I understand that real power loss protection are capacitors on the SSD.
It depends on the terminology, but every SSD ought to have some kind of protection against FTL corruption, or the entire SSD will be 'bricked' resulting in total data loss and a non-functional drive which can only be resolved by perfoming a HPA secure erase procedure. Such a procedure wipes the mapping tables (FTL) and will reset the SSD back to factory conditions.
The FTL or mapping tables are crucial since they store what is on the SSD. If these become corrupted, or inconsistent with what is truly stored on the SSD, the SSD will show corruption or become bricked.
The only exception to this, are SSDs prior to the Intel X25-M such as MTron and MemoRight, which do not have an FTL to begin with. Without an FTL, there is no chance of FTL corruption because it does not exist. But the downside is that these SSDs are slower than HDDs in practice due to very slow random writes.
Then I've read on this forum that Samsung use some kind of firmware level power loss protection that works well when the SSD is used in a normal desktop computer (not in raid). Is this correct and do all Samsung drives have this kind of power loss protection?
That is correct according to my information. But i must stress that i do not work for any SSD company so i have to make due with public information. But it is well known the Samsungs use a technique known as 'journaling' to prevent their FTL from becoming corrupted. This means that upon unexpected power-loss, the SSD reverts to an earlier state. This violates the FLUSH CACHE command and as such Samsung SSDs are not suitable to be used in RAID configurations or other complex storage. At least not where data consistency is a concern.
But what about other SSDs? Are they unprotected or do all modern SSDs have some kind of protection.
Some consumer-grade SSDs have capacitors (Crucial MX100/MX200/M500/M550, Intel 320, S3500, S3700) but most do without. Some may use journaling and some are unprotected, which is why they fail more often.
I don't know if I should consider power loss protection when I decide on my SSD setup? I don't wanna risk corrupting the file system in case of power loss.
If data consistency is a concern for you, Intel is your best bet. Particularly the Intel 320 is the best and most reliable consumer-grade SSD ever produced. It can still be bought at some places, even though it is quite old. Crucial MX200 is my SSD of choice, since it sports DWA - a technique that allows the SSD to gain the performance benefits of SLC NAND instead of MLC NAND, as long as you do not fill the drive more than 50% of its capacity. Only the 250GB SATA version and the 250/500GB M.2 versions of the MX200 support this feature.
A friend of mine, whom I had sold a PC, with an OCZ Vertex Plus R2 SSD, lost the SSD completely when the power glitched during a storm.
That is normal for an OCZ SSD - some of their products had failure rates higher than 50%.
Well, then, your only option is to purchase a UPS so that don't happen.
That is a fallacy. With a UPS you are not protected. You are protected against power failures but you are not protected against unexpected power-loss. The difference is that all loss of power without a prior STANDBY IMMEDIATE command, counts as unexpected power-loss. This includes pressing the reset button, restarting from the BIOS, crashes/bluescreens and even tidy shutdowns are known to cause this by bad power-off timing.
You can read the SMART information and look for Unexpected Power-Loss. The raw value indicates the amount of times the SSD lost power prior to receiving a STANDBY IMMEDIATE command. This may be 40 or even 80 times for a casual desktop system. This does not mean such a user had 80 power failures. In western countries you have a power failure maybe once per every few years or so.
So the idea that a UPS makes capacitors or other form of power-loss protection superfluous, is simply not true.