Let's make a LCD, I mean seriously fantastic one.

We had posted earlier about the music shield, people were interested in the LCD underneath… The bad news #1 is that we can’t source the very same one, the bad news #2 is that the LCD has this color just for shooting from side so it is not real; but we have a good news:

Let’s make one LCD:

  • that have chromatic programmable color,
  • that have customized display pattern for and only for the open source world,
  • that have the size we are comfortable with,
  • that have the character/symbol/logo we love,
  • that can tolerate hacking and ideas,
  • … to be added

The basic principle of a LCD is here:


Reflective twisted nematic liquid crystal display.

Polarizing filter film with a vertical axis to polarize light as it enters.
Glass substrate with ITO electrodes. The shapes of these electrodes will determine the shapes that will appear when the LCD is turned ON. Vertical ridges etched on the surface are smooth.
Twisted nematic liquid crystal.
Glass substrate with common electrode film (ITO) with horizontal ridges to line up with the horizontal filter.
Polarizing filter film with a horizontal axis to block/pass light.
Reflective surface to send light back to viewer. (In a backlit LCD, this layer is replaced with a light source.)
1-5 is what we have talked with a LCD factory for customization, and 6 is we are going to play with. Please leave us your ideas to the forum thead, and get some prototype if adopted. ^___^

Maybe use an rgb led for the backlight?


RGB backlight has been done before, i’ve seen 2x16 LCD’s on eBay with RGB backlighting. While these are nice i’d like to see RGB backlighting with individually controllable leds. This would enable a background fading from one color into another. Just like the fade in the picture.

Another option would be a bright line around the displayable area.

We will surely include a RGB LED back light. How about the content on LCD? Maybe not some standard 16*2 ones, but with some special pattern… I will draft one later.:slight_smile:

If you go with an RGB back light it could be controlled with an Atmel ATtiny45, a trimpot and some simple code like Paul Boardman uses here

It would not be that hard to write the code so that the ATtiny45 changed the backlight color via PWM input from an Arduino either.

As for LCD form factors I would like to see a 24x4 or even a 30x2 or 30x4.

I had a phone (Nokia N73) which had a clock app that ran while the LCD backlight was turned off. In most lighting conditions the contract between the black background and the dim gold colour clock face was enough to read it. Something like that would be ideal for a super low power device such as a solar clock or radio.

How about some kind of holographic display? You can get very expensive full colour ones where you see a different image depending on the angle you look at it from. They use them in some high end cars so that a single monitor can display sat nav data for the driver and a DVD movie for the passenger.

The only other suggestion I have is that if it must be backlit then maybe instead of using just one RGB colour you could put a row of say 5 RGB LEDs along the top and the bottom. That way it would be possible to vary the colour across the display.

Which controller will be used? I could think of a bunch of generic “icons” like battery status etc. I’d prefer i2c interface, controlling backlight color and contrast over the same i2c port would be awesome. Perhaps even a few extra GPIO pins to attach buttons or a simple keyboard matrix all accessible over i2c.

I would prefer a graphical display combined with the generic icons.

You could go with a clock layout like in the picture, but I guess that wouldn’t meet everybody’s needs. Or maybe some combination, you know, 7-segment numerals but with extra segments to be able to display characters as well. Example: stempelshop.nl/letters/lcd.jpg

What about a custom layout service, like for PCB’s? Would that be even remotely feasible (e.g. not incredibly expensive)?

This is a great idea. Do you intend to do a serial/i2c controller for it? There are some reasonable open source/arduino serial libraries for graphic displays. The question is will this be a pixel addressable device, or is it have a bunch of symbols etched in the glass for higher res, but less flexible display. (like many battery symbols that have 4 bits to select which lvl of fullnes to display. Personally I would vote for Pixel addressable, rgb led pwm backlight and I2c controller with graphic library and either a 328 or enough flash for logos, sprites, and fonts.) Plz not a 168 with no external memory :angry:, but then you said hackable so that counts them out :smiley:. A reasonable controller with code for moving around sprites means lot of simple games become simple to do able as well.

I have used some pre-defined displays in the past, but I really want a bit addressable canvas with sprites and or image in flash. Those I have used never had all the symbols I wanted, or where I really needed them. I know it hurts the cost, but it is way more hackable. I can always have the symbol I want with a simple tool to upload and manage the flash/font cache on the device. A few de-bounced switches and a few analog inputs addressable via i2c would make this a tempting front end to many projects. The question is does this push the cost too high?

Most of the suggestions just seem to be for a cheap and powerful graphic display. There are plenty out there already, e.g. the many full colour mobile phone screens that are well documented.

To be different and better than the rest it will need something special. How about layered displays for a nice parallax effect? Or maybe Seeedstudio could be the first company to bring eInk displays to hobbyists. eInk would be awesome for so many projects, since it requires no power to maintain an image and is readable in daylight.

Thanks for the suggestions! :slight_smile:

We will consolidate the suggestions and propose a spec and pattern.

Basically, we don’t want to overlay it with a graphic ???*?? pixel graphic LC but a segment LCD like in the picture. It could be very big number with customized font and lit just like 7 segment LEDs.

Of cause controllable with serial, maybe el electronic brick 4 pin connector for easy control and setup. RGB LED is not an option but must have. :slight_smile:

We have been knocking doors of e-ink suppliers but still not able to get a reasonable cost/MPQ to start with. Can’t wait to have on myself though…

Ah … that was not really clear i think. But in that case i’d go for some variant of starburst.

Large Alpha/numeric can be good. 7 segment is awful limiting, but it good to understand the goals. What are you proposing for a power target, to inform backlighting suggestions? EL can be nice, but rgb LEDs are alot more flexible and I believe you said decided. eInks are interesting, but have funny characteristics and sometimes high write voltages for some of the technologies. Understanding the power goals is a good way to eliminate many options.

So we are talking a large limited digit display primarily for Numbers? Is it limited to 4 digits, or 5?
Time only requires 4, but many other uses do better with a 5th or 6th digit. This also gives a digit for the “:” units specifier if possible. I will confess I really prefer displays that are clearer than a simple 7 digits as the letters/numbers are mangled to really be displayed so I would encourage doing a little better than just straight segments. Big blocky displays are less interesting than they were when we could first build them.

Have you seen the screen on the new Nintendo 3DS? It is 3D without needing glasses.

Thank you guys for sharing your ideas in previous blog and forum topic! Aside from the functionality, Windell from Evil Mad Scientist Lab give us a jaw-breaking Hexagonal idea! I tried to illustrate the idea as below:

Actually it looks great when printed to 2-3cm diameter, which each pixel is about 1*1mm. Every details are uncertain, we might need in-depth consideration to come up the best most balanced way.

In the illustration, I tried to use 4 roll unit of ( 1+6+12+16), good for alphanumeric display or simple icon. If we use 3 roll unit of (1+6+12), it really save some space but some number like “B” will look wield. On the other hands, the controlling would be quite different from a normal graphic LCD.

How do you think?

Interesting idea!

There are many possibilities:

This site also has some interesting details on the merits of various schemes: http://www.balloonhq.com/column/rouse/aug05/

Personally I like the triangle grids because you can make other shapes like hexagons with them.

I really like the Triangular grid idea. That seems like a good compromise between a 7 segment display which looks really bad, and a graphics display that normally cost way too much. Yes a reasonable set of fonts will need to be tested against the layout before it is frozen, but I think it will yield the best compromise. Given that most rendering engines work in terms of triangles it makes al ot of sense. We will need to develop the right set of primitives in a library for driving it, but there are plenty of places to beg, borrow, or copy the algorithms.

The decision will come down to number of triangles to cost clearly if you choose to go this way, how small will they need to be to render non roman characters, in a reasonable form? Line drawing and other shapes are well defined in triangles, and as an earlier post a polygon can always be decomposed into a set of triangles.

Another thought, how close together will the segments be? On cheap character LCDs they are separated by a small gap, where as most decent graphic ones have the pixels meeting at the edges.

I have talked with factory for making the LCD, the Hexagonal idea astonished them, though it would be possible to make, 1mm diameter/pixel would be too small.

Also, there are many details beyond our current experience, we will take a try on RGB backlight with current LCD modules. Then decide the segment and size. Some photos for your reference:

I have wanted to make a weather station for some time. The hardest part is finding reasonably priced and easily interfaced anemometers and rainfall measurers.