PSP 480x272 LCD - ARM µC LPC2478 solution

Hello,

Let us put here some practical info about driving the PSP 480x272 LCD using a ARM µC LPC2478 : inexpensive and built-in LCD interface. No need for a Seiko display controller inbetween. So anybody having success in displaying “Hello” on the PSP 480x272 LCD using a LPC2478 ARM µC is warmly welcome here !

I’m working on the same project, LPC2478 and PSP LCD with MT48LC4M16A2 RAM memory (4Mx 16bit).

I’ve got some difficulties with the EMC config.

Does anybody have some sample code for me?

Many thanks.

Nico

Dear,

Check out the video I did, A LCD graphic controller using a FPGA:

http://www.youtube.com/watch?v=zB6M_l4btpk

Maybe I can help you if you have any doubt about the display.

Cheers.

Clad to see someone has done something with it, however that is WAY to much support circuitry for that LCD. There must be some FPGA guru out there that can made a controller for this LCD with just one chip like the original board did, before that guy left SFE.

“That guy” would be me ;).

Keep in mind that the demo was quite hacked together. The PSP LCD screen was definitely displaying the images and all that, but lots of tricks were used to generate the patterns / SFE logo. The patterns were all procedurally generated, so no RAM was needed. The SFE logo was stored as 4 bit greyscale in the internal ram, and a certain rectangular area was multiplied by a constant to generate the red “flame”.The white background wasn’t part of the stored image.

If there is enough interest in this, I might be persuaded to develop a board with some DDR ram onboard, and the FPGA. The FPGA has nowhere near enough onboard ram to store an entire image using 24 bit greyscale.

What do people think would be an acceptable price point?

–David Carne

I dont know about the rest of the group, for me size & cost is what matters. the board would have to be smaller than the screen itself.

i would just need a simple interface. ~CS, D0-D8, A0-A8.

one could use the A0-A8 to create registers locatons, I.E.

Address 0, X location

Address 1, Y location.

Address 2, pixel color, D0-D8 LB

Address 3, pixel color, D0-D8 HB

Smaller than the screen itself is easily doable. In large quantities I’d move to a BGA packaged FPGA which would decrease the size even more.

Pricing would be based upon a few things:

Price of ram $5-12 depending on qty. Price of FPGA $5-25 depending on features + qty. Price of bare PCB $3-$20 depending on quantity. Keep in mind that those numbers are raw BOM - doesn’t include any profit for myself; nor for the distributors.

That interface seems simple enough, but then you’re looking at least 2 writes / pixel [assuming auto-incrementing X/Y pointers]. That would take a while to fill the screen. What about wider interfaces? 16/32 bits? Or directly mapped framebuffers. Both? How about serial interfaces?

–David Carne

seulater:
out there that can made a controller for this LCD with just one chip like the original board did

Like David wrote FPGAs today don't yet have enough RAM built in for a single chip solution for this kind application. So at least a RAM chip is needed.

Then why is no one offering a FPGA based controller board solution for the display? The issue might be that is there enough interest for it? And would the users find the price acceptable.

David already listed some numbers and in addition you also need an assortment of support components and connectors. Even if the cost is not huge they still add to the BOM. You also want to consider something for the assembly. And maybe some profit for yourself and also distributor if it was offered for sale through SFE for example.

Now if we consider this:

seulater:
cost is what matters.

the numbers might just not add up.

(if you feel your quote was taken out of context, sorry about that, just thought it could be relevant here)

I have built an FPGA based controller board for the display myself that offers fast SPI or up to 16 bit parallel interface through a 40 pin 2.54 mm pitch header. Optional touchscreen support. But it is just an experiment at the moment as I don’t really know if there is more interest into something like that.

I’ve been doing a bunch of research into this, with an eye towards producing a product; I think the minimum saleable price is around $60 USD, and that would be with fairly thin profit margins.

How important is heat dissipation / power efficiency / thickness to people? I can go with a switching PS design, [thin, efficient, bit more expensive], or a linear design [thick, less expensive, hot].

Would people be interested at a $60-$75 price point? How about $80-90 display included? [The screen can be obtained cheaper in bulk].

Do people want touchscreen support?

–David Carne

Just thought I would post my thoughts.

Don’t most solutions that require an LCD of this nature already have a display controller integrated in the uC / supporting chipsets somewhere? Like an ARM or similar? What about the Epson controllers? I think I remember them being able to handle these LCDs with ease.

Again just my thoughts. For someone doing an FPGA solution, a custom controller of this nature may be worth looking at. If it were me, I would even consider a power supply-less board, and if not that, then definitely the SMPS based PS.

++touchscreenSupportCount;

hi folks,

hi busonerd,

there is already a 1,000 units / year market in 2009 - 2010 for a 105x67 mm PCB containing :

  • a switched power supply able to run from 5,0 VDC or 3,3 VDC.

  • a FPGA

  • a DDR or DDR2 connector (VRAM plane)

  • a SHARP LQ043 LCD connector

  • a touchscreen connector

  • a CPU CLK/4 serial interface (18 Mbit/s on a 72 MHz ARM - 20 Mbit/s on a 80 MHz PIC32)

  • a HDMI / DVI interface (bypassing the VRAM plane)

  • one extension A connector good for a matrix keypad (4x4 keys), a serial PC mouse, a serial PC keyboard (all of them together)

  • one extension B connector providing 72 pins (functions to be defined)

  • one hardware interrupt pin to the µC

  • a Linux / Android / Moblin automatic detection (480x272 size) in DVI/HDMI

  • an easy-to-understand GUI when used through the CPU CLK/4 serial interface containing primitives like polygon fill, line draw and character draw.

Selling price to be aligned at 75% of the LCD price. If the SHARP LQ043 LCD is costing 55 eur, then the PCB + FPGA should cost 42 eur. Excluding DDR or DDR2 (to be found at about 15 eur). Total price for such display solution is thus 55 (LCD) + 42 (PCB) + 15 (RAM) = 112 eur.

As an option, it is already possible to opt for a FPGA implementing a ARM µC at 60 MHz.

http://www.actel.com/products/mpu/corte … x#cortexm1

However, this is only a beginning as this kind of ARM processor is still built using standard FPGA cells. The power efficiency is poor. But it is worth testing that way, right now !

Any FPGA already needs an internal µC for programming the gates. In the very close future, any FPGA will also provide a ARM µC available to the user. New 65 and 45 nanometer FPGAs are expected, with such dedicated ARM µC area. This enables the ARM to run at 100+ MHz, maintaining an excellent Mips/watt ratio. Very critical for the DIY and SEMIPRO market is that the NRE (non recurrent expenses) are kept to zero, like any ordinary FPGA. This will change industry. Everybody will love and use FPGAs. There will be many tutorials about how to use those new ARM-enabled FPGAs.

The extension B connector will provide many advanced features like SATA, USB2, AC97 or HD Audio, SDCARD, CANBUS, I2C, PWM, A/D, counters, timers, interrupts and so on. It will thus include everything you need to make a GUI-driven multimedia controlling device ! With the embedded ARM µC option, you can forget about using an external µC !

The market for such combination could be 5,000 units in 2011, then exploding to 100,000 units/year in the next coming years.

Later on, it will be possible to add some hardware video encoding/decoding in the FPGA, along with a CSI2 interface good for a 12M-Pixel CMOS camera module like Sony iU060F.

Such idea of making a FPGA-based PCB acting as a 4.3 inch MOTHERBOARD for the SHARP LQ043 LCD enables us to enter the world of “systems on one chip” in a flexible, progressive and affordable way.

This is the next de-facto standard in the DIY and SEMIPRO world. It is worth trying now.

Instead of continuously reinventing the wheel, using tiny microcontrollers and tiny LCDs, instead of being frustrated by the lack of GUI (now provided by all consumer-grade devices), very soon DIY and SEMIPRO will evolve at no extra cost and base on systems equipped with standard interfaces and operated through a GUI.

Do you know http://www.elektor.com ?

They may be interested, who knows ?

They may “persuade you” starting working on this. Their lab is curently based in Holland but they are expanding worldwide. Hope you get the message.

I definitely know we are on a Sparkfun forum. There may be one “Sparkfun” version loaded with plenty of stuff coming from the Sparkfun culture, and in the other hand, there may be a “Elektor” version loaded with plenty of stuff coming from the Elektor culture. This isn’t really competition. This is Sparkun and Elektor migrating to SoC technology, basing on a common platform that is thus going to be perceived as the next industry standard in the DIY and SEMIPRO market.

Any good idea welcome.

Hi Anderson Pioner,

Would you mind sharing us the implementation using the Altera DE2 for the PSP !

I have bought an PSP and would like to see it work with the my DE1 .

Any help is appreciated!

Thanks,

Ram