MCS 3847 Integrated Amp Noisy Left Channel

[KilgoreCemetery]

I've been having trouble with this home audio amplifier that a friend gave me. It has constant noise on the left channel from the moment the main relay clicks on until I turn it back off. I recently replaced all of the electrolytic capacitors in the hopes that it would fix both this and a humming issue, but the noise is still there. Here are some of the other things I've tried (in no particular order):

Sprayed the potentiometers with Deoxit. They were pretty static-y to begin with. It cleared up the static, but not the constant noise
Adjusting the Volume knob does not increase the volume of the noise
Adjusting Bass, Treble, Balance, Mono/Stereo, Input, Subsonic Filter, Loudness, and pretty much every other knob or switch has no affect.

This is kind of an oddball brand/model and I haven't been able to find a schematic for it. It's JC Penny's Modular Component Systems brand and I think Technics may have actually built it for them, but I'm not sure.

I'm beginning to feel like it has nowhere left to hide and I'm running out of places to look. Can someone give me an idea of where to look next? Any help would be appreciated.

 

[shrtrnd]

I'd be looking at the small (non-electrolytic) caps on the left channel output. That'd be where I start.
You may indeed have a problem elsewhere, but your symptom would have started me out with the capacitors first, like you did. I've just sometimes found cracked or blown small capacitors in the audio output circuits also.

 

[KilgoreCemetery]

Hmm.. So far I've lifted the legs on the left channel MOSFETs, a questionable JBT, and some ceramic caps in the output section, but I still haven't found the problem. Could the relay itself be the issue? It's a 6-pin Daiichi and two of the pins have the noise. I had it out at one point and cleaned the contacts because it clicked more than once after I powered the unit on. I haven't had the problem since and just wrote it up as a dirty contact.

 

[73's de Edd]

KilgoreCemetery . . . . .Attn: of Chief of interment . . .

Looks like your unit is going to be developing its noise right at / within the AUDIO OUTPUT STAGES, since your descripton of the effects at earlier stages, designated there being no earlier stage control over that noise.

That units ACTUAL manufacturer should be NEC of Japan.
AND this depiction below is a power amp only, without a tuner . . . . and your unit WITH its tuner unit should have some knob or trim similarities of the front panel . . . just take note of the knob style and brushed alununum front panel . . . . . as well as the silk screened frontal labeling.

Now we need pics of your internals to see if it used power output transistors or was new enough to be using potted power amps.

NEC Photo

(of their "7000" series of amp . . .with its AUDIO output circuitry, which is akin to your "REBADGED" Penney's MCS3847).

73's de Edd

 

[dorke]

In my view replacing parts that way is a shot in the dark.
These problems should be pin point located.

This kind of faults are best located using a SCOPE,most people don't own one,
so an audio signal tracer is the next thing to use,that can be DIY build.
Since you have one good channel to compare to the task is somewhat easier.

From your description, It is also clear that the problem is in the PA(beyond the volume control pot).

 

[KilgoreCemetery]

Thanks for the info Edd! This unit doesn't have a tuner and I pulled up a picture of a NEC 7000 series and it's definitely the twin brother to the one I'm working on. You could probably switch out the front plate without any modification. Unfortunately, I'm not able to view your attachment, but I can definitely include some from this unit. (edit: Ok, now they pulled up.)

I agree, dorke, I don't like guessing. As for tools, I wish I had a scope, but I have been using a probe to "listen in" to the audio signal, which is how I've decided which parts to lift a leg on.

Looking at the pictures, I believe the problem lies somewhere on the left side of the amp board. Also, the cap is off of the relay currently, because I was giving the contacts a look-over.

 

[dorke]

The final stage is a push-pull "DC-coupled" output and differential amplifier input.

Some possible problems :

1. You should clean the speaker relay contacts with Deoxit soaked paper,rub gently.(blue)
2. Another issue is the the differential amplifier input the 5 pin Tr.(red)
Clean it's bottom side were the legs go into the plastic body,there should be no dirt there.

The nature of the noise isn't clear,if possible record it, and post the audio file.

 

[KilgoreCemetery]

I've already cleaned up the contacts on the relay (that's why the cover is off of it), but I will definitely be checking out those transistors. I don't know anything about posting audio on here, but I could certainly post a youtube video of it.

 

[KilgoreCemetery]

I tried cleaning the contacts on the relay just to be sure and I had it under a little better light this time. I've included a picture of one side, but they're both about the same. Also, I uploaded a youtube video of the sound. It's not great, but it's the best my little phone could do and I hope it gives you a better idea of what is going on. Just think static.

 

[73's de Edd]

Sir KilgoreCemetery . . . . .

I don't know of any EXACT schematic of that unit, with it being one of those exotique French " Jeeeeey Czzeeeeeee Peneeeeeeeee modular units . . . . . . via import . . . au Japan as, actually being a NEC.
I have been using a probe to "listen in" to the audio signal, which is how I've decided which parts to lift a leg on.
If you can hear the noise . . . . lets keep up with that same technique.
But since we might be encountering variable DC voltages also, how about inserting about a 1 ufd paper or poly cap ( no electrolytic . . . they leaks! ) in series for providing some total DC isolation.

I might interpret that audio effect as just a sporadic shot type of staticy noise which occurs at a quite constant rate.

Now if NEC keeps up with their past parts numbering protocol . . .odd assigned numbers are being LEFT channel parts and even numbered parts are RIGHT channel parts.

NOW . . " Reading the chassis" . . via eye and conjunctive neural synapses . . . .

The PINK blocked in area, is where your output stage is voltage sampled at its output to be sure that no AF OUTPUT semis crash, and put voice coil melting voltage to some egg-speen-sive speakers.
That power relay then opens the speakers from being connected, as well, there may be fast acting fuses in series to the speakers.
The ORANGE boxed in area is where peak audio output voltage is sampled to see that no speakers have come loose or are not even being connected, so that when you crank the volume to max, seeking sound . .shutdown.
Also, the current pull of the AF output stages are sampled, to confirm if a shut down is being necesary.

I have perceived the LEFT low level audio input to the power amp as being at the bottom right corner YELLOW A as the LEFT channel input and RED A as the RIGHT channel input. If you "listen in" there, I am expecting you not to be hearing your pesky noise, at that YELLOW A, first test point.
The signal then passes on to be amplified at alphabetically designated / advancing test points, in moving up the page.
YELLOW B and B prime are related to AF coupling and bypassing until audio reaches YELLOW C which is a custom / special dual transistor . . .if its a 2SA798 . . .its an PNP that has shared emitter leads and the two transistors are made on the same semi die so as to get like thermal tracking.
"listen in" to all of its leads to see if your noise origin might start there at YELLOW C.
If not move to transistor YELLOW D and probe all of its leads looking for your noise.
Still quiet, then move to YELLOW E and E prime driver transistors leads to see if they are quiet or not.
Then your last/ final amplifier stages will be the BATMAN ! ! ! . . .no . . .make that Batwing power transistors as
YELLOW F and F prime.
Then YELLOW F and F prime feed thru those white .47 ohm power resistors as YELLOW G and G prime to feed
the left speaker . . . . .notice that the different layout due to ORANGE area encroachment, made them have to spread apart their locations . . . RED G + G' ended up mounted side by side.

Hope that you now have found your noise origin being somewhere within that string of transistors !
( FIO the X trim pots just set the idle currents on the final power transistors. )

If you find a suspicious early stage transistor, the noise may be aggravated / exacerbated via hot or cold.
So you might pull out Mamma Cass's blow dryer and fabricate a cardboard cone to direct warm-not hot air upon a suspect transistor-or -electrolytic cap-or resistor to see if the noise level changes.
Conversely, you can switch to cold air and get a Q tip and frequently dip in alcohol and swab the transistor case and then let the cool air to chill the transistor. Keep wetting the transistor, as the alky evaporates almost immediately.
Now . . .go do-it-to-it

Awaiting further ?' s or my supplying clarifications.
ADDENDA:
I just saw that you had added before I posted this.
You are probing the equivalent point as the LEFT speaker terminal, while the RIGHT channel speaker was quiet.
The continuous noise sort of resembles constant FM radio hiss.
NOW go to the EARLIER gain stages and see if you are missing this noise at the initial
YELLOW A starting point.



MARKED UP REFERENCING OF MCS 3847 . . . . .

73's de Edd

 

[dorke]

Apart from the noise,does that channel work o.k?

What kind of "audio tracer" do you use?(photos?)
should have a DC block and I hope it's a relatively high impedance one,
this is important for probing other parts of the circuit without damaging them.

With a proper tracer you can compare the good and bad channels on a point for point basis.

 

[KilgoreCemetery]

You crack me up Edd. However, while it is quieter, the noise can still be heard on the yellow A. Does this mean I've been barking up the wrong circuit board?

As for going through the letters, I've previously checked the yellow B, C, D, and F, and can confirm that the noise is there. I'll try to confirm the other letters over the weekend, but I'll have my hands full with a 3 year old piranha. Hopefully I can check those and start back-tracking during her nap time.

Those are good questions, Dorke. It's very difficult to tell the audio quality of the left channel because of how loud the noise is with speakers hooked up, but it will play music. I don't think it's distorting the signal so much as layering the noise on top of it.

Also, I don't have any pictures of the tool I've been using simply due to embarrassment. It's one I made from watching a video and is far more functional than it is pretty. It has a probe (POS), ground wire (NEG), a high-voltage, low capacitance, in-line capacitor (that will soon be replaced with a poly or paper one), and RCA connects for output. I don't know what the impedance is on it, but I'll be looking in to that as well.

On a side note, why would it be impedance and not resistance? I thought impedance was more of a variable, where as resistance was a constant?

As always, guys, I appreciate your input and I hope to get you some more readings soon.

 

[73's de Edd]

Sir KilgoreCemetery . . . . .


Also, I don't have any pictures of the tool I've been using simply due to embarrassment. It's one I made from watching a video and is far more functional than it is pretty. It has a probe (POS), ground wire (NEG), a high-voltage, low capacitance, in-line capacitor (that will soon be replaced with a poly or paper one), and RCA connects for output. I don't know what the impedance is on it, but I'll be looking in to that as well.

I could only suspicion it as being some form of audio amplifier . . . I would only relate the
a high-voltage, low capacitance, in-line capacitor
for it then ALSO being compatible with higher voltages being experienced with vacuum tube design of electronics, where up to 400VDC might be encountered . Don't know about that low capacitance, in-line capacitor as I would minimally want 0.1 ufd, or optimally, a 1 ufd, at a DC breakdown voltage greater than you would find in your tested equipment. Also, with its dielectric being paper or one of the poly families.
WITH that DC isolation provided by the input coupling capacitor . . . at the probe input, I could only see any less than desired performance being due to a low gain from use of a low input impedance amp, if that is the design being used as your audio tracer amp.

Back to your noise being at YELLOW B, it also should be at the YELLOW A amp input terminal strip.Should you have a set of speakers or a set of stereo earmuff headphones . . .them being easier to hook up..
If you would then place a jumper connection between YELLOW A and RED A, you should have the same noise coming from BOTH channels.
That then moves the point or origin of the noise as suspicioned to be PAST the volume control and into any post amplifiers before the YELLOW A amp input terminal strip.
Aside thought . . . .have you confirmed that having the balance control to the proper extreme end, will or will not have any effect on this noise, such as is presently confirmed by the volume control.?

A LAST prev maint step would be to resurface those speaker relay contacts , as they show 25 years of arky-sparky speaker switching action.
Get some AUTO-ZONE, etc . . . 1000 grit silicon carbide " wet n' dry" paper and cut into 1/4 or 1/2 in wide strips along the long length of the sheet . . . . and always keep a strip WETTED, while you pull a wet strip betwen a set of contacts,which you are holding pressed/ tensioned together by levering a popsicle stick on one side. Reverse the strips sides such that the next time, that the backing side will be cleaning away the eroded metallic deteris , while additionally burnishing the involved contact, while the other sil card side is eroding the other side down to a fresh and unpitted surfacing. When both sides look good from their resurfacing, then use a folded strip that has the backing exposed outwardly on both sides, then wet it down and use the pulling thru of that clean strip to do a final burnishing of that one set of contacts, Then, move to the other contact sets.

Thassssit . . . .

73's de Edd

 

[dorke]

O.K.
this kind of probe should do if connected to a "line in"(being greater than 10Kohm,as they are)
of an audio amp.

Impedance- meaning anything which isn't purely resistive(like your probe with a capacitor in it).
It is indeed variable(frequency dependent) like your probe is i.e. blocking DC and low frequencies .

Back to the question about the bad channel functionality(apart of noise).
If you feed the same audio signal to both channels(you can force that by activating the MONO state,
if there is one on the amp).
Do they produce the same level of music? distortion?

Since,as you described ,the volume knob has no effect on the noise ,
i.e at min position the noise is still there, it is clear that the problem is in the PA section and not before it.

As I said since you have one good channel to compare to this is probably the easiest way to go .
Let's start:

With volume at minimum.
Be very careful not to short with your probing .
Your probe needs better isolation, so only the very end of it's tip be exposed.

1. Check the DCV(with DMM) of both 5 pin TRs (RED) and compare them pin for pin,what do you get?
2. Probe the noise at same color arrows(noted both channels since I'm not sure which is the noisy one)
what do you get?

BTW,
These 5 pins Trs are hard to find and so are the final Trs in this model.
What is written on them?

 

[KilgoreCemetery]

Thanks for your patience, guys. I have tested A through X from your picture, Edd, and except for one of the yellow electrolytic caps (the one just above "D2-3" on the circuit board), all of the yellow letters have the noise and none of the red ones have it. Today, I'm not hearing any noise on A at all. Not sure if this is user error from before or due to temperature.

Dorke, I checked the ones in red. They have on them: A798 79G and the pinout is B, C, E, C, B. There is no noise on either of the base pins (for either transistor), the collector pins have about -36.5VDC on both pins (both transistors), and the emitter pin for the left transistor is noisy, while the emitter pin on the right transistor is silent.

The voltage for the base pins do differ quite a bit from one transistor to the other, however. On the right one, I get 22mV, but the left transistor fluctuates in the -500mV range. Emitter pin on the right transistor is at 610mV, left one fluctuates in the 300mV range.

This is about as far as I got before the noise vanished. It's been quite a bit warmer today and my workbench is in a partially-heated garage. It's been pretty cold prior to today in here. I'll take some measurements with it working for now and get you more updates as soon as I can.

 

[KilgoreCemetery]

...aaaand it's back.

Edd: I CAN hear the noise at Yellow A. It is much quieter than most of the places I've probed. Jumpering from Yellow A to Red A does allow me to hear the noise from both spots. I can confirm that the balance knob doesn't affect the volume or appearance of the noise. I'll be looking into gently wet sanding the relay contacts.

Dorke: I was in the process of comparing one channel to the other when the noise vanished again.

 

[dorke]

Well,

I don't think it's the relay,leave at alone for know.
It has nothing to do with warming or cooling of the environment either.

You most probably have an intermittent contact/device,your probing makes/breaks that contact .
That is why the noise appears/disappears.

I would focus on the A798 5 pin dual Tr.
As you said, when touching it for measuring, the noise stopped.
Try to get the measurements of the DCV on the it's pins and compare to the other one on the good channel.
Than try poking around it with an isolated "stick" to get the noise back and gently take the measurements in the noisy state.
Please take a closeup photo(no glare) of the noted area(the good channel,but they are the same) so we can try to reconstruct a schematic from it.



.

 

[73's de Edd]

Sir KilgoreCemetery . . . . . . .

On photo post #9 . . . .I can see that in your zest to get back to work on this unit, that you had left your enbalming gloves on

...aaaand it's back. . . . . . . is that being same-same as . . .its baaaaaaaaaack

AND . . . you are such a pleasure to work with . . . . . your thoroughness of feedback . . . .

This new photo covers all sections of the whole unit, with the time yellowed relay cover being resplendently in place.
Looks like your unit is being blackface, while the NEC units are brushed aluminum.
From your given information, if the unit were to be MAKING ITS NOISE, see if taking an isolative / shunting electrolytic capacitor of 47-100 ufd with its negative lead connected to that central ground connenction, being between the YELLOW A and RED A terminals. AND then connect the caps + to YELLOW A and then see if there might be a slight plop from the left speaker / headphone and then listen for that left channel noise presence..
I am fully expecting it to now have diminished to the level of a mouse peeing on a cotton ball.
That clears all of your power amp circuitry inward from that BLUE shielded wire pair into the audio power amplifier stages..
And . . . . our noise source origin should be back towards the front panel of the amp from that BLUE shielded wire pair.

Since we can now see all of the units boards / sections in the new referencing:
The YELLOW RECTANGLE area will be related to power supply and sub power divisions.
Assign the BLUE RECTANGLE as being the tone controls and volume and mute control functions . . . a minor tone control disable /bypass switch. . . .and, at our low levels . . .a not to be heard subsonic filter.

Remove our audio "grounding" test capacitor and have the left channel noise present again and then see if the flipping of theses switches will have any effect on the "tonal colorization" or level of our noise.

TONE on-off
TAPE CUEING . . . in all three possible positions
TAPE MONITOR . . . in all of its 3 possible positions
FUNCTION . . . . in its 3 positions Phono Tape Aux
and finally . . . see if Sir Dorkes suggestion of placing the MODE switch in the MONO position then results in the same as your having temporarily connected YELLOW A and RED A terminals.
The PINK RECTANGULAR area is relevant to the rear low level shielded audio sources, travelling forward to the frontal area, to acquire the switching actions of the FUNCTION and MODE switches.
This is now only leaving the RED RECTANGULAR section with its PRE /or/ POST amplifier portions as the 4 RED SQUARE transistors . . . if I am seeing all of them . . .and there are no small 7 pin black single inline package IC's being involved ( and hidden out of my sight) . . . . . like possibly a set of TA-7136P's. But that circitry, should not be in circuit, if the FUNCTION switch is NOT being in PHONO position.

RE EMPHASIS . . . .
Back when you were getting those voltage variances in the two dual transistors, I believe that was due to INCOMING noise presence on YELLOW A terminal and its being amplified . . . . while RED A had no incoming noise to handle.

I'll be looking into gently wet sanding the relay contacts.

Looking at their degree of pitting it might require that additional popsicle stick, cantilevering pressure . . . looks like they have heavy silver contacts. ( Plus that would be your FINAL . . . Prev Maint act in assuring another 25 + yrs of reliable service. )

TECHNO PHOTO REFERENCING . . . . .

73's de Edd . . . .

 

[dorke]

@73's de Edd ,
I don't think the problem is before the P.A board since the O.P has already checked what you have written at #18
And specifically, the noise is there with Volume pot. at minimum.
He can re-check everything ,but I would expect no results from that.

@KilgoreCemetery ,
In regard to the measurements at #17 ,for clarification,they should be done with volume knob at minimum.

Another thing ,
please post a clear photo of the bottom side of the board(if there is an opening at the bottom).
Area of interest in-red,and larger view in-blue.

 

[73's de Edd]

BEFORE you have to pull the board for bottom photo access . . .BUUUUUGGGGER ! . . . .confirm if the elect cap and its AC signal grounding of YELLOW A terminal doesn't silence the left channel noise .