Electronics Help with LCD required. - *more* Pics added (revived)

I haven't tried powering it on at all yet - worried about frying it.
don't know what power it needs.

pretty sure it's EL backlit - it came with what i think is an inverte(/o?)r.

re controller(s): would like to be able to use a newer one so I can just solder, plug and play (am busy in my final year at the mo - shouldn't even have bought the LCD in the first place!)

 

Your display/controller board combination looks like it could be very well used if you find the signals to the controllers.

The interesting thing about this is that you have two HD61830 controllers there, each driving half of the display I think. So this brings some extra complexity but it shouldn't be more than selecting the correct controller with CS signal. HD61830 has SYNC signal that is for parallel mode to keep them in sync when driving the panel.

As for software support LCDInfo has support for HD61830 displays but it only supports 240x128 displays at the moment as that's the only one I've been able to test. Also it supports only one HD61830. But it should be possible to add support for this module when the hardware side is sorted out.

HD61830 as a controller is "just like all the others", thought it's not the first one I'd recommend for new designs. But here as you already have a controller board with them it should be just fine.

 

Ahh... that's what I thought. May want to check out that EL then. If it's as old as you say it is, the EL might be near it's EOL.

The GFX displays are really dependent on the LCD type and the controller. Most of the older controllers can't handle ANIMATION because they can't run a high enough line clock to refresh the display. Most LCD displays can handle higher refresh rates. The new SED controllers are actually quite good at this now.

As for the serial, it depends on the software you are using. A MAX232 patched to a serial/parallel converter chip, and a bunch of bit banging on the device is all you really need. In fact, it's how most people convert PowerLCD to work over serial (since all that PowerLCD does is a bunch of bit banging).

I actually use the FIO USB serial to parallel converter chip in my display, as it saves me from having to strap on another serial port. You can find pre-built boards from Dontronics if you want (www.dontronics.com). Includes drivers for you too.

That what I was attempting to say previously, was that it'd probably be more effort to reverse engineer that board, than it would be to just strip the LCD out of it and replace the controller with something better like the Toshiba or Epson. You can even use the pin-out diagram to help you determine which lines are where on the PCB on the back of the LCD.

 

i'm not hugely concerned about animation (or any high-refresh requirements).
just want to use the LCD to show Winamp status, and would like to be able to use those buttons to control it.

when i get those circuits i should be able to check whether the backlight works (and the actual LCD itself)

i couldn't find the boards on dontronics - can you give me a link?

 

Sorry, meant to say FTDI not FIO.

The way mine works is this:

USB->FTDI->SHIFT REG (serial/parallel convesion)->SED CONTROLLER->DISPLAY

I suppose you could use something a bit more advanced, like a PIC or TIGER chip in place of the shift register, but it works (just needs a bit more massaging on the software side). In addition, I believe MAX makes some serial to parallel converter chips, along with Micrel and Philips.

I had a similar type of situation, in which I was salvaging a board from a custom design. Problem was that after you get the chip all spec'ed out on the custom design, you then have to work around any other tweaks that they may have added in the form of firmware or interface additions they made (many are just for handling touch screens or buttons). It turned out to be more a pain than it was worth.

The latest controllers from Epson include everything you'd need in a single package now, making the process of re-creating the controller bits so much easier.

It's just a matter of preference I guess..

 

Yeah... Dontronics has a really crappy layout.. you need to really dig through the site to find the parts...

http://www.dontronics.com/dlp.html

The board is the DLP-USB245M USB Adapter. It is a USB to Parallel board using the FTDI chip. Works great.

Here's the link for the datasheet
http://www.dlpdesign.com/usb/dlp-usb245m12.pdf

 

Well, Henry was right about the 2 controllers handling half the display area each. If you were to switch over to a SED133X controller, you would need 2 of them setup the same way since they can only handle a max resolution of 640x200.

Dont confuse the speed of the FTDI chip with a PCs serial port. That FTDI chip can transfer 1.5MBytes per second. Serial port can only transfer (cant remember, but around)196,000 bits per second. Thats more like 24kBytes. Hence the reason why people do not like the serial port for graphical LCDs. And most discussion is about 128x64 screens, not a 600x400 screen. Thats a difference of 8kB vs 256kB.

My advice is this: Your LCD has NO DATASHEET!!!!!! The only reason why you are not SOOL is because you've recieved the controller board to go with it. The HD61830 is good enough until you figure this thing out more. If your looking for a 'plug and play' option, buy another LCD. Otherwise, your only option is:

Forget about the eurocard thingy (the one with all the white plugs in it). Get the adapter as animus said. You will want to keep the inverter, the card that plugs directly onto the LCD, and the card with the 2 HD61830 ICs on it. This card already has the LCD controllers, the external memory, driving (oscillating) circuit, and some interface. Those MC74 ICs are probably buffering the signals from the eurocard interface to the LCD controllers. If you are to get this to work: You will need to read over the datasheets for the HD61380 carefully. You will need to also get the datasheets for hte MC74 ICs also. You will then have to back-trace all connections, from IC to IC and then to the eurocard interface. After this, you will also need to write your own interface for the card by making your own VB program to access the parallel port, or a plug-in for one of the common free LCD softwares out there.

The other option is to sell this to someone else. There is a trading forum in here. This would attract the most attention. If your not willing to go through work to use this display, someone else might.

 

Yup i agree with hazer, it would be a real mess and waste of money to throw away the controller board. If you don't have time, sell it to me i do electrical hacking like people do crosswords, strange init.

 

Notice I said SED 13xx *NOT* 133x series.

Epson has expanded the 13 series well above 640x200. In fact, the newest SED 13xx series handles 1024x768 TFT in 24bit color.

Epson now includes a new series of variants on the SED 13xx which handles 640x480 maximum, with the frame buffer ram built into the package (64k). As a result, you only really need to interface the chip to the LCD and to your PC. I get most of mine through Digikey, but I'm sure there's more vendors out there.

That is what I'm currently using on a couple of 640x400 displays that I use for some home automation stuff I made. It's almost painless now.

I agree with Hazer, that unless you are willing to backtrace all those connections, it'd probably be easier figuring something else out. You may also have some difficulties with those buttons on the side, since they seem to interface between the LCD and the controller. You may find a buch of additional crap on the controller to handle those buttons, and might need to disable/work-around those to get at the LCD (those smaller ICs off to the side of the controller look like they are the ones interfacing the buttons to the bus).

 

You forgot to mention that the S1D13XXX series is a 16bit address&8/16bit data interface. This requiring some interface not directly attached to the parallel port (like multiple latches).

The SED133X is simple enough to get rid of all the hastles, with a simple 8 bit interface for both data and address.

Yeah, the S1D13504, S1D13705, and others are nice chips. But having 40 some input lines does not make things easier for people who can barely understand how the LCD works in the first place. Not to mention, since there is no available datasheet for this particular LCD, how will one go about getting the timing signals required for the higher-end controllers? IE- FLM, Vsynch, HSynch, etc. etc. Not to mention we dont even know if this uses 4 bit (LCD) interface, or 8 bits (4 upper, and 4 lower).

Using anything but the controller card that came with this thing would be a nightmare undeserving of a mono-graphical LCD.

 

Whoa there. I never said building the controller is EASY, I just said it might be EASIER than trying to reverse engineer the controller that *IS* there.

The clock rates wouldn't be hard to figure out, as you have the Hitachi controller sitting there in front of you. Just reverse engineer the clock (the dividers in the Hitachi are fixed) line on that Hitachi. If he can get the unit powered, he might be able to put a freq counter on the CLKIN line and take it from there.

However, if he is as in-experienced as you postulate, than he might just be better off selling it and buying a Matrix Orbital get-up instead.

 

this is getting way beyond me.
can you all vote for what you reckon is the easiest option (but include a guesstimate of the cost).
I would love to be able to buy another controller board, do some soldering, plug it in, run LCDC/LCD Smartie and have some Winamp info - someone tell me if there is any chance of this. (bear in mind that I will be receiving all the circuit information soon(ish))

 

Well, you could either do one of two things:

1. Sell it and buy something pre-built.

2. Purchase another controller board such as this from MicroPak Industries:

http://www.angelfire.com/electronic/micropakindustries/page8.html
(all you would need is the SED add on board, not necessarily the main controller board...the SED board alone can be interfaced via your parallel port to something like PowerLCD reasonably)

to use with your LCD. All you'd really need to do is match up the pinouts on the lcd with the output of the controller.

However, the latter is not the 'cheapest' option. By the time you get done with option #2, you'd probably have spent as much as if you were just to go out and buy a Matrix Orbital kit.

However, you wouldn't learn as much in the process...

 

Just to add: You would need 2 of the SED1335 boards. They only handle 640x200.

The other thing thats a problem here: If you are looking to find a way to make this work with the currently available LCD software out there, I dont think any of them handle 640x400 LCDs. They CAN handle them, but none of them have written the plug-ins for that size....yet. If there is one, please let me know.

Since you are really looking for a simple way to do this, I re-itterate my previous suggestion: Sell this LCD with the boards to someone who wants to reverse engineer it and use the cash to get a more popular LCD that is already written for at least 2 common LCD software.

 

OT: kt3946, you mentioned using the S1D13X (higher end with color support) ICs for your own projects connected to a home system. I would be really greatful to hear how you interface to them. Ive gone over the datasheets a few times. The only thing that keeps turning me off is the amount of inputs they take in order to use properly. Im just wondering how you did it.

Personally, Id like to find a way that uses full-speed USB. Id also like to avoid using the microprocessors that these chips target. Just curious about different techniques that I havent considered yet.

 

USB roll out the high end atmel's

As long as you mean Full Speed USB (ie, 11mbits) rather than High Speed (480mbits) then there are some high end atmels that will do this fine, look at their MP3 player example for a circuit.

Whoever did thouse naming conventions should be shot.

On the advice front, if you get this LCD working as a learning experiance you will become richer than if you had bought a pre-made one I would always use the controller board and tailer the software, righting an LCD plugin is childs play realy, i will be more than happy to whip up ye old IOPM modifing device driver so you can access the parrallel port with ease.

If you wanna buy one, i would avoid matrix orbital like herpis and clymedia. Their graphic ones look so ugly in real life compaired to what you can buy from an electronics surpless like www.eio.com or www.earthlcd.com. Heck, even www.farnell.co.uk have displays with on board controllers considerably cheaper than MO.

 

eh? - my LCD is 120x480. (BTW that makes the pixels .625mm big)

I don't want to sell anything, or get any normal LCD - the whole point is that this is a big, blue backlit graphical LCD. (unlike anyone else's)

I would far rather take your second suggestion and get a new board and do some soldering, but I guess I'll need to wait for the circuits to arrive before I can find out whether that's reasonable.

 

Hazer, the projects I have used them so far have been with buiscuit PC boards so far. So, in a way, I've had some help because most of these board already come with some LCD support. I've seen a few examples from Epson that use a microcontroller alone.

However, what are you concerned about for inputs? The input modes are selectable, depending on what you want for an output type (ie. 4 bit won't work with the 640x400 24bit modes, etc.). I agree the documentation is a bit confusing, and maybe that's the root of the problem.

As far as Full-Speed USB, I'm not sure what you mean. Are you referring to USB 2.0 Hi-Speed as opposed to USB 2.0 low-speed or USB 1.1 speeds? The FTDI chip that I use on one project works at USB1.1 speeds, but since I'm not doing any real animation there, speed really isn't that big of a concern. Really, the only reason you would need a high speed input is if you were doing some color or graphics support.

The thing I like about these controllers is the multi-plane graphics support. Most of the time, the graphics layer is fairly static (just providing lines, and boxes and stuff) with text the big change (showing stats). I suppose that if you want to use it for a gfx equalizer or something would require more speed, but I haven't really played with that at all, however, I have been TOLD by the Epson vendor that the new chips have been specifically designed for that (mfgs have been requesting better gfx support).

 

DOH!

At 480x120, I have no idea why this thing uses 2 of those HD61830 ICs......Im now beginning to wonder if one handles graphics while the other handles text? No, couldnt be. Crosstalk would be horrible.

Hmm. It would be so nice to see what the timing signals for this would be. Since this is the case, a SED1335 board would be much better. You may be able to find a generic plug-in that would handle the resolution for software like jaLCD or LCDstudio.

Since your in the UK (?), I understand that you dont have many alternatives for getting a big graphical display.

At this point, my personal way of going about this would be to get datasheets for the column/row driver ICs directly on the LCD. Find which pins goto the connector. Read the datasheets carefully and figure out how to apply the SED1335 signals to them. Im pretty sure that the timing wont be a big deal (now that I think about it), but you do want to get the connections right. One other problem I just thought of: Most of the older LCDs require a negative driving voltage. Something to consider.

kt3946: I have a couple of those cheap 240x160 Sony LCDs (earthlcd has them for $50, got mine from goldmine electronics for $10). The S1D137XX series would be perfect for driving this thing, but that only handles the timing for me. I would then have to deal with 16 address lines and 8 data lines (S1D137XX takes a 8bit color and internally converts to 9bit TFT). Since the interface is designed for 16bit processors, its kinda changing the hastle instead of getting rid of it. Im trying to figure out a way to get data from the PC to this thing. So far, Ive come up with my own design using EPM7128S PLDs to use a double buffering scheme. The only thing thats a real problem is getting bits from the PC to the memory buffer fast enough. The serial and parallel ports are too slow. I need to transfer 75kBytes per screen. If I want to get close to the 60Hz refresh rate, thats around 4.4MBytes per second. With no overhead. The FTDI chip would only get me about 13 frames per second. Oh well, it would be nice to find a 16bit USB high speed serial/parallel converter.

 

Okay, I pulled open the specs on your Sony LCD and found some interesting things for you:

1. The LCD only does 512 colors, and uses a 9 bit RGB interface (3R3G3B)
2. The Rise/Fall pixel response ratio is 80ms max 20ms typ. [40ms Rise + 40ms Fall] (this is probably black-to-white). It's gray-to-gray time may be MUCH higher.

So, with this in mind...

240x160=38,400 pixels
9 bits per pixel * 38,400 pixels = 345,600 bits per screen
345,600 bpscr / 8 bits in a byte = 43,200 bytes per screen
43,200 bytes / 1024 = 43k per screen

not 75k per screen.

Since the refresh rate is 60hz = 0.167 seconds per cycle, you are looking at a refresh time of 16.7 milliseconds. Since your Pixel Response rate is 20-80ms, you'll be lucky to see a real pixel response beyond 20 hz (regardless of how fast the lcd controller wants it refreshed). This is because the controllers are generic, but the LCD crystals are specific to each design. Just because you refreshed the line driver, doesn't mean that liquid crystal is going to dance for you....

So, using the new data:

345,600 bits per screen * 20 screens per second = 6,912,000 bits per second data rate required.
6,912,000 bps / 8 = 864,000 Bps / 1024 = 843 KBps data rate.

Needless to say, a serial port isn't going to be able to handle this, however, a parallel port, or a USB 1.1 port is still going to have enough bandwidth to handle this data rate. You should be able to make this work just fine.