Question about a Servo setup (locking a laser to a cavity)

  • I
  • Thread starter Malamala
  • Start date
  • #1
184
8
Hello! I am trying to understand the properties of a laser I am trying to lock to a cavity. I would like to adjust its wavelength in order to achieve the locking. I attached below a part of the properties that I am confused about (I emailed the company, too, but I haven't heard back yet). I understand that the locking (in my case PDH locking) works by modulating the laser frequency and using the sidebands to calculate the derivative of the signal and you lock it to the zero of the derivative (I apologize in advance if the naming in this table is not general, in which case I will just wait for the company to reply).

Is "Fast Frequency modulation bandwidth" the frequency at which the laser can be modulated, which means that the servo must produce a dither signal below 100 kHz? I am a bit confused as I found some stuff online saying that "bandwidth" refers actually to how often the feedback is sent to the laser (i.e. how often the laser frequency is corrected to match the cavity), not to the actual modulation frequency.

I am not sure what "Frequency fast tuning efficiency" is, so any advice/reading would be appreciated.

I assume that "Tuning voltage magnitude" is related to "Frequency fast tuning efficiency", but I don't know how.

Any insight into understanding this table would be really appreciated. Thank you!

Screen Shot 2021-07-22 at 7.41.05 PM.png
 

Answers and Replies

  • #2
Greetings,

What you are attempting to do is far from trivial and the potential risks from making a mistake could be catastrophic. Thus I suggest two things:

1. Look carefully at those footnotes in the table you posted.

2. More importantly, be patient and await a response from the manufacturer. Or telephone them directly being well prepared with your questions.

There is too much at stake for you to rely upon answers you might get here, no matter how well-meaning or well-informed those answers might appear.

Best regards,
ES
 
  • Informative
  • Like
Likes Twigg and berkeman
  • #3
Twigg
Science Advisor
Gold Member
587
291
Yeah, its not obvious what those parameters mean, so I'd get a manufacturer engineer on the phone. For example, is tuning voltage supposed to be the voltage on a piezo input? Or is it a current modulation? Unclear. If it was me I might just play around and figure it out, but I don't want to spread the normalization of deviance. Know thyself! :)

The bandwidth you found mentioned online is the bandwidth of the loop filter used to feed back on the PDH error signal. Essentially this is the frequency range over which the servo loop is effective at suppressing fluctuations, assuming the laser's response function has only "positional" signature and no "integral" signature. This has nothing to do with the laser, maybe at second order the laser affects this.
 
  • #4
184
8
Yeah, its not obvious what those parameters mean, so I'd get a manufacturer engineer on the phone. For example, is tuning voltage supposed to be the voltage on a piezo input? Or is it a current modulation? Unclear. If it was me I might just play around and figure it out, but I don't want to spread the normalization of deviance. Know thyself! :)

The bandwidth you found mentioned online is the bandwidth of the loop filter used to feed back on the PDH error signal. Essentially this is the frequency range over which the servo loop is effective at suppressing fluctuations, assuming the laser's response function has only "positional" signature and no "integral" signature. This has nothing to do with the laser, maybe at second order the laser affects this.
Thanks for this! So basically the servo has a frequency at which it modulates the signal for the PDH lock and a frequency bandwidth which shows the frequency range over which the feedback loop is effective. And both of these frequencies have nothing to do with the laser frequency, or the allowed modulation frequency of the laser. Is this right?
 
  • #5
berkeman
Mentor
60,390
10,698
Hello! I am trying to understand the properties of a laser I am trying to lock to a cavity.
Can you post a link to the datasheet where you got the table fragment in your first post? Thanks.
 
  • #6
Twigg
Science Advisor
Gold Member
587
291
Usually the oscillator in a PDH setup is set to a fixed value and then left alone. The exact value isn't really critical. It's a "set and forget" kind of parameter, so quoting a bsndwidth on it is a little extra.

If I had to guess based on limited info, I'd guess you have yourself a diode laser and the "fast frequency modulation bandwidth" refers to the fastest frequency that the piezo-transducer in the grating can react to. That's a guess though, need more info to give you solid answers. The "sensitivity" appears to be how many MHz the cavity shifts per volt applied to the piezo. The "magnitude" would be the acceptable input range of voltage before you let the magic smoke out. I dunno, just guesses.

Got a copy of the laser manual?

If the above hypothesis sounded like Greek, I would encourage you to read about the Littrow configuration for diode lasers.
 
  • Informative
  • Like
Likes Klystron, DaveE and berkeman
  • #7
184
8
Usually the oscillator in a PDH setup is set to a fixed value and then left alone. The exact value isn't really critical. It's a "set and forget" kind of parameter, so quoting a bsndwidth on it is a little extra.

If I had to guess based on limited info, I'd guess you have yourself a diode laser and the "fast frequency modulation bandwidth" refers to the fastest frequency that the piezo-transducer in the grating can react to. That's a guess though, need more info to give you solid answers. The "sensitivity" appears to be how many MHz the cavity shifts per volt applied to the piezo. The "magnitude" would be the acceptable input range of voltage before you let the magic smoke out. I dunno, just guesses.

Got a copy of the laser manual?

If the above hypothesis sounded like Greek, I would encourage you to read about the Littrow configuration for diode lasers.
@Twigg @berkeman here is a link to the laser I am talking about. I can't seem to find the actual manual online. If there is a way to share files here I can share the PDF of the manual (it is just a few pages).
 
  • Like
Likes Twigg and berkeman
  • #8
184
8
Usually the oscillator in a PDH setup is set to a fixed value and then left alone. The exact value isn't really critical. It's a "set and forget" kind of parameter, so quoting a bsndwidth on it is a little extra.

If I had to guess based on limited info, I'd guess you have yourself a diode laser and the "fast frequency modulation bandwidth" refers to the fastest frequency that the piezo-transducer in the grating can react to. That's a guess though, need more info to give you solid answers. The "sensitivity" appears to be how many MHz the cavity shifts per volt applied to the piezo. The "magnitude" would be the acceptable input range of voltage before you let the magic smoke out. I dunno, just guesses.

Got a copy of the laser manual?

If the above hypothesis sounded like Greek, I would encourage you to read about the Littrow configuration for diode lasers.
@Twigg thank you for this! I do understand what you mean and this is really helpful (still waiting for them to reply, tho). Just to make sure, by "the fastest frequency that the piezo-transducer in the grating can react to" you mean how fast the piezo can move. I assume this is related to how fast the feedback from the servo can be sent back to the laser i.e. for 100kHz, the correction signal would be sent once every 10 micro seconds. Is this right (or at least this is how fast the laser can respond to the servo signal)? However, is this related to the modulation frequency i.e. can the laser be modulated at a higher frequency than this or does this number reflects the limit on the modulation frequency, too? As far as I understand, the modulation (at least in a EOM) works by adjusting the index of refraction of a medium by applying some voltage to that medium, so it shouldn't be related to changing the size of the cavity. But I am not sure I got that right.
 
  • #9
Twigg
Science Advisor
Gold Member
587
291
I have to be honest, when I saw that this was a 20mW 1064nm laser for a "request quote" level of pricing, I almost choked. Then I saw that beautiful 5kHz linewidth :oldbiggrin:

Since it's a 1064, I have no idea if it's a diode or if it's a YAG crystal. In either case, there will be a cavity around the gain medium (essential ingredient of a laser). That cavity defines both the center wavelength and the linewidth of the laser (the cavity is more restrictive than the lifetime of the lasing transition). So, by changing the length of the cavity (as with a piezo-transducer), you change the frequency of the laser. It has the same effect as an EOM, but it changes the distance not the index of refraction (both result in phase shifts / modulation). The PZT can continuously change the frequency of the laser until you see a "mode-hop", at which point the frequency jumps to a more convenient longitudinal mode (convenient for the laser, not for you!! many a four-letter-word have I uttered at a mode hop).
 
  • #10
Twigg
Science Advisor
Gold Member
587
291
Ah, sorry I forgot the question about bandwidth. No, the servo bandwidth and the modulation bandwidth are not the same. The piezo can't move in response to any frequency. There is a limiting rise time for that motion. My guess is that there's an amplifier between the modulation port and the piezo itself (as I seem to recall that piezo's require somewhat high voltage). That high gain amplifier is probably what limits the bandwidth of the modulation. If not the amplifier, than it's the piezo itself not being able to expand/contract infinitely quickly.
 
  • #11
184
8
I have to be honest, when I saw that this was a 20mW 1064nm laser for a "request quote" level of pricing, I almost choked. Then I saw that beautiful 5kHz linewidth :oldbiggrin:

Since it's a 1064, I have no idea if it's a diode or if it's a YAG crystal. In either case, there will be a cavity around the gain medium (essential ingredient of a laser). That cavity defines both the center wavelength and the linewidth of the laser (the cavity is more restrictive than the lifetime of the lasing transition). So, by changing the length of the cavity (as with a piezo-transducer), you change the frequency of the laser. It has the same effect as an EOM, but it changes the distance not the index of refraction (both result in phase shifts / modulation). The PZT can continuously change the frequency of the laser until you see a "mode-hop", at which point the frequency jumps to a more convenient longitudinal mode (convenient for the laser, not for you!! many a four-letter-word have I uttered at a mode hop).
Thanks a lot for this! So I have a question about this signal modulation (it is my first time working with lasers in practice, I am sorry if my questions are very silly!). I understand that before locking the laser I do a swipe of a given frequency region, such that I can see the peaks corresponding to when the laser is mode-matched on the cavity. For example if my cavity FSR is 100 MHz, and taking the value of 100 MHz/V from the table above, if I were to apply a signal of, say 3 V amplitude, with a frequency of 1 Hz, I would scan a region of 300 MHz of my cavity every second, and I should see 3 peaks. At the same time, for a PDH lock, I need to modulate my laser frequency at a high frequency (usually of the order of MHz) in order to get the derivative of my signal to which I lock. Here is a figure with what I mean (that figure is for some absorption lines, not locking to a cavity, but they reflect what I mean).

So, it looks as if I need to modulate my laser both at 1 Hz or a small frequency used to swipe the region of interest (orange line in that figure), but also at a large frequency (at least several times bigger than the cavity linewidth) in order to obtain the derivative used for the error signal (cyan curve). How can I modulate the same laser at 2 different frequencies? How is this done in practice? Thank you!
 
  • #12
Twigg
Science Advisor
Gold Member
587
291
Here is a figure with what I mean
Ah, seeing a sat abs curve again brings back memories :oldbiggrin:

So, it looks as if I need to modulate my laser both at 1 Hz or a small frequency used to swipe the region of interest (orange line in that figure), but also at a large frequency (at least several times bigger than the cavity linewidth) in order to obtain the derivative used for the error signal (cyan curve)
You do this with a summing circuit. On on input of the summing circuit, you add the "modulation" signal that's used to generate the derivative of the cavity reflectance, and on the other input you add a "scan" signal. When you're looking for the line you want to lock to, you unlock the servo loop and you turn on the scan waveform. When you've found where you want to lock, you turn off the "scan" and you crank up the feedback gain on the servo. (Don't forget to turn off the feedback before you scan, or you're going to burn hours fiddling with an optical isolator and troubleshooting a weird jumpy spectrum. Ask me how I know :headbang:)

Sometimes, if you have a servo box that was expressly built for this purpose, the servo will contain the adder circuit and/or the scan waveform generator. Look for anything that says "sweep" or "scan" or "ramp" on the front panel.

Also, you don't want the modulation depth to go past a cavity linewidth. You want the modulation to stay confined to the region where the PDH signal is linear with frequency.
 
  • #13
184
8
Ah, seeing a sat abs curve again brings back memories :oldbiggrin:


You do this with a summing circuit. On on input of the summing circuit, you add the "modulation" signal that's used to generate the derivative of the cavity reflectance, and on the other input you add a "scan" signal. When you're looking for the line you want to lock to, you unlock the servo loop and you turn on the scan waveform. When you've found where you want to lock, you turn off the "scan" and you crank up the feedback gain on the servo. (Don't forget to turn off the feedback before you scan, or you're going to burn hours fiddling with an optical isolator and troubleshooting a weird jumpy spectrum. Ask me how I know :headbang:)

Sometimes, if you have a servo box that was expressly built for this purpose, the servo will contain the adder circuit and/or the scan waveform generator. Look for anything that says "sweep" or "scan" or "ramp" on the front panel.

Also, you don't want the modulation depth to go past a cavity linewidth. You want the modulation to stay confined to the region where the PDH signal is linear with frequency.
Thanks a lot! You explain this so much better than what I find online!

So, basically the summing circuit turns the normal oscillation of the laser (say electric field), ##E_0e^{i\omega_o t}## to something like this ##E_0e^{i(\omega_o t + \beta_1 sin{(\omega_1)t} + \beta_2 sin{(\omega_2)t}\)}##, where ##\omega_1## would be the scan frequency (say 1 Hz) and ##\omega_2## the modulation frequency (say 1 MHz). Is this right?

And when I switch to the locking phase (i.e. I stop scanning) would the path used before to send the 1Hz scanning signal be now used to send the correction signal to the piezo, or is there a 3rd path doing that?

The device that I am using is actually this. It has an output ("RF output") that generates a 4 MHz signal for modulation and a different one ("SERVO OUTPUT") used in the ramp mode (and it seems like the same output is used to send the correction signal during the lock phase, too). However, given that there are 2 different ports, it means that I have to build the summing circuit myself or check if the laser does that on its own (the laser seems to have only one BNC port), right?
 
  • #14
Twigg
Science Advisor
Gold Member
587
291
You explain this so much better than what I find online!
It's the byproduct of making a lot of dumb mistakes o0) but thanks!

Weird side note: this manufacturer uses the terminology "dither signal". That's what I normally call the modulation signal, but I thought it was bad form to call it that for a cavity reference. Huh! Tl;dr: dither signal = the RF modulation signal you use to get the derivative of the cavity reflectance in PDH.

So, basically the summing circuit turns the normal oscillation of the laser (say electric field), E0eiωot to something like this E0ei(ωot+β1sin(ω1)t+β2sin(ω2)t\), where ω1 would be the scan frequency (say 1 Hz) and ω2 the modulation frequency (say 1 MHz). Is this right?
For clarity's sake, you use a sawtooth waveform instead of ##\beta_1 sin(\omega_1 t)##. That way the scan of the cavity frequency is nice and linear (whereas the sine wave levels off and gets curvy towards the tops and bottoms). But yes, you have the right idea! Note that ##\beta_1 \gg \beta_2##, so the fast modulation isn't visible on the same scale as the scan. If you DO see your fast modulation on the scan waveform (you'll see the scan will look fuzzy on the oscilloscope), try listening for a high pitch whine coming from your laser. That's your piezo begging for death because you're driving it way too hard.

And when I switch to the locking phase (i.e. I stop scanning) would the path used before to send the 1Hz scanning signal be now used to send the correction signal to the piezo, or is there a 3rd path doing that?
There's a lot of flexibility in how this part is handled. In most of the ones I've used, you turn off the scan waveform, either by a switch or by turning the amplitude down to 0. I checked the servo box you linked, but the friggin picture of the front panel is in potato vision. If you have the box in hand, let us have a look at the front panel and that'll help us help you.

I noticed they have two stages of integrating gain. Watch out for saturation ("clipping" in electrical engineer speak).

I also noticed the max ramp amplitude for that servo box is +/- 5V, whereas the laser's modulation limits are +/- 4V. Might be worthwhile making yourself a little comparator alarm circuit. Trust me, it sounds silly, until you break something. Then you're silly. Actually, some diode laser controllers come with alarm settings for that, but it's usually just for the diode current.

Some servo boxes have two modes of operation: one where they generate the scan/ramp for you and another where you give it a TTL signal to synchronize it with an external ramp you make from your own function generator. I don't think I've ever seen anyone NOT use their own function generator. I think it's a trust thing :oldbiggrin:
 
  • #15
Twigg
Science Advisor
Gold Member
587
291
It has an output ("RF output") that generates a 4 MHz signal for modulation and a different one ("SERVO OUTPUT") used in the ramp mode (and it seems like the same output is used to send the correction signal during the lock phase, too). However, given that there are 2 different ports, it means that I have to build the summing circuit myself or check if the laser does that on its own (the laser seems to have only one BNC port), right?
Whoops, sorry. I missed this.

I would assume the servo box does it for you, and will automatically switch over. The stuff I was talking about is when you DIY it.
 
  • #16
184
8
It's the byproduct of making a lot of dumb mistakes o0) but thanks!

Weird side note: this manufacturer uses the terminology "dither signal". That's what I normally call the modulation signal, but I thought it was bad form to call it that for a cavity reference. Huh! Tl;dr: dither signal = the RF modulation signal you use to get the derivative of the cavity reflectance in PDH.


For clarity's sake, you use a sawtooth waveform instead of ##\beta_1 sin(\omega_1 t)##. That way the scan of the cavity frequency is nice and linear (whereas the sine wave levels off and gets curvy towards the tops and bottoms). But yes, you have the right idea! Note that ##\beta_1 \gg \beta_2##, so the fast modulation isn't visible on the same scale as the scan. If you DO see your fast modulation on the scan waveform (you'll see the scan will look fuzzy on the oscilloscope), try listening for a high pitch whine coming from your laser. That's your piezo begging for death because you're driving it way too hard.


There's a lot of flexibility in how this part is handled. In most of the ones I've used, you turn off the scan waveform, either by a switch or by turning the amplitude down to 0. I checked the servo box you linked, but the friggin picture of the front panel is in potato vision. If you have the box in hand, let us have a look at the front panel and that'll help us help you.

I noticed they have two stages of integrating gain. Watch out for saturation ("clipping" in electrical engineer speak).

I also noticed the max ramp amplitude for that servo box is +/- 5V, whereas the laser's modulation limits are +/- 4V. Might be worthwhile making yourself a little comparator alarm circuit. Trust me, it sounds silly, until you break something. Then you're silly. Actually, some diode laser controllers come with alarm settings for that, but it's usually just for the diode current.

Some servo boxes have two modes of operation: one where they generate the scan/ramp for you and another where you give it a TTL signal to synchronize it with an external ramp you make from your own function generator. I don't think I've ever seen anyone NOT use their own function generator. I think it's a trust thing :oldbiggrin:
Thanks a lot for this! The laser company replied, and you were totally right in terms of parameters. They show how fast I can modulate the laser, how much the wavelength changes per voltage applied and the range of voltages I can apply.

I attached below a picture of the actual servo. I emailed the servo company these questions, but I thought it's worth asking you, too, as you are just as right and a lot faster (I will try the ideas with an oscilloscope first to make sure I don't break the laser). I would actually want to scan my laser frequency at a lower frequency that the one they provide in ramp mode (they use 500 Hz). You said I can use the Ramp TTL port. Does it mean that I can just create the signal I want with a signal generator, plug that signal in the Ramp TTL port and that will replace the 500 Hz default signal?

About the voltage of the servo, in the ramp mode that is not a problem. Keeping the amplitude below +/- 4V is enough to see the modes of my cavity. However I am a bit worried about the locking phase. The voltage produced by the servo in this phase is between +/- 12 V. In principle that voltage used should be very small (below 1V?) as that is just a fine tuning used to follow the cavity modes closely. But I am worried that if the lock is lost, the servo might send a voltage much bigger than +/- 4 V to relock it. Is there a way to avoid this? I was thinking to use a voltage divider, that would send to the laser only 1/3 of the voltage generated by the servo, but I am not sure how this would work during the fine tuning. I assume that also the fine tuning voltages will be divided by 3 and that might affect the lock?

Actually, another thing I am not sure about is how does the servo know how much voltage to send (I assume that different lasers have different values for MHz/V parameter). I understand that the PID is used for that, but from my basic understanding that should still require a lot of fine tuning in order to adjust the weights for the 3 path (actually 4 in my case as the loop is a PIID).

servo.jpg
 
  • #17
Twigg
Science Advisor
Gold Member
587
291
Because this laser has such a narrow linewidth, you might want to be careful adding protection circuitry to the modulation input. 5 kHz linewidth and 100MHz/V means that a mere 50 ##\mu##V will move the laser center wavelength by a whole linewidth.

The most common form of voltage protection is a pair of Zener diodes. I'm stuck on my phone until Monday or I'd draw you a circuit. Check out the Wikipedia page on Zener diodes (https://en.wikipedia.org/wiki/Zener_diode) and scroll down to the "waveform clipper" subsection in the "uses" section. That would do the trick, but it will add noise, which will effectively widen the laser linewidth. I don't recall how to calculate the noise in nA/sqrt(Hz), but you multiply that by the impedance of the modulation input (probably 50 ohms) to get the voltage noise (and by extension the freauency noise in Hz/sqrt(Hz)).

I only see a coarse and fine gain on the front panel. Are there more controls for the gain paths?

Edit: I wrote microhertz earlier when I meant microvolts. Fixed.
 
Last edited:
  • #18
184
8
Because this laser has such a narrow linewidth, you might want to be careful adding protection circuitry to the modulation input. 5 kHz linewidth and 100MHz/V means that a mere 50 ##\mu##V will move the laser center wavelength by a whole linewidth.

The most common form of voltage protection is a pair of Zener diodes. I'm stuck on my phone until Monday or I'd draw you a circuit. Check out the Wikipedia page on Zener diodes (https://en.wikipedia.org/wiki/Zener_diode) and scroll down to the "waveform clipper" subsection in the "uses" section. That would do the trick, but it will add noise, which will effectively widen the laser linewidth. I don't recall how to calculate the noise in nA/sqrt(Hz), but you multiply that by the impedance of the modulation input (probably 50 ohms) to get the voltage noise (and by extension the freauency noise in Hz/sqrt(Hz)).

I only see a coarse and fine gain on the front panel. Are there more controls for the gain paths?

Edit: I wrote microhertz earlier when I meant microvolts. Fixed.
Thanks for this! Do you mean that due to this high sensitivity to low voltages, the locking might not work that well? Shouldn't the servo account for that sensitivity? That was basically related to my previous questions: given that all lasers are different, shouldn't the servo somehow adapt to the laser sensitivity when correcting the frequency i.e. if I want to adjust the laser by 1KHz, and one laser need 100 mV while another one needs 100 ##\mu##V, shouldn't the servo account for that on its own, or do I need to build an external circuit to connect the laser to the servo for each laser I have?

There are no other controls for the gain paths.

I have one more question which might be silly but I couldn't find a clear answer online. I managed to reach a point where I see the modes of the cavity using a photodiode and an oscilloscope (while scanning the laser frequency). The photodiode I have (this one) has a A/W vs wavelength plot so I know the A/W for my wavelength. Also I see the amplitude of the cavity modes on my oscilloscope, which is about 500 mV. How can I use these 2 values A/W and V to calculate the actual power the diode receives? Is there a general formula, or does it depends on the type of oscilloscope? Thank you!
 
  • #19
DaveE
Science Advisor
Gold Member
1,737
1,376
The photodiode I have (this one) has a A/W vs wavelength plot so I know the A/W for my wavelength. Also I see the amplitude of the cavity modes on my oscilloscope, which is about 500 mV. How can I use these 2 values A/W and V to calculate the actual power the diode receives? Is there a general formula, or does it depends on the type of oscilloscope?
The current output from the photodiode is converted to voltage by the termination (or load) resistor you have supplied.
 
  • #20
184
8
The current output from the photodiode is converted to voltage by the termination (or load) resistor you have supplied.
Thank you for your reply. I am sorry I just got started and I am not sure I know the terminology. I see 2 resistor values listed by the oscilloscope: 50 ##\Omega## and 1 M##\Omega##. My oscilloscope is set to 1 M##\Omega## so does this mean that an amplitude of 1V is equivalent to ##1/10^6=10^{-6}##A?
 
  • #21
DaveE
Science Advisor
Gold Member
1,737
1,376
Thank you for your reply. I am sorry I just got started and I am not sure I know the terminology. I see 2 resistor values listed by the oscilloscope: 50 ##\Omega## and 1 M##\Omega##. My oscilloscope is set to 1 M##\Omega## so does this mean that an amplitude of 1V is equivalent to ##1/10^6=10^{-6}##A?
Yes, exactly. But those diodes are intended to operate into a lower resistance load, so I'm not sure you can trust the responsivity data they provide. Maybe, it will still work, but more in a photovoltaic mode than photoconductive mode. Linearity (over range of incident power) is better with lower load (termination) resistance. Try 50ohms, that's more what they were designed for, although for low power you might not get a big signal.

There are lots of good resources on the web to learn about photodiodes. Here's one:
http://www.osioptoelectronics.com/application-notes/an-photodiode-parameters-characteristics.pdf
 
  • #22
Twigg
Science Advisor
Gold Member
587
291
Do you mean that due to this high sensitivity to low voltages, the locking might not work that well? Shouldn't the servo account for that sensitivity?

I was responding to this question:
But I am worried that if the lock is lost, the servo might send a voltage much bigger than +/- 4 V to relock it. Is there a way to avoid this? I was thinking to use a voltage divider, that would send to the laser only 1/3 of the voltage generated by the servo, but I am not sure how this would work during the fine tuning. I assume that also the fine tuning voltages will be divided by 3 and that might affect the lock?
(Sorry, I should've quoted it in that reply.)

Using a 1/3rd voltage divider like this is a type of voltage protection circuit (a circuit that protects something sensitive, like the modulation input, from seeing too many volts). I was suggesting that you may want to be judicious in your choice of voltage protection so you don't introduce electrical noise to the modulation input which will then broaden your laser linewidth. For an example of what NOT to do, imagine you created a 1/3rd voltage divider with a 2Meg series resistor and a 1Meg shunt resistor. Thermal noise alone will produce ##70 \mu V## RMS noise over that 100kHz bandwidth. That's enough to broaden your laser by 50% of it's spec'd linewidth. (Low resistance is your friend for low noise applications!)

if I want to adjust the laser by 1KHz, and one laser need 100 mV while another one needs 100 μV, shouldn't the servo account for that on its own
Eh, it's more you than the servo. You setting the servo's setpoint to a known spectral feature is what determines the laser frequency at the end of the day. Servo's are great, but they aren't going to steal your job :oldtongue:

There are no other controls for the gain paths.
You probably were not meant to fiddle with the integrating gain in that case. The only other place I could think it would be would be on one of the side panels. Look for holes in the side panels, with trimmer potentiometer screws visible inside. Even if you do find it, I would assume that you don't need to mess with it unless you see weird slow drifts in your lock voltage. And even then, I would look for optical feedback before I started messing with the integrating gain.
 

Related Threads on Question about a Servo setup (locking a laser to a cavity)

  • Last Post
Replies
9
Views
296
  • Last Post
Replies
7
Views
17K
  • Last Post
Replies
11
Views
2K
  • Last Post
Replies
12
Views
162
  • Last Post
Replies
10
Views
4K
Replies
2
Views
2K
  • Last Post
Replies
1
Views
783
  • Last Post
Replies
3
Views
2K
  • Last Post
Replies
11
Views
4K
  • Last Post
Replies
1
Views
1K
Top