DGT Serial E-Board: First Impressions and Setup

I'm lucky to have access to a DGT E-Board for possible use in the scholastic tournaments and the Space Coast open in April.

Of the serial, USB, and BlueTooth forms of the board, serial was chosen because of its proven track record in large tournaments and ability to expand out to many boards, with a longer distance than USB allows. The boards are somehow connected in a bus with the software reporting "DGT Bus Protocol". I'm used to serial being a two-party point-to-point arrangement, so in a later post I'll try to explore exactly how multiple boards share the serial lines.

Around $1100 purchased:

• Serial Tournament e-Board (board and pieces)
• DGT-3000 clock
• "Connection Material" for 1st Tournament e-Board

Beware, I'm a critical consumer, especially when a considerable amount of money has been dropped for something. The first thing that I didn't like was that the DE-9 female connector and plate is not flush mounted with the side of the board. It's sticking out quite a bit, ruining the finished look.

Like the DGT North American Clock, the DGT3000 feel light and hollow to me. The former seems to get great reviews on Amazon and the chess subreddit, but can't help but feel that their build is inferior to the stocky Saitek Mephisto. And it came with a single AA battery, when two are required!

The user manuals are just inkjet printed pages stapled together, but this is 2017 so the latest PDF's from the website obviate these physical copies. Here is the one titled Installation Manual:

The wiring diagram shown here doesn't match the cables that come with the kit (see below for full parts list). After verifying that I have the latest Rev 1603 manual from the DGT website section containing manuals I nearly gave up before finding that it's the LiveChess software manual that depicts the serial setup a bit better:

But what is the 10252 "connection box"? There's none in the $300 "Connection Material" kit:

• 2 cat5e patch cables: 20m and 5m lengths (ValueLine brand) $15
• 4 RJ45 splitters (ValueLine brand) $5
• a 12V/1A AC adapter (Sunny brand, SYS1460-1212) $20?
• USB to RS232 (EasySYNC ES-U-1001-R10) $26
• male DE-9 to male cat5e and male 3.5mm stereo jack $20?
• female DE-9 to male cat5e and female power jack $20?

Those connectors sound like some custom work, but how complicated can they be? In model trains, Lionel sells a similar connector including power supply for $38. I couldn't resist peeking inside:

Some type of power regulation stuff? A cap, diode, and resister? It's beyond me, but it's not just spliced wires, as I naively assumed. Anyways, it's up to you whether the price is fair, but a penny pinching buyer may be able to assemble this kit themselves. Here are some pictures of the first connections:

The power cable in the top right leads to a wall outlet, and the male USB in the bottom right leads to my computer. Here's is a wiring diagram that includes my estimate of how additional boards would fit in. Connectors in green, wires in purple:

Now that the board and clock are wired, let's get the computer hooked up. While other software can communicate with the board, that's out of scope for this post where I'll use the intended LiveChess software from DGT. Upon plugging the board in, Linux reports:

[125867.211017] usb 2-2.2: new full-speed USB device number 8 using uhci_hcd

[125867.320186] usb 2-2.2: New USB device found, idVendor=0403, idProduct=6001

[125867.320190] usb 2-2.2: New USB device strings: Mfr=1, Product=2, SerialNumber=3

[125867.320191] usb 2-2.2: Product: US232R

[125867.320192] usb 2-2.2: Manufacturer: FTDI

[125867.320193] usb 2-2.2: SerialNumber: FTYQGZEA

[125867.352676] usbcore: registered new interface driver usbserial

[125867.355062] usbcore: registered new interface driver usbserial_generic

[125867.355198] usbserial: USB Serial support registered for generic

[125867.362643] usbcore: registered new interface driver ftdi_sio

[125867.362731] usbserial: USB Serial support registered for FTDI USB Serial Device

[125867.362796] ftdi_sio 2-2.2:1.0: FTDI USB Serial Device converter detected

[125867.363611] usb 2-2.2: Detected FT232RL

[125867.363614] usb 2-2.2: Number of endpoints 2

[125867.363616] usb 2-2.2: Endpoint 1 MaxPacketSize 64

[125867.363617] usb 2-2.2: Endpoint 2 MaxPacketSize 64

[125867.363618] usb 2-2.2: Setting MaxPacketSize 64

[125867.366861] usb 2-2.2: FTDI USB Serial Device converter now attached to ttyUSB0

So from your computer's perspective, there's just your typical FTDI emulated serial bus on /dev/ttyUSB0. Now say what you want about Java, its use in LiveChess allowed it to run on Linux with no steps other than unzipping it and invoking the runtime:

a@ubuntu:~/Downloads/LiveChess-1.4.8$ java -jar ./livechess-1.4.8.jar

I'm very pleased that DGT supports Linux directly, not just by accident (because its software is Java). There's even tips about creating udev rules so that you don't have to run LiveChess as root to access the serial device (section 3.4 of the DGT LiveChess manual).

And I'm also very happy that DGT doesn't burden the user with licenses and key files and other hurdles like registration. The software is available for download immediately to any visitor of the website.

The GUI enumerates serial ports and finds the board by ID, which is the product serial number from the side of the board. Direct and simple!

There's a little bit of boring steps you need to do in order to actually get a game started, like creating a tournament, players, and a round, but when all that routine is done, I'm happy to report that it works:

One question I was very interested in where the intelligence lies to determine what is or isn't a move. If the board sees simply the position of the pieces, then during the act of sliding a bishop across a diagonal or castling, it will see multiple board states that should collapse into a single reported move. Does the board maybe wait for the pieces to settle? If so what's the sampling rate and would it work for blitz? Does it even know whose move it is?

The answer is that the board captures and remembers every unique board state it can, and does not appear to know the player turn. The LiveChess software performs the abductive reasoning to determine which chess moves would have explained those board states.

This is evident in the "moves" and "memory" tab in LiveChess when you try to trick it by, say, walking thing king around the board before settling it on g1 before moving rook to f1. LiveChess will show all those intermediate moves, and condense it to O-O at the end.

The next step is trying to stream a live game to chess24, but that's for another post.