3 Responses to “SDHC Class 6 lifetime”

1. 

   Ilya Elenik
   Jan 08, 2009
   

   SDHC manufacturers promise at least 10000 rewrite cycles. eBoostr updating cache every hour, so it should work at least ~500 days continuously. It should work much more if consider the builds canceled due to non idle computer or the time when computer powered off.
   But this is only a theory

   
   
2. 

   freshmetal
   Jan 08, 2009
   

   Ilya san
   Thanks for the response!
   Maybe the way of using as a cache memory was way out of manufacturers’ thought. I never thought using a memory card as a RAM
   If I use Toshiba or Panasonic or high end Sandisk, maybe it wont die in short term but its way too expensive. I can buy 7 Silicon Power SD cards in same price…

   
   
3. 

   kosimov
   Jan 09, 2009
   

   Here are some lengthy ramblings of my old brain which came to me when I saw your message here….

   I am a hardware/software design engineer (BSEE). I have designed with Flash memory (the kind used in USB thumb drives such as eboostr uses). I have also designed memory chips to meet published specs (though that was decades ago!). I have known of many cases where a chip’s specs were published well in advance of the design of the chip even BEGINNING!

   (here is a real-world true example which I experienced: I would be told to begin designing a chip and, when I would ask for the specs, I would be given the data book for chip, which had been published and in circulation for, sometimes, a year or two! The specs were published, and some designers had used the chip in new designs based on the published specs, before the design of the chip was even started!)

   Even though the designer will finally make the chip meet the published specifications, the designer may run into some problems meeting the spec absolutely, and there may be one or two of the specs which are met – barely – but that may also introduce “quirks” which are not mentioned in the literature which later designers refer to when designing products using the chips. And so on, etc etc etc……there are quite a few other factors which can make the absolute maximums in published specifications for a chip a little too optimistic, or which introduce traits into the chip which will show up some time later when they are designed into a special situation in which that trait becomes important, even though it is not a published spec, …. etc etc

   I have learned to take published specs “with a grain of salt”, and I am very careful to stay well clear of maximum specified limits. For example, in this case, even though the chip company “guarantees” that the flash memory will work for a minimum of 10,000 cycles, there is no way to measure that when the chips are tested, because it would take much too long, and, more importantly, it would “wear out” the chip; testing it 10,000 times would either cause it to fail at 10,001, or at least, to be close to failure when it was delivered to the customer. The chip is probably designed to work well beyond 10,000 cycles, so the company is comfortable claiming it will go at least that far.

   So, the specifications are a kind of “ball park” number used to “indicate” that, when testing a large number of prototypes of the chip, that number was achieved regularly.

   Flash chips are more critical in this area than are some other chips, as they are a type of ROM which is electrically erasable (in fact, they were originally called “EEPROM”s, or “Electrically Erasable Programmable Read Only Memory”. They are meant mostly to be read many times, and written to from time to time, but were not necessarily intended to be used as RAM. This is because the erasing process uses a voltage which is a little stressful for the device, compared to the voltage/power used to operate the device for repeated reading of its contents, which is of course much less stressful.

   Eboostr takes advantage of the availability and low price of large flash memory devices and uses them as “RAM”, (Random Access Memory), which is written to and read from a lot more than is the norm for flash memory. I think it is a great idea, very clever, and I was happy to see it work so well. I don’t have very up to date computers and anything which will make them perform better is OK with me! I am doing some design work, after which I hope I can afford new computers, but for now, eboostr and a couple other add-ons are the best I can do.

   However, I was using a 2 GB SD card, one of the better brands, as the eboostr memory, and it recently failed also. I was having problems getting all 2 GB of memory to show so eboostr could use it, then, it died completely. I noticed it was quite warm when I took it out of the USB card reader. I thought at the time that it must have been ready to fail anyway, and this failure was merely a coincidence. Now that I see your message, I begin to wonder, though. I wonder if anyone has done any testing to see how much the temperature of the chip rises if it is used with eboostr? Shouldn’t be too hard to do, but unfortunately, I now don’t have anything I can test with. I will replace the failed memory with a “thumb” drive and do some testing when I have time, but that won’t be for a week or more. In the meantime I am only using system memory with eboostr, so that is safe, I suppose. If it turns out that eboostr does cause early failure of USB drives, I imagine it would not be too much trouble to design one which functions properly with eboostr but uses regular RAM instead of flash ROM. I would do that if the problem turns out to be real.

   I apologize for the very long message, but I thought I should make it detailed and establish that I know what I am talking about, and for me at least, that required the preceding details. If I learn anything more about this “problem” (which would first be to discover if eboostr causes failures in flash memory, then, why), I will let you and the others here know what I find. It would be a shame if eboostr did have such a problem as it is a great product idea, but again, there are ways around it.

   One other thing just popped into my mind: if Windows uses the USB flash drives with Ready boost (I don’t know how well that works as I have never used it but I assume it works OK), I would think that they would have examined this issue and would not have supplied software which uses the flash memory in the same manner as eboostr, if there was any chance it would destroy the memory chips. Hmmmm… well, I guess I will go ahead and test when I can, and see what I learn. My previous experience with flash ROM in communication systems, computer main boards, etc., which held the BIOS or operating system for those products, will come in handy. Now I have to figure out what chips they are using and then ……… (continued next time, same time, same channel)