It has been a frequently asked question regarding how to interface to an embedded LCD panel using the Beagle Board, given that the OMAP3 LCD interface pins are not brought to the header. At BarCampHouston3, I got to meet Bill Wood who is looking to build a one-off long-lasting laptop. He introduced me to a module available from Digi-Key that enables hardware hackers to utilize a variety of (typically smaller) LCDs than what are enabled by using off-the-shelf DVI-D monitors. While it costs about $140, it still seems like an interesting option for getting something a bit more custom without building your own board.
BB_LLL.pdf (25.9 KB)
On similar lines some one can even try putting a converter for DVI to VGA, I was talking to some folks but nothing materialized yet.
The THS8200 can be used to do this,
http://focus.ti.com/docs/prod/folders/print/ths8200.html
Regards,
Khasim
Jason Kridner wrote:
It has been a frequently asked question regarding how to interface to an
embedded LCD panel using the Beagle Board, given that the OMAP3 LCD interface
pins are not brought to the header.
You'll be fixing that with the 'C' revisions, right?
b.g.
There are no plans att he momen to fix this on the Rev C board. It would require a total re-layout of the board which would increase the overall size of the board substantially as well as impact the integrity of the signals. Adding buffers would also increase the cost. I will look into a way to do this, but at the moment there is no easy solution.
Gerald
FYI, I would also be very interested in this feature and think there are
many others who would too - it is the one missing feature that I could
really use.
Ed.
Gerald Coley wrote:
There are no plans at the moment to fix this on the Rev C board. It would
require a total re-layout of the board which would increase the overall size
of the board substantially as well as impact the integrity of the signals.
Adding buffers would also increase the cost. I will look into a way to do
this, but at the moment there is no easy solution.
I know it's a major deviation from the design goals of the B series, but
I think this particular enhancement might be worth it if the costs and
other factors can be managed.
If the aforementioned DVI-to-TFT conversion board is confirmed to work,
however, then I can live with that! Is anything lost when you go that
route as opposed to going right to TFT on the beagleboard itself?
Of course, adding this capability to Beagleboard makes it look
increasingly like a reference design and/or OEM product, rather than a
product demonstration--- and that in itself might be justification for
leaving it as a feature found on systems designed for that purpose.
Has Bill Wood confirmed that this board works with the beagleboard as
intended?
b.g.
I understand. There is a way to purchase a DVI-D to LVDS board for $40. But it limits you to a LVDS panels. As the output is 1.8V on the OMAP3, buffers need to be added to convert it to a higher voltage to support a typical LCD panel.
If we added $25 in cost to the board to support this feature, is it worth it? Would the community pay $175?
Gerald
Would the $25 provide LVDS as you describe, or direct access to the LCD signals? Personally, I would be paying $175 right now to buy one if I could get to the direct access to the LCD signals, even if I needed to add my own buffers, and even if the board size increased slightly. As it is I am still sans-beagle. I think many of the hobbyists in the community would appreciate this too, as it really opens the doors to a huge range of embedded possibilities aside from a desktop/development system.
If you’re worried about losing users who are not interested in the feature, how about another version of the board? If you could put all the connectors onto pin headers in that version – even better, and that might help reduce the cost compared to all the standard connectors.
I know the beagle board isn’t intended as an OMAP development system, and hence it is missing some features of the OMAP, but any other kind of peripheral can be sensibly added using SPI or some other serial bus – e.g. keyboard matrix decoders, digitizer interfaces, cameras etc, while it seems silly and counter productive to go from raw digital LCD signals to DVI-D back to LCD signals again.
Otherwise I’ll probably be waiting for cogent/gumstix to produce their OMAP products L
Ed.
Not confirmed yet... still waiting on parts (haven't even received my
Beagleboard yet ... backordered).
Not confirmed yet... Waiting on parts... (Haven't even received my
Beagleboard yet... backordered)
Personally, I prefer keeping the BeagleBoard as cheap as possible. In
my own case, it would have been nice to have had a VGA interface
instead of DVI-D. I had to buy a DVI-D monitor, which cost more than
the BeagleBoard but now I have a really nice wide-screen monitor that
does both VGA and DVI-D. My alternative was to buy a DVI-D to VGA
converter box but those start about $200 so you may as well buy the
monitor these days. (I'm guessing that DVI-D to VGA converter boxes
are expensive mostly because they don't sell a lot of them. They're
probably only $20 in components. However, once you put it in a box
and add a power supply, it adds up quickly in low volumes.)
Here's my suggestion for what it's worth. Since the DVI-D (HDMI)
connector and the expansion connector are on the same edge of the
board, it would be quite practical to make low-cost plug-in expansion
boards, one that converts DVI-D to VGA and one that converts DVI-D to
LCD. The expansion connector provides +5V power. It's challenging to
find a surface-mount male HDMI connector, but it may be possible to
take the type of connector that's used in HDMI cables and solder its
pins to both sides of the expansion card.
John
I agree it seems silly, but what I pushed Gerald to design for us was a system that went with standard interfaces. My feeling is that greatly reduces the support burden for a community that I hope will grow to be very large. While I’d love for the cost of all OMAP development platforms to come down, what I was looking to minimize was the cost to give the biggest wow factor. For me, 42" television monitors and DLP projectors provide a bigger wow than tiny LCD panels. To keep the cost as low as possible, we needed to choose one over the other. If there is enough demand, then I’m sure someone can make the board that would provide the direct LCD interface, and it seems like there is enough demand for that…
My feeling is that waiting on Cogent or Gumstix is the right answer. Or, you can go with LogicPD today, if they work for you. (Digi-Key OMAP3430 mobile zoom kit part number is 296-23351-ND, OMAP3530 medical zoom part number is 460-3478-ND). I’m guessing that the problem with LogicPD is that it isn’t easy to buy just the SOM-LV and the full kits are $800+.
In general, I think we should keep Beagle close to what it is in the future. It provides standard bus expansion–not ideal for the hardware hobbyist, but lower cost for the software hobbyist who is potentially pulling from his existing stock of PC peripherals. Hardware hackers can get to the “easy” busses, but for more in-depth activities they would really need to use one of the system-on-module or other single-board-computer solutions.
In the meantime, something like the $130 DVI-D to LCD panel conversion board is something that strikes a balance of now and not totally excessive in cost–and doesn’t require fragmenting the Beagle community with boards that are quite different in function.
Feel free to tell me why I’m wrong and that we’d sell 10X as many Beagle boards by adding the LCD connector. 
This doesn’t include the costs of we-already-did-it-the-other-way and this-would-impact-the-sale-of-the-existing-design.
If it were to change, it would be on a post-revision-C design. Rev C is just to add the USB EHCI host port.