[vcf-midatlantic] much ado about chording keyboards

[Jonathan Gevaryahu]

On 11/29/2019 4:40 AM, Jonathan Gevaryahu wrote:
> On 10/1/2019 10:24 PM, Jonathan Gevaryahu wrote:
>> On 9/29/2019 1:14 PM, Eric Rangell via vcf-midatlantic wrote:
>>> Thanks Jeff.
>>> To give context to this discussion, I built a chording keyboard for 
>>> my Replica 1 and Apple 2+ by interfacing an Arduino to the keyboard 
>>> socket and using 5 arcade momentary switches from Micro Center.  I 
>>> coded it so that my fingers chord a hex nibble and the thumb strobes 
>>> it into the Arduino, and the LED shows whether you are chording the 
>>> high nibble or low nibble.  It is possible to recode it for other 
>>> chording mechanisms and add switches to choose a mode.
>>> Eric
>>>
>>>
>> Eric, that reminds me: did you ever get a chance to probe out all the 
>> outputs of the GI "AY-5-3600-931-B6" (second sourced as SMC 
>> "331-0931-B KR3600-070" ) from the early apple2/plus keyboard, for 
>> every possible keyboard input, shift+input, ctrl+input and 
>> shift+ctrl+input combination?
>>
>> I.e. what data in B1-B10 (pins 14 down to 5) is returned, if any, for 
>> every possible unique X-bus output (on pins 40 down to 32) and for 
>> each unique X-bus output, every possible Y-bus input (on pins 26 down 
>> to 17); this whole process gets repeated 4 times, with neither shift 
>> (pin 29) nor control (pin 28) asserted, shift asserted, control 
>> asserted, and both shift+control asserted.
>>
>> See 
>> http://www.bitsavers.org/components/gi/_dataBooks/1980_GI_Microelectronics_Data_Catalog.pdf 
>> page 3-23 (pdf page 134)
>>
>> I'm trying to improve the accuracy of the apple2 keyboard emulation 
>> in MAME/MESS.
>>
>> As far as I'm aware, the 331-0931-B uses the very last mask option on 
>> the list, where pin 4 is AKO and pin 5 is B10.
>>
>> I'm guessing these tests will require an arduino mega to read the 9 
>> X-bus outputs in a fast loop and sequentially supply a 10-bit y-bus 
>> input when the x-bus reaches the correct 'count' being probed, and 
>> immediately after that either poll the AKO pin, and if it goes 
>> active, immediately read the 10-bit B bus, or just read the 10-bit 
>> bus and assume it detected a key-down.
>>
>>
> As it turns out, Sean Riddle decapped one of these, so it should be 
> possible to create an exact replica on a microcontroller, or out of an 
> ay-5-3600-pro...
>
> and it also turns out, about a week ago, someone dumped the eprom from 
> a clone apple2/2+ keyboard which DOES use an ay-5-3600-pro plus an 
> eprom, see 
> https://www.applefritter.com/content/clone-keyboard-eprom-data-intact
>
> So between those two things, it should be very possible to make a 
> replacement part/board for burned out SMC KR3600-070 or AY-5-3600-070 
> or 331-0931-B parts.
>
Whoops, I forgot the link to Sean Riddle's decap: 
https://seanriddle.com/kr3600.html

-- 
Jonathan Gevaryahu
jgevaryahu at gmail.com
jgevaryahu at hotmail.com