> Anyone remember when you could make a crude image sensor by ripping the > lid off a D-RAM chip?
> -- Jim L.
Yeah but I have a feeling it was SRAM rather than DRAM (man desperately trying to remember what an NMOS static ram design looks like... Failing miserably... ho hum off to google to find out)
From article <992441379.19026.0.nnrp-12.3e31f...@news.demon.co.uk>, by "Peter Ibbotson" <d...@harvestme.co.uk>:
> "Jim Large" <jim.la...@foo.boo> wrote in message > news:3B26087F.32AAF0D@foo.boo... >> Anyone remember when you could make a crude image sensor by ripping the >> lid off a D-RAM chip? > Yeah but I have a feeling it was SRAM rather than DRAM ...
No, it was DRAM. SRAM is insensitive to light with the lid off, unless the light is awfully bright.
DRAM, on the other hand, stores data capacitively. This is why such memory requires refreshing, and the thing to note is that the refresh rate required to force the memory to retain its information depends on the leakage rate. The leakage rate, in turn, depends on the light level, which is zero inside the properly closed chip, but is definitely nonzero if you rip off the lid and focus an image on the chip.
So, to use the DRAM chip as an image sensor, write all 1's (or all 0's, depending on how many inverters stand between you and the chip), and the wait a bit and see which bits are still 1 (or 0). The bits that changed were in brighter light than those that didn't. Do this once and you get one bit per pixel. Do this again, with a different waiting time between writing the 1 and reading it back, and you'll get more information about the light level. Do it enough times (with different waits between write and read) and you get full greyscale info for each pixel.
Most DRAM chips automatically refresh the bits you read, so you can't just write the 1's once and then read again and again to see when the 1 decays to 0. The reason they do the refresh on read is that the actual read operation discharges the capacitor enough that it's effectively erased by being read -- in that regard, DRAM is like core memory. Doug Jones jo...@cs.uiowa.edu
> No, it was DRAM. SRAM is insensitive to light with the lid off, unless > the light is awfully bright.
I'll have to go away and do some research on this. Please note I'm not saying that you're wrong just that I need to go and drag out some of my old chip design notes (most of my actual design work was CMOS based and I have a feeling that only NMOS worked for this) and refresh my tired old gray cells (along with looking up how CCDs work etc). More search engine work for me then.
> "Douglas W. Jones,201H MLH,3193350740,3193382879" <jo...@cs.uiowa.edu> wrote > in message news:9g7v00$bvm$1@flood.weeg.uiowa.edu... > > No, it was DRAM. SRAM is insensitive to light with the lid off, unless > > the light is awfully bright. > I'll have to go away and do some research on this. > Please note I'm not saying that you're wrong just that I need to go and drag > out some of my old chip design notes (most of my actual design work was CMOS > based and I have a feeling that only NMOS worked for this) and refresh my > tired old gray cells (along with looking up how CCDs work etc). More search > engine work for me then.
: > Anyone remember when you could make a crude image sensor by ripping the : > lid off a D-RAM chip? : > : > -- Jim L. : >
: Yeah but I have a feeling it was SRAM rather than DRAM (man desperately : trying to remember what an NMOS static ram design looks like... Failing : miserably... ho hum off to google to find out)
No, it was DRAM.
The operating principle was that light falling on a DRAM cell caused the capacitor to discharge faster. So you wrote all '1's to the DRAM and then read the chip back a few ms later. The bits that read as 0s corresponded to cells that had light falling on them.
You could vary the sensitivity a bit by changing the delay between writing the '1's and reading out the chip.
The problem is that readout is destructive on DRAM. Most DRAM chips refresh the cell being read, which is what you want for data storage, but in the case of the 'image sensor' you effectively started a new exposure. You couldn't make multiple reads to see which cells had more or less light falling on them.
What was generally done was that several 'exposures' were taken. You wrote all 1's, waited a bit and read out the chip. The cells that were 0's were the ones that were brightly illuminated. You then wrote all 1's again, waited a bit longer, and read out the chip again. The extra 0's came from cells that were illuminated but not brightly enough to discharge in the shorted time. Do it 8 or 16 times and get a 3 or 4 bit greyscale image.
Obviously it only worked for static subjects.
Another problem was that there was address decoding logic, etc, down the middle of the chip. So you had 2 light-sensitive regions with an insensitive band between them. The resulting images had the middle missing :-).
Jim Large (jim.la...@foo.boo) wrote: : Steve O'Hara-Smith wrote:
: > : > Ceramic pack chips could be split along the seam to reveal a fully : > functioning chip (with about a 70% success rate after practice). [...]
: Anyone remember when you could make a crude image sensor by ripping the : lid off a D-RAM chip?
Anyone else remember the IS32 'Optic Ram' which, IIRC was a 64K bit DRAM chip that somebody had flipped the lid off and stuck a (quartz?) window in its place.
Used in cheap(ish) 'digital' cameras in the mid 1980s.
> "Douglas W. Jones,201H MLH,3193350740,3193382879" <jo...@cs.uiowa.edu> wrote > in message news:9g7v00$bvm$1@flood.weeg.uiowa.edu... > > No, it was DRAM. SRAM is insensitive to light with the lid off, unless > > the light is awfully bright. > I'll have to go away and do some research on this. > Please note I'm not saying that you're wrong just that I need to go and drag > out some of my old chip design notes (most of my actual design work was CMOS > based and I have a feeling that only NMOS worked for this) and refresh my > tired old gray cells (along with looking up how CCDs work etc). More search > engine work for me then.
CCDs are basically analog devices, usually organized as shift registers that shift analog quantities along. -- Chuck F (cbfalco...@my-deja.com) (cbfalco...@XXXXworldnet.att.net) http://www.qwikpages.com/backstreets/cbfalconer :=(down for now) (Remove "NOSPAM." from reply address. my-deja works unmodified) mailto:u...@ftc.gov (for spambots to harvest)
In <9g5au0$k6...@ausnews.austin.ibm.com>, on 06/12/01 at 03:02 PM, gla...@glass2.lexington.ibm.com said:
>CPUID = FF64072074700000 >[1] Ok, for those not in the know,
And IIRC for a few of us, the leading FF indicates that it is running a form of VM.
-- Julian Thomas: jt . jt-mj @ net http://jt-mj.net remove letter a for email (or switch . and @) In the beautiful Finger Lakes Wine Country of New York State! Boardmember of POSSI.org - Phoenix OS/2 Society, Inc http://www.possi.org -- -- Windows: an Unrecoverable Acquisition Error!
| Julian Thomas wrote: |> gla...@glass2.lexington.ibm.com wrote:
| | >CPUID = FF64072074700000 | | >[1] Ok, for those not in the know, | | And IIRC for a few of us, the leading FF indicates that it is running a | form of VM.
Yes, and also the ACP (and I presume, TPF) hypervisor does this also.
> : > Anyone remember when you could make a crude image sensor by ripping the > : > lid off a D-RAM chip? > : > > : > -- Jim L. > : >
> : Yeah but I have a feeling it was SRAM rather than DRAM (man desperately > : trying to remember what an NMOS static ram design looks like... Failing > : miserably... ho hum off to google to find out)
> No, it was DRAM.
> The operating principle was that light falling on a DRAM cell caused the > capacitor to discharge faster. So you wrote all '1's to the DRAM and then > read the chip back a few ms later. The bits that read as 0s corresponded > to cells that had light falling on them.
What's really tough is to rip the top off of a core memory plane and then use it as a camera. It's a close trade-off between an image that is bright enough to pass the curie point of the cores, and the melting point of the copper wiring in the cores. Also there is the problem of test subjects either wilting, getting sun burned, or their clothing catching fire during the exposure, because of the bright lighting required.
Somewhat back on topic.
( Note: the catching fire part was not so bad as it sounds, for it often enhanced the photograph's composition. ) -- not authorised for sale on a stand alone basis /\>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>\/ /\ I may be demented \/ /\ but I'm not crazy! \/ /\<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<\/ * SPAyM trap: there is no X in my address * || attatch FLAME here || \/ \/ X
a...@p850ug1.demon.co.uk (Tony Duell) writes: > Anyone else remember the IS32 'Optic Ram' which, IIRC was a 64K bit DRAM > chip that somebody had flipped the lid off and stuck a (quartz?) window > in its place.
No one "had flipped the lid off". It was made that way. I doubt that the window was quartz, as there wasn't any motivation to pass UV light.
>In <9g5au0$k6...@ausnews.austin.ibm.com>, on 06/12/01 > at 03:02 PM, gla...@glass2.lexington.ibm.com said:
>>CPUID = FF64072074700000
>>[1] Ok, for those not in the know,
>And IIRC for a few of us, the leading FF indicates that it is running a >form of VM.
>-- > Julian Thomas: jt . jt-mj @ net http://jt-mj.net > remove letter a for email (or switch . and @) > In the beautiful Finger Lakes Wine Country of New York State! > Boardmember of POSSI.org - Phoenix OS/2 Society, Inc >http://www.possi.org > -- -- > Windows: an Unrecoverable Acquisition Error!
You are, of course, correct. I'm running VM/ESA-370 Feature on it. You can find the real CPU ID via one of the Diagnose instructions (Diagnose X'00'?). The real CPU ID is:
0164072074700000
The leading byte indicates the particular submodel for those machines which have different submodels.
Dave
P.S. Standard Disclaimer: I work for them, but I don't speak for them.
CBFalconer <cbfalco...@my-deja.com> wrote: > CCDs are basically analog devices, usually organized as shift > registers that shift analog quantities along.
Yep, I first came across them used as analogue delay lines (less noisy and better quality than the spring and transducer method) and commonly referred to as bucket brigade devices. One friend at the time liked to describe them as working by lining up the electrons one by one and counting them through a slight exaggeration but a nice image.
-- Directable Mirrors - A Better Way To Focus The Sun
freddy1X <fredd...@indyx.net> wrote: > What's really tough is to rip the top off of a core memory plane and > then use it as a camera. It's a close trade-off between an image that > is bright enough to pass the curie point of the cores, and the melting
Hmm, core plane plus Mu metal box = crude magnetic field camera ?
-- Directable Mirrors - A Better Way To Focus The Sun
On Wed, 13 Jun 2001 16:25:25 +0100, Peter Ibbotson <d...@harvestme.co.uk> wrote: >Please note I'm not saying that you're wrong just that I need to go and drag >out some of my old chip design notes (most of my actual design work was CMOS >based and I have a feeling that only NMOS worked for this) and refresh my >tired old gray cells (along with looking up how CCDs work etc). More search >engine work for me then.
Someone sold DRAM parts with clear covers. There was a Circuit Cellar story in Byte about making a DRAM image sensor. The odd geometry (for imaging) of the memory array was a little limiting.
> a...@p850ug1.demon.co.uk (Tony Duell) writes: > > Anyone else remember the IS32 'Optic Ram' which, IIRC was a 64K bit DRAM > > chip that somebody had flipped the lid off and stuck a (quartz?) window > > in its place.
> No one "had flipped the lid off". It was made that way. I doubt that > the window was quartz, as there wasn't any motivation to pass UV light.
Yes and no. Someone discovered it worked by prying off the lid and wrote it up in the magazines. So many people were trying it afterward a manufacturer made a run of DRAMs with quartz windows (I think it was Micron in it's early days) because the quartz packaging was in place for EPROMS. Commercial DRAMs with windows would have been a nightmare, imagine program memory changing according to ambient lighting. As I recall solid state imaging sensors were being used then, the E G & G Reticon CCD imagers come to mind, but their cost was astronomical (and I think that was their primary use too, astronomical telescopes). Jack Peacock
On 12 Jun 2001 15:02:24 GMT, gla...@glass2.lexington.ibm.com
<gla...@glass2.lexington.ibm.com> wrote: >[1] Ok, for those not in the know, the 7470 CPU model corresponds to a P/370 >card, which is a S/370 on a card that plugs into a PS/2 and gives you a >full and complete S/370 (with a very minor exception or two, such as 4K storage >keys). Of course, it's been obsolete for quite some time, having been >replaced by the P/390 cards (full ESA/390 on a card), although I've heard a >rumor that those may have been withdrawn from marketing.
That "minor exception" keeps it from running MVS 3.8, at least natively. (I haven't tried it under VM on mine.)
The P/390s have indeed been withdrawn from marketing; IBM's approach to anything under 60 MIPS these days is emulation.
Once upon a time, Jack Peacock <peac...@simconv.com> said:
>Yes and no. Someone discovered it worked by prying off the lid and >wrote it up in the magazines. So many people were trying it afterward a
I remember an article and program in one of the Atari 8bit magazines (probably A.N.A.L.O.G. - they seemed to do more of the hardware stuff) for building a scanner out of a dot-matrix printer, a DRAM chip, and I guess some type of light source (or maybe you just used a good lamp). The RAM chip was hooked up via the joystick ports (which made a nice 8 bit bi-directional parallel interface) and the program would send escape codes to the printer to move the head (and mounted chip) across the page, reading the bits from the joystick port.
I never got around to setting it up myself however. -- Chris Adams <cmad...@hiwaay.net> Systems and Network Administrator - HiWAAY Internet Services I don't speak for anybody but myself - that's enough trouble.
> Well, no, and I was hoping that no one was going to ask, since by doing that, > I give away the secret[1]:
> CP Q CPUID > CPUID = FF64072074700000
> [1] Ok, for those not in the know, the 7470 CPU model corresponds to a P/370 > card, which is a S/370 on a card that plugs into a PS/2 and gives you a > full and complete S/370 (with a very minor exception or two, such as 4K storage > keys). Of course, it's been obsolete for quite some time, having been > replaced by the P/390 cards (full ESA/390 on a card), although I've heard a > rumor that those may have been withdrawn from marketing.
On 12 Jun 2001 15:02:24 GMT, gla...@glass2.lexington.ibm.com <gla...@glass2.lexington.ibm.com> wrote: [1] Ok, for those not in the know, the 7470 CPU model corresponds to a P/370 card, which is a S/370 on a card that plugs into a PS/2 and gives you a full and complete S/370 (with a very minor exception or two, such as 4K storage keys). Of course, it's been obsolete for quite some time, having been replaced by the P/390 cards (full ESA/390 on a card), although I've heard a rumor that those may have been withdrawn from marketing.
Jay Maynard wrote: > The P/390s have indeed been withdrawn from marketing; IBM's approach to > anything under 60 MIPS these days is emulation.
How does the performance of the P/370 and P/390 compare to a Hercules? -- / Lars Poulsen - http://www.cmc.com/lars - l...@cmc.com 125 South Ontare Road, Santa Barbara, CA 93105 - +1-805-569-5277
On Fri, 15 Jun 2001 00:29:48 -0700, Lars Poulsen <l...@cmc.com> wrote: >How does the performance of the P/370 and P/390 compare to a Hercules?
There are three versions of the P/390: Micro Channel and PCI P/390s, and a PCI P/390E. Hercules on a reasonably fast CPU (Pentium-class >800 MHz) will run at speed comparable to a P/390, and a multiprocessor host at 1 GHz or faster can approach P/390E speed.
JP> Yes and no. Someone discovered it worked by prying off the lid and JP> wrote it up in the magazines. So many people were trying it afterward a
Yep, exactly the way dry cell battery chargers got started. Some bright spark discovered that if you treated them right you could recharge zinc/carbon and alkaline batteries (which alkaline ???) rather than explode them. The design found it's way into electronics magazines years before they started to be manufactured. And I *still* sometimes see batteries with 'Do not recharge' warnings.
-- Directable Mirrors - A Better Way To Focus The Sun
|
