So, you are all right and embarrassingly so! I royally screwed up the power on the VIA and ACIA.
The library for those parts have the power pins as a secondary, optional, part. This usually isn't a problem, as the VCC and GND connections should still happen automatically. It just so happens that ...

Here's a development that makes even less sense: Removing the VIA exacerbates the problem. The line is stuck low more frequently and for longer. Once the VIA is reinserted, it improves significantly. The code is in no way reliant on the VIA; It is written to but never read.
I think I'm just gonna ...

As far as I can tell, the chip is seated properly. No pins are bent, missing etc. I'm away from my desk at the moment, but I'll double check later.

My analyzer has very little sample memory, so I ran it slower. At faster speeds it shows the osc running at 1.8432MHz.

The regulator and caps are in ...

You are right on the money, Garth. We have some kind of AC problem, not logic.
As a shot in the dark I put a 4.7K pull-up on the line. Now it works perfectly. That being said, I have no idea why that works. Nothing is wired to TXD beside a passive header and a non-populated inverter footprint ...

I've attached a picture of the analyzer's output. I tweaked the code so PORTB is high during the delay function and low while waiting on the UART. This is shown on Channel-0.
As you can see, the output sticks low, the UART spins a bit, soon reports it's ready for another character and then garbles ...

I'm using a R65C51P4 @1MHz, with CTS tied to ground.
The terminal program I use is pretty minimal and doesn't support showing frame errors. My logic analyzer, however, does. I slowed the UART to 1200 baud and capture it at 50kHz. Here's an example of what it shows.

num error RxD data ASCII
0 6
1 ...

I got a 1.8432MHz can oscillator on the UART, but it still isn't working. It's set up for 19200 baud, 8N1. Here's a sample of what I receive with my ftdi cable:
48 65 6c 6c 6f 20 57 6f 72 6c 64 21 48 65 6c 6c Hello World!Hell
6f 20 57 6f 72 6c 64 21 48 65 6c 6c 6f 20 57 6f o World!Hello Wo
72 6c 64 ...

I'm not too experienced with git, but my MEEPROMMER fork is overhere .
I also threw together a repo for this project here . Just schematics and a bare bones example program. A mouser BOM is coming.

The board lives. New VIA is in place and it can read a NES controller while blinking leds. Also, the ...

A much delayed update:
Boards have been in for a while. I ordered a batch of the MEEPROMMER boards as well.
They gave me a lot of trouble. Tried to use old, but compatible, atmega88 chips for it. The Arduino IDE, with atmega88 as target, compiles and programs without error, but the chips lack the ...

Yeah the MC1489 would have worked if it routed, but the boards are already made. I looked for devices with the same footprint and proper tolerances, but I found nothing. If I can't slap on a few passive to fix it, I'll call it caput.

To clarify I've added the schematic of the serial circuit. I want to connect this to a device similar to this

So the idea is to have TTL level serial, but make the board RS232 level tolerant.

If that doesn't make sense or can't easily be done, no worries. The microvga header breaks out all of the ...

If you need an inverter, use something in the 74 families. RS-232 standard output is a minimum of ±5V, and inputs are to be sensitive down to ±3V and must tolerate ±25V without damage. Negative voltages are part of the RS-232 spec..
Ok. That doesn't really answer my question. I don't think I've ...

The 1.2 and 1.7V is the TTL-side inputs to the line drivers, not the RS-232 side.
I don't know about that. 1.2 and 1.7 are listed under "RS232 INPUT" while .8 and 2 are under "TTL INPUTS". Am I looking at it wrong?

I'm not sure what you meant by "two inverters, two transistors."
Just a normal ...

Boards are made and on their way. Pic from the manufacturer attached.

IIRC I've been able to use it at TTL levels and have it play nice. It uses a SP213ECA as a level shifter. It's RS232 input low voltage threshold is +1.2v, high is +1.7. I think the MAX*232 line is similar.
I still need inverters though. Whats the smallest footprint for two inverters, two transistors ...