
- noise figure: 11dB
- VSWR 2:2 (in and out)
- 2 to 500 MHz
- gain 24dB
- output power: 5dBm min. (1dB compression)
- noise figure: 2.9dB
- VSWR 1.5:1 (in) and 1.8:1 (out)
- 0.1 to 500 MHz
we replaced the mount for the combining beam splitter in the beat setup as it caused a large, broadband peak in the spectrum around 1.4kHz. The new mount is one of the old, fixed turning mirror blocks they used in initial LIGO at LLO as far as i know. After replacing the mount the peak is entirely gone. I've used two springs instead of one to increase the pressure. We could not determine the resonance frequency of the new mount. Tapping the mount excites only known mechanical resonances from the surrounding mirror mounts. Tara posted a plot for comparison before and after replacing that mount (see here). He also has prepared a nice plot combined with a drawing which mount corresponds to which resonance we see in the spectrum. We will use this to start reducing (or even eliminating) those resonances starting with the most dominant ones close to 1kHz
Attached a copy of the drawing.
Did a little bit of peak hunting to clear our frequency span of interest from those massive mechanical resonances we currently have. After replacing the combining beam splitter mount we got rid of the 1.4kHz peak already. Yesterday i've focused on the mounts within the beat setup, but not the periscope, as we already know that this is very unstable and we will take care of that soon. I didn't want to replace things, just know where which stuff comes from.
I've found (only) one mirror mount which is currently clearly visible in our noise spectrum . Tapping the other mounts or damping the front plate or springs does not change the spectrum (at least i don't see any changes). Tapping (even slightly) is very difficult anyway as you also excite all the mounts surrounding your DUT, especially the periscopes and your whole spectrum changes and it's hard to figure out which is your primary resonance you are looking for. So i prefer damping it with a large piece of rubber and than compare it with a spectrum taken before with a reasonable.amount of averages.
Anyway, i found only one mirror mount (out of six) which i could clearly identify in our current noise spectrum. It's one of the mirrors right in front of the combining beam splitter.
Below a comparison before and after damping the front plate of the mirror mount. Resonance frequency is 544Hz. I have to check but i think we can replace this one with a non-adjustable turning mirror.
We still don't know where the 1.1KHz stuff is coming from.
from Measurements of Earth-station delay instabilities using a delay-calibration device measured at 70MHz
other data (e.g. for RG223/U) can be found here : http://tmo.jpl.nasa.gov/progress_report/42-99/99E.PDF
or here : http://ivs.nict.go.jp/mirror/meetings/v2c_wm1/phase_stability.pdf
or here: http://tesla.unh.edu/courses/ece758/Handouts/cable-specs.pdf
Type of cable Temp coef (ppm/K)
RG-223 -40 to -100
Belden 9913 -21
Andrew FSJ4-50B -2 to +6
Andrew LDF2-50 -8 to +6
Andrew LDF4-50A +7 to +16
And for the LMR-240 which i would buy for future cable delay lines ~24ppm/K
measured the delay for the old cable (RG58): dPhi=180deg, df=600KHz
1.67ns/ft (value from datasheet: 1.53ns/ft)
typical values for other cables using the following dielectric materials:
Dielectric Type Time Delay (ns/ft)
Solid Polyethylene (PE) 1.54
Foam Polyethylene (FE) 1.27
Foam Polystyrene (FS) 1.12
Air Space Polyethylene (ASP) 1.15-1.21
Solid Teflon (ST) 1.46
Air Space Teflon (AST) 1.13-1.20
We removed the periscopes in beat path and use breadboard setup instead. There are higher broadband noise in the beat around 100 - 3kHz. At least, the peak around 800 Hz is slightly small, since the contribution from the periscope in the beat path was removed.
We are trying to get rid of individual mechanical peaks in the beat signal. One of the major peaks comes from the periscopes and the associated mirror mounts. So instead of using periscopes to bring the beam height down, we use breadboard setup to bring the whole beat path up. This new setup gets rid out the periscopes, and 4 mirror mounts on it.
Frank has a solid work assembly file showing how the setup should look like. However, the beam coming out of the cavities are not exactly on the whole pattern, and without the periscope we have no way to steer the beam to the designed path. As a temporary solution, I use a post mounts which are clamped on the board, not screwed.
Fig1: The beam path is not on the whole pattern on the table, so the block mirror mount cannot be used.
fig 2: current breadboard setup. The mirrors behind quarter waveplates are mounted on a regular post and clamped down on the board. If I use the block mount, the beam won't hit the mirror. Note: I use the previous beam splitter post (with 1.4kHz resonant peak) for now, because we don't have anything that fits right now.
fig3: Comparison between original setup (in blue,with periscope) and breadboard setup (green). Both signals were taken with 2kHz tuning range, gain 200.
I think the broadband noise might come from the scattering on the PD. THe spotsize on the PD is not much smaller than the PD. I haven't found the appropriate lens for modematching yet. The peaks around 1kHz-2 kHz also seem to be more than the regular setup. This will be investigated.
started with some simple calculations for replacing the PLL with a delay line. Started with modeling the loss in the cable depending on frequency and length (separate matlab-function for different cables).
Below some first plots for our current "situation" (which probably changes in the near future but that's what it is right now) having a beat note @ 160MHz , 5dBm from the PD and an ZHL-1A amplifier (16dB gain) and a 4-way splitter (for two delay lines with different cable length):
files are on the svn in "CTNLab\simulations\noise_budget\delay_line_readout".
The optimum sensitivity is reached when the decrease in output signal is compensated by the increase in 2*pi*tau, which happens with a total loss of 8.68dB (factor 1/e) of the cable.
We don't win with adding delay if we make the cable longer, even if we increase the power going into the cable! That also explains why i had such a poor sensitivity with the 500ft of RG58 (which had 33dB of loss). Using 15ft of cable instead would have given the same sensitivity!
UPDATE 2/14/2012@8pm- files on the SVN contain now also data for RG405 and LMR-400 !
now as we know that the optimum loss of the delay line is 8.68dB we can calculate the optimum cable length.
optimum length for 160MHz are:
cables which introduce more delay for the same amount (8.68dB) of loss are better.
Now, we compare the minicircuits low-pass filter SLP-200 (datasheet) with the cables.
so we could add 22 filters for an optimum total delay/loss ratio. Total group delay would be 132ns.
If we compare now with the delays we get from the cables we see that even the simple RG58 gives us 50% more delay for the same loss ( and the price for the cable is the same as a single filter).
Using RG142 instead we get almost a factor of 2 more sensitivity and even more using lower loss cables.
So i don't see an advantage using those LP filters instead of cables.
I checked the mechanical peaks in breadboard setup. We get rid of the peak at 800 Hz from the periscope and a big peak around 200 Hz. However, there are some new peaks popping up which are not identified yet.
As the beams are not on the whole pattern, Frank suggested that I move the whole board and clamp it down instead. In order to do that, I had to remove the QWPs. I clamped it down with steel clamps around the legs. After that I inserted some damping posts (see below pictures). It turned out that the rubber damping does not help much at this point the noise spectrum does not change at all between with or without the rubber. I think it is because I cannot insert enough rubber between the board and the post, as I slid them in after I clamped the board. I should have placed the damping posts in their places then clamped down the board.
[add clamping fig]
fig: beat noise in different setup: Blue: beat path with periscope, Magenta: beat with breadboard setup, Green: with damping on mirror mounts. The data were taken with 2kHz tuning range, gain 200.
Since only the beat path that has been changed, all the peaks that popping out are contribution from the breadboard setup.
So things that are changed are:
There are a few new peaks due to the breadboard setup which have not been identified yet. It is very hard to check, since tapping with slight force already excite the peaks of the mirror mounts around 1 kHz. Once the mounts are damped, other peaks might be easier to be found.
Mixer will be driven very hard to saturate it. To operate the mixer in the required saturated mode, the RF signal level should be at least:
so if we use the right (optimum) cable we would have ~8dBm, which should be perfect for a level 13 mixer.
Let's see if we can confirm the calculations...
reduced our RG58C/U cable length to optimum value (134.2ft) and characterized it. Below the confirmation that it is what it should be.
loss is 8.6dB, delay 206.4ns
We still working on the breadboard setup. There will be several things we have to modify for the setup.
Now we can mount the mirrors on the three screw blocks, with beam on the hole pattern.
[add details about what to modify]
Note: I have been looking at frequency only around 100 - 3 kHz. Here is the beat for broadband with other noise. The seismic noise in the plot is for floated table, I'll edit that, but it is pretty similar for what we had before. So the breadboard setup has not introduced any noise at low frequency that shows up in the beat yet.
==Some intros about this measurement:==
We are using a mixer as a phase detector. Usually, mixers are not optimized to be used as phase detectors (that provide DC output), they do have 50 ohms impedance output. However, when we use a mixer as a phase detector, the output is DC, and the output behaves like a current source. We want to know how load impedance will affect the readout system, so we measured the output and varied the load impedance.
==Setup and Result==
The setup is shown here. We checked how much we can gain in the cable-delay readout scheme by using a different (higher) load impedance. Used the standard setup with a Marconi as the source, modulated the frequency with a triangular signal and looked after the LP filter with the scope for different termination impedances after the mixer. Plot is in arbitrary units as only the relative change is relevant. Signals are triggered at slightly different times.
The result shows that with higher load impedance, we have better sensitivity (steeper slope). This is as expected from a current source with load impedance ( V = IR). At this point we are using 500 Ohm load impedance in our regular setup because we are not sure if the higher load impedance will introduce any extra noise.
Started characterizing the cable-delay setup with the right length of cable (134ft of RG58 for 160MHZ). After checking the change in sensitivity with load impedance i've changed the load to 500 Ohms (instead of the usual 50 Ohms). I think an additional low-impedance path for the 2f has to be put in parallel later (to have proper 50Ohms @ 2f) to not get it reflected at the input of the low-pass filter back into the IF port of the mixer. (see first schematic).
However, the following simple setup has been used for the measurements:
I've measured the output signal vs different LO power levels while keeping the RF signal strength constant (8.29dBm) to find out the optimum signal strength in terms of size (not noise at this point!).
The following plots show the result:
tried to measure the frequency noise of the Marconi using the delay line. Setup is identical to the schematic posted in entry #832.
I've set the LO power to 13.64dBm as it is close to optumum value. The mixer output is terminated with 500Ohms. The slope is 1.1145MHz/V.
Measured the noise at the output using the SR785 and a SR560, gain 1000.
Plot shows the following:
once again measured the Marconi noise with the delay line - this time without the amplifier (so using a 7dBm mixer instead of 13dBm) and at two different frequencies, 20MHz and 160MHz. Still have no clue where the flat noise floor is coming from which we've seen in previous measurements (see elog #833).
The measurement at 20MHz (left graph) was taken with the frequency tuned so that the DC offset is close to zero (0.1mV). The measurements show a consistent 1/sqrt(f) noise level at low frequencies, independent from the marconi phase noise. And again, the Marconi noise for 1k input range can't be measured.
The right graph shows the result at 160MHz, but this time with a slight DC offset, so that LO AM couples into the measurements. The slope is similar to the one seen at 20MHz, but with clearly more features which come from amplitude noise of the Marconi. The situation at high frequencies is the same, the phase noise of 10k range can be seen, for lower input ranges not.
Next i measured the coupling from AM into the signal for different DC offsets (only at 160MHz), this time again for the original mixer and power levels (the one we want to use in our setup). As before at zero DC offset we are insensitive to AM and the noise floor is somewhat flat. With increasing offset the coupling from AM into the measurement becomes more dominant and looks identical to the coupling which can be seen on the upper right plot. This looks similar at other frequencies but i didn't save those.
The question now is: If the (almost) white noise floor is not thermal noise, amplifier noise etc. and not from AM, where does it come from? Any ideas?
I've tried the following thing, but nothing worked:
Beat signal is back now, but we have not measured the spectrum yet since the temperature is still drifting fast (10kHz/ min). We plan to measure it tomorrow afternoon.
The optics for RCAV are aligned. The visibility is ~95% without adjusting the lenses for mode-matching. I do not redo the mode match yet because we probably have to move the chamber for installing the air springs very soon. For ACAV, the beam is aligned, with visibility only ~ 80%. I also adjusted the cable length for PDH locking so that the error signal looks symmetric (Frank added the 4-ch splitter for demodulating RFAM, so the phase shift changed a bit. All reflected beams are properly blocked with razor blade blocks.
The beat frequency is ~ 188MHz instead of the nominal value of 160 MHz because the temperature is not settled at the set point yet. It takes longer than before because of the copper shields around the cavities. We expect it to be stable by tomorrow afternoon.
We expect it to be stable by tomorrow afternoon.
What about my dance party?!