You are very welcome Chris. I am glad it has restored your machine's sound.

I notice that you do not have a connector to the "dot power". I assume you have a low voltage display (probably ColorDMD) so you should remove F601 and F602 to prevent the high voltage section from being energized.

I have been working through some big board builds. The one outstanding one I picked up today/yesterday. I will be working on that in the coming days but will likely overlap the dot matrix display with it.

I started building the DE board despite not knowing if/when my package sent with FedEx would arrive. It showed some progress on Wednesday when it showed up in Memphis! That's right west coast to Memphis and back to west coast. You have to love FedEx. They love their Memphis hub. I have seen this before so I suspect the recent weather events in the center of the continent were the reason for the delay. I have also seen packages end up in Oakland and then up the west coast.

tracking.jpg

Needless to say the package arrived today. It contained the 74LS126 and 74LS273 that I needed to get some functionality on the board. It also had a resistor network in a value that I needed. I was still missing one other value (due to arrive tomorrow = Friday) but the board is pretty much complete without it so I substituted a similar value (it is a pull-up / current limiting resistor so its value is not critical). When I'm very close to being able to test a board I will do what's needed to get the answer!

So I got my answer.

boards_121.jpg

There are two assumptions that follow on from that.

1. Passing on the bench is a very good indicator that the board will work in a machine. Boards that pass on the bench work in a machine unless the machine has a wiring error (a fault not due to a fault in the board).
2. The alphanumeric display exercises all (except comma1+2, comma3+4 and blanking) signals. Dot matrix displays are driven through the 26-pin ribbon cable so if the alphanumeric display passes the dot matrix display data will be correct.

Having said the above I now need to find a dot matrix Data East machine to actually test the board in. I have a Laser War so I can test Laser War that has special support for that game only. I think an optional alphanumeric Data East machine would be nice but necessary as Laser War and dot matrix will cover most every base.

Now to find that dot matrix Data East machine ...

In my copious amounts of free time I took a closer look at this the other month. I went to the Olympia Pinball Museum (free plug for Welby) a while ago to help with a Junk Yard and saw a Flipper Football.

I took a look at the display and noticed that it's going to present a problem. Vishay squeezed 16k pixels into what was 4k pixels in roughly similarly sized physical dimensions. This means that pixel size and spacing (pitch or stride) is different (smaller).

The specification actually documents the above dimensions.

• 128x32 (4k pixels). The pixels are circular and spacing is equal horizontally and vertically. The pixels are 1.27mm in diameter and spaced 2.54mm horizontally and vertically.128x32.jpg
• 256x64 (16k pixels). The pixels are oval (or rectangular) and spacing different horizontally and vertically. The pixels are 0.635mm by 1.016mm and spaced 1.27mm horizontally and 1.69mm vertically.256x64.jpg

The 128x32 display is readily made because there are 8x8 blocks and 16x16 blocks that "fit" the measurement criteria.

• The 8x8 blocks are 1.9mm in diameter and spaced 2.5mm horizontally and vertically.
• The 16x16 blocks are 1.7mm in diameter and spaced 2.5mm horizontally and vertically.

The 256x64 display would require a new block to be made that meets the measurement criteria in order for the 16k pixels to fit in the assigned "viewing area". Neither the 8x8 nor 16x16 blocks will work as they are just too large to fit. Consequently I need to work with the block manufacturer to see if it is even possible to make a block that meets the measurement criteria and how much it will cost to have that custom block made. I haven't contacted the block manufacturer. I intend to do this at some stage in the near future because I need more 16x16 blocks made (I had a small number made for prototyping). I do not have a time frame because I want to build out the 16x16 block display first before committing to purchasing more 16x16 blocks.

Yes. I corrected my post above from circumference to diameter. It's a diameter specification.

The 8x8 is 1.9mm and with 16 blocks lined up the viewing area is larger than the 16x16 that is 1.7mm with 8 blocks lined up. Although not by much. The lens covering the LED in the block is visibly larger in the 8x8 than the 16x16.

The Vishay specification says the dots are 1.27mm in diameter. When I put a 16x16 block next to a plasma display the plasma dots are visibly smaller than the 16x16 block. What's odd is that it visibly appears the difference between the plasma to 16x16 is smaller than the difference between the 16x16 to 8x8. Which is odd because the nominal measurement difference is 0.43mm to 0.2mm. When I look at a measuring tape I can't the difference between 0.2mm so I suspect that the specification of 1.7mm diameter is a little larger than it actually is. It might be closer to 1.5mm - i.e. closer to the middle of the difference between the 8x8 and the plasma.

The dot size is: plasma < 16x16 < 8x8.

I think you'll definitely want the 16x16 if you're looking for an LED replacement to a plasma. As I said above the dots are visibly bigger but not much bigger. Without the 8x8 blocks to compare it would pass for a very similar size. It's not easy to capture an image of the comparison due to parallax error. I think to have any chance of removing the error I would need a tilt-shift lens.

I plan to try to get an 8x8 as well as a 16x16 panel in a machine and take some pictures so it can be seen how it will look. It's in the pile of work.

This is the comparison between 8x8 and 16x16 in WHITE.

It is VERY hard to capture due to the high dynamic range required.

• 8x8: The blocks show a gap between them. It's possible to change block layout so they are closer but the closer they are together the harder they are to insert when building the board and the harder they will be to service if/when a block goes bad. The dots are also visibly larger versus the 16x16 when you look at the black on white images.
• 16x16: The blocks do show a gap but you have to look hard and it is only more easily seen in the black on white image. The gap is due to soldering job on the blocks. It is possible to keep the gap minimal if you solder one in and then align each subsequent block. There is also variance in how close the block sits to the board due to the molded plastic of the block. The dots are visibly smaller than the 8x8. You can see the amount of black between each dot in the black on white images.

White on black images.

• 8x8boards_122.jpg
• 16x16boards_123.jpg

Black on white images.

• 8x8boards_124.jpg
• 16x16boards_125.jpg

I definitely like the 16x16. I will try to get images from a real machine (probably Terminator 2) when I next get a chance.

Tonight was machine verification night. Two posts for two separate topics.

The first is the Data East CPU board. I went to test it the other day in a friend's Tommy. It worked great - as expected. No images.

I tested it tonight in Laser War. Laser War is the odd machine in that the SRAM is located at a different address in the memory map so I wanted to verify the jumper settings before declaring victory. So here it is ... working in my Laser War.

boards_126.jpgboards_127.jpg

I will update the first post with availability and bare board pricing. I had five boards made so there are four boards available - either bare or complete. It is compatible with all Data East games.

Testing moved to the DMDs. Several different colors and block sizes. Also shown is a filter. The Terminator 2 I have came with a red plastic filter to make the orange plasma look red. I put that back in to see what would happen to the white display.

Unfortunately I only have the red filter. The brightness can be adjusted low to produce something akin to gray - I guess.

Display images are difficult to capture on a digital camera due to the high dynamic range required. In real life the displays are sharp, crisp and bright. If you're not into ColorDMD and want an upgrade this is a good solution. The high voltage DMD "buzz" is no longer present when using these displays.

The white DMD is BRIGHT. I think it's beautiful and might consider replacing a lot of my displays with it. Brighter than the original plasma and low voltage!

The orange DMD is brighter than the original plasma.

The base brightness is set using a 1.8k resistor with a pot in the final path. Higher brightness levels can be achieved by substituting a lower resistor value if desired. If the brightness is excessive the pot can reduce it to the desired level. The low voltage for this test was taken directly from the J116/J117/J118 connector - directly from the +5V and not the +12VU with regulation. On the master board the LED and logic voltage is drawn from the same source but it should be possible to draw them from different sources to alleviate any potential drop in the +5V source to reduce potential for watchdog resets. The datasheet for the LED driver say it has a maximum output supply voltage of +17V but +12VU is unregulated (nominally at +12V but often measured at +14V) and it would probably be better to regulate it although the driver IC is constant current output. I would need to experiment to see if this is truly possible.

• The board itself.boards_128.jpg
• The white DMD with the red filter.boards_129.jpgboards_130.jpg
• White DMD using 16x16.boards_131.jpg
• White DMD using 8x8.boards_132.jpg
• Orange DMD using 16x16.boards_133.jpg
• White DMD 16x16 fit.boards_134.jpg
• White DMD 8x8 fit.boards_135.jpg
• Orange DMD 16x16 fit.boards_136.jpg
• 8x8, 16x16 and original plasma size comparison. Parallax error compensation is required.boards_137.jpg

I plan to look at the datasheet some more. I intend to revise the master board so I'm thinking if I don't explicitly create a prototype board I might just put the option for a different supply for the LED and a jumper to allow the current implementation.

None of the Data East manuals seem to document the -4 version. The PinWiki mentions a -3 board with a -4 label but the schematics in the very last manuals (Baywatch and Batman Forever) only document the -3 version. If anyone has images or knows of the differences I can certain add any if there are differences but without the information there's nothing I can do.

Williams boards used these three pins for three signals.

• Comma for player 1/2.
• Comma for player 3/4.
• ~BLANKING to disable the display if the blanking signal is stuck.

In Williams games which only had numeric displays the BCD indicates the digit for player 1/2 and player 3/4 but did not contain information for the comma. When the switch was made to alphanumeric segments instead of BCD data the ~BLANKING signal was moved to the 26-pin ribbon cable connection and the comma signals were rendered obsolete.

Data East essentially duplicated (they may have actually licensed although I don't know anything about the history) the electronics - including the 3 extra signals in 1J3 (CN3). Since Data East games started with Laser War and are all at least alphanumeric the signals were never used. They are re-purposed as additional features. I don't know enough about Data East games to know for sure which games use the signals.

The reproduction board includes the circuitry for the signals. If that's the only difference then the board is compatible (provided a 12-pin header is installed instead of a 9-pin header).

de_cn3_10-12.jpg

It's compatible. PM if you're interested in one. I have boards available.

For WPC-89 Williams switched coin door interface boards with Star Trek (TNG). They switched it again with WPC-95 but only changed one connector (and key pin configuration).

I try to keep all boards in stock as much as possible but I am discovering that demand is 150% unpredictable. The busier I am the harder it is for me to track everything going on and stock counts sometimes (rarely) suffer.

I have never seen one populated in any board I have ever seen. I try not to spend time on things that I don't know anything about, have never seen used and have no way of verifying correctness.

From what I have seen boards were often designed with options that were never used or they remained static (fixed configuration) through the lifetime of board usage. I generally remove them because it's extra complexity that must be verified and confusing for any end user.

If you have information on what the RS323 connector is for that would be great.