Precision constant optical power laser driver

[Cloud]

I need to have a laser at the exactly the same (max +-1%) optical power over the span of 2-3 hours. Currently I am using a simple driver which uses output from PD in laser as feedback. But I am still having problems because optical power drifts for cca 5-8% in hour and it settles in about hour and a half.
What do you recommend to improve? I was having idea for keeping laser at the constant temperature...

[CM800]

In order to have stability, you will absolutely need a stable temperature.

you should use a TEC for this with a suitable controller. If you try this, you should get a lot closer to your needs.

[ajb]

When you say that optical power drifts, do you mean that the sensitivity of your PD is changing, causing the output to drift even as the PD feedback is held constant by the driver?  Or is some other part of the driver drifting so that the PD feedback is NOT held constant as it should be?  What sort of power level?  Is the PD integral to the diode package?  What are you using to measure the actual power to determine that the output is in fact drifting?

[LaserSteve]

Have you tried a beam splitter and external PD?   

Steve

[David Hess]

Maybe the simple driver is not using the feedback photodiode correctly or well.

[danadak]

Something like a PID controller. The below happens to be a
fan controller but you could mod it to handle power in diode
and manage the T issue. PSOC 5LP might be best way to
approach, it has a 20 bit delsig, 12 bit SAR, can handle diff
signals, at very low level (read current measurement off a
shunt).

http://www.cypress.com/documentation/application-notes/an89346-psoc-4-intelligent-fan-controller


Board to use ($10) -





http://www.cypress.com/documentation/development-kitsboards/cy8ckit-059-psoc-5lp-prototyping-kit-onboard-programmer-and


Regards, Dana.

[grouchobyte]

In order to have stability, you will absolutely need a stable temperature.

you should use a TEC for this with a suitable controller. If you try this, you should get a lot closer to your needs.

Using a TEC to control or keep temperature stable will add unnecessary complexity, depending on the laser package, mechanical and thermal constraints, etc.

You have not disclosed any details about the Photodiode that monitors laser optical power. If it is integral to the laser package then it’s likely your APC control loop does not track properly. External PD monitoring is problematic and can contribute to inadequate power tracking.

Ideally the optical power will be captured by the PD and accurately fedback to the error circuitry in a linear fashion so proper tracking can be achieved. In the real world things sometimes don’t work as they should because there are design considerations that are easily overlooked or assumed, like self heating, thermal package drift, CTE errors, etc

If you want some additional help, please provide specifics about your design,.....ie: APC schematic, PD config, physical setup, etc

We can probably help you better after you consider PD loop error sources

@grouchobyte