IBM 5150/5160 - Hypothesis about continuous tone (beep)
I can see a continuous tone (beep) happening on some IBM 5150/5160 motherboards where the POST (for whatever reason) does not execute.
How I can see it happening is (and I may be wrong):
Refer to the IBM 5160 diagram at here.
Very shortly after power-on, the 8255 chip gets reset, via its reset pin. Per the Intel 8255 datasheet, at reset, the pins of ports A/B/C on the 8255 go into input mode. In the diagram, that corresponds to the PB0 and PB1 pins being inputs. With the circuitry being TTL, that means that the top input to the AND gate will float high (and the GATE 2 pin on the 8253 chip will float high).
In the Intel 8253 datasheet is, "Prior to initialization, the MODE count, and output of all counters is undefined." So, at this time, the OUT 2 pin of the 8253 chip could possibly be alternating. If so, those alternations are going to go through the AND gate and on to the speaker.
The difference between that and a 5160 motherboard where the POST is being executed, is that the POST then reconfigures the 8255 from its defaults: Per steps 5 and 6 of the POST at here, the 8255 gets programmed such that the PB0 and PB1 pins on the 8255 become outputs, and the PB1 pin set low and the PB0 pin set high. A low PB1 is going to stop any (if any) OUT2 alternations from getting to the speaker.
( FYI: Later, the PB0 pin gets set low as well. On my multimeter, I see that happening just before video appears on-screen. )
Just a hypothesis.
If ever I catch one of my 51xx motherboards emitting a continuous tone (beep) from the speaker, I will be sure to investigate with test equipment to see if my hypothesis is correct.