need a new lcd display

Most parallel 8 bit LCD's can be configured in 4bit mode you will still need a clock and an enable line however.

Ed

wich

from

any

available to

something

Thanks, im getting one of these, actualy 2 incase I break one I learned my lesson,

its quite a bit larger than what I have but might be usefull for other things such as graphicaly displaying a signal, it has 4 data bits, 2 clocks and 2 other signals, but theres no explanation of how the data and control lines are used on the datasheet or their website, any idea if I'm right in assuming the 2 clocks are for col and row like the display I have ?

Colin =^.^=

This looks like a graphic panel... They're a pain to use without a control IC, but I can tell you what the signals do according to a past experiment of mine... First thing to cover is the architecture of the common and segment drivers: The common drivers are effectively a shift register operating as a serial to parallel converter, whos outputs go into voltage-converters and on to the LCD panel. The Row (Segment) drivers contain a shift register also in serial to parallel mode, whos outputs go into a latch with the outputs of the latch through voltage conversion to the LCD.

The Common Driver is provided with CL1 as clock, FLM as data, and M as an AC drive for the voltage conversion. It's parallel outputs are used to enable 1 line at a time of the panel, and CL1 moves the line down each time a new line is ready to be drawn. FLM must be held active for one CL1 pulse at the start of the frame to switch on a line to subsequently scan down the screen.

The Row (Segment) Driver is provided with CL2 as clock to the shift register, CL1 as clock to the latch, and the data lines which may be

1bpp or more if your panel does greyscale. A line is clocked into the shift register with CL2 from the data lines, and presented to the latch inputs. When the line is complete, the CL1 pulse causes the latch outputs to match the inputs, and a new line can be clocked into the shift register without losing the old line which is being output to the LCD by the latch.

so... M is an AC drive to alternate the drive voltage on the liquid crystal and must be supplied with squarewave to avoid damaging the liquid crystal. Ususally covers one cycle every 2 frames so each frame is drawn with a reverse voltage to the last.

CL1 is the line clock, and should be clocked after every complete line is loaded onto the display.

CL2 is the dot clock, and is used to load line data

FLM is the frame sync, and must be active for the first CL1 pulse, and inactive until the 1st line is drawn again

D0-D3 are however many bits-per-pixel raster data. If you have 4-bit greyscale, then each CL2 will be one pixel. If your display in monochrome, it will represent 4 pixels with each CL2 (1bpp).

It's quite a lot to describe... if you have trouble feel free to ask me more! If you are lacking a control chip, I might be able to help you design a controllor for your display with 74-series logic like I did! I've still got the schematic somewhere, but it seems no-one in their right minds designs such circuits so if you've no control chip you might prefer to give up!

Good luck. Nick.

Many thanks for the detailed explnation :)) the panel I have atm is an old epson 256x64 bit graphic panel too, I already have done the software to drive that directly from a dspic30f4011 (but I managed to fry it and I only had 1) from your explnation it should be almost the same with :- xcsl->cl2, yscl/lp -> cl1 din -> flm fr -> M xecl -> not needed

and just hope the edge to edge relationships are the same.

The only real draw back is the amount of ram most pics have isnt enough to memory map the display so the character rom has to be looked up each bit, but with 30mips this isnt too bad.

Colin =^.^=

I'd be a little concerned din might be raster data unless you have other pins for that... not sure what xecl is... an x-clock enable line? But the good thing is you can experiment to your hearts content to find out, as all the lines are inputs :)

Myself and a friend have experimented with programming an 8051 clone (actually a DS5000T) to drive large graphic panels, and found the timing to be hopelessly slow! perhaps your lower resolution will let you get away with it. we were running panels with resolutions around

640*200, 1bpp mono; and the DS5000T, even overclocked to 33MHz was not capable of mapping external RAM and CG Rom onto the display without the scan being painfully visible!

Good luck :) Nick.

My epson panel had 4 row chips, and needed an extra 4 clock pulses on xecl to clock in each row, dont know why din is labeled so, but it is a once per frame pulse, there are 4 data lines just like this one. the dspic30f runs at 30mips (120mhz internal clock), I got one line of data out in about 20us, wich left enough time to process other stuff inbetween lines, written in C too, although paying very carefull attention the the generated code, this shall be quicker (per bit) with no xecl, just depends on the setup times etc. the timing was a bit critical, with a scope probe on the data lines the display lost a few bits here and there.

I was looking out for small mcu with enough ram to do it bitmapped, if need be, or might be a good idea anyway.

If the panels are nice enough I might use it to display something like a waveform, for some future project anyway.

thanks Colin =^.^=

Well ive got it wired up plugged in and programed, but the only thing that can be seen is a single line of pixels moving down the screen when the scan rate is reduced to a crawl, as I speed up the scan rate it gets fainter untill it disapears, even stil at a relativly slow rate.

The pixels do however apear to be correct, ive just made a simple chequorboard pattern, so I must be doing something right, the vlcd is -20v to -25v, vdd 5v etc, the contrast voltage seems to be best at 0v, and of course /doff at 5v ive messed around with the order of the edges of the various clocks etc, but with no improvment, and of course adding an increasing delay inbetween each transistion. im not sure if cl1 should be ative while cl2 is low or high or during the h-l or l-h transition but I tried them all lol. while it is low seems to be ok and the easiest.

Is there anything else that I might of missed ? I hope ive not ended up with another dead panel :s

thanks Colin =^.^=

I've had another look over your datasheet, and it's a little cryptic in places and lacking good information to answer little riddles easily :) I've never encountered the 'M Modulation' block in any panels I've looked at... There are two lines going into it though, and it makes me wonder if the M line needs a different waveform to what I'm used to. I assume it would be CL1 and M that supplied the modulation circuit, in which case it's possible the display arranges to 'paint' each line in an alternating voltage... You could try odd duty cycles on the M input and see what happens. But it may be arranged to take the normal, equal-duty M signal and generate alternating line scans from it. If the image appears and then fades out, then the liquid crystals are getting a DC drive, and are likely to be damaged (but hopefully still useable). I'd also suggest something more interesting than a chequorboard - I've had displays work fine with that, but turn out to be drawing stuff on the screen in a really bizzare way :)

Also, I'm not sure from the datasheet of the contrast voltage range (I'm tired and might not be thinking right). It looks to me like the Vop column in the voltage specifications must refer to to the Vo voltages for nice contrast at various temp. ranges, in which case contrast would be negative. It's been a while since I hooked up one of my panels, and I have a suspicion at least one was a negative Vo too, but driving it negative always makes me nervous! I'm sure I've blown a panel by giving it -ve Vo before... Have you been testing it from 5v to

0v, or 5v to -25v ?

Final plan, another risky one... raise the Vee (or rather lower) to -30 to -35 (not over the spec. 'MAX'). I say risky, as it's the highest contrast voltage I've heard of yet, but the datasheet appears to suggest it's possible. Also, my panels are really old, and all pulled from ancient XT "laptops", so perhaps yours requires a higher drive voltage than did mine. Bear in mind I'm guessing at possible solutions, one of which *could* damage the liquid crystals, and two of which

*could* damage the control ICs. I'd be tempted to try in your shoes with two panels to play with- but dissapointed to be reduced to only one!

Good luck - hope it doesn't end in smoke! Nick.

your not bloody kidding lol !!!! I think its disgracefull to not put enough information at least somewhere on their site, I gues an email to them is in order.

If I understand it and from experience of the last panel the ac signal is just to stop damage from long term DC votage and should still see something, unless its also used to actualy draw the line or something. if it cant even display the chequorboard I doubt it wil display much else either. (nor will it be an great deal of use lol)

When the scan stops the line stays there and gradualy fades out. im not sure ive got the correct dc power up sequence, some data sheets are rather picky about how vee,vdd,von etc are brought up, although vee is generated from vdd so it cant come up for a short while.

ive just found a few other data sheets wich give the timings, and have the same name for the signals :-

FLM stands for first line marker wich is nice to know. not sure what M could stand for.

PS ive just found the data sheet on one of the chips inside the panel so that shoud shed some light, lc4901,lc4930

although it doesnt seem to tell me anything new. at least I can probe a few of the pins.

Ive tried 0-5v as that seems to be common on other data sheets ive seen, im a bit nervous that negative voltage might blow it up too ! maybe I might try makeing it negative -5v through a 1Mr or something. however it is more visible at 0v than 5v.

Ive gone to about -27, wich if I understand it coresponds to a vop of 32 actually disconecting the -vee mad no diference at all, its as if its not even using it I might trace the signal on the panel, its not taking hardly any current at all either.

worse thing is if I destroy both of them, I dont want to plug the other one in and find its the same, then I wont know if ive got 2 panels wich I dont know how to use or 2 dead ones. although its new theres a chance it may be dead for one reason or another.

thanks and thanks for the help so far, Colin =^.^=

I'm less tired now, though not by much! -35V would almost certainly end in smoke! With a 5V supply, the max negative would be 30V (still seems pretty high to me).

Your problem still sounds like a contrast issue, and again I'm wondering if the Vo pin should go down to -ve voltages. On some of my panels the I'm sure the Vo pin works best with a potentiometer to adjust from 0v to Vee, and it supplies an emitter-follower on the panel presumably as a current-boost. 1M might be a little big to try and expect to work due to the low current... the emitter-follower probably needs more current. But I'd certainly only try -ve on one panel to see if it worked! And then only if I didn't mind losing one :)

Unforunately I can't think of any way of ensuring you don't destroy the panel, and can only suggest that if you do try it you should cross your fingers!

I'll not be able to post for a time from sunday, as I'm off onto the north sea to join in the tests of a holographic camera, but I'll look in again when I get the chance!

Good luck, Nick.

Yes you were right it was negative voltage for vo, I tried with a 3M3 then 1M then 100k wich was ok. I posted in a new thread on this here and sed too. many thanks :)

I get a max refresh rate of 140hz with my mcu, wich isnt bad, it shld let me do plenty of other stuff inbetween lines. with a vee of just -20 volts its ok.

Im gona try and make a transformerless EL driver now.

good luck with the camera sounds a lot of fun !

Colin =^.^=