Forums

Force LCD monitor "white"/"black"

Started by Don Y July 10, 2014
On Fri, 11 Jul 2014 21:20:38 -0700, josephkk
<joseph_barrett@sbcglobal.net> wrote:

>On Thu, 10 Jul 2014 17:53:13 -0700, Don Y <this@is.not.me.com> wrote: > >>Hi George, >> >>On 7/10/2014 1:04 PM, George Herold wrote: >>> On Thursday, July 10, 2014 3:52:55 PM UTC-4, Don Y wrote: >> >>>> No, when I refer to having just the "panel", I mean "just the >>>> glass" >>> >>> Hi Don, Knowing almost nothing, I'd guess there are some different >>> technologies. I certainly remember a prof who did liquid Xtal stuff >>> back at Uni. He used an electric field which caused the crystals to >>> line up or not line up and this resulted in more or less scattering >>> of the light. >>> >>> So the short answer is maybe an electric field. Perhaps you can >>> "scuff up" some high voltage on a carpet and put your finger near an >>> lcd. >>> >>> (Why the question.. and why not start at wiki?) >> >>Sorry, I had that "basic" understanding. I've used "simple" >>LCD displays (single backplane) and very small (7x95) multiplexed >>glass. Driving the former is easy -- drive the backplane with >>an AC signal and invert the signal for any segments/indicators >>that you want to "light". The latter was much more complicated >>and involved multi (voltage) level drive to ensure the columns >>in the "selected" row were driven properly while NOT driving >>any of the dots in the unselected rows (which, obviously, saw the >>same column drives that were being presented to the selected row!). >> >>[IIRC, Hughes made some devices that did this.] >> >>My question was intended to elicit how much *larger* panels >>are driven -- presumably more complex than the small multiplexed >>glass I had used. > >Rather hellishly so. There are about 2 million full color pixels in a >standard HD display (1920 by 1080) so figure maybe 10 million connections >to the glass (unless there is a lot more smarts in the glass) minimum.
Huh? Do you really think there are 10M connections *to* the glass? There are, perhaps, on that order of connections _on_ the glass but that's an entirely different thing. The lines/rows are multiplexed. Yes, and modern panels have smarts on the glass.
>Of course if you want to drive something like that in an "on-off" mode a >fast AVR or PIC could do the job of creating the analog or DVI video >signal.
Not fast enough, without some accelerator hardware.
Den l=F8rdag den 12. juli 2014 15.08.17 UTC+2 skrev k...@attt.bizz:
> On Fri, 11 Jul 2014 21:20:38 -0700, josephkk >=20 > <joseph_barrett@sbcglobal.net> wrote: >=20 >=20 >=20 > >On Thu, 10 Jul 2014 17:53:13 -0700, Don Y <this@is.not.me.com> wrote: >=20 > > >=20 > >>Hi George, >=20 > >> >=20 > >>On 7/10/2014 1:04 PM, George Herold wrote: >=20 > >>> On Thursday, July 10, 2014 3:52:55 PM UTC-4, Don Y wrote: >=20 > >> >=20 > >>>> No, when I refer to having just the "panel", I mean "just the >=20 > >>>> glass" >=20 > >>> >=20 > >>> Hi Don, Knowing almost nothing, I'd guess there are some different >=20 > >>> technologies. I certainly remember a prof who did liquid Xtal stuff >=20 > >>> back at Uni. He used an electric field which caused the crystals to >=20 > >>> line up or not line up and this resulted in more or less scattering >=20 > >>> of the light. >=20 > >>> >=20 > >>> So the short answer is maybe an electric field. Perhaps you can >=20 > >>> "scuff up" some high voltage on a carpet and put your finger near an >=20 > >>> lcd. >=20 > >>> >=20 > >>> (Why the question.. and why not start at wiki?) >=20 > >> >=20 > >>Sorry, I had that "basic" understanding. I've used "simple" >=20 > >>LCD displays (single backplane) and very small (7x95) multiplexed >=20 > >>glass. Driving the former is easy -- drive the backplane with >=20 > >>an AC signal and invert the signal for any segments/indicators >=20 > >>that you want to "light". The latter was much more complicated >=20 > >>and involved multi (voltage) level drive to ensure the columns >=20 > >>in the "selected" row were driven properly while NOT driving >=20 > >>any of the dots in the unselected rows (which, obviously, saw the >=20 > >>same column drives that were being presented to the selected row!). >=20 > >> >=20 > >>[IIRC, Hughes made some devices that did this.] >=20 > >> >=20 > >>My question was intended to elicit how much *larger* panels >=20 > >>are driven -- presumably more complex than the small multiplexed >=20 > >>glass I had used. >=20 > > >=20 > >Rather hellishly so. There are about 2 million full color pixels in a >=20 > >standard HD display (1920 by 1080) so figure maybe 10 million connection=
s
>=20 > >to the glass (unless there is a lot more smarts in the glass) minimum. >=20 >=20 >=20 > Huh? Do you really think there are 10M connections *to* the glass? >=20 > There are, perhaps, on that order of connections _on_ the glass but >=20 > that's an entirely different thing. The lines/rows are multiplexed. >=20 > Yes, and modern panels have smarts on the glass. >=20 >=20 >=20 > >Of course if you want to drive something like that in an "on-off" mode a >=20 > >fast AVR or PIC could do the job of creating the analog or DVI video >=20 > >signal. >=20 >=20 >=20 > Not fast enough, without some accelerator hardware.
not fast enough for DVI,=20 but if he only wants all white/black he doesn't need to drive pixels all=20 he need is a clock ~25MHz, hsync/vsync and all the color bits tied high/low= =20 to get white/black an AVR/PIC etc. could do that -Lasse
On Thu, 10 Jul 2014 12:16:28 -0700, Don Y wrote:

> Hi, > > Anyone know how the drive to an LCD panel works? Specifically, how one > can force the display to appear "all white" or "all black" WITHOUT > relying on the electronics already present in the monitor? > > E.g., imagine having *just* the LCD panel and being able to operate it > as a "light gate" (full on, full off). > > Thx!
How about examining an auto-dark welders helmet?
On 7/11/2014 1:17 AM, Tauno Voipio wrote:
> On 11.7.14 03:56, Don Y wrote: >> On 7/10/2014 1:02 PM, Tauno Voipio wrote:
>>> You might be able to tie all rows together and all columns together, >>> and drive the rows and columns with antiphase square waves. The >>> hard part will be to find the proper voltage levels. >> >> I would suspect that would work on "dumb glass" (i.e., where the >> rows and columns were just electrodes screened on the glass). >> >> But, I thought larger panels had some active components *in* (on) >> the glass to effectively treat the display as a storage device >> that is "refreshed" with new data, periodically (?) > > For monochrome panels, the usual chip was KS0083. It is a 80 bit > shift register with LCD output level shifters.
This seems similar to the Hughes (?) devices I used ~1980 (though they were DIP40). IIRC, the output was "multivalued" -- not just a simple (2 level) translator -- to ensure DC stayed off the glass.
> For a 320 by 240 display, there were 4 KS0083's for columns and > 3 chips for rows. The data was fed serially to the column chips, > with the rows clocked once for each full row. I used a refresh > rate of 14400 row/s, giving 60 frames/s.
Yes, similar to the Hughes interface -- though I was driving a much smaller display (7x95). OTOH, a fair bit of overhead for a machine with a 2MB/s bus! But, how does this all relate to current technology? My understanding is that new panels are treated almost as a sort of write-only DRAM. The electrical interface serving to store a charge in each "accessed row's" column cell instead of just presenting an instantaneous voltage across each cell.
Hi Joseph,

On 7/11/2014 9:20 PM, josephkk wrote:

[attrs elided]

>>>> No, when I refer to having just the "panel", I mean "just the >>>> glass" >> >> Sorry, I had that "basic" understanding. I've used "simple" >> LCD displays (single backplane) and very small (7x95) multiplexed >> glass. Driving the former is easy -- drive the backplane with >> an AC signal and invert the signal for any segments/indicators >> that you want to "light". The latter was much more complicated >> and involved multi (voltage) level drive to ensure the columns >> in the "selected" row were driven properly while NOT driving >> any of the dots in the unselected rows (which, obviously, saw the >> same column drives that were being presented to the selected row!). >> >> [IIRC, Hughes made some devices that did this.] >> >> My question was intended to elicit how much *larger* panels >> are driven -- presumably more complex than the small multiplexed >> glass I had used. > > Rather hellishly so. There are about 2 million full color pixels in a > standard HD display (1920 by 1080) so figure maybe 10 million connections > to the glass (unless there is a lot more smarts in the glass) minimum.
No. There would be about 3000 (electrical) connections to said "glass" (that the electronics would have to "twiddle"). The 2 million cels are "driven" by cleverly manipulating the timing of voltages impressed on those row and column drives. In simple, non-multiplexed glass, you drive the "common" with a (relatively) fixed frequency square wave. Any segments that you want to "illuminate" (misnomer) are driven with a signal that is 180 degrees out of phase. Of sorts, this is an AC version of driving an LED (which expects to see a particular polarity of DC to illuminate the indicator). On (old) multiplexed glass, the common signal is replaced by a "common per row" while the columns are driven in/out phase to cause their "illumination". (think of a single row as a "simple, non-multiplexed LC display) Where this gets tricky is when you look at the field across some other cell (pel) that is NOT in the selected row! I.e., one of the "electrodes" at that cell is bouncing around with whatever column voltages want to be impressed FOR THE CELL IN THE CURRENTLY SELECTED ROW (which is NOT "this one"). The other electrode (row) is doing whatever it needs to do to keep this row from being "selected". To ensure the (not selected) cell does not illuminate, you want to keep the field across it at "0" -- despite the fact that one electrode is dancing around willy-nilly as required for the other cells in that column. *New* glass (active matrix) puts a Q at each cell *on* the glass. That transistor effectively *stores* the value ("amount of twist") for that particular cell. In a sense, you've created a DRAM and are now tasked with "refreshing" it before the charge leaks out of each of those Q's.
> Of course if you want to drive something like that in an "on-off" mode a > fast AVR or PIC could do the job of creating the analog or DVI video > signal.
Hi Lasse,

On 7/12/2014 6:25 AM, Lasse Langwadt Christensen wrote:
> > but if he only wants all white/black he doesn't need to drive pixels all > he need is a clock ~25MHz, hsync/vsync and all the color bits tied high/low > to get white/black
No, that requires some electronics along with the "glass". I want to know what the *glass* sees!
> an AVR/PIC etc. could do that > > -Lasse
=20
> I want to know what the *glass* sees!
Which glass? It probably depends on the manufacturer. As you and others p= ointed out, there are many ways of driving LCD/LED screens. They have dedi= cated controller for the specific panel technology and it is unlikely that = you can use (or even get) the glass without their controller.
On 7/12/2014 10:02 AM, edward.ming.lee@gmail.com wrote:
> >> I want to know what the *glass* sees! > > Which glass? It probably depends on the manufacturer. As you and > others pointed out, there are many ways of driving LCD/LED screens. > They have dedicated controller for the specific panel technology and > it is unlikely that you can use (or even get) the glass without their > controller.
FORGET *GETTING* THE GLASS! Does the (un)availability of the glass invalidate the question? I can't go to the corner store and buy "lightning" -- does that mean I can't inquire as to its properties?
On 12.7.14 19:32, Don Y wrote:
> On 7/11/2014 1:17 AM, Tauno Voipio wrote: >> On 11.7.14 03:56, Don Y wrote: >>> On 7/10/2014 1:02 PM, Tauno Voipio wrote: > >>>> You might be able to tie all rows together and all columns together, >>>> and drive the rows and columns with antiphase square waves. The >>>> hard part will be to find the proper voltage levels. >>> >>> I would suspect that would work on "dumb glass" (i.e., where the >>> rows and columns were just electrodes screened on the glass). >>> >>> But, I thought larger panels had some active components *in* (on) >>> the glass to effectively treat the display as a storage device >>> that is "refreshed" with new data, periodically (?) >> >> For monochrome panels, the usual chip was KS0083. It is a 80 bit >> shift register with LCD output level shifters. > > This seems similar to the Hughes (?) devices I used ~1980 (though > they were DIP40). IIRC, the output was "multivalued" -- not just > a simple (2 level) translator -- to ensure DC stayed off the glass. > >> For a 320 by 240 display, there were 4 KS0083's for columns and >> 3 chips for rows. The data was fed serially to the column chips, >> with the rows clocked once for each full row. I used a refresh >> rate of 14400 row/s, giving 60 frames/s. > > Yes, similar to the Hughes interface -- though I was driving a > much smaller display (7x95). OTOH, a fair bit of overhead for > a machine with a 2MB/s bus! > > But, how does this all relate to current technology? My understanding > is that new panels are treated almost as a sort of write-only DRAM. > The electrical interface serving to store a charge in each "accessed > row's" column cell instead of just presenting an instantaneous > voltage across each cell.
Actually the output was 6 levels, to cater for the multiplexing and DC avoidance. See the KS0083 datasheet, it is still available in the Net. -- -TV
On Sat, 12 Jul 2014 10:33:07 -0700, Don Y <this@is.not.me.com> wrote:

>On 7/12/2014 10:02 AM, edward.ming.lee@gmail.com wrote: >> >>> I want to know what the *glass* sees! >> >> Which glass? It probably depends on the manufacturer. As you and >> others pointed out, there are many ways of driving LCD/LED screens. >> They have dedicated controller for the specific panel technology and >> it is unlikely that you can use (or even get) the glass without their >> controller. > >FORGET *GETTING* THE GLASS! > >Does the (un)availability of the glass invalidate the question? >I can't go to the corner store and buy "lightning" -- does that >mean I can't inquire as to its properties?
You answered the question with a question - not a good way to get information.