TT no sound

[Steve]

Check if chips related to sound have 5V and clock.

Also Dogs post at the beginning...

I'm not sure if you have done this yet, but is the PSU outputting correctly? Are you getting proper +5, -5, +12, -12?

[jacolover]

Maybe when you have time, look in TOS see if there is any mention of this switch in there ?

Good idea! I just did -- for TOS 3.06.

The DIP switch 8 does not enable/disable any hardware. It merely influences the bit for DMA sound in the _SND cookie. The logic is inverted -- as noted in Cyprians post (should have read it more carefully :( ). So, when the switch #8 is on, the bit indicating DMA sound capability in the _SND cookie is off and vice-versa.

Like I said, this does not directly influence any hardware. But software that takes into account the _SND cookie might refuse to do DMA sound when the switch is on (like in Cyprian's machine).

But let's get on-topic again. If there no sound at all (not even PSG sound, like keyclicks) in the TT there's something else amiss. @jacolover, do you have a scope? In that case for troubleshooting I suggest getting the schematic from dev-docs.org and tracing the sound from the PSG (in case of keyclicks) to the output. In that way, you should be able to see where it's faulty.

Thanks czietz, that makes the most sense right now, the schematic I have is barely readable, not sure if there is a better copy somewhere else. I assume right now the first step is to check the YM chip for output using the keyboard to see if, 1) that is working and 2) trace from there forward to see where I am loosing the signal. If I had to guess, it's going to be further up the line as I have no sound from games or the keyboard and at some point the DMA sound should kick in, loosing both tends to indicate a circuitry area where both signals go into a device/circuit with neither coming out. I also guess that I need to make sure something in the software isn't inhibiting or turning off the sound, any suggestions on that? With the hard drive I have everything in German, which is hard to follow, and the floppy isn't working on this machine, although the floppy does works using the one in my Mega STe. Thanks!

[czietz]

Atarian Computing has good quality schematics for board revs. A & C: http://www.atariancomputing.com/downloads/schematics. Also dev-docs has multiple TT schematics ranging from badly scanned to high quality: https://docs.dev-docs.org/htm/search.php?find=_4

I would boot with no hard disk or floppy disk, because that's probably the easiest way to ensure that it's not a program/ACC muting the sound. Then you need something to generate sound. A diagnostic cartridge (like Exxos sells) has tests for both PSG and DMA audio. However, you can test at least PSG audio simply by repeatedly pressing a key and tracing the key-click sound.

From memory: The PSG has three output channels that are tied together and then go into an input of the LMC1992 mixer chip. From there the audio signal goes -- probably through some filter cap and amplifier -- to the speaker and to the audio jacks.

I agree with you that losing both PSG and DMA sound points to a common fault, such as the mixer chip. But tracing the signal should give you certainty.

[jacolover]

Thanks czietz, I will give that a try in the next day or two and report back. Happy Holidays!

[jacolover]

OK, was able to check out the YM and it does output pulses when the keys are pressed, those same pulses show up at the LMC1992 chip on both pins 4 and 26, but there is no output on pins 14 and 16 (or 18) my schematic is so blurry hard to tell the rt channel output. Can't find a clock pin, is there a clock on the 1992?. At any rate it doesn't look good for the 1992, any ideas where to buy a replacement? Thanks!

[czietz]

Before considering replacing the LMC1992 you should check if it receives (correct) Microwire commands from the TT Sound Shifter. Iirc, the LMC1992 starts muted after power-up. Sound is enabled by TOS during boot by sending multiple Microwire commands.

Microwire is an unidirectional serial protocol, much like one-way SPI with an enable, a clock and a data line. See the LMC1992 datasheet https://docs.dev-docs.org/htm/search.php?find=LMC1992 for pinout and the exact description of the format.

Looking at an old logic analyzer trace, I see that TOS 3.06 transmits the following commands:
0x4E8 <= Max. Volume (0 dB)
0x554 <= Max. Left Front Fader (0 dB)
0x514 <= Max. Right Front Fader (0 dB)
0x486 <= Treble Response Flat
0x446 <= Bass Response Flat
0x401 <= Select INPUT1

[czietz]

Also, I now looked into the TT schematic. The signal path is a bit more complex than what I recalled from memory in my previous post.

PSG audio gets into the LMC1992 on pins 4 and 26. Assuming INPUT1 has been selected via Microwire, it comes out again on pins 8 and 22 to get mixed with DMA sound in an external op-amp (U804, at least in the schematic I'm using). From there the signal goes into the LMC1992 again on pins 9 and 21, is subject to (Microwire-programmed) volume and bass/treble control and then comes out on pins 14 and 16. Thus, be sure to check the entire signal path.

[jacolover]

thanks again, czietz! How would I do that, i.e., check that the microwire commands are coming out of the sndshifter to enable the psg signal (keyboard presses in my testing) to be sent on to the outputs? Would I see a high on pin 1 of the 1992 when pressing the keys?

[czietz]

The Microwire commands I referred to are only sent once after reset (while TOS boots). Meaning that you should see activity on data (pin 1), clock (pin 2) and enable (pin 3) of the LMC1992. I don't know the idle state of the data and clock signals afterwards, but enable should be high. EDIT: To clarify: enable should be high, when no commands are being sent, e.g., during normal operation of the machine. Of course, enable must pulse low while commands are being sent during boot.