Independent industrial and scientific laser service · New Jersey Mon–Fri 8:00–18:00 ET service@phasexlaser.com
Technology

SHG, THG and harmonic generation diagnostics

Wavelength-conversion stages are blamed for more faults than they cause. This page sets out how a conversion stage is actually tested, what the temperature behavior really tells you, and when the crystal genuinely is the problem.

First questionWhat fundamental reaches the stage
Second questionIs the phase-matching condition held
Last questionIs the crystal itself damaged
Start here

One measurement decides whether this page applies to you

Second-harmonic generation converts part of a fundamental beam into light at half the wavelength. It cannot produce more converted power than the fundamental delivered into it, and in a real system it produces a fraction of that fundamental, set by intensity, crystal length, phase matching and beam quality.

That single constraint settles most conversion-stage questions before any hardware is opened:

What the evaluation establishes first

Our evaluation establishes the fundamental power actually arriving at the input of the conversion stage, measured on a defined path with the fundamental separated from the converted output.

That result decides where the work belongs. A healthy fundamental with disproportionately low converted output is a conversion-stage question, and the rest of this page applies. An absent or strongly suppressed fundamental moves the work upstream, into the seed, the drivers, the amplifier chain or the fiber path.

The arithmetic that usually resolves it

Where a converted reading and a fundamental reading are of the same order, the two numbers cannot both describe a working conversion process fed by that fundamental. There are three ways that resolves, and our evaluation tests each rather than assuming one:

  1. Both readings are at the detector noise floor. The honest interpretation is that there is effectively no output at either wavelength, and the conversion stage is not the subject at all.
  2. The measurement setup is producing invalid numbers. A fiber-coupled meter on a free-space beam collects an unknown fraction. A sensor calibrated at one wavelength misreports another. Repeat with calibrated free-space detection and wavelength-selective optics before drawing any conclusion.
  3. The readings describe different optical paths. Residual pump leakage, scattered light or a second port can produce a number with no relationship to the fundamental reaching the crystal.

What the oven temperature actually tells you

Crystal ovens are one of the most consistently misread signals in scientific laser diagnostics, in both directions.

Flat temperature is not a fault

A temperature-stabilized oven exists to hold its setpoint. Holding it is success, not failure. When significant optical power passes through the crystal, the absorbed fraction creates a heat load that appears as a disturbance the controller has to work against, and you can sometimes see that disturbance when the drive current changes. When little or no fundamental arrives, there is no meaningful load, and the temperature sits flat.

So a crystal temperature that does not move when the current command changes is consistent with an absent fundamental. On its own it is weak evidence of a conversion-stage fault, and treating it as the primary symptom sends the investigation to the wrong end of the machine.

What a genuine thermal fault looks like

  • Temperature that will not reach setpoint, or takes far longer than normal to get there
  • Temperature that oscillates or hunts around setpoint
  • Temperature that drifts with ambient conditions rather than holding
  • A setpoint that can no longer be commanded, or a readback that is physically implausible
  • Conversion efficiency that changes with a small temperature offset in a way it did not previously, which points at the phase-matching condition rather than the controller

Those are TEC, driver, sensor, wiring and control-loop questions, and they are repairable at component level in most systems, including obsolete ones.

Real conversion-stage faults

When the fundamental has been confirmed healthy and the converted output is still low, these are the causes worth working through, roughly in the order they are worth testing.

01

Phase-matching condition lost

Crystal temperature offset from its correct operating point, or an angle that has shifted. Frequently recoverable without touching hardware, once the correct condition has been found experimentally.

02

Polarization state wrong at the crystal

Conversion is polarization dependent. A rotated or degraded polarization state upstream, including stress in a fiber path or a failed waveplate, reduces conversion without damaging anything.

03

Surface contamination or coating degradation

Deposits on crystal faces or on the input and output optics of the stage, and degraded anti-reflection coatings. Common in systems that have run in unfiltered laboratory air for years.

04

Crystal damage

Bulk or surface damage from a previous overdrive event, or from operation with a degraded beam. Genuine, but far less common than it is assumed to be.

05

Focus and beam condition at the crystal

Conversion depends on intensity. A degraded spatial mode or a shifted focus reduces efficiency while the average fundamental power looks unchanged.

06

Separation optics after the stage

Dichroics, filters and beam dumps that separate converted output from residual fundamental. A degraded separator can remove the converted light you are trying to measure.

07

Oven, TEC, sensor and control electronics

Thermal control hardware faults as described above, including sensor drift that leaves the crystal at the wrong actual temperature while reporting the right one.

08

Mechanical shift in the mount

A crystal mount disturbed by shipping, thermal cycling or previous service work, changing the angle or the beam position on the crystal.

How we test the stage

  1. As-received condition report. Configuration, labels, connections and shipping condition recorded and photographed before anything is changed.
  2. Configuration verification. Every module label read separately. What the system is takes precedence over what the record says it is.
  3. Bring-up and infrastructure. Supplies, cooling, interlocks, protection states and control communication established, and readback values checked against physical measurement.
  4. Measurement and fault localization. Output characterized by wavelength on a defined path, the optical pulse train verified, and the chain divided until the fault sits in one subsystem.
  5. Written findings and your approval. What was measured, what was found, what remains unknown, and the repair options with their cost and risks. Nothing is repaired, substituted or modified until you approve the scope in writing.
  6. Repair, validation and return. Work inside the approved scope, then electrical, optical, timing and functional validation, a written service report and documented packing.

Spectral measurement is used throughout to confirm that the light being measured is the light you think it is. On a stage producing weak output with residual fundamental in the same path, that confirmation is not optional.

OPO and OPA subsystems

Optical parametric oscillators and amplifiers share the same first principle: they cannot produce more than the pump delivers, and their output depends on a phase-matching condition that has to be held. The same diagnostic order applies, with additional questions specific to these systems:

  • Pump condition at the module input, measured rather than assumed
  • Signal and idler behavior across the tuning range, and whether the range has narrowed
  • Tuning mechanism condition, including motorized stages, encoders and their control
  • Cavity condition in an OPO, including alignment and optic degradation
  • Timing and synchronization between pump and seed in synchronously pumped arrangements

We evaluate OPO and OPA subsystems as part of a system evaluation. On systems where factory alignment data is unavailable, we state in the proposal what can and cannot be re-established, and we do not promise restoration of a tuning range we cannot verify against original acceptance data.

What can be repaired

Routinely repairable

  • TEC modules, drivers, temperature sensors and thermal-control wiring
  • Control electronics and the interfaces that command the oven
  • Cleaning and, where appropriate, replacement of degraded input, output and separation optics
  • Re-establishing a lost phase-matching condition experimentally
  • Polarization-control hardware upstream of the crystal

Case by case

  • Crystal replacement, where the cut and coating are available from a crystal supplier, the mount can be reassembled to the required tolerance, and the phase-matching condition can be re-established without factory alignment data
  • Custom mount, oven or driver engineering where the original assembly no longer exists

Where we stop

  • Sealed factory assemblies that cannot be opened without destroying the alignment they contain, where no realignment path exists
  • Restoration to original published conversion efficiency where the original acceptance data is unavailable. We document the measured condition before and after instead.

Scope and safety

Class 4 work is planned and executed under a project-specific laser safety plan. The required engineering controls, personal protective equipment, beam management and facility conditions are defined in writing before any energized optical work begins, and the infrastructure that plan calls for is provided within the project scope. Where a laboratory's own controls are involved, final hazard analysis and approval stay with the institution's LSO or EHS function.

Independent service

PhaseX Laser Services is an independent third-party service provider and is not affiliated with, authorized by, or endorsed by any manufacturer unless expressly stated in writing. Brand names and model names are used only to identify equipment for which customers may request evaluation. We do not guarantee a repair before evaluation, and we do not certify a system to original published specifications without the original acceptance data and the corresponding standards.

Related pages

Next step

Send both numbers: converted output and fundamental.

Tell us what you measured at each wavelength, with what instrument and how the beam reached it. On conversion-stage enquiries that information tells us whether the question is here or upstream, and it costs nothing to send.

Request service
Contact

Talk to an engineer, not a call center

Phone
(551) 379-6083
Email
service@phasexlaser.com
Bench
239 New Road, Suite B210Parsippany, NJ 07054
Hours
Monday–Friday8:00–18:00 ET
Request service Call Email