10/31/2023: Localization News - Dead of the Brain 1!

No, NOT a trick, a Halloween treat! Presenting the Dead of the Brain 1 English patch by David Shadoff for the DEAD last official PC Engine CD game published by NEC before exiting the console biz in 1999! I helped edit/betatest and it's also a game I actually finished in 2023, yaaay! Shubibiman also did a French localization. github.com/dshadoff/DeadoftheBrain
twitter.com/NightWolve/PCENews
Main Menu

PCE GT audio problem - after cap change!

Started by Spirantho, 02/22/2013, 02:44 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Spirantho

Hi everybody,
I've been working on repairing a PCE GT.
It had no sound at all - fixed that nice and easily by changing the caps in the corner of the board (the 3 100uF and 2 33uF ones).
For a while - it worked perfectly. But there was a lot of interference...

So... I decided to change the 470uF radial capacitor, as it wasn't SMT and quite easy to get to....

.... and since then I've had no sound from the speaker.

Headphones work with both channels, no problem.
If I wire up pin 11 of the ribbon (the "Speaker+" wire, in other words) to the left-hand pin of the headphone socket, I get nice loud sound again - albeit probably just the left channel. Connecting to the right pin gives faint sound (probably the R channel). Through the headphones, both channels are good and strong, remember.

Now normally, the speaker+ (pin 11) is wired directly to another pin on the headphone socket. I'm guessing that what's supposed to happen is that when the headphones aren't plugged in, then both L and R channels are connected to this pin, which then goes to pin 11 and thereby the speaker. I'm thinking the headphone socket, therefore, is a bit kaput.

Therefore, could someone please help by taking a look at what the connections are supposed to do? The L channel connects to the op-amp (via a 100uF cap), the R channel connects to the op-amp's other output (again via a 100uF cap), and the presumed Mono channel connects to the speaker terminal via pin 11 of the ribbon. In a working PCE GT, what is the resistance between the L and the Mono, and the R and the mono? On mine, it's very large. Does the resistance change depending on whether a headphone plug is plugged in or not?
There are also three pins at the back of the socket. Looking at the socket from the front (the pins are only visible from the bottom of the PCB), on mine I get two pins connected (pins 2 and 3 I think it was), and pin 1 isolated. Is it this way with the plug inserted or not?

If anyone could help with this diagnosis I'd be very grateful! Ordinarily I'd try to substitute a part from another machine but they're very expensive!

Thanks!

Keith Courage

the headphone jack switches from pin 1+2 to 2+3 when you either insert or remove the headphones. Maybe the switch that goes between the two is bad. Try bridging either pin 1+2 or 2+3 and see if the speaker gets loud again.

Spirantho

Thanks for the reply! That confirms what I suspected.

I fixed it!
I'm not certain that the pins at the end were at fault, though, in the end... at least, not  directly.
The pins at the end are connected to the A+ input and B+ input on the op-amp respectively. I suspect this controls the mono or stereo output.

I eventually tracked the fault down to the left channel's plate inside the socket. What happens is that when there's no plug inserted, the left plate is quite heavily sprung by its shape, so that it makes contact with the centre plate, which is connected to the speaker. The end of the left plate is connected to the switch at the end of the stereo socket, which controls the pins 1/2/3 as we called them above. When the left plate moves because of insertion or removal of the plug, it moves the switch from 1/2 to 2/3 (and makes it mono).
The problem was that the left plate was moving the switch, but the switch has a stop in it so that the switch can't move too far - if it did this, the left arm would "fall out" of the switch, as because the plate bends, so it turns inwardly towards the centre of the socket and away from the switch. The switch was therefore stopping the left plate before it could get to its fully sprung position - in this "half-sprung" position, the connection wasn't being made any more.

The solution? I just bent the left plate end very slightly such that the plate needed to move further internally to move the end of the plate the same distance. That way when the plate was fully travelled on the switch, it was internally bent further, giving more spring and therefore a better connection.

I'm not sure if that makes sense to anyone - but it may help someone some day who has the same problem!
Thanks again for the advice!