V9938 with SCART at LCD-TV

This day I tested my V9938 RGB-output-circuit at the SCART-input of Toshiba and Samsung LCD-TVs and it works perfect:

V9938 with RGB output

With the 9129 it's possible to write with maximum speed to VRAM if the screen is off.

This is not possible with the V9938. I don't know why? - I have to take a look in the datasheet.

Now I write with some NOP's to VRAM and it works...

 

 

After I viewed some old philips MSX2 schematics I built these emitter followers at the V9938 outputs.

 

Modifications to VG8235:

 

- NPN-transistor for the csync-signal with 1k basis-resistor and 220 ohm emitter resistor.

- I changed the R,G,B emitter resistors from 100 to 220 ohms.

 

-> This works fine with my 1084s. It has 150 ohms input resistors. Most TV's have 75 ohm input resistors. There it should work too.

 

Maybe this doesn't work with all monitors / TV's, but I have a perfect picture :-)

In the SCART-connector you have to connect CSYNC to pin 16 and pin 20!

First test of Yamaha V9938

Yesterday the "hard to find" 21.477 MHz crystals arrived from Aliexpress :-)
Today I soldered the SDIP64 adapter PCB and tested the V9938 with the R65C02 @ 3.58 MHz (Yamaha cpuclock =1/6).
I connected the 32k SRAM with the latch - like at the 9129 - and it works!

I connected the composite out with an simple emitter follower to my Commodore 1084S.
You see - there's something wrong - it looks a little bit shaked ;-) - but I have a picture!
Tomorrow I will check the SRAM connections and test the circuit with a 16-bit DRAM.

First VDP-testing with 65c02

After some succesful VDP-testing with the ATmega 328p I decided to test it with the real ;-) LC64.
I connected a 32k SRAM with a 74F374 latch clocked by RAS at the DRAM-interface of the VDP, because I had some problems with real DRAM at the breadboard. (The "Steckschwein" developers had the same problems)
Today starts the first tryout with the TMS9129 and the 65c02 on the breadboard. I clocked my Rockwell CPU with the clock-output from the VDP (10.73MHz / 3).

I created three 8x8 characters to show something on my 1084s.
Do you like my Smileys?









This is CSW to the VDP (upper signal) and an adress-line.
The timebase is 40ns/cm.
The CSW has 26ns delay because of the 15ns-GAL-adress-decoder and an OR-gate for combining with R/W.
The 9129 wants the adress-line stable 30ns before / 30ns after CSW goes low.

I think it's OK.





Here is CSW with a data-line (lower signal).
The timebase is 40ns/cm.
The data-line is stable 175ns before / 40ns after CSW goes high.
The 9129 wants the data-line stable 100ns before / 30ns after CSW goes high.

This is perfect.

 

Better breadboards

After I had a lot of problems with some low-quality breadboards I buyed some new boards.
I think the best boards are in the MB-102 package with in the middle cutted power lines at the sides:

I got other breadboards from different chinese dealers with very low quality. The contacts have very low force:

Now I rebuilt the LC64 with only one 65c22 and a 128k-SRAM to minimize the wireing.
I added adress-buffers to add later the AVR for the TFT.

SRAM replacement for 9129

For more stability I built the SRAM-replacement from " Tom LeMense ".
But because I have the 9129 (not the 9929) I only need one 74F374 latch clocked with RAS.
With the SRAM the VDP works much better.

I made a test-picture with a small AVR-program to adjust the colors with the resistors in the RGB converter.

First VDP 9129 test with AVR

I just buyed some TMS9129 and wanted to test them for later use in my homebrew design.
After many problems I got a picture with SCART at my old Toshiba-LCD-TV.
The DRAM is very instabile at the breadboard.

 

I wrote a loop that fast rewrites the characterset in 16 colors to the screen and tested it on my Commodore 1084s monitor. I think the problem is the dram-refresh at the breadboard. The screen flickers because of the rewrites - but it works!

To convert thy YPbPr output to RGB I built the interface from " Peter Zevenhoven ".



 

Problems with WDC W65c02

I made my new PLCC-adapter with 4 jumpers switchable between Rockwell/CMD and WDC.
After I got four W65c02-10 I tested them all on the breadboard.

Everytime I had stability issues with my LC64 on the breadboard I changed the WDC CPU to the Rockwell - and all is OK!

• The WDC is not TTL/NMOS-compatible
• W65c02 has a more critical timing - it's for 10 MHz.

Also I tryed the BE-input to disable the outputs with the AVR. Than the AVR read the screen-memory from the RAM and printed it to the TFT.

But the WDC runs not stable in the system and I decided to use the good old Rockwell and adress-buffers with OE. The Rockwell 4MHz CPU runs without any problems @ 5 MHz - that's enough for me and the fast 90ns-eprom.

GAL-Programmer finished

My GAL-programmer is ready to burn!
I made a schematic and a 10x8cm single-sided PCB layout with Eagle CAD.
There are sockets for the 16v8 / 22v10 DIP and a wide 24-pin-socket for the 22v10-PLCC-adapter.
I added a 74HCT245-buffer for better printer-port voltage compatibility.
The 15V-burning-voltage is provided by a TL497 switching regulator.

Thanks to ChaN (1997/1999): Simple GAL-Programmer