A standard application in laser diode test is the characterization of laser output across broad temperature ranges, usually from 0 °C to 85 °C. Rapid changes to and fast stabilization of laser diode case temperature suggests increased production throughput because of quicker laser characterizations.
What is required to quickly change a device’s temperature to either a cold or hot extreme and maintain it? In the case of temperature control with thermoelectric coolers, this normally requires a TEC capable of creating a large temperature gradient between its cold and hot sides besides pumping a large quantity of heat. TECs of this type require current on the order of 5-10 amps so as to pump tens of Watts of heat. A telecom diode laser for instance may only create 1-2 W of heat at most, so why is such a large heat pumping capacity needed? Heat pumping capacity must be larger than what the device under test generates because in order to reach the required extreme temperatures, the heat that must be pumped into or out of the device comes from several sources.
In general, the total heat flow necessary to maintain a specified temperature can be described by the equation:
||Qtotal = Qambient + QTEC + Qload
When cooling below ambient, heat must be removed from the laser so all terms of Qtotal are negative. When heating above ambient, heat must be pumped into the laser which results in Qambient and QTEC to be positive and Qload to be negative since the heat is already being produced inside the laser. The heat transfer Qambient occurs from heat being lost to the environment when Tload > ambient and heat being absorbed from the environment when cold. This transfer is through conductive, convective and radiated means. The conductive, convective and radiated heat can be quantified as shown in Equations 2, 3 and 4.
||Qrad = σe1 Α1 (Τ14 - Τ24)
A full description of these terms can be found in App Note #1 - Controlling Temperatures of Diode Lasers and Detectors Thermoelectrically on the ILX website and their definitions are not required for the present discussion.
The point of interest concerning the above equations is that the heat lost or absorbed is proportional to the temperature difference between the device being regulated and the environment. In several cases, the main heat transfer mechanisms will be conductive and convective suggesting that as the temperature becomes colder or hotter, a larger quantity of heat will have to be moved. This in turn requires a higher power TEC and a controller to operate it.
The purpose of this article is to give information for quick temperature cycling of a thermal load across wide temperature extremes using an ILX Lightwave LDT-5980 120 W temperature controller. The focus will be on regulating a 1.5 W thermal load at temperatures of 0 °C, 25 °C and 85 °C. Since enhancing the LDT-5980’s PID control loop can be a time-consuming job, even with the instrument’s Auto-Tune feature, examples of PID constants will be provided that allow the LDT-5980 to quickly change and stabilize the thermal load’s temperature.
A fixture was built to allow the case of a standard 14-pin butterfly laser mount to be temperature controlled. A Marlow Industries® model DT12-8-01 thermoelectric cooler was placed between an OFHC (Oxygen Free High Conductivity) coldplate and a finned heatsink (aluminum). The coldplate top was built as a pedestal to allow the butterfly package to be secured to it without requiring extra exposed surface area. The coldplate base was sized to allow full contact with the Peltier’s top surface, so that thermal transfer to the TEC could be increased.
A 10 kΩ thermistor was installed in the coldplate pedestal for feedback to the temperature controller. A 94 CFM fan was fitted approximately ½” away from the fins (because of space constraints within the mount) to offer airflow for convective cooling.
The thermal load made up of a 14-pin butterfly laser package with an internal TEC and thermistor. To mimic a 1.5 W heat load, the internal TEC was enabled with a constant current of 1.1 A. This had the effect of driving the internal temperature 40 to 60 degrees below the case temperature.
Note: Thermal joint compound is NOT used between the coldplate and the test load. This was done to mimic a more productionized environment where shortened test time is most vital. Refer to Figure 1 for details on the test fixture.
Figure 1. Test Fixture.
The test procedure used was designed to mimic the characterization of a laser diode as L-I-V tests were run at 0 °C, 25 °C and 85 °C. A LabVIEW™ program was written to exploit this performance, the current limits for heating and cooling were set to the maximum current suggested by the TEC manufacturer. For the Marlow DT12-8, this limit is 7.4 Amps. The PID and current limit values used to attain each setpoint are shown in Table 1.
So as to reduce oscillations, the PID values illustrated were enhanced from those acquired from the instrument’s autotune feature.
Table 1. LDT-5980 Control Parameters.
|Cooling Current Limit
|Heating Current Limit
ILX Application Note #14, references the fact that maximum current may not be the ideal choice when a TEC is being used to cool an object. This subject was examined and for the parameters of this test, was found to provide negligible benefit. The only modification noticed was an increase in the time required to cool down to the setpoint temperature. If extended periods of time are required with the temperature at 0 °C or colder, a current limit below device maximum is suggested to prevent the heatsink from saturating (thermal runaway).
The graph signifying the internal temperature in Figure 2 illustrates the temperature lag experienced because of thermal resistance between the internal components of the laser and the case. The thermal lag shows the point that if stabilization of the internal temperature is vital, a dwell time must be built into the test.
Figure 2. Case Temperature vs. Time.
As can be seen in Figure 3, the case temperature changes ~0.1 °C when the 1.5 W thermal load is applied. The changes, however, do not go beyond the ±0.2 °C temperature tolerance band for this test. The quantization of temperature observed in graphs with expanded scales is because of the approximate 2 Hz measurement update rate of the instrument.
Figure 3. Effect of Thermal Load on 25 °C Setpoint.
Figure 4 shows similar behavior taking place at 0 °C. There is an approximate 0.1 °C increase in the case temperature because of the enabling of the heat load but it is returned back to the setpoint in less than five seconds.
Figure 4. Effect of Thermal Load on 0 °C Setpoint.
Figure 5 illustrates how enabling the 1.5 W thermal load influences the case temperature at the 85 °C setpoint. Here, the external temperature actually surpasses the 0.2 °C tolerance band for several seconds.
Figure 5. Effect of Thermal Load on 85 °C setpoint.
Record the data and configure the temperature controller for each step displayed below:
- Enable temperature controller output with setpoint of 25 °C.
- Allow laser case temperature to stabilize to 25 °C ± 0.2 °C.
- Enable 1.1 A to internal TEC to generate a 1.5 W heat load.
- Wait 30 seconds to simulate the L-I-V data gathering.
- Disable 1.5 W heat load.
- Change temperature controller setpoint to 0 °C.
- Allow laser case temperature to stabilize to 0 °C ± 0.2 °C.
- Apply 1.5 W heat load.
- Wait 30 seconds.
- Disable 1.5 W heat load.
- Change temperature controller setpoint to 85 °C.
- Allow laser case temperature to stabilize to 85 °C ± 0.2 °C.
- Apply 1.5 W heat load.
- Wait 30 seconds.
- Disable 1.5 W heat load.
- Change temperature controller setpoint to 25 °C.
- Allow laser case temperature to stabilize to 25 °C ± 5 °C to simulate the cooling-down needed to allow an operator to safely remove the laser from the fixture.
Temperature Cycle Results
The test data in Figure 2 reveals the coldplate (case) temperature and the laser’s internal temperature as a function of time. The effect of disabling and enabling the 1.5 W heat load is not fully obvious in this data because of the resolution of the vertical scale. The sharp variations in the internal temperature are because of the internal TEC being enabled and disabled.
The data shows how swiftly the setpoint temperatures can be attained to allow data collection. A summary of the time to modify temperature is provided in Table 2. The total test length can be considered to happen in less than five minutes. It is crucial to note that production setups with undersized temperature controllers can take tens of minutes to attain the same temperatures.
Table 2. Time Required to Reach Temperature Setpoints.
||Approximate Time Required
|25 °C → 0 °C
|0 °C → 85 °C
|85 °C → 25 °C
This example reveals that to maintain temperature stability, the L-I-V test must either be accelerated to finish before the temperature surpasses the tolerance window or greatly slowed down to allow thermalization between each current ramp step.
As a final test, the laser diode was subjected to the same temperature profile as before but with time allowed for the internal temperature to steady to the setpoint temperature. This was performed to see how much extra time would be needed beyond case temperature stabilization before a temperature-critical test could start. Figures 6 - 9 show this procedure at each temperature setpoint. In all cases, an extra 30 to 60 seconds are needed to allow the internal temperature to stabilize.
Figure 6. Laser Internal Temperature Stabilization at 25 °C Setpoint.
Figure 7. Laser Internal Temperature Stabilization at 0 °C Setpoint.
Figure 8. Laser Internal Temperature Stabilization at 85 °C Setpoint.
Figure 9. Laser Internal Temperature Stabilization at 25 °C Setpoint.
In brief, the presented data reveals that quick temperature shifts of case temperature are possible using a high power temperature controller such as an ILX Lightwave LDT-5980 Temperature Controller. Through a sensible choice of PID coefficients mentioned in Table 1, altering temperature between the setpoints of 0 °C, 25 °C and 85 °C can quickly take place. In most instances, the actual time needed to go from one case temperature to stabilization within ±0.2 °C at another temperature can occur within 60-90 seconds. This situation shows that the L-I-V characterization of a laser diode can occur at three temperatures within five minutes.
It must be kept in mind, however, that the time needed for the internal temperature to stabilize will be considerably longer. This is due to thermal resistance between the case and the internal components of the laser package. If the internal temperature is to be stabilized as well, several extra minutes may be required to attain this. The total test time will also rely on the temperature tolerance needed.
Other ILX Lightwave Application Notes may be of interest when configuring a test station for fast temperature cycling. These notes include:
App Note #1 - Controlling Temperatures of Diode Lasers and Detectors Thermodynamically
App Note #14 - Optimizing TEC Drive Current
App Note #20 - PID Control Loops in Thermoelectric Temperature Controllers
This information has been sourced, reviewed and adapted from materials provided by ILX Lightwave Corporation.
For more information on this source, please visit ILX Lightwave Corporation.