Hi Daryl,



There's not much difference, scans are exactly 64µs with a 4.3µs HSync. Obviously there's quite a few more scans (312) and equalization is very similar (with 1 more half-sync). And also crts are quite forgiving so when my first US customer to build his FIGnition tested the NTSC code he found that it worked first time, except I'd added too much of a front porch, but that was easy to correct. And PAL performance is very similar because the extra scans compensate for the lower frame rate

Like you, I use sleep to synchronise the scan data: the Sync routine's sate machine alternates between kFrameSyncScanLine (which sleeps the MCU after the HSync pulse) and kFrameSyncScanGen which calls the actual scan generator (which is in assembler) for both bitmap and character modes. Sleep is more efficient than explicitly synchronizing by clock cycles.

Personally I think it's great fun doing video with an MCU, especially as MCUs are now fast enough to compete with early video generators and in many ways are more flexible!

-cheers from julz

More progress:

I have completed coding the firmware. This included support for 4 jumper blocks to be read to select 1 of 4 inputs:

RS-232 (default)
SPI
Parallel
Serial (fed from 6522 Serial shift register)

And 3 outputs:
LCD (4 bit parallel with shift clock, latch clock, and frame clock)
PAL composite
NTSC composite (default)

I have the command structure written as well.

I am waiting on a 24MHz oscillator to test the NTSC mode with a 6MHz dot clock. This will make the graphics display more square, at a cost of horizontal resolution (from 320 to ~280).

I have included in the command structure a way to change the timing values so that users can supply their own clocking and configure their own timing. These changes are temporary and will need to be reconfigured at every power-up but will be easy to do.

I will have to test each of the 4 input sources and verify the outputs switch appropriately too. Then test some more to try to flush out any anomalies.

Daryl

I am getting closer to being done. I have added the option to use a 16Mhz clock and full 320x200 in composite mode with rectangular pixels; or, a 24MHz clock with reduced resulution but square pixels.

USART communication is fixed at 19200,N,8,1 for either clock.

Thanks to Arlet's post here:
viewtopic.php?p=18494#p18494
... I tried my EasyCAp USB capture device and was able to set both NTSC and PAL modes. With it, I was able to test my timing parameters for both systems.

I have prototype PCB's due in this week that will facilitate the testing of the other input modes. I plan to use my SBC-4 to provide the SPI, SER, and parallel inputs.

I checked back with the company I purchase my Apollo LCD and found this one:
http://www.halted.com/commerce/ccp21599 ... -80815.htm

It's similar to the Apollo and for $20 is a great deal. It would require an LCD power supply (-20vdc and +5vdc). It's not clear what the CCFL inverter requires for an input.

Once I get the IO code verified, I'll write a little demo to show it off.

Daryl

I was looking to find posts on this particular topic and your post came up first on Google search and finely i have found your post that is even slightly related to what I need Nice post. i have very impressed to see this post. i have benefited from the information you give. nice work keep it up.

_________________
james321

Thanks. I've put this aside for the time being and want to play more with a VGA interface. I may come back to this and finish it at a later date.

You won't achieve 256 visible lines on an NTSC composite signal. You only just get 256 visible lines on a PAL 625-line display. 192 or 200 lines on NTSC is possible but I think even 240 might go off the screen.

_________________
Shift to the left,
Shift to the right,
Mask in, Mask Out,
BYTE! BYTE! BYTE!

I am considering resurrecting this project and have found an LCD module that might fit the bill.

It is a 5.7” display with LED backlighting and the -28v generator is built-in. That means that the whole thing can be powered from a single 5v supply.

Data and sample photo can be seen here: http://www.vslcd.com/Products/1016.htm

The cost is ~$60 each for 10 units. However, shipping from Hong Kong was quoted at $90 for 10 units, bringing the cost per unit to $69. On top of that, there is about $30 for the AVR driver/interface board.

To recap the features, there are several 8x8 text fonts including DOS, C64, CBM, Aple][, etc. You can also download your own custom font and save it to EEPROM. Text can be scrolled in all directions. It does not wrap around but rather falls off the edges. There are also several graphics commands such point, line, box, filled box, circle, filled circle – all with set and reset function. Text and graphics use the same layer, so placing text over a graphic will erase the graphic under it. This is also a single-buffered display.

User can interface through SPI, 65C22 parallel data, 65C22 serial shift register, or USART. Video output can drive an LCD, NTSC composite video, or PAL composite video. Input and output selection is made through jumper options.

So my question is, would anyone be interested in having a 320x240 monochrome text/graphic LCD display for their platforms, at a cost of about $100 including LCD, or $30 driving a user-supplied composite video device?

Thanks

Daryl

_________________
Please visit my website -> https://sbc.rictor.org/

Daryl, I've used NHD in the past, although it is different than your experiment, I highly recommend NHD for portable App's. NHD is most helpful & they have forums from which to pose questions. This may be able to bring down costs.

_________________
65Org16:https://github.com/ElEctric-EyE/verilog-6502

I like their displays too. The closest one I could find with 5v supply is this one:
NHD-320240WG-BxTFH-VZ#

It lists for $77. These also have a controller that would allow it to be placed directly on the sbc data bus (taking bus speeds into consideration). I didn't read the datasheet enough to determine what, if any, built-in drawing commands are supported. It could be a less-costly solution.

It would be nice if they sold a plain 4-bit display without a controller - that would fit my module perfectly.

thanks!

Daryl

_________________
Please visit my website -> https://sbc.rictor.org/

Daryl, I'm interested. I have a couple of Sharp LQ6NC02's I've been thinking I should send you if they'll help you with your development. They're 5" diagonal. One of them still has the protective cellophane film over the glass. I only have one ribbon cable for them, but it has the manufacturer and style number printed on it so you could probably find more. This ribbon is from Parlex Corp., AWM style 20566, a 20-conductor flex circuit that's about .83" wide.

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

Hi Garth,
According to this:
http://forums.parallax.com/attachment.p ... ntid=88660

Those are a color LCD with 2 CCFL lamps (those require a 350-400 VAC driver). They do appear to accept a composite video input however. I'm not sure how you will generate the CCFL power. I have one supply that runs from 12v that generates a CCFL output, but I cannot find it for sale any longer.

I already have a 5" composite LCD screen that I have been testing on. I am considering buying 1 of those vslcd modules to experiment with. A single unit would run me $75 including shipping.

thanks!

Daryl

_________________
Please visit my website -> https://sbc.rictor.org/

The first thing that comes to mind is to use an op-amp oscillator followed by an audio power amplifier IC followed by a small power transformer used backwards, ie, that say a 6V secondary becomes the primary and the primaries in series would be the 230V output. It could handle the somewhat higher voltage (290-370V, 330V typical, according to page 8) just fine at the higher frequency. It probably wouldn't be as efficient at the higher frequency as it would be at the 60Hz it's designed for, but it will work.

Anyway, I see now that it's a different bird, not really applicable for this project, but they're available if you move on to do a color one later.

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

Since I already had the prototype board constructed, I went ahead and bought one of those LCD's to experiment with. I will keep my eye out for some cheap, surplus displays that will work as well.

I try to video the new LCD once I get it going.

Daryl

_________________
Please visit my website -> https://sbc.rictor.org/

I forgot to mention the open circuit voltage is usually around 800-900v. Here is a better solution:
A TDK CXA-L10A CCFL inverter powered by a 5VDC source and costing $11.14 from digikey.

http://www.digikey.com/product-search/e ... 45-1621-ND

Daryl

_________________
Please visit my website -> https://sbc.rictor.org/

I saw a tear down video the other day, saying that the cold cathode supply needs to arrange a super high voltage to strike the tube at power up. He pointed out a capacitor - I think it was in series - which apparently caused the necessary initial transient.