RoadTest: ROHM 8-Channel Multi-Rail DC/DC Converter Board
Author: fmilburn
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?: Other manufacturers supply similar boards for testing their products.
What were the biggest problems encountered?: No major issues. Several small issues such as print size on the silkscreen were noted. Some of the procedures in the accompanying Application Note are not possible without modification to the board such as removing parts.
Detailed Review:
Introduction
Thank you to Rohm and element14 for selecting me for the Rohm 8-Channel Multi-Rail DC/DC Converter Board (REFRPT001-EVK-001) RoadTest. The primary 5V DC to DC buck regulation and secondary 3V LDO regulation are similar to requirements I anticipate needing in future projects.
The RoadTest was intended to be performed to a level that demonstrates the capability of the board and components but not to the level of a design or datasheet validation. Comparisons were made to specific sections of the associated Application Note but with some differences that have been noted in the report.
For example, the Rohm Application Note gives results with the LEDs and supervisory ICs removed in some instances. I elected not to do this since my RoadTest application did not contemplate modifications to the board.
The board has many capabilities and multiple configurations. Not all aspects and in particular not all configurations of the board were evaluated.
Unboxing
The board arrived double boxed and undamaged. As I've come to expect from Rohm, the inner packaging was excellent with padding and what appears to be an antistatic sleeve.
The PCB is marked PCB0220 Rev A and the quality of fabrication on my 4-layer board is excellent. The test pin openings are quite small. They are large enough for mini-grabbers and my oscilloscope probes but small enough for me to have some difficulty threading the needle. The lettering was small enough to be difficult to read also.
There is minimal documentation inside the box but the documentation provided on the Rohm website is excellent. The primary links are provided at the bottom of this RoadTest.
Description and Overview
The REFRPT001-EVK-001 is a power tree solution reference design for automobile infotainment devices and advanced driver assistance systems (ADAS) electronic control units (ECUs). Thus, we expect superior component quality, safety features, and documentation. The REFRPT001-EVK-001 provides all these attributes.
The system block diagram is shown below.
Credit: REFRPT001-EVK-001 User's Guide
The primary conversion from a typical 12V input (9V minimum to 16V maximum) is done with a BD9P105EFV-C buck converter to 5V and a BD9P205EFV-C buck converter to 3V3. The 5V primary then provides secondary 1V25 output with a BD9S201NUX-C buck converter and 3V3 output with a BD00IA5MEFJ-M LDO regulator. The 3V3 primary provides secondary 1V8 output with a BD9S201NUX-C buck converter, 1V5 output with a BD9S300MUF-C buck converter, and 1V output with a BD9S400MUF-C buck converter.
A BD3904MUF-C supervisory IC provides additional voltage monitoring, WDT, reset, and built-in self-test.
Rather than copying monumental amounts of data to the RoadTest, links are given in the paragraphs above to the individual IC datasheets.
The design operating conditions (as opposed to operating conditions given by the datasheet) for the components in the REFRPT001-EVK-001 are shown below.
Credit: REFRPT001-EVK-001 User's Guide
The figure below labels the location of components.
Credit: REFRPT001-EVK-001 User's Guide
As noted above, the documentation provided by Rohm is excellent and additional links are given at the bottom of this RoadTest. The documentation includes a schematic, parts list, simulations, and PCB layout.
Since the board showcases some of Rohm's DC/DC converters, it is worth taking a quick look at one of them. The BD9P105EFV-P 5V converter was arbitrarily selected. This chip does the primary conversion down to 5V and operates at around 2.2MHz. The output voltage accuracy is given by the datasheet to be ±1.75 % (-40 °C to +125 °C). There are two different typical applications given in the datasheet. The one used in the REFRPT001-EVK-001 is "without Discharge Function" as given by Figure 4 in the datasheet as shown below.
Credit: Credit: BD9P1x5EFV-C Series Datasheet
While the REFRPT001-EVK-001 uses this circuit with VOUT_DIS tied to the ground, it also allows connection to VOUT as shown by the circled excerpt from the User Manual below.
Credit: Selected portion of the schematic in REFRPT001-EVK-001 User's Guide
The purpose of showing this is not to delve into the BD9P105EFV-P converter or the other DC-DC converters but to observe that Rohm appears to have provided users with options for exploring different applications outlined in the datasheets.
The instructions in the User Manual are limited but Rohm also provides a detailed test report with their results for the REFRPT001-EVK-001. The instructions below are what is provided in the User Manual.
The Configuration and Mode Control of the board is easily done with jumpers. Tables 3 and 4 in the User Manual describe the settings. Modes are described in the datasheets for the individual components.
The remainder of this report describes how my tests were designed and the results.
Test, Equipment, and Methods
The RoadTest was not done with calibrated lab-grade instruments and is not intended to be a verification of the datasheets or a full test of the REFRPT001-EVK-001. Instead, it explored the capabilities of the board and performed tests sufficient to determine the usefulness of the board and documentation. Engineers wishing to use the design will have specific testing requirements for their specific application likely not covered here.
The primary test equipment used for the testing is listed below:
Dynamic load tests were limited to one channel at a time with this equipment. Most tests were limited to 12V input and the influence of temperature was not examined. High loadings were not obtained in all instances. Test Methods are described at the beginning of each test.
The tests included:
The nomenclature in the system block diagram was used in the tables and figures below for recording test results.
Basic Test - no load
Procedure: The board was powered up at 9, 12, and 16V with no load. The power-on LEDs (D11 and D21) were checked. The voltage at the output of all 8 channels was checked against the expected output with the bench multimeter.
Results:
Observations: The results are as expected.
Basic Test - 10 Ohm resistive load on the primary converters
Procedure: The board was powered up at 12V with 10 ohm resistive loads on both the primary DC/DC converters. The power-on LEDs (D11 and D21) were checked. The voltage at the output of the primary channels was checked against the expected output with the bench multimeter. A rough overall efficiency using the bench output readings (without 4-wire measurement) was determined.
Results:
Observations: the voltage output is as expected. Note that the 3V3 output has dropped slightly (still well within datasheet norms) and that overall efficiency is about 82%. Eyeballing the datasheets one might expect around 75% efficiency at these currents so these all seem good. More accurate measurements were taken later in the RoadTest.
Startup Behavior
Procedure: The board was powered up at 12V with the electronic load on the DC/DC converter set to constant current at the level indicated in the results. The oscilloscope was connected to the output of the converter being tested and adjusted to show the entire trace at startup. The high-resolution acquisition mode was selected. The electronic load was then turned off and the oscilloscope placed in single-shot capture mode. The electronic load was then turned back on and the screenshot recorded.
The test setup is shown below.
Note that the electronic load tester used has a lower limit of 1V5 which limited the channels tested.
This test differs from the one in the Rohm REFRPT001-EVK-001 Application Note in that it uses the enable pin to initiate the test while the board was already enabled and a load suddenly applied in this test. The tests also differ in that LEDs and supervisory ICs on the board were removed in the Rohm Application Note but left in place for this RoadTest.
DCDC_P5V Results powered up to 1A:
Observations: The voltage rose to 5V without overshoot. The full time to rise was 4ms. The results are in line with expectations. The waveform from the Rohm Application Note is very similar.
DCDC_P5V_S3V3 powered up to 1A:
Observations: This is a 3V3 LDO regulator. The Application Note reports a flat curve without the bump and about the same rise time. The differences may be due to the supervisory ICs.
DCDC_P3V_PSW powered up to 0.1A:
Observations: The voltage rose to 3V3 without overshooting. The full time to rise was 3.2ms. The bump in the RoadTest results may be due to the presence of the supervisory ICs.
DCDC_P3V_S1V8 powered up to 1A:
Observations: The voltage rose to 1V8 without overshooting. The full time to rise was 10ms. The waveform from the Rohm Application Note is similar but shows only 5ms to obtain full rise.
DCDC_P3V_S1V5 powered up to 0.5A:
Observations: The voltage rose to 1V5 without overshoot. The full time to rise was 5ms with two small bumps. The waveform from the Rohm Application Note is similar but with a rise time of 3ms.
Voltage Stability
Procedure: The board was powered up at 12V with the electronic load on the DC/DC converter or LDO being checked. The current was then increased in increments of 100 mA using the electronic load and the output voltage was recorded with the bench multimeter until the DC/DC converter or LDO cut out.
DCDC_P5V Results:
Observations: The voltage output is stable out to 1.14A and then drops 4.4% at 1.16A. The voltage accuracy is given to be +/- 1.75% per the datasheet. The reset for the 5V power tree and power good for the 5V power tree LEDs come on as expected. The voltage output stability is as good as or better than expected.
The Application Note extends the curve out to 2A, but the REFRPT001-EVK-001 User Manual states that the maximum is 1A. Otherwise, the plots are similar.
Voltage Ripple and Noise
Procedure: Only DCDC_P5V was recorded although several others were tested. The oscilloscope probe was placed into the test pins using the ground spring instead of the flying lead on the oscilloscope. The board was powered up at 12V with the electronic load at 1A on DCDC_P5V and a representative single-shot recorded.
DCDC_P5V Results:
Observations: There was some difficulty in maintaining good contact on the test pins with the ground spring clips in place on the probe. Peak to peak ripple is something over 4mV and representative. The Rohm Application Note was made with an output current of 2.0A but peak to peak ripple is similar. Ripple was observed on several other channels with similar results but not recorded.
Overall Efficiency
Procedure: Only DCDC_P5V was tested. DCDC_P3V was disabled so it would not draw current. The board was powered up at 12V with the electronic load at 0A on DCDC_P5V. The current was then increased in increments of 100 mA using the electronic load, the output voltage recorded with the bench multimeter, and the current from the bench power supply recorded using the meter on the power supply.
DCDC_P5V Results:
Observations: The efficiency is somewhat lower than the plot in the Application Note. The difference is largely due to the powered LEDs on the board and the presence of the supervisory ICs which were left in place. In any case, it is reasonably close. Note that once again the curve runs all the way out to 2A The plot from the Application Note for DCDC_P5V is given below with the RoadTest outcome added in red for convenience.
Credit: REFRPT001-EVK-001 Application Note with RoadTest outcome added in red by author
It is not clear to me how the secondary DCDC converters were tested directly in place without modifications to the board as there isn’t an obvious way to measure the current.
Power Monitoring
Power monitoring is done on both the 5V and 3V3 trees. It consists of power good, XRSTOUT, and WDT status. LEDs make the status visible. In the video below the input voltage is reduced and the load increased until the power monitoring signals a failure.
Observations: Limited testing was done but power monitoring worked as expected. While pin headers are provided, indication with LEDs was useful and did not require additional equipment or a microcontroller.
Component Temperature
Procedure: Only DCDC_P5V was tested. The board was powered up at 12V with the electronic load at 1A on DCDC_P5V and allowed to run for 30 minutes. A low-resolution MLX90640 IR Thermal Camera was used to get a visual and was checked with a thermocouple measurement using the bench DMM.
DCDC_P5V Results:
Observations: The red spot in the center is the DCDC_P5V and the orange spot to the right is the associated inductor. The hot spot is reading 32.2 degrees C on the thermal camera and 32.0 was read using a thermocouple. The current was 1A. The room free air temperature was 20.7 degrees using the thermocouple for a difference of around 11.5 degrees C.
The Application Note recorded a temperature difference of 24.5 degrees C with roughly 1.3A throughput. This is partially explained by the higher current in the Application Note and also my rough measurements but is larger than expected. In any event, it demonstrates that the IC does not run hot.
Summary and Conclusions
This was an interesting RoadTest and satisfied my curiosity about what goes into a design of this type. The REFRPT001-EVK-001 has the following positive attributes:
There were no major issues encountered during the RoadTest but the following observations were made:
And finally, a critique of my efforts. While my test equipment does not achieve lab certification quality it was adequate for the RoadTest as proposed. Where it was lacking was the ability to do multi-channel testing and measurement simultaneously. My original thought in the RoadTest application was to use purely resistive loads for all channels under test except the one attached to the electronic load. It turned out the appropriate high-wattage resistors were not on hand to do this.
Thanks for reading. Comments, suggestions for improvement, and corrections are always welcome.
Links
