Searching \ for 'IDE or Memory Card use for high speed data storag' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/mems.htm?key=data
Search entire site for: 'IDE or Memory Card use for high speed data storag'.

Truncated match.
PICList Thread
'IDE or Memory Card use for high speed data storag'
1999\07\21@183035 by William K. Borsum

flavicon
face
Greetings:
I've been following the PIC to FDC conversations a bit, and have a related
question:

I have an application where I need to log data at the rate of up to 200
K-bytes per second, to a depth of 120-160 Mega-bytes.
Part of the problem is that it is memory is broken up into a series of ring
buffers that are sequentially  over-written till an event occurs, so flash
or eeProm with a short finite life--or holes in the memory because of bad
blocks--won't work.  Only viable alternative I've come up with is sRam, but
the largest chips only appear to be 512Kx8, which means 200+ chips to get
the total memory I need.  Oh yes--environment sees shock and vibration to
100+ G's, and the whole thing needs to fit in a can about 3.5" in diameter.

Question: does anyone have a sneaky alternative that will work--like a
micro-hard drive that will survive?  If so, how would the interface work?
Has anyone implemented an IDE interface, or interface to a sRAM PCMCIA card
on a PIC?

Kelly
****************************************************************************
********
All legitimate attachments to this email will be clearly identified in the
text.
William K. Borsum, P.E.
OEM Dataloggers and Instrumentation Systems
<spam_OUTborsumTakeThisOuTspamdascor.com> & <http://www.dascor.com>

1999\07\21@211218 by Mark Willis

flavicon
face
William K. Borsum wrote:
{Quote hidden}

 sRam in PCMCIA form is *hideously* expensive, about $133 for a 4Mb
card, and you don't often SEE larger SRam cards than 4Mb.  40 of those
wouldn't be a good answer for you, I'd assume!

 Now a laptop (or 1.7" PCMCIA) HDD, that's a possibly do-able answer,
and more like $15-$20 for a 200Mb or so laptop HDD.  About 1 year
service life (if you leave the drive motor spinning constantly), shock &
vibration are a potential problem but these drives do handle some decent
forces (I suspect if you sprung it inside the can, with a little added
mass & some air damping or something, you could make it work, MAYBE -
depends on the amplitude of the vibrations.)

 Also, PCMCIA Flash Cards might be do-able (Sandisk makes 175Mb and
larger cards), except they'd get worn out by your application, right?
Same for laptop Flash drives (SanDisk makes these, act just the same as
regular laptop drives but they wear out after lots of writes.)

 Have you considered a set of battery-backed 72-pin SIMMs, or one or
two DIMM memory modules?  Seems to me that using that setup may be a
really good answer for you here (The storage media's fairly cheap, data
bandwith's NOT a problem, etc.)  Could sample the data to a PCMCIA drive
for downloading, or some such, perhaps.  That might be good (you'll want
a GOOD socket that's not going to vibrate loose, though these types of
sockets should be OK, I'd think - maybe tape the metal latches in place
to prevent any possible shaking loose of the RAM PCB's.)  I'd wonder if
DIMMs would stay in place well, here, have to help them stay in place
I'd think <G>

 Mark

1999\07\21@213306 by Wagner Lipnharski

picon face
That's the solution Mark, four 64MB memory cards can give him 240MB
easily, with a good and speedy interface, good support mechanics (why
not consider micro soldering?), and a nice battery system, you can have
close to one GigaByte inside that CAN space you have.
Wagner.

Mark Willis wrote:
[snip]
{Quote hidden}

1999\07\21@213322 by Thomas Brandon

flavicon
picon face
First of all, good luck, I think you'll need it.

Not sure on sRAM but I don't think any moving parts based hard disk is gonna
cut it. For one there's the 100G the other thing is that PC hard disks (the
more shock resistant laptop drives may be better) don't like running when
they're not level (or at least used too). I'm not sure how sensitive the
latest drives are (of course you won't find a recent drive <2Gb) but not
long ago as little as 5 degrees of slope could screw an IDE drive.

Unless 100G is a misprint I think about the only option is solid state.
Sandisk (http://www.sandisk.com) make a few solid state mini storage products.

They have solid state hard drives ranging from 32Mb-440Mb in 2.5" form
factor. They also have compact flash cards the size of a matchbook in from
4-96Mb. Both are PCMCIA and ATA compliant and

However they are probably not suitable due to the following:
1) Rewrites
       Both products list data reliability as < 1 non-recoverable error in
10 to the 14 bits read. On a 160Mb drive this would allow approx. 77000
rewrites before a single bit failure is expected. Along with this is the
problem of bad blocks however I believe they have implemented some sort of
hardware bad block table (can't really remeber, check the specs, in any case
I've seen info on dealing with the bad blocks and it didn't look too
hideously hard. Also, I think this was mainly to do with bad blocks present
at manufacturing).

2) Vibration
       You mention the product will see 100+G of shock and vibration. This
could be a big problem. Keep in mind these are both completely solid state
so unless there poorly manufactured they *should* be typical of other solid
state devices. In terms of shock, you're fine. They are rated for 2000G max.
of shock while operating (exactly the same if not operating suprise
surprise). However they are only rated at 15G peak to peak max. vibration
and so you may have problems here. Granted this could be due to their
manufacturing. But If a solid state flash drive can only take 15G then I
can't see why a solid state SRAM will take significantly more.

{Original Message removed}

1999\07\22@003533 by Steven Keller

flavicon
face
Check out http://www.m-sys.com

Steve


-----Original Message-----
From:   pic microcontroller discussion list [.....PICLISTKILLspamspam@spam@MITVMA.MIT.EDU] On
Behalf Of William K. Borsum
Sent:   Wednesday, July 21, 1999 5:28 PM
To:     PICLISTspamKILLspamMITVMA.MIT.EDU
Subject:        IDE or Memory Card  use for high speed data storage

Greetings:
I've been following the PIC to FDC conversations a bit, and have a related
question:

I have an application where I need to log data at the rate of up to 200
K-bytes per second, to a depth of 120-160 Mega-bytes.
Part of the problem is that it is memory is broken up into a series of ring
buffers that are sequentially  over-written till an event occurs, so flash
or eeProm with a short finite life--or holes in the memory because of bad
blocks--won't work.  Only viable alternative I've come up with is sRam, but
the largest chips only appear to be 512Kx8, which means 200+ chips to get
the total memory I need.  Oh yes--environment sees shock and vibration to
100+ G's, and the whole thing needs to fit in a can about 3.5" in diameter.

Question: does anyone have a sneaky alternative that will work--like a
micro-hard drive that will survive?  If so, how would the interface work?
Has anyone implemented an IDE interface, or interface to a sRAM PCMCIA card
on a PIC?

Kelly
****************************************************************************
********
All legitimate attachments to this email will be clearly identified in the
text.
William K. Borsum, P.E.
OEM Dataloggers and Instrumentation Systems
<.....borsumKILLspamspam.....dascor.com> & <http://www.dascor.com>

1999\07\22@014928 by William K. Borsum

flavicon
face
At 09:21 PM 7/21/99 -0400, you wrote:
>That's the solution Mark, four 64MB memory cards can give him 240MB
>easily, with a good and speedy interface, good support mechanics (why
>not consider micro soldering?), and a nice battery system, you can have
>close to one GigaByte inside that CAN space you have.
>Wagner.
>
>Mark Willis wrote:
>[snip]
>>   Have you considered a set of battery-backed 72-pin SIMMs, or one or
>> two DIMM memory modules?  Seems to me that using that setup may be a
>> really good answer for you here (The storage media's fairly cheap, data
>> bandwith's NOT a problem, etc.)  Could sample the data to a PCMCIA drive
>> for downloading, or some such, perhaps.  That might be good (you'll want
>> a GOOD socket that's not going to vibrate loose, though these types of
>> sockets should be OK, I'd think - maybe tape the metal latches in place
>> to prevent any possible shaking loose of the RAM PCB's.)  I'd wonder if
>> DIMMs would stay in place well, here, have to help them stay in place
>> I'd think <G>

What is power consumption for a 64M dRAM SIMM??  or are sRAM SIMM's
available in equally large sizes?
Is there a hardware solution to refreshing dRAM? Totally transparent to the
PIC?
I want to deal only with data and address lines--i.e. set the address and
write the data.
Kelly

****************************************************************************
********
All legitimate attachments to this email will be clearly identified in the
text.
William K. Borsum, P.E.
OEM Dataloggers and Instrumentation Systems
<EraseMEborsumspam_OUTspamTakeThisOuTdascor.com> & <http://www.dascor.com>

1999\07\22@014942 by William K. Borsum

flavicon
face
At 06:10 PM 7/21/99 -0700, you wrote:
{Quote hidden}

Flash in any form will not work for the ring-buffer applications--just gets
over-written too many times too quickly.  Also, don't want to deal with the
holes, memory management issues, etc.

I would like to hear from anyone with mini-hdd experience.  Yes, I can
isolate the drive from shock and vibration to a certain extent--don't know
the profile yet.  Also, IDE interface management issues--has anyone done a
PIC-IDE interface and is the code available?

Simm's and Dimm's--possible--but are they available in sRAM?  If so, from
whom? and in reasonable chunks that can be built up into the 100+ Megabyte
range?

Again, the problems associated with normal dRAM and its power consumption,
refresh, etc. may be too much for this project.  The data is very costly to
acquire and any risk of loss is too much.  I also have volume and weight
limits on the batteries--but would consider it if there were dRAM
controllers that could handle all of the memory management functions in
hardware and be totally transparent to the the Pic. Anyone have any
experience along these lines??

Kelly



****************************************************************************
********
All legitimate attachments to this email will be clearly identified in the
text.
William K. Borsum, P.E.
OEM Dataloggers and Instrumentation Systems
<borsumspamspam_OUTdascor.com> & <http://www.dascor.com>

1999\07\22@020719 by Wagner Lipnharski

picon face
I believe that you would need another microcontroller to do the memory
management, while the PIC would only do the data gattering and transfer
it to the second processor.  Power will be an issue, doesn't matter if
you use hard disk or pure silicon storage... memory chips *always*
consume great share of energy.

You are not opening much information, but +100G is only obtained in
certain environments and one of them is at pure bright sun, so if it is
the case why not use a nice plastic flexible solar cell to keep battery
size reduced?  Also, a secondary memory system, to store the valid data
on flash, while the full-time data collecting memory system, you can
call it "cache", would be a DRAM SIMM, smaller, less power.

Also, remember radio data link to download your data... GigaHertz is
here to flow wideband in highspeed, a transmitter and battery could be
smaller and light weight.

--------------------------------------------------------
Wagner Lipnharski - UST Research Inc. - Orlando, Florida
Forum and microcontroller web site:  http://www.ustr.net
Microcontrollers Survey:  http://www.ustr.net/tellme.htm

1999\07\22@053818 by Steve Lawther

flavicon
face
Kelly,

I have to admit that 64Mb DRAM chips is probably your best bet.
for the ring buffers,
(SRAM would be better if the ring buffers were small)

Then after trigger, dump data straight to a PCCard flash disk,
(or Compact Flash if 96MB were enough / you could use compression)

Then when complete logging, dump the ring buffer to flash disk,


http://www.sandisk.com/oem/compact_specs.htm

(they're spec'ed to 2000G - thou you need to lock it into the connector!!)

sadly this would mean you have to bite the bullet on:-
1)  DRAM refresh (although a ring buffer would by definition be just
an address increment which it is easy to generate refreshes)
2)  Bad sectors on the flash disk,
(store a note of the bad sectors in RAM on power on.)
3)  DRAM power consumption, although at 200K/sec, this will be
low (<10mA) if you use the low power, slow refresh versions.

any more help, just shout!

Steve Lawther




--------------------
DISCLAIMER

This email and any attachments are confidential and intended
solely for the use of the individual or entity to whom they are
addressed. Any views or opinions expressed are solely those of
the author and do not necessarily represent those of WPI Husky
Technology. If you are not the intended recipient, be advised
that you have received this email in error and that any use,
dissemination, forwarding, printing, or copying of this mail is
strictly prohibited.

1999\07\22@171822 by Barry Baldwin

flavicon
face
Hi there,

I just returned from and an Embedded Solutions conference and I picked
up some literature on a device that
might be what you're looking for.  It's called chipDisk and it's an IDE
compatible Flash-Disk.  The largest
one they specify on their literature is 80 MB so maybe you could use
two.  It states that it is absolutely
insensitive to Percussion, vibration, concussion, temperature variation
and magnetic fields.  Their web address
is http://www.jumptec.de.

SanDisk also makes flash disks that have an IDE interface.

Anyway hope this helps,

Barry

{Original Message removed}

More... (looser matching)
- Last day of these posts
- In 1999 , 2000 only
- Today
- New search...