This is an update to my SP0256-AL2 Allophone Speech Synthesiser project. To recap: The board presented here is designed to reduce the number of IO pins required to interface the vintage GI SP0256-AL2 to modern microcontrollers. The speech chip has 6 address lines that enable the user to access any one of 64 allophones. Using these 64 allophones it is possible to mimic any sound that makes up human speech. There are 3 control lines: ALD, SBY and RST. ALD, or Address LoaD, tells the SP0256 that there is a valid address on the allophone address bus and SBY (LOW = Busy), or StandBY, is an output that is used to signal that the SP0256 is busy. The final control line is RST (Active LOW), ReSeT. The address lines are setup via a serial to parallel converter (74HC164), you would then test the SBY line in a loop until it is HIGH, then pulse the ALD line low. Then loop round load the next address etc.

In this version (3B) some tracks have been re-routed, I've removed an unnecessary 100n multilayer ceramic, split the speaker off on a separate 2-way JST socket and added a 1K SMD resistor in the SBY line. The main reason for which is to protect the SBY line if it is accidently connected to an output from the host microcontroller (which would certainly damage the SP0).

I have attached a copy of Chiptalk to this page, this is a text to speech converter/emulator for the SP0256 from 2003 (Tested and working fine under Windows 10 and 11). Type normal text into the box on screen and the program will say it, and print the corresponding string of allophone addresses that you can add to your source code as data (Please see Sample Arduino code below).

Chiptalk is also available on my old website: Do a search for 'purezx weebly'.