Namakichi TA-3A Power Supply: I don't quite understand this circuit

[RedShoeRider]

Good Day, All:

I'm sort of at a loss to understand part of a schematic for a Nakamichi TA-3A that I'm repairing.

The problem child outputs three voltages from the logic power supply: +10, +5, and +4.5. I have +10 and +5....but the +4.5 is just not there. A couple tenths of a volt, at most. In walking though the schematic, this is what I don't understand:


My confusion is about the link between Q401 and Q402. I think I'm reading this wrong, as to my eyes, I have 0.78v on the collector of Q401, fed through R405, which is just 470 ohms, but comes from 4.65v from Q401's base? Even if Q401 was supplying Q402, I'd have a 4 volt drop across a 470 ohm resistor? In-circuit testing bears this out: the collector of Q402 has 4.6 volts coming in.

Help!?
Thank you, all!
-Chris

[Andy Watson]

The voltages quoted on the schematic suggest that Q401 and Q402 are used to turn the power on to whatever is connected to the 5.31V line. Q402 is being driven hard into saturation. If you measured only 4.6V on the collector of Q402, either Q402 is defective, or (much more likely) something is loading the 5.31V line.

[mij59]

Q402 is used as a switch, the collector is the output, when Q401 is conducting the output voltage 5.31V is switched on.
Where is the +4.5V supply rail in the schematic ?

[SeanB]

Q402 is a PNP transistor, used here as a switch, so the collector voltage should be the 5V3 as indicated. Base is pulled down hard by Q401, so yes Q401 at 0.7V is correct, there will be some 8mA or so flowing in R405, which will turn Q402 on quite well.

If you are only getting 4V6 from the 5V regulator U401, check the input is not low, which could be your problem ,the 10V supply being faulty, either a open circuit diode in the poser supply or a failed capacitor. If the regulator is getting really hot then there is a downstream low resistance which is causing the regulator to run in current limit.

[RedShoeRider]

Ah! That makes a lot more sense now. Thanks, all!

U401 isn't getting hot. Nothing is getting hot at all on the board. That would at least have given me a clue where to look.

Here's the whole power supply circuit:


If I have this correct....the top half of the circuit provides the standby power to the logic of the stereo, the +4.5 in particular is used to run the logic. On startup, once the controller says that everything is ok, it puts power on pin 3 of CN9 to kick on Q407, closing the relay to the amplifier transformer, and away we go.

Cheers,
-Chris

[Tomorokoshi]

1. Check for the 11V at the input of U401.
2. What is the output of U401?
3. What is the "ground" of U401?
4. U401 regulates to 5.46V by floating the ground up by the forward drop of D404.
5. Are there any loose solder connections on U401 or D404?
6. Q402 provides 4.5V due to the Vce drop.

[RedShoeRider]

Thanks for the help!

1. +14.01v in on U401 (I assume this is because it's not under load? Somewhere else in the manual it mentions that values given are for when it's fully up and running)
2. +5.58 out on U401
3. +0.665 to "ground" on U401
4. Ok, I get that. Just about over my head, but I get it.
5. No. I reflowed them just to be sure. U401 is cold as a cucumber.
6. Ok, I get that, too. He's registering 0.23v rather than 4.5v on the emitter.


Cheers,
-CH

[KE5FX]

1000 uF on the output side of a 7805, with no protection diode.  I hope somebody brought marshmallows.

[RedShoeRider]

It's funny you mention that, KE5FX.

I repaired a Nakamichi TA-2A a few months back that did exactly that. The 1000uf died shorted, which caused the 7805 to heat up to some very interesting temperatures. I damn near burned my hand on it after the "Gee, that smells hot" moment. A new 1000uf, new 7805, and away we went.

[mij59]

I think your using the wrong definition of collector and emitter, the emitter is the lead with the "arrow" on it.

So if Q402 is not running hot, the emitter voltage is 5,46V and the collector voltage of Q402 is about 0,23V then Q402 is faulty.

[RedShoeRider]

I'll fully admit- I'm a neophyte when it comes to reading schematics, particurally transistors.
I have a lot to learn (and, usually, the patience to learn it): PNP vs. NPN screws me up every time.

Agreed: it looks like Q402 is toast. I'll replace him when I get home from work today.
The odd-ish thing is that I checked Q402 and pulled it pulled out of the circut earlier in testing, and by a multimeter check, it looked like it was ok. But, then again, it's cheap for a new one rather than questioning the old one.

Thanks very much!
-Chris

[KE5FX]

It's funny you mention that, KE5FX.

I repaired a Nakamichi TA-2A a few months back that did exactly that. The 1000uf died shorted, which caused the 7805 to heat up to some very interesting temperatures. I damn near burned my hand on it after the "Gee, that smells hot" moment. A new 1000uf, new 7805, and away we went.

Chances are good that the 1000 uF cap didn't fail first.  7805s are thermally protected and won't ordinarily fail from overcurrent conditions alone.  But when you put a huge capacitor on the output of one of these regulators, you have to make sure that the input voltage can never fall below the voltage at the output terminal.  That will kill them in a hurry, and, worse, it will do so in a manner that causes a low-resistance condition from input to output.

Say someone turned the amp off during a loud musical passage.  The 10,000 uF reservoir capacitor might drain rapidly into some other circuitry, or into the speakers... or maybe it's just old and has lost most of its capacitance.  Regardless, if the circuitry downstream of the 7805 doesn't immediately drain the 1000 uF cap, then the conditions are right to drive the 7805 backwards.  When this happens, the 7805 will pass its input voltage along to its load the next time the amp is powered on.  Sounds like the 1000 uF cap may have shorted out due to the overvoltage condition and saved the downstream components, so it did you a big favor.

Moral: there is no reason to ever put a 1000 uF cap on the output of a 78xx regulator, but if you do -- or if you're working on equipment where someone else did -- consider adding a 1N400x diode across the output-input terminals (cathode to input).  The magic smoke you save may be your own...

[RedShoeRider]

Wow. Thank you!
The scenario you paint is a completely possible case, so, one diode added in! Why would Nakamichi, or anyone, put that cap there at all, then? I assume it was just to smooth out the output of the 7805 when things were cranked up, but with the big 10k caps upstream from there, why would that 1000uf need to be there?

As for my unit.....Q402 and Q401 were replaced last night. I didn't have a lot of time to test, but their replacement made no apparent difference. The output of Q402 is still 0.2v.
More digging tonight

I will get this thing to sing "Merry Christmas"......even if it's in March!

[mij59]

Did you check for a short in the +4.5V rail ?
Could also be a broken copper trace, please post a photo of the pcb.

[RedShoeRider]

Most importantly: Thank you to all who helped me understand what on earth was going on in the power supply. I learned a lot in a short amount of time!

I added a protection diode to the 7805; easy enough to do with everything pulled apart.

It turns out the problem wasn't anywhere near where I was originally looking. It went like this, and I realize I'm about to reference parts/pins that aren't in the schematic that I posted. It's more for someone, somewhere in the future, who might have the same problem, as it looks like this part of the circuit is the same in all of the Nakamichi TA series amps:

Q402 was replaced, but still not working with a base voltage of 0,6v. I lifted the base pin and put 1,35v on it, and voila, Q401 started working correctly. So, working upstream, I went though the next pile of control transistors and found exactly nothing odd. But, it did lead me to U503, the MPU, which was outputting some strange voltages. In looking at Vcc, the problem became apparent. Vcc is supposed to be 4.95v, but it was reading 2.2v. Oh, odd. But I know my +5 was good coming into the board. There wasn't much upstream of the Vcc pin, just a diode (D504 (1N4148), and a cap, C511 (470u/10v)).....a diode who was only partly working. Instead of passing 5v though, it was passing 2.2v though, and getting a little warm in the process. The cap was trashed, too. 10v, 470uf filter cap, but acting like a resistor to ground, so I guess it was pulling a ton of current though it and beat the diode to death. Replacing both of them....and the little puppy turned right on.

I'm assuming some of the oddness that I saw was the MPU sort of working, but that 5v controlled the transistors that turned the +4.5v supply rail on, but being at 2.2v instead of 5 made the transistors upstream act oddly rather than just dead if it had gone completely open.

After adjusting the idle current and testing it on my junk speakers, the first thing that it sang out: Bicycle Race by Queen. Might as well start the party right!

Again, my thanks to you all.
-Chris