Reducing the voltage from a servo

[kelly0303]

Hello! I am trying to control a laser frequency (locking it to a cavity) using a servo (implementing a PID loop). The servo can output voltages between -10 V and 10 V while the laser can only take between -4 V and 4 V. What is the best way to make sure that the voltage to the laser stays below 4 V? The simplest would be to just use 2 resistances such that I can divide the output from the servo 40% and 60%, that way I am sure the 40% branch can only do between -4 and 4 V, but I wanted to double check if this is enough. Thank you!

 

[DaveE]

Sounds good to me. There are other choices, but this is definitely the simplest.

 

[berkeman]

The servo can output voltages between -10 V and 10 V while the laser can only take between -4 V and 4 V.

What is the datasheet output impedance of the servo? What is the datasheet input impedance of the laser?

 

[TonyStewart]

Separate +/-4V regulators would be best with an adjustable current limiter to regulate intensity. The load is non-linear so R dividers will not work.

 

[DaveE]

The load is non-linear

How do you know this? He's told us nothing about the laser inputs, except the ±4V limitation.
Hence this:

What is the datasheet output impedance of the servo? What is the datasheet input impedance of the laser?

Anyway, it doesn't have to be linear, just monotonic. The question isn't about the transfer function it's about absolute maximum limits.

Separate +/-4V regulators would be best with an adjustable current limiter to regulate intensity

PS circuits (regulators) and current limiters would be unusual in a signal path. Or maybe you meant zener diodes?

Clamping circuits (which I think you are getting at) will work, but a divider makes a better mapping of ±10V to ±4V. Then a saturated controller doesn't have a big dead band to slew through in recovery (like locking) to have an effect on the frequency. If you are going to have a range limit or clamp, it should be at the integrator to avoid wind-up issues. Everything downstream should be matched to that range with appropriate scaling.

 

[TonyStewart]

How do you know this? He's told us nothing about the laser inputs, except the ±4V limitation.
All diodes LEDs and LASER included are non-linear loads. This is common knowledge to Electronics experts.Anyway, it doesn't have to be linear, just monotonic. The question isn't about the transfer function it's about absolute maximum limits.
For a voltage divider to not exceed 1% load regulation error, the source R must be < 1% of the load which is not-linear escpecialy during "lasing".

PS circuits (regulators) and current limiters would be unusual in a signal path. Or maybe you meant zener diodes?
NO The servo error amplifiers would run off the constant voltage regulators to prevent overvoltage.
Clamping circuits (which I think you are getting at) will work, but a divider makes a better mapping of ±10V to ±4V. Then a saturated controller doesn't have a big dead band to slew through in recovery (like locking) to have an effect on the frequency. If you are going to have a range limit or clamp, it should be at the integrator to avoid wind-up issues. Everything downstream should be matched to that range with appropriate scaling.

You can have limiters before the final stage if you know the impedance drop and know how to make a servo feedback amplifier also have gain but limited or clamped output and must also be current limited. There are many solutions, and this problem description is insufficient for design specs. and deserves more thought in the question in terms of measurable design specs.

Here is one example of a servo 4V Laser driver amplifier 45A .
Examine the quality and detail of the specs. The internal design would have even more specs.
https://www.analogtechnologies.com/document/AAS45A4V2.pdf

 

[DaveE]

You can have limiters before the final stage

Of course you can. But that may not be the best design. It depends on what you need and how you do it. Neither of which are known in this case.

Here is one example of a servo 4V Laser driver amplifier 45A .

Which would have little to do with frequency locking to a cavity as the OP asked about. It may be better to think of this question as a PLL, not a power supply.

Not all Lasers are diode lasers, laser tuning has many different mechanisms, for diodes it's typically temperature based in the cooling system (TECs and such). But these are nearly always separate from the main energy source, i.e. the power control. Unless clearly specified otherwise, I would assume that the user would want (essentially) constant power while frequency tuning.

So...

Examine the quality and detail of the specs. The internal design would have even more specs.

Yes, if you're referring to the OPs system requirements. That's my overall point here; we simply don't have nearly enough information to make any intelligent comments yet. Much of this thread is pointless until we hear more.

OTOH, if you mean that there are a lot of specs and design details for the laser power supply, then yes, I agree. I've designed many for a variety of laser technologies in my previous work. But, again, they are only peripherally related to laser frequency control.

 

[TonyStewart]

@kelly0303 can you do this forum a favour and include all links to the Laser, Cavity and interface requirements you have?

Is this relevant? https://www.wikiwand.com/en/Pound–Drever–Hall_technique

 

[DaveE]

@kelly0303 can you do this forum a favour and include all links to the Laser, Cavity and interface requirements you have?

That's kind of a lot of effort to ask for. Entirely appropriate, probably necessary, for the interface, if he still has questions. The laser and cavit(ies) aren't needed at this level. We weren't asked to design a servo system, we were asked about a simple interface issue.

Still, I'd bet he's solved this and moved on to other issues. After all, he started by suggesting a simple, likely appropriate, solution. I doubt that his job description entails entertaining kibitzing engineers on the web. We'll see; if the OP is happy then so am I.

 

[TonyStewart]

The question seems to be overlooking some major issues and needs to be upgraded with details and contents or closed as unanswerable without necessary specs for this LASER before further comments.

 

[Baluncore]

The question seems to be overlooking some major issues and needs to be upgraded with details and contents or closed as unanswerable without necessary specs for this LASER before further comments.

@TonyStewart
If you don't like other member's responses, you should just walk away and let it be.
Censorship of other's interpretations is not your job, nor is it good for PF.

 

[TonyStewart]

@TonyStewart
If you don't like other member's responses, you should just walk away and let it be.
Censorship of other's interpretations is not your job, nor is it good for PF.

Elevating the quality of content ought to concern everyone. That is what I politely requested before the lack of response.