Home Cinema active Subwoofer malfunctioning

Status
Not open for further replies.
I got some 100pF caps (for C2 on PCB1 and C9 on the pre-amp PCB2, in case), as well as a new 100µF/50 cap for C13 (or C16 or whatever it is, the one you suspect is bust), and... a 500 ohm trimpot ! Time to get back to work and do some tests etc !
 
Did you remember to get a 3.3K to replace the R14, 3.9K resistor?

Do you plan to switch gears for a while and see if we can do something about the lack of a delay while keeping Q7 (E-C) shorted?

Your ? diode designation is probably D1. You forgot the second LED on the output.
 
Last edited:
Did you remember to get a 3.3K to replace the R14, 3.9K resistor?

Yep - ready to go in.

So, I just changed C2 on PCB1, and the "thump" is partly gone, or at least seems to be reduced... Would that make sense? A simple 100pF cap that was, at least in part, responsible for all of this? There is still a noticeable surge, but nothing as violent as the original videos I posted...

However there is an audible mains hum, still, that could be an interesting area to look into...

Do you plan to switch gears for a while and see if we can do something about the lack of a delay while keeping Q7 (E-C) shorted?

Sounds good to me - I have the turnpot ready too, either works.

Should I change C16 and C9 on the pre-amp PCB just to be safe and rule those out? Should I also change R14 on PCB1 to its 3.3k replacement?

Your ? diode designation is probably D1. You forgot the second LED on the output.

Good stuff, thanks, will do some edits!
 
Last edited:
However there is an audible mains hum, still, that could be an interesting area to look into...

Interestingly, the hum is present as soon as the AC is on, whether the amp is on stand-by, auto, or ON mode... And it's not the transformer noise, it's definitely a hum going through the speaker !

4. You need to short the source input with a resistor to be a valid test for hum. The closer the Hfe of Q1 and Q2, the better the hum rejection. Is your measurement an AC voltage?

Am getting sidetracked! I can focus on this at a later point...
 
Last edited:
The voltage should increase across the capacitor starting when audio is detected. Use
You meter on a fixed range. Don't autorange. You have a 100 uf cap to substitute here
because you can't test it with your meter.

OFF mode : C16 = 0V
AUTO mode : C16 = 0V, and then progressively increases, continuously.... The voltage just seems to climb slowly, regardless of source plugged in, or not.
ON mode : C16 = 14.7V (with or without music)

test #3
Look at the voltage across the cap, as you apply audio. Use a fixed voltage scale on your meter, not auto-range.
Look at this voltage when you apply music. It should steadily increase.
With music applied, it jumps straight up to 14.7V.



D5 = 0.52V whether amp is on, off, or auto (very slight variation, 0.516 ON, 0.519V OFF, if that accuracy means anything at all...



BTW, I changed C16 out of precaution.

EDIT : Those tests have been done with Q7 not shorted, the heatsink gives me a couple of minutes to do the tests before things get hot.
 
Last edited:
Wow!! Too many things to deciefer.

C2: It could help a little since it defines the bandwidth of the amp. If you think of a square wave applied at startup and a square wave has "infinate bandwidth" in theory, then restricting the BW reduces the slew rate.

I don't like your C16 tests.
C16 Auto mode is not good. Need to think.

I'd like you to do one C16 test over:
Switch to OFF for at least 3 minutes: Put meter on C16, fixed range. Switch to ON. Look at your meter AS you switch to on. There has to be an increasing voltage.

D5: fine.

HUM: Not sure how to troubleshoot without a scope.
E on Q15 to ground. DC voltage and try to measure AC ripple, fixed range. Start with a high scale and work down manually ranging.
E on Q6 to ground. Same

Basically I'm stumped on that one.

You can try lifting the end of C1 that points to SG of the connector. Clip the end you lifted to ground. That will tell us if it's in the preamp or power amp section.
Gain matching Q1 and Q2 will have an effect on common mode noise.

For fun, you can always put the bias pot in. Won't hurt.

HUM: Waiting for determination of what section. Power or pre-amp.
Delay: Waiting for C16 ON test (probably not done right). Consider what to do next.
POT, Resistor: It will give you something to do.
Match Q1, Q2: MIGHT reduce hum
 
Last edited:
Wow!! Too many things to deciefer.

Sorry yes I get carried away...!

I'd like you to do one C16 test over:
Switch to OFF for at least 3 minutes: Put meter on C16, fixed range. Switch to ON. Look at your meter AS you switch to on. There has to be an increasing voltage.

No change, this is weird. Fixed range, AC off for 10-15min... I look at C16 voltage:
OFF = 0
AUTO = voltage climbs steadily, although AC is OFF??? 0.14V and climbing.
ON = 0.57V roughly, once again, AC off !

AC ON
AUTO = climbs around 0.001V per second
ON = 14.7V

Gain matching Q1 and Q2 will have an effect on common mode noise.
That could be fun to do once the rest is sorted out. I have the principles on the last post you wrote on page 13, I can focus on this and start questionning when I get there!

You can try lifting the end of C1 that points to SG of the connector. Clip the end you lifted to ground. That will tell us if it's in the preamp or power amp section.
Will do.

For fun, you can always put the bias pot in. Won't hurt.
Sounds good to me, I'll wait for some recommendations on plugging that in, not sure how to proceed exactly, need to look up classification of the pot and figure that that!
 
Another Measurement: Power supply decay
Look at the +12 and -12 voltages at the connector or C13 or C14 or anywhere else convienient.
About how long does it take for these voltages to decay to about 1/2 their value.


POT:

Hopefully you picked up something that is close to this: https://media.digikey.com/photos/Bourns Photos/3006P SERIES.jpg
There are others that are about 1 cm square. All which are multi-turn which I prefer. A single turn will work too. Their are lots of varieties that will work.

POTS or Potentiometers (at least the expensive ones) are marked CW, CCW and the wiper is indicated. The middle terminal is the wiper. Consider a single turn panel pounted pot will have three terminals. With the shaft toward you, CCW, W and CW are the implicit markings when the terminals are at the top. This means that when the POT is turned to the extreme CCW position, CCW and W are the lowest resistance ( ~0). When the POT is turned in the extreme CW position, The CW and W terminal have the least resistance.

R14 has to be replaced with a 3.3K with a 500 ohm variable resistor in series. I would expect, that as the resistance of the series combination increases, the voltage across Q7 (E-C) increases. So with the POT full CW, use your meter to find 500 ohms between the wiper and (CW/CCW). It will be CCW. The end of the resistor that went to Q7 can go to CCW and the wiper to the existing base connection. Since we don't know, set the POT to 250 ohms before applying power. Monitor the voltage between Q7 (E-C) and see if it increases when you turn clockwise. It's customary when POTs are used in this fashion is that the unused terminal be connected to the wiper. So, CW will be connected to the wiper in the final configuration.

This is a little different and simpler than I described before.

Once you get it that far, set the bias to about 2V and we'll take it from there later.
 
Last edited:
Are we having fun yet?

I looked at your data and I think:

1. Q2 is leaky, or
2. Something is causing Q2 to turn on slightly or intermittantly such as Hum in pre-amp section.

So, curious about the pre/post test and the power supply decay test.
and

With no signal (inputs unplugged), look at pin #1 of the op amp in both AC and DC mode. From now on for any measurements that are AC use units. e.g. 5 mV AC or 5 mV. The latter will be assumed DC.

Back burner: Later when this thing is hopefully tamed. I'd like you to possibly insert a 1K resistor between the AUTO terminal of the switch and the R44/R40 junction. This should improve the ON position thumping.

I can't tell which direction the Emitter direction arrows are pointing on your revised schematic. Q1 is NPN. Q2 is PNP from the datasheets.
 
So, I added a 3.3k and 500ohm trimpot to the extra PCB I had setup for the caps, linked the two in series and replaced R14. Then... I noticed I had already replaced R14 with a 3.3k resistor?! Not sure if I had misunderstood you in any of the earlier posts, but it was already a 3.3k resistor, instead of 3.6k. Now I'm thinking, if you are right about increasing resistance equating to increasing Q7 voltage, then this this setup will be, at best, equal to the 3.3k, and at worst... 3.8k. I assume I'd have to get a 3k resistor instead...?

I'll reassemble the thing anyhow and do some testing.
 
Last edited:
And the drama continues! I thought I'd look at voltages at Q7 E-C, with the 500 + 3.3k combination, in both most extreme conditions (3.3k and 3.8k). Realising I would probably have to change the 3.3k resistor I didn't reassemble the PCB on its heatsink, and thought rapid measurements would be fine. Big mistake (and I'm a slow learner apparently!), I smoked Q10, Q11, and another 1A fuse. Before burning them, I noticed the following:

"R14" (500@min + 3.3k) at 3.3k yielded Q7 E-C = 2.5V. I didn't have a chance to increase it much before it burnt out, I think increasing the resistance does indeed increase voltage. So on top of new Q10-Q11, I'm going to get a 2.7k and a 3k resistor to play around with...

I looked at your data and I think:

1. Q2 is leaky, or
2. Something is causing Q2 to turn on slightly or intermittantly such as Hum in pre-amp section.

I tested Q2 out of circuit, I think it's fine. If in doubt I can probably get a spare, and substitute it anyway... However, when I plugged the whole thing back together, with AC on, before burning the transistors, I did notice something strange. With the RED diode, so stand by mode, there was a major hum going to the speaker. Now, if I touched part of the PCB, metallic parts, the hum would sometimes disappear - was I earthing it? Another interesting observation, when I forced it ON, the hum disappeared; this got me sidetracked and I then proceeded to burn my Q10-Q11...

So I'm off to get some replacements, again, a few more resistors, a few more 1A fuses... And some talent if I can find some on sale.

On a side-note, I also progressed on the improved PCB map, got most of the circuit deciphered, so that'll be fun to look at.

With no signal (inputs unplugged), look at pin #1 of the op amp in both AC and DC mode. From now on for any measurements that are AC use units. e.g. 5 mV AC or 5 mV. The latter will be assumed DC.
Another item on my to-do list, once I have replaced the transistors !
 
I'm still rather worried with this Q7 voltage thing. If I decrease Q7 E-C voltage, then we'll get less heat off Q10 and Q11?

I tried replacing my 3.3k resistor with a 3k resistor, on my extra PCB with the 500ohm turnpot. After replacing Q10 and Q11, I found Q7 E-C = 2.6V, when the "R14" replacement is set to 3 ohms! I thought reducing the resistance would help reduce that voltage; or is it because of R13, at 3.3k?
 
Leakage is going to tough, if not impossible, to check with the diode test. Need the other tests. It would be E-C leakage. It might show up on ohms.

OK, I said to set the POT to half-way.

Since I had some time to think. On Q7, the voltage between B-E (R14) will change the the conduction. Assuming the current through R13 and R14 is the same, then R14 being a larger resistor means a larger voltage between B-E and therefore the base of Q7 would get more current as R14 gets larger. If Q7 conducts more, then the volateg drop across Q7 (E-C) gets smaller. So, I believe, as R14 gets larger, there is more bias to Q7 and the chance of overheating increases.

So, the hFe of Q7 and the value of R14 controls the actual bias. It must be R14 that gets bigger and not R13.
The system must also be "balanced" If R13 and R14 were identical AND the voltage wasn't enough to turn on Q7, you should have a reduced voltage drop across Q7(E-C).

At Zero bias, the voltage between the junction of R11/R15 and R12/14 should be half the supply voltage of 37.5V. You can always pull Q10 and Q11 again.

"R14" replacement is set to 3 ohms!
Don't you mean, the R14 replacement set to 3K?
 
Leakage is going to tough, if not impossible, to check with the diode test. Need the other tests. It would be E-C leakage. It might show up on ohms.

Ah... Good to know. So "blanket replacing" might be a good idea here.

OK, I said to set the POT to half-way.

Well I figured I better be safe than sorry.... Although that didn't work out too well!

Don't you mean, the R14 replacement set to 3K?

Oh yes, sorry, definitely 3k!


Slight setback. I think I shorted something, again? This time the power LED just about has the time to go RED before it goes completely dark, the 1A fuse burning out... I suspect this could have been caused by the Q7 side of R14 getting in contact with the transformer: my "extra PCB" slipped and the R14 extension had a bit of resistor touch the metallic casing of the transformer. Not sure if that could be the cause. But at least Q10-Q11 aren't burnt out, the LED has (barely) the time to go red before dying out with the fuse...

For the moment I removed my extra PCB, replaced C1 and C10 with the original, lower quality caps, and replaced R14 with the previous 3.3k resistor, just so that I would not risk doing anything else stupid and blowing up any more fuses. I haven't attempted to switch the amp back on since, I think I need a rested mind before I attempt anything else. I'm (slowly) learning that I really have to be rigorous but also think things through before testing them with electronics...

Clearly my ambition outweighs my talent...
 
Last edited:
Blanket replacing: Still some tests to do.

That's why I said half way. Any other time I would have said fully CCW (minimum bias).

3K and 3.3K is disportionate. With 20/20 hindsight, I'd now say the same value as the other resistor. The same might not be good enough though.

You never built the AC series light bulb gizmo. It would have saved some agrivation.

Clearly my ambition outweighs my talent...

This is one way to learn. It's effectively how I did.

It's also part of:

If you don't have time to do it right, when are you going to have time to do it over?, but maybe that doesn't apply here.

You can always unsolder ALL of the parts and replace them with new ones <G>?

NO MORE PLAYING UNTIL YOU MAKE THE LIGHT BULB GIZMO!! Deal? Don't even think about it.

Look on the bright side: At least it doesn't hum anymore.
 
Last edited:
Quick questions: What's your plan? What are you going to do first?

About to go and buy a small sprocket for an old 25W bulb I found in a cupboard. Incandescent bulbs are actually hard to come by in Europe now, with energy saving rules in play.

So I plug it in series between the transformer generated current and the PCB?

Look on the bright side: At least it doesn't hum anymore.

Indeed; and it will be brighter with a lamp gizmo.
 
So I plug it in series between the transformer generated current and the PCB?

Put the bulb in series with the outlet on the wall and the primary of the transformer. Basically, In series with or in place of, the fuse.
 
Status
Not open for further replies.
Cookies are required to use this site. You must accept them to continue using the site. Learn more…