Simulating resistance for buttons

MAS3:
It looks to me like you did not completely understood the schematics.
You connected pin 9 to what looks to be a pull up resistor.

No, I got that. I figured it was a pullup resistor, but something I thought of today was that
there was a 470pf cap across the leads before the resistor ladder. So, I was wondering if the combination
of the 4.7k ohm resistor (inside the radio) with the cap (at the resistance ladder) was, in fact, functioning as
a low pass filter? And, since I interrupted the +5V line and didn't include a cap on the radio side of the cut wire
if I added it back in would it work? I'll keep looking.

MAS3:
That is actually what is does, but only until a key is pressed.
In case a key is pressed, it is part of a divider.
The keys have an array of resistors, and that key that is pressed (in case multiple keys would be pressed, the one most to the right will set the value) connects to 0 volt (GND).
So in case the 5th key is pressed, you can add 1K5 + 1K + 1K + 680 + 470 (i'm guessing because i can't read all values) so that would become 4650 or 4K65.
You will have an array of resistors forming 4K65 and an other resistor of 4K7.
These two values have 5 volts across, and at their joining point you will have about 2,486 volts.

2,486?!? I assume it is actually 2.486 (less than 2 and a half) volts?

MAS3:
This is quite close to a value actually measured by yourself.
You cannot just make this measured value and insert it instead of the resistor/key.

I would just connect pin 9 to the input of the unit, without this 4K7 resistor to 5 volt.
Thus, put your analog output directly to the input (that arrow pointig out) but not to the existing resistor.

I had thought of that. The resistor is inside the radio, and I cut a trace on the board that I thought would do exactly that, but I still have 5V out, so I assume I cut the wrong trace. I'll fix that tomorrow. I figured if I could keep the 5V output from getting to the point where the radio looks for the signal from the SW buttons, and feed my analong/PWM output directly to that point that it should work. If I use a LPF then I might be able to return the trace inside the radio to stock and do everything outside the unit. That would be a plus.

MAS3:
You still need to make the PWM output an actual analog output, because else it's still unlikely to work.
And of course you always need to connect the GND of your Arduino, to that of the unit you want to control.

I have a 7812 taking the voltage in the car down to a little under 12V so it doesn't exceed the input voltage capability of the Arduino.
Grounds are all common to the car.

Jay