Package thermal resistance of Allegro MicroSystems current sensor with built-in conductor

Narrow down by specifying conditions

現在2147件がヒットしています。check

Design Analog Sensors Allegro

Junction to Ambient Thermal Resistance (RθJA)

In semiconductor devices, the die temperature should not exceed the maximum junction temperature listed in the datasheet for extended periods of time. For Allegro current sensors with integrated conductors, the maximum junction temperature is 165°C, while some competing products are limited to 150°C. If the temperature exceeds the 165°C limit listed in the datasheet, there is a risk that the characteristics of the pn junction of the device will change, affecting the IC's performance and the long-term reliability of the part. The thermal resistance of the package (RθJA, the thermal resistance from junction to ambient) is the industry standard for determining the die temperature under specific conditions. Knowing the RθJA of the package and the ambient temperature, the junction temperature can be determined for the power dissipated in the package.

Typically, RθJA is based on a JEDEC standard board and is useful for comparing the relative thermal performance between packages. However, for certain applications, the standard RθJA number can be misleading, because the heat spreading is highly dependent on the PCB, the ambient environment, and the package construction. Especially for high power applications, power-optimized PCBs and specialized packages such as Allegro current sensors are often used. For power components, RθJA should be determined based on a thermally optimized PCB for comparison purposes, or ideally based on the board used in the application being evaluated. In this article, we will cover the following:

1. Allegro conductor embedded current sensor package constructions (MA, MC, LA, LZ, EZ, LH) and their impact on thermal metrics
2. Industry Standard Thermal Index
3. Thermal measurement methods to quantify metrics for specific applications
4. Heat generation measurement results of the Allegro current sensor evaluation board

Figure 1: Allegro packages discussed in this article (provided by Allegro MicroSystems)

Package Structure and Thermal Model

Allegro conductor integrated current sensors are available in two basic package types:

■ Die-up structure (MA, MC and LH packages)
In this construction, the die is placed directly in the current path to sense the magnetic field generated by the current in the primary inner conductor (IP loop). There is a polyimide insulating layer between the die and the conductor. In this configuration, wire bonds are used to connect the secondary side of the device (VDD, GND and I/O) to the leads (see Figure 2, the wire bonds are shown in white).

■ Flip-chip construction (LA, LZ and EZ packages)
In this structure, the top of the die is placed closer to the sensor Hall plate, enhancing magnetic coupling. In this configuration, solder balls are used to connect the secondary to the leads (see Figure 3, where the solder ball connections are shown in blue).

package	Footprint (mm)	IP Resistance (mΩ)	size (^mm2)	Constitution
MA	106.1	0.85	10.3 x 10.3 x 2.65	Die Up
MC	146.9	0.27	11.3 x 13 x 3.01	Die Up
LA	106.1	0.68	10.3 x 10.3 x 2.65	Flip Chip
LZ	29.3	1	4.89 x 3.9 x 1.47	Flip Chip
EZ	16	0.1	4 x 4 x 1.45	Flip Chip
LH	8.6	1.6	2.9 x 2.97 x 1	Die Up

Table 1: Overview of Allegro integration packages

Figure 2: Die-up internal structure with wire bonds shown in white. *Courtesy of Allegro MicroSystems

Figure 2: Die-up internal structure with wire bonds shown in white
* Provided by Allegro MicroSystems

Figure 3: Internal structure of a flip chip with solder balls shown in blue. *Courtesy of Allegro MicroSystems

Figure 3: Internal structure of a flip chip with solder balls shown in blue
* Provided by Allegro MicroSystems

For most standard IC components, the main heat source is the die itself, and the temperature of the heat source (Q) is considered the junction temperature. For current sensors with conductors, the main heat source is the IP loop, which can dissipate several watts of power, which is orders of magnitude larger than the power of the die itself, which typically does not exceed 75mW, and can therefore be ignored in this analysis.

The primary heat source in the package is the IP loop, which is separated from the die by a polyimide insulation layer, a high-voltage insulating material, as shown in Figure 4. The polyimide insulation layer (also called polyimide insulation tape) is a good thermal insulator, so the temperature of the heat source cannot be considered the same as the junction temperature, which is typically the assumption for RθJA. The thermal resistance of the gap (RGAP) can be greater than the thermal resistance of the surrounding mold compound (RMC2). This affects the temperature of the top of the case (TC). In some cases, the temperature of the top of the package can be equal to or lower than the junction temperature (TJ).

Figure 4: Overview of package thermal resistance *Provided by Allegro MicroSystems

On the other side of the die there is the thermal resistance of the electrical interconnect (RINT) that connects the die to the secondary side of the copper lead frame, which has a thermal resistance (R2nd). The gold wire bonds in a die-up package configuration have poor thermal conductivity through thin wire bonds. Flip-chip packages use solder balls and have short metal interconnects that have good thermal conductivity from the die to the copper lead frame. Therefore, in most applications, flip-chip packages provide better heat dissipation from the die to the PCB. This assumes that there is minimal airflow or heat dissipation through the top of the package. Note that in high voltage applications, dissipating heat through the top of the package is generally not recommended, as it significantly reduces the creepage and clearance distances required for high voltage isolation.

The thermal model of the Allegro current sensor is much more complex than a standard semiconductor because it has two lead frames with different thermal resistances and a gap between the heat source and the junction. This thermal model is provided to show the difference between the standard semiconductor model and the Allegro current sensor model (see Figure 4). This model is too complex to be used in a real application. Experimental data is required to accurately measure the RθJA of the Allegro integrated current sensor.

Heat Index Definitions

RθJA is a lumped thermal resistance to account for the heat flow from the power dissipated in the package between the circuitry on the die and the ambient environment. Because RθJA is a single lumped parameter, it incorporates the entire thermal system including the die, package, PCB and heat dissipation to the ambient environment. Changes in the PCB and environment will affect the value of RθJA. Caution should be used when considering RθJA in thermal analysis and misuse of this parameter can result in misleading data.

RθJA is typically based on a two-layer PCB layout as specified by JESD51. For current sensors, the JEDEC standard board is insufficient as it cannot handle the high currents that the Allegro integrated current sensor packages are designed for. The RθJA defined here is based on the Allegro evaluation boards used in this application note, which are optimized to carry high currents. RθJA is highly PCB and environment dependent and will be compared across different packages later in this article.

Another common set of metrics used to determine junction temperature are RθJC (junction to case) and RθJB (junction to board). These assume a simple model of the package and are independent of the final application (see Figure 5). This model simplifies the complex 3D structure of Figure 4 and is practical for system-level thermal simulations. The RθJC and RθJB metrics can be misleading for Allegro current sensors because they assume ideal heat flow from case to ambient and board to ambient. Allegro provides these thermal resistance values for several packages and allows users to use these metrics in their own thermal models and simulations.

Figure 5: Simple model for thermal simulation *Provided by Allegro MicroSystems

RθJC is the package thermal resistance between the junction and the top of the package (or convection to ambient) and is useful for calculating heat dissipation in low voltage applications. This metric is often determined with very low thermal impedance to ambient and a copper slug on top of the package as specified in JESD51-1 (see Figure 6). RθJB is the package thermal resistance between the junction and the PCB and is useful for calculating heat dissipation through the board and convection to ambient. This metric is often determined with very low thermal impedance to ambient and a copper enclosure as specified in JESD51-8 (see Figure 7). Simulation results for RθJC and RθJB for several Allegro current sensor packages are shown in Table 2.

Figure 6: RθJC model
* Provided by Allegro MicroSystems

Figure 7: RθJB model
* Provided by Allegro MicroSystems

package	R_θJC_(simulated)	R_{θJB (simulated)}
MA	14	14
MC	15	7
LA	10	8
LZ	23	12
EZ	70	1
LH	155	19

Table 2: Allegro integrated package simulation overview

RθJC and RθJB are idealized for system-level simulations and do not lend themselves to experimental measurement.

Since RθJC is a function of the surface area of the top of the package, the LH package, which is the smallest package discussed in this application note, has the highest RθJC. The MA and MC packages have a larger polyimide insulating layer and the resistance to the top of the package is greater than the LA package, which has a smaller tape.

RθJB is a function of the metal used to spread the heat from the die to the evaluation board. The EZ packages have the lowest RθJB because they are directly connected to the PCB through the bottom of the package. The MC packages have a thicker leadframe material than the MA packages, which gives them a lower RθJB. The LA packages have a lower RθJB because they use secondary leads and have solder balls that help with heat dissipation. The LZ packages have fewer leads. The MA and MC packages do not benefit from the heat dissipation provided by secondary leads.

ΨJT is a practical thermal metric that indicates the temperature difference between the junction temperature and the maximum temperature of the top of the package. Due to complex heat flow paths, ΨJT must also be determined based on the final application. Once determined for the application, ΨJT can be used to experimentally determine the die temperature by measuring the temperature of the top of the package under different load conditions. ΨJT is not a resistance and can be negative if the top of the package is hotter than the junction because the main heat source is not in the same location as the die.

ΨJB is a thermal indicator of the temperature difference between the board and the junction, but it is not very useful for current sensors with integrated Allegro conductors because, due to the package construction mentioned above, heat dissipation to the board creates different temperatures on each side of the part.

Junction Temperature Measurement

To determine RθJA and ΨJT and get an accurate indication, laboratory measurements of the die and case temperatures are required in a bench setup that approximates the final application.

Measuring the die temperature

Because the die is embedded in a plastic molding compound, the only way to accurately measure the die temperature is by direct measurement. One technique is to use an electrostatic protection (ESD) diode from VDD to ground. The voltage across an ESD diode from VDD to GND, or between an unconnected pin and ground, varies linearly with temperature for a given current. This property can be exploited to determine the ratio of temperature change to voltage change (ΔV/ΔT) using the setup shown in Figure 8. This does not give an absolute temperature, but rather an indication of the change from a known ambient temperature.

To determine ΔV /ΔT, inject a known current (typically around 1mA) through an ESD diode from GND pin to VDD pin. Measure the voltage change across the diode at two different known ambient temperatures (for best accuracy, a large temperature change between the two known ambient temperatures is recommended, e.g. 25°C and 125°C). The current sensor is powered off and no current is flowing through the IP loop. Allow sufficient time for the die to reach thermal equilibrium before taking measurements.

Once ΔV /ΔT is known, the change in die temperature from the current ambient conditions can be determined by injecting a current into the IP loop and measuring ΔV. This provides the most direct measurement of die temperature versus applied current; however, this method cannot be used during operation and is intended for engineering evaluation only. You should measure ΔV/ΔT for each Allegro current sensor part number, but part numbers of the same type will share common characteristics.

External Temperature Measurement

A thermocouple can be used to measure the top case temperature (TC). Thermocouples are thermally conductive and can absorb heat, lowering the measured temperature, so care should be taken when using them. Also, to ensure reliable contact, the thermocouple must be glued to the top of the package, which also increases heat dissipation through the thermocouple. It is recommended to use the smallest possible thermocouple and the smallest amount of thermal epoxy to minimize heat dissipation. The advantage of using a thermocouple is that the measurement can be made while the current sensor is electrically operating or remotely in a temperature-controlled chamber.

Alternatively, TC can be measured externally using a thermal camera. This method is easy to use to measure the case top temperature in an ambient environment, but difficult if the current sensor is inside a temperature-controlled chamber or other enclosure. To get more accurate results using a thermal camera, care should be taken to reduce the effects of emissivity of reflective surfaces (see manufacturer documentation for camera specifications). Allegro used a thermal imaging camera to measure RθJA and ΨJT.

Allegro Current Sensor Evaluation Board Results

Allegro Current Sensor (ACS) Evaluation Boards (EVBs) are available for most package types. As shown in Figure 8, the test setup for the ACS EVB for the MA and LA packages is shown. These PCBs are 6-layer construction with 2oz copper with via-in-pad for the IP pins to maximize copper heat dissipation through the PCB. This reduces the heat generated by the current on the PCB and helps with heat dissipation by keeping the temperature difference between the current sensor and the PCB as high as possible. Testing was done using 2 AWG wire, which also provides an additional heat dissipation path.

Figure 8: Bench setup *Courtesy of Allegro MicroSystems

To demonstrate the relative thermal dissipation characteristics of the packages, 3W of power was dissipated in each package. For the EZ package, 1.5W was used as 3W was beyond its operating range and the PCB could not handle that current. This was done by measuring the voltage drop in the IP loop and setting the current to produce the desired total power in the package. Thermal images of the board and package are shown in Figures 10 to 12. See Table 3 for a summary of the currents used.

package	IP Loop Drop (V)	Current (A)	Power (W)
MA	0.0532	56	3
MC	0.037	82	3
LA	0.064	47	3
LZ	0.053	56	3
EZ	0.012	124	1.5
LH	0.075	47	3

Table 3: Allegro integration package test summary

The MA and MC packages show hot spots on the primary side of the package (the IP loop side) with very little heat being spread to the other side of the package. The LA, LZ and LH packages show a more uniform heat distribution across the package because the solder balls spread the heat more efficiently to the other side of the leadframe. The MC package has a leadframe that is twice as thick as the other packages and spreads heat more efficiently into the PCB as evidenced by the higher temperatures on the board.

The LZ and MA packages required the same 56A to achieve 3W in the package. The LZ package has lower resistance than the MA package and better heat dissipation through the solder balls to the other side of the package. However, the MA package is larger and has more area and thermal mass for heat dissipation. These tradeoffs balance each other and show that the heat dissipation characteristics of a package are a complex combination of materials and geometry. The EZ package shows that it spreads heat more effectively to the PCB and dissipates less heat through the top of the package than the leaded package with its larger body. The RθJA and ΨJT of the Allegro current sensor packages on these boards are shown in Table 4. The LA and LZ packages have lower RθJA and ΨJT than the MA and MC packages because the solder balls spread heat more effectively from the die to the PCB. The ΨJT of the LA and LZ packages is close to 0°C/W or negative, meaning that the temperature of the top of the package is the same as or lower than the die. The RθJA of the EZ package is higher than other packages due to its small package size and excellent heat dissipation to the PCB.

package	R_θJC_(measured)	R_{θJB (measured)}
MA	20	2.4
MC	19	2.4
LA	19	0.5
LZ	16	-1.7
EZ	55	7.5
LH	33	3.7

Table 4: Allegro Integrated Package Measurement Summary

Careful consideration of heat dissipation on the PCB is important to minimize the die temperature for average heat dissipation. Maximize the PCB copper trace area and trace thickness for the current path to the IP loop and use via-in-pad to minimize the die temperature rise. In addition, the current carrying wires and interconnects that deliver current to the PCB are also important factors that affect the die temperature. The size of the wires and PCB trace area dedicated to heat dissipation can affect the die temperature by more than 20°C.

Figure 9: MA package *Provided by Allegro MicroSystems

Figure 10: MC package *Provided by Allegro MicroSystems

Figure 11: LA package *Provided by Allegro MicroSystems

Figure 12: LZ package *Provided by Allegro MicroSystems

Figure 13: EZ Package *Provided by Allegro MicroSystems

Figure 14: LH package *Provided by Allegro MicroSystems

Conclusion

Determining the die temperature requires careful consideration of the thermal characteristics of the package, PCB and overall system to keep the die temperature at or below the maximum specified junction temperature of 165°C. Industry standard metrics like RθJA should be used with caution as they are often applied to specific conditions and can be misleading in certain applications. The results presented here are useful for comparing packages and illustrating the different effects that package construction can have on thermal performance.