has anyone had any experience with making their own LCD displays? Something like a calculator display but much much simpler
I've been reading up on the subject and think I've got an idea about how to fabricate it all but I'm having problems finding a supplier of the actual Liquid Crystal chemical. Its rather annoying as I've got a supplier of ITO films and polarising film but cant find the last piece of the puzzle!
I was evaluating this professionally 25 years ago or so, so I have a clue as to what's involved. Sadly, most LCD makers consider their liquid crystal to be a proprietary mix, and won't release the formula, nor even sell the material.
That doesn't mean, of course, that you can't come up with your own from readily available/published materials, but be aware that your display probably won't perform as well as a commercially made one. You'll need a nematic type of liquid crystal material.
You've mentioned the ITO films, which are applied (diffused into) the glass plates which form the "bottle". You've also mentioned the polarizing film. But, you haven't mentioned the alignment layer that's applied to the interior of the bottle. Fortunately, this is usually a high molecular weight polyvinyl alcohol. Unfortunately, it may not be easy to find, and the exact type may depend upon the liquid crystal you're planning on using. Without this, the relaxed liquid crystal may take on a random orientation (e.g., won't work quite right).
You also haven't mentioned the epoxy used to glue the bottle sides together. But, fortunately, epoxy is usually easy to get.
You'll also need some glass spheres/rods approximately 10 micrometers in diameter, used to space the glass plates that comprise the bottle the right distance apart. Without these, the spacing between the glass plates may vary, and this will cause objectionable effects.
You'll also need a way to get the liquid crystal material into the bottle. Capillary action may get most of it in there, but you'll need to get the air out of the way first, so that the liquid can be sucked in. This may mean a vacuum pump, or it may just mean orienting the bottle correctly, with the correct entry/exit slots.
Don't forget to epoxy the openings closed once the bottle is filled.
For a small panel, you can make electrical contact with the ITO patterns on each surface of the bottle with some type of clip. The old production way of doing it was by using a isotropic conductive rubber strip (called a "zebra strip", due to the alternating stripes of conductive and insulating rubber). The newer production way of doing it was via a flex cable bonded to the ITO pattern on the glass.
For production panels, the old way of doing it was to bring the ITO pattern from the top glass plate to the bottom one with silver filled/conductive epoxy beads loaded between overlapping areas of the ITO. But, for small panels, it may be adequate to use two clips, one for the front panel and one for the back panel.
For a front view panel, you'll also need a brushed aluminum backing layer that can be glued to the back polarizer. Or, you can go for a see-through, or projection type of display.
although its google-translated its still fairly clear, his final piece also looks pretty good! Theres some smudging at the bottom of one of them but if that happens in mine I dont really care, it'll be artistic :p
Is it important that the spacing across the two glass plates is at 10um? If it were larger would it still work but require a larger current to change the crystals orientation?
My plan is to try and create a curved lcd screen (set into a bracelet so it wont be flexible), building up from a shapeways printed bracelet. The screen itself wont be detailed, probably just large blocks or maybe even the entire screen blanking out to start. The effect I'm going for is artistic rather than technically perfect, so smudging and imperfections are ok
would it be possible to reclaim the liquid crystal from old LCD's? If i were to grind down the epoxy holding the plates together would I be able to scrape the LC from the screen?
As for driving the display I think an ATTiny with a timer running at ~60hz and outputting a square-wave on one pin and inverse of that wave on another pin should do it?
I thought that maybe connecting wires to the ITO using a conductive epoxy/glue might work?
Does this sound feasible? Even if its not, I still want to try this out :)
Yes, that's a good explanation of how one works/is-built. I had neglected to mention the rubbing. The ones I was familiar with used PVA as an alignment layer, although Silicon Dioxide (SiO2) could also be used. The rubbing creates the microchannels that cause the liquid crystal material to "lay down" in the correct orientation.
The spacing is somewhat critical, but not overly so for a direct drive device. It becomes MUCH more important if you're going to multiplex the display. However, for something simple, like a direct drive display, it won't be too much of an issue. However, if your spacing gets to be too large, then you'll need a high voltage to cause the display to operate.
The reason the spacing is important is that the electric field potential across the display depends upon the separation between the electrodes as well as the applied voltage (e.g., Volts per meter). A larger separation means a lower electric field. And, since it's the electric field potential that causes the liquid crystal to switch states, a variation in it will adversely affect the threshold where the device changes state.
So, for example, if it takes 3 Volts across a 10 micrometer spacing, that equates to an electric field potential of 300,000 Volts per meter (Hmm, that sounds a little high; maybe my memory is getting faulty?). But, if you increase the thickness of the liquid crystal layer to 100 micrometers, then to get the same 300,000 Volts per meter electric field strength, you'll have to apply 30 Volts across the display.
Note that the liquid crystal exhibits a non-linear effect. A little voltage across it doesn't cause it to do anything. A bit more voltage still doesn't cause it to do anything. Even a bit more voltage doesn't cause it to change. But, then, just a tiny bit more voltage causes it to suddenly switch states. That non-linear effect is quite important for multiplexed displays, where the voltage between an on pixel and an off pixel can be different by as little as one percent.
There are also some second order effects and other non-linearities associated with liquid crystal displays. There's a "pre-tilt" angle that can influence the switching point, although that's mostly only of concern in multiplexed displays. Temperature comes into play, too.
There's also the issue of how steep the twist angle is, and that relates to the thickness of the device. For a 10 micrometer thick layer, the twist is normally 9 degrees per micrometer, giving you a total twist of 90 degrees over the 10 micrometer thickness of the display. However, one of the advanced technologies was "super-twist", which employed a 270 degree twist, which required a much faster twist angle. Unfortunately, I've forgotten most of the aspects of this in the 25 years or so since I last worked with LCDs (and, be aware that my knowledge is somewhat dated).
If your display thickness isn't the same all the way across, variations in the thickness, accompanied by the steepness of the twist angle, can cause some weird effects. For that matter, one possible weird effect is that the liquid crystal can twist in one orientation in one part of the display (e.g., clockwise), and twist in the other orientation in another part of the display (e.g., counterclockwise). I've forgotten how this is usually prevented.
You may be able to salvage the liquid crystal from a dead display. However, be aware that liquid crystal displays use a VERY small amount of liquid crystal material, somewhere between a half of a drop and a drop. Of course, since you're going to be making a liquid crystal display with the salvaged material, that may be acceptable.
Your drive method sounds reasonably appropriate. Note that liquid crystal displays are voltage controlled devices, and, as such, require VERY small currents to operate (typically in the microAmp, or even picoAmp, range or so, which is almost entirely due to the capacitance). However, the voltage that is applied MUST be an AC voltage, with no DC component present. The presence of a DC voltage will cause an electrochemical reaction, which is highly undesirable, and which can ruin the display in a period of a few weeks (based on some unpublished empirical tests I conducted back in the early 1980s).
The conductive epoxy/glue should work well. Since liquid crystal displays are voltage operated devices, with VERY low current requirements, moderate values of resistance in the conductors or joints is usually not a problem.
The largest problem I can see is the curved design. Trying to get two pieces of glass reasonably flat with respect to each other around a curve may be problematic. It is, of course, theoretically possible, but the practical issues may prove to be difficult to overcome. On the other hand, approximating a curve with piecewise linear segments should work well.
Good luck with it, and please let us know how it goes.
Oh, I probably should have mentioned more of the temperature effects, too. If you get a liquid crystal display too cold, the liquid crystal will revert to being a solid, and won't exhibit the phase change with the application of voltage (e.g., it'll be frozen), but it should start working when it warms up again. If you get one too hot, the liquid crystal material will lose it's nematic/crystal properties, and become an ordinary liquid, which won't have the optical rotating ability. It should start working again when it's cooled down, but you may run into issues with the twist pitch being wrong or reversed in portions of the display, which can be optically annoying.
Hi Dave, thanks again for all the information, even if nothing comes of this its still a damn interesting read!
you said : "For that matter, one possible weird effect is that the liquid crystal can twist in one orientation in one part of the display (e.g., clockwise), and twist in the other orientation in another part of the display (e.g., counterclockwise)."
would this not show as a wierd pattern of dark and light across the display? If so that might produce an interesting effect and look rather nice. As I said this is more of an artistic piece than an attempt to produce a well defined display, so blemishes and wierdness are perfectly ok :)
You also mentioned PVA. is the stuff used in LCD production just a stronger version of the standard wood-glue? How would that be applied if I were to test it?
I'll certainly try reclaiming the crystal from some dead panels I have, theyre good for nothing else! I guess the amounts used in displays were low when you mentioned "capillary action"..
I might go for the non-curved approach to start with, although if i can source enough material (and can get enough working prototypes..) then I might try it later on.
Just out of interest, did you ever try making any yourself?
The change in twist orientation is more noticeable as a contrast ratio shift than anything, usually with changing viewing angle.
PVA is PolyVinyl Alcohol. not PolyVinyl Acetate, which is what the wood glue is. PolyVinyl Alcohol is a high molecular weight alcohol, which has some crystalline properties, and some adhesive properties, too:
It's normally rubbed on the inside surfaces of the bottle, and rubbed to orient the molecules in the desired direction. Since the PVA has an adhesive property, it sticks to the glass. Since it's composed of long molecules that can be oriented (e.g., crystalline), it causes the nematic liquid crystal molecules in contact with it to orient themselves in the same direction.
The amount of liquid crystal used is extremely small. Let's see, a 11 inch by 15 inch display, that is 10 micrometers thick will require about .011 milliliters of liquid crystal material (if may math is correct). That's not very much, well under a "drop" (close to one-fifth of a drop!):
As for making any, no, I haven't personally tried it. At the time, I was an engineer in an advanced technology group in a Fortune 50 company, and we decided that it wouldn't be cost effective to do our own LCD manufacturing; that it would be better to buy our displays from companies that specialized in LCD manufacturing. I would have certainly liked to have tried making some, but it would have been a development dead end, and would have required a significant investment in technology and development at a time when the technology was changing rapidly.