Trying to understand serial to parallel circuit

This is why I like this forum above others; always great help and folks are encouraging rather than saying what a dummy you are (even when I make dumb mistakes like I did here not finishing mapping everything!)

beacon_dave said:

Looks like your schematic is missing an inverter gate as well.

Wait, I am counting 8 traces going to the 74LS14

pin 1 (inverter 1A) -> DB9 pin 2
pin 2 (inverter 1Y) -> 74LS164 pin 8 (CLK)
pin 3 (inverter 2A) -> DB9 pin 4
pin 4 (inverter 2Y) -> Centronics pin 1
pin 10 (inverter 5A) -> Centronics pin 13
pin 11 (inverter 5Y) -> DB9 pin 6 (joystick port trigger input)
These 6 are on my schematic, and then there are
pin 14 (Vcc)
pin 7 (GND) - these 2 I didn't know how to show as KiCad only gave me inverters for the 74LS14, no options for GND or Vcc pins..

Is there another inverter? I did get the labeling wrong for inverter 5A/5Y on schematic, it says 5/6 when it should be pins 10/11.

beacon_dave said:

Looks like Out 5 heads to the 9-pin connector.

I think this is correct though?

beacon_dave said:

The 8 data lines are there and mapping to the correct printer pins. 

Thank you writing out this mapping! Wow, I really screwed up the mapping the first time I did this (it was rather late). I was getting the numbering mixed up thinking the lines were odd numbers 1-17 in the row closest to the board edge, then 2-18 in the next row, then 19- , and finally 20-36 in the last row. No wonder I couldn't get it mapped correctly. Now that I see it goes 1-2-3-4 and follows the ribbons "down", and then starts over at 5-6-7-8 etc.

Ok, mislabeled inverter then. Three of the six are used.

Connectors can be a little bit tricky with their numbering schemes at times. Ribbons are a bit more predictable, although here it looks like you have a 9-way ribbon terminated in a 10-way PCB connector so need to watch out for things like that.

It looks like you should have a more sensible looking schematic now. 

opalko said:

0) Can someone explain how this whole interface works!?

Clocked serial data comes in via the 9-pin Atari interface into the shift register. After input 8 clocks the shift register holds 1 byte of data and the printer is strobed to tell it to look at the 8-bit data in parallel. There are some additional control lines for the computer to reset the printer and for the printer to tell the computer that it is selected and that it is ready to receive more data or not. The cycle repeats, 8bits clocked in, 1 byte clocked out. Some signals are active high and some active low, so logic inverters are used to flip the signal as required.

The 'magic' pretty much happens in the shift register, so take a look at Karen's video on shift registers (serial in parallel out mode). 

opalko said:

1) How is the data being read on the printer side?  Is the printer polling every data pin on the parallel port at once regardless of how many are actually being used?

Once the shift register has 8 bits clocked into it, the data bus has 8 bits set. The printer is strobed by the computer to tell it that the data is now ready to be read. All 8 data pins are being used, they hold a byte value in binary with high and low voltages representing 1's and 0's. If the printer is in plain text mode you will find that the data is ASCII values for each character being printed. There are some 'magic values' that tell the printer to change its mode rather than printing the character. You need to refer to the printer's manual to see what those values are. Some are pretty standard, others are printer specific.  

opalko said:

2) Since there are really only 2 address data pins (3 & 5) being used here, it doesn't seem a vast improvement over serial?  But I guess because the Atari only has pins 1-4 for I/O and 1 of those sends the data to the 74LS164?

I think we can probably skip over this one due to the errors found in the original schematic. Probably less of a performance issue and more of a compatibility issue as the Atari didn't have a Centronics parallel interface and a lot of printers didn't have a serial interface, so a convertor was required just to get the two to talk. Also with home printers, speed was often limited by the speed of the printer mechanics. Some had larger memory buffer options that allowed you to dump the data into the printer and leave it to do its own thing whilst you used the computer for something else, but that bumped up the cost. Some commercial line printers did however go really fast. 

(...which reminds me a531016 how about a tear down of an old line printer for an episode of The Electronics Inside ? )

Re: you have a uni-direction printer cable. When the printer is busy it sets 'BUSY' active and the computer stops sending data. conversely, when the BUSY is NOT ACTIVE the computer will send data. The same is true for the ON-LINE Signal the must be active for anything to work.