While browsing through the latest ICs released in the market by major semiconductor manufacturers, I came across this cute little buck regulator by Fairchild Semiconductor. I really don't do written reviews (because it kind of feels like I'm advertising the company's product) but I'll make this one an exception. I mean, how can I just ignore this teenie weenie monster of an IC (I'll explain why I see it as a monster later) that I just came across with.
FAN53202 or TinyBuck (mind if I call it by its family name for the rest of the article) is "a highly efficient step-down switching voltage regulator that offers best-in-class load transient and a programmable output voltage" according to FS. Now, the terms "highly efficient" and "best-in-class" are highly arbitrary marketing terms (that sometimes vexes me a lot, really, because engineers don't make comparisons on a whim without reference points). Plus, I didn't want to believe that TinyBuck, which is 1.6mm by 2mm in size, could offer superior efficiency and load transient characteristics as compared to its competitors so easily, so I took a closer look at its data sheet (link available below).
Download TinyBuck's DataSheet
DataSheet AnalysisBased on the data sheet, the input voltage range is from 2.5V to 5.5V, which is typical. TinyBuck is programmable through an SPI or I2C interface that can operate up to 3.4 MHz. There is an internal soft-start to avoid shocking sensitive internal circuitry with a sudden VIN transient, undervoltage lock-out (UVLO - I'm guessing around 2.5V-2.9V), TSD of 150 degrees Celsius (a little lower than ordinary), and SCP protection. The default frequency is 2.4MHz and VOUT is trimmable from -0.6 to 1.3875V in 12.5 mV steps. It is claiming a "Best-in-Class" load transient again (ugh, I'll dig deeper into this) and a continuous output current ability of 5A (wow). The quiescent current is 60uA in PFM mode, a little higher than expected. Finally, a 91% efficiency (I'm guessing this is the maximum achieved efficiency by TinyBuck because the efficiency curves are different for different modes of operation. All of this is crammed in a 1.6mm. by 2mm. wafer-level chip scale package ( so tiny <3 , I hope I don't accidentally inhale this thing).
Now it is time to digest. Shown below is the block diagram and pin configuration of TinyBuck.
|Figure 1. Block diagram of FAN53202|
|Figure 2. Pin configuration of FAN53202|
To power up TinyBuck, we set EN to high, bias VIN with 3.6V [the datasheet didn't really indicate a typical voltage but I'm pretty sure it will operate correctly at 3.6V], set my I2C slave address to C0
and off we go. From the register map, by sending 01h/00h to 01h/3Fh, I should be able to trim VOUT.
It looks like the modes are up to date (they've integrated the Auto and PFM modes). Oh, and the slew rate is adjustable.
So on the testbench, when I probe the SW pin, I'm expecting a 2.4 MHz switching frequency to appear on my scope. The power up sequence should go smoothly.
Highly Efficient, just like a typical switching regulator ought to be
|Figure 3. Efficiency curves for FAN53202|
From the efficiency curves, TinyBuck was obviously evaluated in Auto/PFM mode (Remember that PWM mode is inefficient at low loads, and the hump near the end of the curve accounts for the transition from PFM to PWM - i.e. Auto mode). TinyBuck indeed has a maximum efficiency of 91%, but don't expect it to be 91% efficient at a range of 1 mA to 10 mA load.
The "...continuous output current ability of 5A" - Wow!In order to analyze the claim that TinyBuck can indeed withstand 5A of continuous load, I've extracted the load regulation curves of the chip.
Below are the load regulation curves of TinyBuck.
|Figure 4. Load regulation curves of FAN53202|
"Best-in-Class" Load Transient Response, pretty much looks like itFinally, I review the transient response of TinyBuck, which looks very promising. Below is the transient response of TinyBuck.
|Figure 5. Transient response of FAN53202. Probably measured with a Tektronix MSO.|
A pulsed load of 1A to 1.5A would typically yield a voltage deviation of around 200mV or more. But from the scope measurement (mind you, I think a Tektronix MSOXXXX series was used XD Hahaha... one of my favorite scopes by the way, I have to write all about it in another article...) the voltage only changed by around 30-50mV!! Amazing! If you get a typical statistic of load transients of switching regulators, you'd get 30-50mV at loads of 10-100mA with 1uA rise time and fall time.
But wait, looking at the right side of the CH1 label. The channel is bandwidth limited? Well, I'm not really sure if band-limiting the voltage probe would have a BIG effect on the transient response but I'm sure there is, though quite negligible (in the order of a few mV perhaps).
In the end, TinyBuck does live up to its claims, based on current standards. I would like to see more reviews of TinyBuck by actual commercial users in the future.