Tired of hard drive Crashes!! Use Flash instead!!

[Slick27]

Have successfully fired up a saber 750 with a A2100 control using flash memory module, which replaces the IDE hard drive. Ran all day yesterday without a problem. Have a edm tech that has been using these modules in edm machines that were running hard drives and has had no problem in over a years time.

No moving parts or spinning platters at 3600 rpm waiting for the heads to crash and everything lost when you least expect it!! It will costapprox $50.00
And you can do it yourself!

Let me know if you need more info

Mark

[scudzuki]

Do yourself a favor and make a few copies of that memory card.
Flash memory has a limited number of write cycles before it becomes unreliable and eventually fails entirely.

It is also not infallible. I have seen several types of flash memory fail prematurely; usb flash drives and various memory cards.

Joe

[maver1ck]

I looked into the flash memory option years ago, and decided to stick with the hard drives when I found out how quickly the millions of read/write cycles the flash memory are rated for get used up on the A2100.

As scudzuki suggested, make extra copies or better yet, keep the original hard drive in a safe place so you will be able to make additional copies as you need them. Flash memory is an inexpensive alternative, but from everyone I've known using it with the A2100, you need to plan for the failure because it will happen eventually (although I've made the same comment about the hard drives on the A2100).

And I'll add one more note since these hard drives have a bit of a bad reputation. I work on machines with the A2100 that are 15 years old, still running production every day on the original hard drive. I doubt anyone will still be running on the same flash memory in 15 years.

From my experience, these hard drives rarely fail mechanically but are susceptible to having the data on them corrupted. This can happen from bad workstation boards, bad power supplies and from power outages or operators that cut off control power while the hard drive is writing critical data.

[Mr Bonavita]

Mark I need that info, we are trying to recover a hard drive out of an Arrow 1000
ACRAMATIC 2100.

We'll use the flash memory as a back up. The hard drive that crashed has been working in the machine for 14 years, so we will replace it with another hard drive.

By the way any help or information on how to recover the hard drive will be greatly appreciated.

[mike_Kilroy]

Have successfully fired up a saber 750 with a A2100 control using flash memory module.......Let me know if you need more info
Mark

Mark, sounds like u r willing to share the how to? If so, can u post it here? I am sure a lot of folks would like to know how to do this just in case, as well as perhaps someone can expand on it to make it a 2nd device for real backup in case of a HD failure or something....

thanks

[PinnacleMachine]

I ran flash for a while but was finding that I could only get 4 months out of a card, so I went back to a HD. I think I'll look into a solid state HD instead.

[mike_Kilroy]

good info! 4 months.... from this thread I searched about flash and SS HDs and think I convinced myself they arre one and the same! the SS HD uses flash just like other flash...... read somewhere like 80,000 writes typical.... also read there are 2-3 levels of quality of the flash chips; the cheap ones like we all buy for usb sticks and then ones that cost more to mfgr but can last 3-10x longer maybe? fuzzy to me how one would figure out what they were buying unless the specs for the HD state typical or min no of writes.....if you learn more, share please!

[swarf_rat]

OK, an old thread but I have been looking at machines with the Acramatic so I'm looking at this for the first time.

Flash memory is a type of structure called NAND flash, referring to the logic used in the bit cell. This type of memory has a limited number of writes before failures occur, like most electrically alterable nonvolatile memory. However, NAND flash has an additional and unique failure mode: If a cell is read too many times, the data is destroyed (a "read-many" failure). Most other NV ram cells do not have this failure mode.

To deal with it, the chips themselves have error correction built in, along with a reporting mechansim. A device that is designed to use them properly will look at the error data and if there are too many errors being reported in a certain block, will attempt to re-write the block to restore the data. If the re-written data does not verify on read, then the block is spared in the directory and the data written elsewhere. A device that is not designed to use them will not have some or any of this correction mechanism.

Now, hard disk drives have many of the same issues with errors and sparing of blocks. Both hard disks and NAND flash are block sectored devices. But a hard disk does not have a known read-many failure mode. In theory a sector can be read an infinite number of times without destruction.

The reason this is material is that data (programs, the os, boot blocks) is read many more times than it is written typically. Orders of magnitude more times.

Notebooks and other portable devices designed around NAND flash use programming techniques to get around these problems, like varying where things are written over time, checking the errors reported, etc. An old NT OS is not going to do that. So it may be a bad candidate for a flash drive, unless the drive is considerably more intelligent than they seem to be.

[scudzuki]

Actually NTFS (the default file system on Windows NT) uses TTS.
It hotfixes the drive when it detects bad clusters, moving the data from failing allocation units, then reading the written data before commiting the write. It adds a bit of overhead to the FS vs lets say FAT but the data integrity is worth it.

Lets face it, hard drives were never meant to last multiple decades.

Joe