RDK-877 Power Reference Design Kit for LinkSwitch-TNZ ICs RoadTest

Table of contents

RoadTest: Power Integrations Home Appliance and Industrial Power Reference Design Kit

Author: flyingbean

Creation date:

Evaluation Type: Power Supplies

Did you receive all parts the manufacturer stated would be included in the package?: True

What other parts do you consider comparable to this product?: No

What were the biggest problems encountered?: High voltage domain signal measurement

Detailed Review:

1. Unboxing P.I. Power Reference Design Kit Package

The mailed package was from Newark. I liked the picture of the development kit: flyback from your cell phone to the coffee machine, which means air-gaped LinkSwitch-TNZ topology and zero cross detection (ZCD) can be used by a smart phone or a smart house from this photo.

{gallery}OOB

Newark Mailed Package

Mailed Package

RDK-877

RDK-877 Package

The size of the development board is small enough to fit into a normal power adaptor case as one demo in Figure 1.

RDK-877 Size

Figure 1. Small Footage of RDK-877 Reference Design

2. RDK-877 Development Kit Documents

There are three main documents (Ref[1], [2], [3]) for me to learn about LinkSwitch TM-TNZ LNK3306D IC part and the reference design, RDK-877. The first impression of the roadtest is the documents from P.I. are well presented with a lot of signal waveforms, pictures, and photos. So many designers with/without deep power design background can jump into a project using LinkSwitch TM-TNZ  family chips right away.  

LinkSwitch-TNZ circuit design from Ref[1] is for non-isolated power supply, however, the general guidance of the design procedure is a good reference for isolated power supply as well. I really like the table/sheet approach to list most of steps for selecting design parameter as below.

RDK-877 presents a DCM(discontinuous conduction mode) reference design. Ref[1] illustrates the comparison between MDCM(mostly discontinuous conduction mode) and CCM(continuous conduction mode) as below.image

RDK-877 picks DCM as the reference design due to its higher efficiency and lower cost advantages.

The development board looks very clean on both sides.  The overall PCB layout demonstrates the safety design guidelines for high AC power domain. I opened RDK-877 PCB layout design file, RD_877_REVB.brd, and checked  the cross-section report. There is 8mil distance between the top and bottom layers as below.

 image

Here is the typical spacing constraint I used to validate power domain PCB layout:

image

The bottom of the PCB assembly is a great demonstration of the safety design.

RDK-877_bot

Figure 2. RDK-877 PCB Layout

3. RDK-877 Development Kit Roadtest Items

Ref[2] lists  the minimum performance of RDK-877. I planned to evaluate the items inside the marked rectangle area.

image

The LNK3306D device U1 integrates the power switching device, oscillator, control, startup, and protection functions. R8, R9, R14 are key design parameters for AC line zero crossing detection as the reference design board below. The red “X” marks present my bench testing probe location.

image

Figure 3. RDK-877 Schematics and RoadTest Probing Location

In addition, I did measure ZCD signal and soft start-up on RDK-877.

3.1 Power Efficiency Evaluation

a. Resistive Load

I used high power resistors as the load of 12VDC power supplied by RDK-877.  The overall bench testing setup is presented below.

PWR Effy BenchTesting Setup

Figure 3. Power Efficiency Measurement with Resistive Load

RDK-877 does provides a set of test points to make my bench testing much easier than I expected as Figure 4.

RDK-877_TP

Figure 4. Test Points on RDK-877 Board

I did spend a couple of weeks to find a reliable way to measure the real power from AC line. Finally, Hantek CC-65 AD/DC current clamp meter with Rigol MSO5074 power quality feature provided me a reasonable result for the roadtest. Hantek CC-65 AD/DC current clamp meter is a low-cost meter for my hobby projects. To validate the accuracy of Hantek CC-65 AD/DC, I connected a True RMS Flute multi-meter in serial at AC115 line before AC115V_L test point.  Here are the snapshots of real PWR measurement from AC115 line as the input power of RDK-877 board.

Case01: 3% resistive load

image

        Ch1: AC115V line voltage

        Ch3: ZCD

        Ch4: AC115V  line current

        Purple color waveform: AC115V  line power from Rigol power quality measurement. The real power for AC115V I_rms=6mA is around 225mW. The power factor is around 0.31.

Figure 5. AC115V Real PWR with 3% Resistive Load

Case02: 83% resistive load

image

         Ch1: AC115V line voltage

         Ch3: ZCD

        Ch4: AC115V  line current

        Purple color waveform: AC115V line power from Rigol power quality measurement. The real power for AC115V I_rms=86mA is around 6.5W. The power factor is around 0.64.

Figure 6. AC115V Real PWR with 83% Resistive Load

Here is the final measurement report for resistive load on my bench testing. I used Flute 177(calibrated until 2024 April 3rd) for AC_115V line current True RMS measurement at the same time and recorded the data into the spreadsheet below for an adjustment input PWR reading if Hantek CC-65 AD/DC current clamp meter reading is not good enough.

image

The overall power efficiency vs load is illustrated below.

image

Figure 7.  RDK-877 PWR Efficiency with Resistive Load

I attached some of the measured PWR efficiency data (PWR_Ef_Adj) vs Ref[2] Figure 9 as below for a better visualized comparison for my bench testing.

image

My power efficiency measurement looks reasonable for AC115V line, however, the accuracy might not be high enough compared to P.I. lab’s professional measurements as Ref[2] due to my hobby level equipment or I did not soak RDK877 long enough for the measurement.

b. Zener Diode Load

I used 3.5V/5W Zener diodes (Onsemi 1N5333B) as emulated lighting LEDs for the PWR efficiency measurement.  The emulated high current LED lighting subject is presented at Figure 8.

Zener Diode Load

Figure 8. Zener Diode Load for Power Efficiency Measurement

I configured the load at 100% from Zener diodes and presented you another snapshot of the real power measurement from AC115V  line.

Case03:  100% load from Zener diode configuration

image

          Ch1: AC115V line voltage

         Ch3: ZCD

         Ch4: AC115V  line current

         Purple color waveform: AC115V line power from Rigol power quality measurement. The real power for AC115V I_rms=104mA is around 8.4W. The power factor is around 0.67.

Figure 9. AC115V Real PWR with100% Zener Diode Load

My bench testing data on the PWR efficiency measurement for Zener diode load is as follows. I assumed that 500mA is full load current.

image

image

Figure 10.  RDK-877 PWR Efficiency with Zener Diode Load

I combine the measured PWR efficiency data (PWR_Ef_Adj) from Figure 10 and Figure 7 vs Ref[2] Figure 9 as below.

PWR_effy

The orange “X”s mark the PWR_Ef_Adj from Zener diode loads and the red “X”s mark PWR_Ef_Adj from resistive loads. None of my bench testing can get better power efficiency results than P.I. lab report.

3.2 RDK-877 Soft-start Check

Soft-start can gradual turn on of an electronic power supply to avoid stressing the components by the sudden current or voltage surges associated with initially charging capacitors and transformers. Soft-start also make the overall system more safe and reliable. P.I. LinkSwitch-TNZ family ICs do provide such features for better power supply performance.

Case01: 3% load from cold start

Here is one snapshot under 3% load for illustration of the surging signal from AC115V line while power-on RDR-877 board. RDR-877 has not been powered over 12 hours before this start-up.

3%load_startup

Figure 11.   RDK-877 Start-up at 3% Load

The soft start-up time is around 11ms, which is longer than 256 cycles of 33kHz which is defined at Ref[3] datasheet for LNK3306D switching clock. The surging AC115V line current can be up to 2~3 A.

ZCD signal looks stablelized after the first cycle of AC115V line from this snapshot.

Case02:  100% load from cold start

Here is one snapshot under 100% load for illustration of the surging signal from AC115V line while power-on RDR-877 board. RDR-877 has not been powered over 12 hours before this start-up.

100% star-up cold

Figure 12.   RDK-877 Start-up at 100% Load

The soft start-up time is around 9ms, which is longer than 256 cycles of 33kHz which is defined at Ref[3] datasheet for LNK3306D switching clock. The surging AC115V line current can be up to 5~6 A.

ZCD signal looks stablelized after the first cycle of AC115V line from this snapshot.

Case03: 100% load from warm start

More cycles on 100% load after resetting the board are presented on Figure 13. This start-up was tested from a reset cycle, which means RDK-877 started to work after a few minutes of power-off from previous power-on status.

100% Load Warm Startup

Figure 13.   RDK-877 Start-up at 100% Load after Power Cycle Reset

The soft start-up time is around 15ms, which is longer than 256 cycles of 33kHz which is defined at Ref[3] datasheet for LNK3306D switching clock. The surging AC115V line current  is not visible.

ZCD signal looks stablelized after the first cycle of AC115V line from this snapshot.

3.3 ZCD Signal Measurement

One of the applications using LinkSwitch TM-TNZ LNK3306D is ZCD. As the datasheet mentioned, LinkSwitch TM-TNZ LNK3306D can accurately indicate when the sinusoidal AC line is at zero volts. This signal can be used to minimize switching stress and system in-rush current. If a power distribution system has relays, ZCD can be used to control how to switch relays for smart home products in terms of the phase of AC line.

There is around 4kHz jitter from LinkSwitch-TNZ oscillator design referenced (Ref[3]) due to the need to minimize EMI emission as below.

image

ZCD signal jitter was analyzed via Rigol 5074 oscilloscope as below.

image

Figure 14. ZCD Zero-cross signal jitter

The measured jitter TIE is around 137us, which used PLL recovery method for the measurement. Since LinkSwitch-TNZ datasheet already mentioned there is around 4KHz -> 250uS frequency jitter, 137us TIE jitter is inside the margin of the design. Then RDR-877 does provide a decent approach to measure ZCD signal without distortion.

 The measured ZCD signal delay from AC115V line at 100% load.

{gallery}ZCD_latency

 zcd_100%load

ZCD at 100% Load

 ZCD Rising Edge Latency

ZCD Rising Edge Latency

 ZCD Falling Edge Latency

ZCD Falling Edge Latency

The overall ZCD latency vs AC115V at 100% load are:  220us at rising edge and 80us at falling edge. I think VR3/Q1 and VR4 are the dominated parts defining ZCD latency performance.

Conclusions

RDK-877 presents a great reference design for P.I. LinkSwitch-TNZ IC, LNK3306D. The datasheet, application guide, lab experiment setup and measurement data from P.I. are well documented and very informative for designers. The overall performance of the power efficiency, and ZCD timing from my bench testing did match with the datesheet and design notes well enough. LinkSwitch-TNZ can provide an efficient approach for AC line zero crossing detection. RDK-877 provides an easy way to output ZCD signal for more applications without distorting ZCD signal.

In summary, RDK-877 did a great job to provide a reference design for better using P.I. LinkSwitch-TNZ family parts.

References

[1] Application Note AN-98 LinkSwitch-TNZ Family, Buck and Buck-Boost Design Guide, 2022

[2] Reference Design Report for a 6W Isolated Flyback Power Supply with Lossless Generation of AC Zero Cross Signal Using LinkSwitch TM-TNZ LNK3306D, 2022

[3] LINK33x2-7D LinkSwitch-TNZ Family Datasheet, 2021

Anonymous