Beovox 3000-2 (2402) - Secondary Power Supply issues...

[Johnny_Law]

From the depths of my "I'll fix it someday" queue ... A Beovox 3000-2 Type 2402 with multiple, serious issues (related? no idea).

DC offset is OK (0V) on both channels.

There are two separate hums, one in the L channel unaffected by volume (Hum 1), the other in both channels that increases with volume (Hum 2).

The Tape input works on the R channel and the L channel, but there is random scratchiness that comes and goes, and the L channel has the background Hum 1. You can hear the FM radio faintly in the background, only on the R channel.

-Left channel Hum 1 & Hum 2 is present but much less pronounced (i.e., music plays and sounds fine except quiet parts where you notice the hum & scratchiness)
-Hum 1 is ...
--Not affected by Volume
--Not affected by Loudness
--Affected by Lo Filter (gets much quieter)
--Affected by L & R Mono switches (disappears completely when either or both depressed)
--Remains even with the signal path from the pre-amp to the amp section removed (C416 and C499)

In Tape mode, across L speaker output at lowest volume (same with / without 8ohm dummy load connected), a strange 60Hz wave:


In FM, the L channel Hum 1 is gone, but Hum 2 is present ... it quietly gets stations and you can faintly hear them in both channels, but garbled and very quiet underneath Hum 2.
--Hum 2 increases in intensity with Loudness
--Hum 2 tone affected by Bass & Treble
--Hum 2 moves L&R with Balance

With nothing connected ...
Phono 1 has two yet entirely different hums in each channel, the L hum lower pitched than the R hum, each increase with volume
Phono 2 has a quieter hum on both channels that increases with volume (with sounds like a normal Phono input when the ground is disconnected)

So I went to the power supply to poke around.

Here's across the main rectifier bridge output, a 120Hz jagged AC wave:



These are new issues for me. Is that a normal output from a full bridge rectifier? I don't think it is ... right? Almost like one diode in the bridge might be shorted and another might be open? FWIW I see ~53VDC across the 7000uF filter capacitor (which is new).

This is all being done via a DBT w/ a 53W bulb which stays dim during testing.

And in the middle of measuring stuff, that handheld multimeter all of a sudden broke. Turns on but won't measure correctly and won't read the internal test tone either. D'oh.

Thanks for any input / advice / suggestions...

 

[Johnny_Law]

Ok - so after much diagnosis and tail-chasing (including an incident with a missing chassis panel resulting in a strange hum from the amplifier here) ...

I'm now ONLY chasing a 60Hz volume-dependent hum at the speaker outputs. The other hums were from the bad rectifiers (solved) and then a grounding issue (solved).

I've narrowed the issue down to the secondary power supply, which uses a 23V AC transformer winding to feed a smaller B30C350 rectifier that is smoothed by a single capacitor, and then regulated by two transistors TR17 and TR18 to produce a 15V power supply which is used by the tuner and pre-amp.

I've copied this secondary power supply circuit out of the service manual below:



Currently it's on a DBT and only getting 16V AC at the rectifier. As a result, instead of 15V DC I am seeing 10.2V DC.

When measured across C145, I get this, which seems questionable ...



But here's the real problem ... this is measured from the positive side of C145 to ground, in other words, this is showing up AT the 15V DC power supply!



In terms of individual part diagnosis, I have replaced the following with new devices:

-The rectifier with a B40C1500G
-C145 (with a Vishay 470uF / 40V axial)
-TR17 with a new BC212
-TR18 with a new 2N2219
-ZD148 with a new 9V zener diode

This gets fed into the amplifier, which produces at the speaker outputs a 60Hz wave that is amplified in proportion with the volume. This is the signal at maybe 1/3 volume across an 8ohm dummy load:



If I remove the power supply feed to the pre-amp, it disappears completely regardless of volume.

So it seems clear this is a problem, and it's coming from the 15V power supply. But everything in the circuit is new or tests fine, the grounds are all good ... so I am scratching my head about how a 0.27Vrms signal out of the rectifier could possibly turn into a 7.6Vrms 60Hz wave coming out of this circuit?!

I would LOVE any pointers or suggestions as to what to check / try next ...

 

[Johnny_Law]

I was wondering what the positive DC signal SHOULD look like in this circuit, so I modeled it out in LTspice.

Looks like the positive DC signal should be relatively stable - in my model it is a little off of spec (~19.2V) but I think this proves that I should NOT have that giant 60Hz waveform...

 

[Johnny_Law]

Suspecting the new rectifier was shorting out under load and letting AC through, I replaced it ... but it made no difference.

Still getting that weird 60Hz wave ...

Anyone have any ideas as to next steps to diagnose?

 

[cademan]

Something nearby maybe?

Light, fan, fridge, surge protector etc.

 

[gslikker]

Could it be because of the DBT there is not enough voltage to have the circuit work properly over the whole sinewave half cycle?

BTW you have no load current or a resistor representing a load specified in your emulation, and the schematic says 100k for R2 while the simulation drawing shows 1k.

 

[Johnny_Law]

D'oh! Model circuit fixed, and looks much closer to the spec voltages ... Thanks!

 

[Johnny_Law]

In other news ... some wild diagnosis today.

First crazy thing...

So in the circuit I pasted above, I deleted one additional part: the ON and FM strength meter light bulbs.

They are fed from the same 22V AC winding, reduced via a 220ohm resistor.



This bulb circuit uses two brown wires soldered to the PCB at the rectifier.



On a whim, I de-soldered one of the brown wires from the PCB ...

SUDDENLY, NO MORE 60HZ VOLTAGE!

It was GONE!

So ... why? My theory: back in 2016 when I bought it, the bulbs were burnt out. So, I replaced them with LEDs. But ... I didn't have any wedge T1.5 base LEDs, so I just bent the leads and shoved them into the sockets. Well ... I think I left too long a lead sticking out of one of them, and that lead must have been making contact with the chassis. This would have grounded the 22V AC signal which then would of course show up in the DC power supply.

But ... not out the woods. The 60Hz hum was gone, but ... like @gslikker you suggested, neither TR17 nor TR18 were turning on. I was suddenly measuring ~0V instead of 15V.



Measurements: [EDIT: this was wrong! TR17C was measured AFTER the circuit started conducting, I just didn't know it at the time...]



BUT ... after taking the above measurements, all of a sudden everything activated! And finally, the sweet, sweet stable 15V appeared!

 

[Johnny_Law]

Then, I turned the unit off and back on again. Dead. Only after I touched my red DMM lead to various parts of the circuit, eventually it turned on and back to 15V.

At this point I fed some music into the Tape input and was greeted with nice, clean, hum-free sound from the headphones!!!

But, off then on and back to a dead circuit.

SO: why isn't the circuit activating when I turn it on? What suddenly starts conducting when I'm poking around?

I can't say that touching any one specific thing with my DMM probe is the thing that does it, but it seemed like TR17C / TR18B is what might have done it more often than not.

Hypothetically if that is the case, then perhaps R151 is intermittently open?

Will investigate further tomorrow but would love any other suggestions ...

 

[gslikker]

Measuring -1.1 on the base of the NPN transistor and its emitter being -18, suggests the base to emitter connection measures open due to transistor fail or a bad (solder?) connection.

Schematic suggests -11.4 and -12, which is the around 0.6 VDC difference a base to emitter junction should look.

 

[Johnny_Law]

Thanks - it appears that it must have switched on mid-measurement before. Here are the corrected measurements. Yellow before it starts conducting, green after it starts conducting (except for the AC input, the yellow highlighting doesn't mean anything there, it's in the ~17.6V AC ballpark either way).

Note that I've confirmed the trigger point: [edit woops] TR17B or TR17C. If I measure that with my DMM probe, the PS will immediately or soon thereafter turn on.

Note also that everything highlighted green (which is every single component) has been removed and verified to be operational.

The rectifier B30C350 is a new B40C1500.
TR17 was originally a BC212B and is now a new BC212B.
TR178 was originally a BC119 and is now a new 2N2219 (this is a TO39 package device).
ZD148 is a new 9.1V zener diode, I forget the part #, but the original one works fine and the circuit does the same thing with the new or original one in place.
C145 is a new Vishay 138AML axial 470uF / 40V.
The resistors are either original or were replaced by metal film before my time. All measure correct values out of circuit.
All grounds are good.

I also plugged it directly into 120V and not the DBT, and it did the same behavior although the voltages were a hair higher than through the DBT.



So I'm totally at a loss now.

At first, TR18 does appear to be conducting given the B-E drop. However, TR17 is not conducting, nor is ZD148.

So how could TR18 be conducting if TR17 is NOT conducting?

And why would a DMM probe on [edit woops] TR17B or TR17C cause TR17 and ZD148 to all of a sudden conduct???

 

[gslikker]

I think the circuit needs the proper current offtake and the +15V "users" may overload the supply. As long as the voltage over it is very low you will not see the 2n2219 heating up either.
So if you have a lot of offtake it may not sufficient to switch on, so the load is to be checked somehow (maybe a 1 ohms resistor in series with the 2n2219 and measure millivolts over it)

OTOH, you also may have a problem with the transformer itself, still having only 17VAC.

You could check that just taking TR18 out , switch on and measure voltage over the capacitor or rectifier. (do not forget to discharge the cap before putting in the transistor again).
Having the unit off, you can measure whether the 100k resistor (including traces where it goes) measures OK while the transistor is out (one way your ohmmeter may be affected by TR17 then try other way)

 

[Johnny_Law]

Thanks for sticking with this ...

I think the transformer is OK. When plugged into the wall directly, I see 21.2V AC (compared to the spec 22V).

In fact I am listening to music with my headphones right now with the unit plugged straight into the wall, after I got TR17 to turn on by touching my DMM probe to TR17C. Note that it doesn't work right away, it takes a few moments of poking at it to switch on. But once it does, all is good and everything sounds great.

So I think we can safely say the transformer is OK ... do you think?

And focus on the issue of why TR17 and ZD148 don't start conducting until I poke at TR17B or TR17C.

Worth noting: I've wiggled TR17 physically to confirm it isn't a loose physical connection. No matter how much wiggling, it won't switch on unless I touch C or B with my DMM probe. It also works if I use other metal tools like my small forceps. So I wonder if the DMM probe is actually transferring some charge and that's enough to switch it on? In which case, maybe I should try a different device in place of the current new BC212...?

 

[Johnny_Law]

... and I did just that, put in a new 2N3906 at TR17 just to see what happened, and same behavior.

Rats.

Put the BC212 back in and will continue to scratch my chin here ...

 

[gslikker]

maybe lower the 100k resistor to some 47k or so to force the transistor conduct earlier, further I have no ideas other than too much load.

 

[Johnny_Law]

Thanks ... I will go look at each of the places that uses the 15V power supply and see if there are issues with any of them (like maybe a corroded adjustment pot in the FM section) that would present as too large a load to the 15V supply...

 

[Johnny_Law]

First attempt: No dice. The 3000-2 is known for developing open adjustment pots. So I de-oxited all the pots connected to the 15V PS in the FM circuit, but no change. Still gets stuck until I poke TR17B or C with something metal.

I guess I could try to come up with a safe value for the equivalent load, and just wire in a resistor from the +PS to ground and see what happens. Do you think that's a good idea @gslikker ?

 

[Johnny_Law]

Thinking about this ...

It seems very odd that I'm getting almost all of the voltage on the negative side (-18.8V) and almost no voltage on the positive side of C145 (0.06V) at startup.

As soon as the rectifier sees its first positive AC wave, C145 should start to charge and current should flow through R150 and R151 to ground. The voltage developed across R150 is what activates TR17 which, as soon as TR17E starts to draw current through R146, develops the 9V breakdown voltage for ZD148, as the current from TR17C activates TR18.

BUT ... none of that is happening, and I'm seeing that -18.8V drop across TR18 E to B as -18.1V. So: does this mean it's possible the 2N2219 has developed an E-B short under load? If it's sucking all the current from C145, maybe that's why there's practically no current at all across R150??

 

[gslikker]

to me it still looks the 15V power section of the amp is drawing too much current, during startup and takes away the power from the transformer.

other option would be if it is still on DBT and the transformer sags because of that.

 

[Johnny_Law]

Thanks ... one step at a time.

First thing: I measured the resistance from the +15V PS to ground with the unit off, and it's 285ohms.

I plugged the equivalent load resistance into LTSpice and it should still activate...



If I reduce the load to 1ohm, it fails to activate - I think this is what you were suggesting as a potential cause @gslikker



So ... maybe if, when powered on, something shorts the +15V PS to ground, maybe that's what is happening??