In this article, we discuss power hold-up and some energy storage management devices in the context of SSDs.

SSDs and Why They Need Power Hold-Up Circuitry

Most, if not all, SSDs (solid-state drives) employ some sort of power hold-up scheme using an energy storage system. The power hold-up is used to protect the NAND memory during an unexpected power loss. The term “unexpected” is the key word; during a normal SSD power down (technically referred to as a controlled shutdown), like when you properly turn off your Windows computer by clicking on “Start” and then “Shut down,” an SSD’s power hold-up system isn’t utilized.

This is the case because during a controlled shutdown a special command—called a “standby immediate” command—informs the SSD controller that a power down is forthcoming. Once the standby immediate command is received, all data in the SSD’s buffer is written to nonvolatile NAND cells. However, according to the Data Center Journal, “In the event of a sudden voltage drop, if the host fails to transmit the ‘standby immediate’ command to the controller, all unwritten information will be lost, resulting in corrupted files.” In other words, if an SSD unexpectedly loses power, then the data that’s in the process of being written to the NAND—referred to as “the data in flight”—is lost (aka corrupted).

It’s during this unexpected power loss, or “sudden voltage drop,” when the hold-up power from the energy storage system is engaged. Once the power hold-up circuitry has been activated, the stored energy will be available allowing the SSD to finish writing data to the NAND. The image below shows one SSD manufacturer’s block diagram of power failure and backup power.

 

Figure 1. Block diagram of an SSD’s power failure data protection scheme. Image courtesy of Hexus.net

 

Some Available Energy Backup Options

Most, if not all, SSD energy storage backup systems utilize capacitors as the actual energy storage “tanks.” Based on various SSDs that I’ve seen online, the choice capacitors used for energy storage are tantalum caps, polymer tantalum caps, or good old-fashioned ceramic caps. However, it should be noted that there are strong opinions on which type of capacitor is “the best” capacitor. In my experience, the best capacitor is one that offers a good compromise between cost, availability, reliability, energy density, and physical size, specifically in the z-axis for allowing SSDs to be low profile.

 

Figure 2. Tantalum capacitors are used for energy storage management. Image courtesy of tweaktown.com

 

Figure 3. This SSD uses ceramic caps to provide power hold-up during a power-loss event. Image courtesy of tomsitpro.com

 

Whichever capacitor is chosen, it’s imperative that the capacitors are very robust, have long lifetimes (while being energized at their expected voltage level…this is key), and have good to excellent performance characteristics at the same advertised temperatures range as the end product has. And because the energy of a capacitor uses the square of the voltage (see equation below), the voltage of the capacitor—that is, the higher the voltage rating the better (as long as the capacitance value isn’t derated too much at the elevated voltage level)—can be more important than the actual capacitance value of the capacitor.

 

$$Energy\ stored: W = \frac{1}{2} CV^2 \ (joules)$$

 

If you’re planning to use a supercapacitor as your energy storage device, the LTC3350, from Linear Technology, may prove to be a good capacitor-charging IC. This IC is advertised, according to its datasheet, as a “backup power controller that can charge and monitor a series stack of one to four supercapacitors.” As a side note, supercapacitors have been used on SSDs for power hold-up, however, because they contain a liquid electrolyte, as described in this article, they are not always the first choice for providing energy storage for SSDs.

 

Figure 4. A typical application of the LTC3350. Image taken from the datasheet.

 

If you’re looking for, perhaps, a simpler approach by use of a reference design, TI offers their PMP30046, which is advertised as being an “Enterprise SSD Backup Power Reference Design.” The schematics to this reference design can be found here.

 

Figure 5. PMP30046 enterprise SSD backup power reference design board. Image courtesy of TI.com

 

Another option to consider is the MP5505A from MPS. This backup energy IC is characterized, according to its datasheet, as a “lossless energy storage and management unit targeted at the solid-state and hard-disk drive applications.” This 3mm x 4mm QFN-20 IC has a wide input operating voltage range from 2.7V to 7V and is capable of charging the energy storage capacitors up to a voltage (VSTRG) of 35V.

 

Figure 6. MP5505A typical application. Image taken from the datasheet.

 

In Conclusion

When it comes to memory applications, specifically the use of NAND flash in SSDs, energy hold-up can be a critical part of the design. And when designing a power hold-up/energy storage management system, it’s important to consider which capacitor(s) to use—of which depends up on the environmental conditions of the final product—as well as choosing the best energy-charging and energy-release device (IC). Although only three such options were listed in this article, it would behoove you to perform due diligence on other such devices that are available.

 

Comments

0 Comments