Aspect Ratio: We decided that 16×10 was the best aspect ratio for our media panels. A 16×10 screen’s height is 62.5% it’s width, whereas a 16×9 screen’s height is only 56.3% it’s width. This may not sound like a big difference, but having so much more horizontal space than vertical space means that the 16×9 (while great for movies) doesn’t do as good a job handling photos or apps. By comparison, iPad sports a 4×3 aspect ratio, making it’s height 75% it’s width–a great ratio for a device you use up close for web browsing and email. For our media panel, we want something between 16×9 and 4×3. 16×10 seems to fit the bill, and is what Apple has been using for a number of years on MacBook Air and MacBook Pro.
Pixel Resolution: Once you’ve decided on an aspect ratio, choosing dot resolution becomes much easier because you won’t have too many choices. Let’s look again at what Apple has decided to do on MacBook Pro to see if we can learn anything. The 15″ MacBook Pro ships with a maximum resolution of 1440×900 and the 17″ MacBook Pro (before it was discontinued) had a maximum resolution of 1920×1200. While it is possible to get a 15″ 1920×1200 lcd panel, things do get a little bit small when you put that size resolution on a 15 inch screen. And putting the 1440×900 on the 17″ looks a bit big and doesn’t maximize the extra screen real estate we have. Ok, so we’ll choose 1920×1200 for our 17″ screen and 1440×900 for our 15″ one.
Weight: Despite the fact that our media panel is largely fixed in place and we don’t carry it around like a tablet (or even a laptop), having something light has a lot of advantages. A light screen can be much more easily put in a wall frame, desk frame, or even on the refrigerator without specialized and bulky hardware. Also, our goal with the media panel isn’t to replace the TV in the living room, or any of the other computing devices/tablets/phones already in the home. We’re looking for a slim low cost solution that can supplement the existing screens with a single appliance-like purpose. For this reason, we have focused on laptop LCD screens as opposed to desktop monitor or television LCDs. A 15″ laptop LCD screen can weigh as little as 1 pound (and newer ones weight even less) and a 17″ laptop LCD screen weighs just 1.5 pounds.
Dimensions: While there are great uses for media panels of all sizes, we are going to stay clear of both small screens and large screens in our initial project. Small screens have a harder time concealing additional circuitry, are less good as group devices, and might confuse the casual observer as to how they are different from tablets and smart phones. Large screens, on the other hand, quickly get expensive, heavy, and more like furniture than an appliance. For our purposes, we will pick one 15″ screen and one 17″ screen. Beyond that, the most important dimension for us to consider is thickness, since the height and width are largely known once you have aspect ratio and dot resolution (assuming you have a reasonable pixel pitch). However, a 17 inch LCD panel for a desktop monitor can have a thickness of 11.5mm, almost twice the thickness of a 15 inch or 17 inch laptop LCD display. 6.5mm (or 0.25 inches) is a good thickness and gives us lots of flexibility to add circuitry behind the screen or a a touchscreen glass in front of it and still have a nice slim panel.
There is a lot you can read about LCD technologies that we won’t go into here, but safe be to say the twisted nematic screens with a lower color depth (like 6 or 8 bit) are significantly less expensive, so we’ll focus our search there so we can keep the cost down. If you’re not concerned about costs, there are some amazing screens to be had with not just 24 bit color but deep color: 30, 36 or even 48 bit color.
The 17″ LP171WU3′s optical specs show a 6 bit 262k color TN display with a contrast ratio of 600:1 and a brightness of 200 cd/m2. The 15″ B154PW01′s optical specs show a 6 bit 262k color TN display with a contrast ratio of 400:1 and a brightness of 300 cd/m2. Overall, the 17″ seems slightly better than the 15″, but both should acceptably meet the needs of our media panel.
Electronics, Signaling and Control Board Considerations
Since we don’t plan to battery power our LCD screen, we’re less concerned with the power demands of the screen than we might otherwise be. That said, our expectation is that the media panels will run continuously day and night, so being as power efficient as possible makes the device a better global citizen and saves on electricity bills.
The larger concern is the signaling interface for the board and whether we will be able to find controller boards that will work for it. Most LCD panels communicate with a version of Low Voltage Differential Signaling (LVDS) called Flat Panel Display Link (FPD-Link). You cannot simply plug a bare LCD panel directly into the HDMI or RCA video out on your Raspberry Pi (nor, once it’s documented, the DSI video out). FPD-Link is a mechanism for transferring the RGB video interface. To take HDMI from your Raspberry Pi (or any other device) and display it on the LCD panel requires (as of today) a two chip solution to do the appropriate conversions. In addition, unless the LCD panel is a newer LED model, you also need a cold cathode fluorescent lamp (CCFL) inverter to operate the backlight.
The difficulty in obtaining and configuring an LCD control board is probably the reason there aren’t more people building DIY media panels. Most a la carte laptop LCD panel purchases are headed for the laptop screen replacement market, where all this circuitry is already there in the laptop and it’s simply a matter of swapping out the panels and hooking things back up.
Since most LCD control boards are custom configured components of a laptop, there aren’t really standard plug and play solutions. LCD controller prices at lcdparts.net are $70-$90, and you can expect to pay even more (if they will even sell it) to buy a single quantity unit from a specialized player like Digital View. There are some smaller niche players like Chalkboard Electronics in Malaysia offering a solution for $39 and Frank Industries in Germany offering a solution for 70 euro, but my experience has been that they seem perpetually out of stock since they are only doing small runs.
Keep in mind that you have to buy exactly the right board for your specific LCD panel and you must have it correctly programmed so it will work. We found the best solution was from a company in Guandong, China named NJYTouch. They produce a number of development/testing boards that will work with a variety of older common LCD panels for as little as $29 per LCD controller. This includes the LCD controller board, control switch, programming for your panel, and the correct LVDS cable and inverter. Power adapter and cord is sold separately, but a standard 12v 4a LCD adapter with a 2.5mm internal diameter and 5.5mm external diameter will work fine–no need to have a power supply shipped from China as well.
Here is a video of us hooking up our Raspberry Pi to the 17″ LCD panel using one of NJYTouch’s LCD controller boards:
Taking into consideration the goals from our mechanical and optical considerations factored with which LCD panels we can easily get an inexpensive LCD control board for further narrows down the choices and pushes us toward slightly older slightly more expensive LCD panels (but still only about $75 each fromlaptopscreen.com).
All in all, there was a significant learning curve to understand all the options, make good choices and get a solution that we believe will ultimately work (both figuratively and literally). Takeaways at this stage are:
A more generic, easily configurable, less expensive and smaller HDMI/DVI to LVDS bridge solution is needed. The Raspberry Pi board is big enough as it is…and most of the LCD controller boards available for purchase are closer to dev boards and include features and capabilities not really needed in a production system.
An LCD controller for an LED screen vs a CCFL screen (what we got) might reduce overall costs since the LED screens seem to be cheaper. Unfortunately, none of our sources had a controller that would work for these newer screens. If you know of someone who does, please let us know!
The LCD panels we bought were matte finish…but having used them for a few days it is clear that they should be glossy finish to really do a good job with pictures and video.
DIY LCD screens are definitely doable, but require a non-trivial amount of research, patience (shipments from China) and coordination to not end up with a solution that doesn’t work.