IBM 5150 - Continuous tone (beep) from speaker

[modem7]

Background

See my web page at [here].

Update

Finally. I accidentally caught one of my IBM 5150 motherboards (type 16KB-64KB) doing this. I accidentally had socket U33 empty.

Over the course of the motherboard being powered on for about 30 minutes, the tone varied from continuous, to intermittent, to nothing, back to intermittent, etc.

I noted that:
- I could always get the tone back by pressing down on the 8253 timer chip.
- When the tone was present, it would abruptly stop as soon as I put my oscilloscope probe on pin 16 of the 8253 timer chip. That happened every single time.

A video of that behaviour is at [here].

Reference diagram at [here]. Going in with an oscilloscope:

- U95 (open-collector driver), output pin 6 had an alternating waveform on it - see [here].
- U63 (NAND gate), output pin 11 had an alternating waveform on it - see [here].
- U63 (NAND gate), input pin 12 had +1.65V on it. That is neither HIGH nor LOW, but U63 must have been treating it as HIGH because of what was on pins 11 and 13. (TTL floating HIGH voltage ?)
- U34 (8253 timer), output pin 17 had an alternating waveform on it - see [here].
- U34 (8253 timer), input pin 16 was unable to be measured, because the tone stopped as soon as I put the oscilloscope probe on pin 16.

 

[Malc]

Interesting, I once had an XT 5160 motherboard that appeared dead except for a horrible loud continuous ' Screech ' from the speaker, It turned out to be the U18 BIOS ROM, I replaced the U18 with an Eprom i burned and the motherboard worked perfectly fine after that.

 

[daver2]

I think what you have done is to work your way back to either a faulty U36 or the fact that U36 has not been programmed correctly by the BIOS.

The default behaviour for the 8255 is for all the I/O ports to default to INPUT when the IC is reset (i.e. all of the outputs will be floating).

Dave

 

[modem7]

I think what you have done is to work your way back to either a faulty U36 or the fact that U36 has not been programmed correctly by the BIOS.

The default behaviour for the 8255 is for all the I/O ports to default to INPUT when the IC is reset (i.e. all of the outputs will be floating).

I get the impression that you did not read the link that I pointed to in the 'Background' section of post #1.
What I believe that I have done is proven the hypothesis in that link.

A video of that behaviour is at [here].

Looking at the video, some readers may believe that my 8253 timer chip is faulty in a way. But my 8253 is still conforming to the Intel 8253 datasheet. I believe that Intel would say something like, "You have the GATE2 pin floating high, a 1.193 MHz clock coming in to the CLK2 pin, and the 8253 has yet to be initialised. And in our datasheet is, 'Prior to initialization, the MODE count, and output of all counters is undefined.' In your particular case, the uninitialised 8253 is dividing CLK2 by 10000, outputting 119.3 Hz".

Interesting, I once had an XT 5160 motherboard that appeared dead except for a horrible loud continuous ' Screech ' from the speaker, It turned out to be the U18 BIOS ROM, I replaced the U18 with an Eprom i burned and the motherboard worked perfectly fine after that.

Yes. For some (repeat: some) 5150/5160 motherboards, the continuous tone/noise will be heard if, for whatever reason, the portion of code in the POST that initialises the 8255 chip, does not execute. And so, for affected motherboards, lots of possible triggers. My motherboard is in that category. For example, I can easily trigger the continuous tone by removing either the 8088 CPU or the BIOS ROM.