Repair SMPS power supply from Bose subwoofer

[2PAC Mafia]

Hi guys,

in the past I already had to repair this unit with problems at power supply. I had to replace these components:


U101 = L6598D
R101 = 10 ohm
R102 = 33 ohm
Q101 = STP9NB50FP (installed STP9NK50Z)
Q102 = STP9NB50FP (installed STP9NK50Z)
D101 = K6 (MMBD914L1) (installed KA2 = MMBD914-7-F)
D102 = K6 (MMBD914L1) (installed KA2 = MMBD914-7-F)

This time, again the equipment with problems at power supply so I replaced :

R101, R102, R103, R104, Q101, Q102, ZR101 and U101

After replacing them I see how at pin 12 of U101 are only coming 4V so I tested the resistor divisor R121, R122, R155 and R156 and all are right.

I tested the output of U101 only connecting 11Vdc to see if it works and it does. Now I don´t know why I get only 4V so the U101 is not starting up the power supply.

I attach the schematic:

http://www.restoretronic.com/descargas/sd254165_1_3_02 Shematic Fonte .pdf

May be you can give some idea.

 

[Tony Stewart]

Vs pin 12 is driven from R ladder input and bootstrap cap C112 from output, If neither can reach 12V then the drain current on pin 12 is too high for the source impedances given.

 

[tomizett]

Can I clarify, when you tested U101 did you remove it from the circuit and test it seperately, or did you apply 11V to it in-circuit? If you tested in-circuit, did that force the power supply to start up? Also, do you know what current it drew - the datasheet suggests about 2mA.

Do you think that the power suppy has runs for a very short while as soon as power is applied? If that's the case then I suspect that the "run" power supply (C114-D103) is not working.

In any case, this supply seems a likely source of trouble. What impedance do you measure between pins 12 and 10 of U101? It should be very high. If you remove D103, does the voltage come up further?
I'd start by checking D103, C117, ZR102, D108, D112 and C114 for leaky diodes and open-circuit caps. Check Q103, too, in case the gate has gone short.
Do you know the purpose of Q103 and friends (is it some kind of PFC circuit)?

Let us know how you get on.
(Mods - perhaps this belongs in "repair" forum?)

 

[tcmtech]

If it was me I would rip out the stock power supply and replace it with a higher quality independent system made in backwoods China being if you have ever looked at any Bose equipment in detail and cross referenced their parts to anything you will find that they are usually made out the cheapest crap they can get away with. Just look at most of their commercial sound speakers that are in enclosures that don't have removable covers over the speakers. They are the cheapest looking and built paper coned stamped low grade metal framed POS components you will ever see backed up by a a ton of signal processing to make them sound sort of reasonably good in most controlled environments.

Really. Try running bose speakers on a good audio system by themselves. They suck big time without the heavy signal processing or try running some good high quality speakers off a bose system. They too will sound terrible due to the excessive amount of unnecessary signal processing that is being fed to them.

Honestly given the amount of rebuild work you have done to that power supply already to keep it functioning it's apparently built about as well as the rest of their stuff.

 

[2PAC Mafia]

Hi,

Can I clarify, when you tested U101 did you remove it from the circuit and test it seperately, or did you apply 11V to it in-circuit? If you tested in-circuit, did that force the power supply to start up? Also, do you know what current it drew - the datasheet suggests about 2mA.

I only connect power to the IC in-circuit), without 220Vac coming into the circuit, just to check the PWM output signal. In any case I replaced by another one and it had the same behavieur.

In any case, this supply seems a likely source of trouble. What impedance do you measure between pins 12 and 10 of U101? It should be very high. If you remove D103, does the voltage come up further?
I'd start by checking D103, C117, ZR102, D108, D112 and C114 for leaky diodes and open-circuit caps. Check Q103, too, in case the gate has gone short.
Do you know the purpose of Q103 and friends (is it some kind of PFC circuit)?

If I measure impedance (Z) with LCR meter I get 17 ohm, if I measure resistence I get 18,7 K.

If I open the D103 line I get same voltage, around 4V.

I have checked one by one: D101, D112, D108, D103, D105, ZR101, ZR102, C126, C120, C117, C118, C114, R106, R109, R110, D111, ZR103.

I don´t know the function of Q103, it seems some kind of extra protection for Op + input.

 

[2PAC Mafia]

Hi,

If it was me I would rip out the stock power supply and replace it with a higher quality independent system made in backwoods China being if you have ever looked at any Bose equipment in detail and cross referenced their parts to anything you will find that they are usually made out the cheapest crap they can get away with. Just look at most of their commercial sound speakers that are in enclosures that don't have removable covers over the speakers. They are the cheapest looking and built paper coned stamped low grade metal framed POS components you will ever see backed up by a a ton of signal processing to make them sound sort of reasonably good in most controlled environments.

Really. Try running bose speakers on a good audio system by themselves. They suck big time without the heavy signal processing or try running some good high quality speakers off a bose system. They too will sound terrible due to the excessive amount of unnecessary signal processing that is being fed to them.

Honestly given the amount of rebuild work you have done to that power supply already to keep it functioning it's apparently built about as well as the rest of their stuff.

This unit is installed on a yacht, with communication cables going through walls and very difficult to replace so they wanted to repair the unit again as Bose technicians didn´t want to repair it the first time and the unit is obsolete. Now they have the owner on board as it´s summer time...

 

[2PAC Mafia]

Checking the schematic I guess:

The first startup of IC should be through R156, shouldn´t be? But I don´t understand why it takes the voltage from AC line...

When it starts up then the voltage is stabilized at ZR102.

Is it possible to open the D108 and put an independent 12Vdc source? What do you think? Can I break anything at circuit doing that?

 

[tomizett]

Yes, C118 should charge, slowly, through R156 etc until it reaches the turn on threshold voltage - the controller IC will then switch on and current will be supplied through C114. I can't see why the start-up path is from the AC side - it's usually post-rectifier, or you'd expect a diode in that resistor chain. It must be something todo with whatever Q103 does.

So, if the fault persists with D103 out then it looks like something local to U101. 18.7k doesn't to ground sound too bad though...
My next move would be to start taking things out and see at what point the voltage is allowed to rise (making sure you don't allow too much voltage on C118). Try removing U101 and placing a 12v zener between where pins 12 and 10 go - see if you get up to 12v. You may want to tie the gates of the MOSFETs while you do that to make sure they don't float high and turn on.

You *should* be able to connect a power source directly across pins 12 and 10, assuming your supply is isolated from ground - just double check everything first. Don't exceed the regulation voltage of the IC (see the datasheet).


And TCM... don't be such a killjoy! I know they're crap, but at least we can make them as crap is the day they left the factory...

 

[JimB]

And TCM... don't be such a killjoy! I know they're crap, but at least we can make them as crap is the day they left the factory...

ROTFLMAO

JimB

 

[Les Jones]

When you say that you have checked the capacitors exactly what did you check. Capacity, ESR , leakage ?

If I measure impedance (Z) with LCR meter I get 17 ohm, if I measure resistence I get 18,7 K.

As C118 is connected between pins 12 and 10 I would expect it to show a low impedance unless the LCR meter uses a very low frequency. I suspect C118 might have a high ESR I have seen a few switch mode power supplies not starting when the smoothing / decoupling capacitor on the control IC had failed with high ESR.

Les.

 

[Tony Stewart]

The Bose design 3 stages of regulation.
1st it uses an offline method of pre-regulating using a phase controlled triac with brown-out protection with Q101 for the DC-DC regulator which provides the minimum dropout for the LDO to drive with ripple free-power for only 1.5A at +/- 8.7V with a dropout of 3V

So why not replace the whole supply with a cheap dual DC-DC of +/12V @1.5A?
then drive the LDO's $50

 

[2PAC Mafia]

Hi,

When you say that you have checked the capacitors exactly what did you check. Capacity, ESR , leakage ?

My LCR meter is checking for capacity and ESR, I don´t have any instrument to check leakage.

As C118 is connected between pins 12 and 10 I would expect it to show a low impedance unless the LCR meter uses a very low frequency. I suspect C118 might have a high ESR I have seen a few switch mode power supplies not starting when the smoothing / decoupling capacitor on the control IC had failed with high ESR.

I measured and it seems good, 2 ohm ESR, 9,1 uF 0,128 D. In any case I can replace it to be sure it´s not causing the problem...

The Bose design 3 stages of regulation.
1st it uses an offline method of pre-regulating using a phase controlled triac with brown-out protection with Q101 for the DC-DC regulator which provides the minimum dropout for the LDO to drive with ripple free-power for only 1.5A at +/- 8.7V with a dropout of 3V

So why not replace the whole supply with a cheap dual DC-DC of +/12V @1.5A?
then drive the LDO's $50

For me that is the last option, it´s intalled on a big heat case and interconnected to 2 more boards inside, then everything screwed at back side of subwoofer case.

My last step was to remove D108 to open the power source, connect 11V at C117 and complete power supply switched on getting 9,6V and -10V as outputs. Would you design an auxiliar 12V power source for starting up the power supply?

 

[2PAC Mafia]

I found also C119 out of tolerance and it has to do with bootstrap but after replacing it I still have the same behavieur.

I only have this idea now for not wasting much more time looking for the start up failure, regulate 12V to connect at C117, what do you think? how would you do it?

 

[Les Jones]

I think the ESR of C118 is OK. (I checked a few 10 uF 25v capacitors I had and they varied between 1.2 and 2,6 ohms.) I would just check the leakage resistance using a power supply, a current limiting resistor and a multimeter on the current range. At first I could not understand how the startup circuit worked with R121 being connected to the AC side of the bridge. (I was falsely thinking that the AC would be centred on PCOM but I eventually realised that the waveform would have the negative crest at zero volts (With respect to PCOM) (Actually -0.7 volts due to the forward voltage drop of the bottom left diode of the bridge.) the positive crest would be about 340 volts (Assuming 240 V mains) Have you checked the value of R121,R122, R155 and R156 in case they have gone high in value ? It would probably be quicker to replace C118 than measure it's leakage current.

Les.

 

[2PAC Mafia]

Yes, I already checked the resistor chain, this was one of the first things I did.

 

[tomizett]

Les has a good point - I also wondered if C118 could be leaky. Try removing it and see if the 18.7k you measured earlier increases or not.

It's good news that you did manage to start the power supply with an external supply. You should be able to do the same with D108 in place. If you feed in your voltage with a diode in series with your power supply (the new diode's cathode meeting the cathodes of D108 and ZR102) then you are safe from feeding voltage into your external supply, and you should be able to remove your external supply once the circuit has started and see if it keeps working.

Tony, I wonder if you are mistaken about the outputs from the this supply - I think that me main rails come from the points labeled +v and -V, and the linear regulators just supply the low-power signal stages. Voltages aren't marked, but surely it's more than 12V? Or is it me that's missing something?
Also, I don't think that Q401 is doing any phase control - looks like it's just switching the voltage-doubler action for mains voltage selection. The AVS12 seems to be a dedicated IC for this purpose.

 

[Les Jones]

Hi Tony,
I think tomizett is correct. I believe the voltage between V+ and V- is about 51 volts. (+ and - 25.5 volts) Here is the reasoning. A fraction of the voltage between V+ and V- is compared with a reference voltage of about 6.8 volts. (6.2 volts (Zr300) + the vbe of Q301) The ratio of R305 and 306 defines the output voltage. This would be too high just to supply the input to the two linear regulators. (Also if only the main amplifier was powered from just the + and - 10 volts output at about 1.5 amps it would be a very low power subwoofer.)

Les.

 

[2PAC Mafia]

Hi,

Les has a good point - I also wondered if C118 could be leaky. Try removing it and see if the 18.7k you measured earlier increases or not.

I replaced it just in case but same behavieur, also same resistence.

What do you think about regulate 12V from main bridge rectifier and feed them at C117?

 

[2PAC Mafia]

Hi,

It's good news that you did manage to start the power supply with an external supply. You should be able to do the same with D108 in place. If you feed in your voltage with a diode in series with your power supply (the new diode's cathode meeting the cathodes of D108 and ZR102) then you are safe from feeding voltage into your external supply, and you should be able to remove your external supply once the circuit has started and see if it keeps working.

Yes, only with some seconds with my power and then I remove my power and power supply stays on.

 

[Les Jones]

Could ZR101 have gone short circuit and the 18.7K reading between pins 10 and 12 be the 20 K resistor R109 ?

Les.