Repair ZP120


The amp has no power. Is there an internal fuse for this unit?

Has anyone taken one apart?

372 replies

Hello, I replaced the components and it works again. Thanks very much

 

Hi Tim, thank you. I have now ordered two of them. Can I destroy them by changing the AC voltage from 230V to 115V on the 230V grid?

Yes that could cause a lot of damage.  The usual rectification generates the input times the square root of 2 as DC voltage.  SO 115V input translates to about 162V.  The switch on the back adds a voltage doubler to the rectifier circuitry.  With a 115V input, that produces 320V internally, which drives the high voltage circuitry.  If 115V is selected but the input is 230V, then it would internally generate 640V DC.  That could break a lot of things.  All of the power passes through the thermistors, and they blew up in protest of the voltage.  The question is if they went before anything else ‘downstream’ died or not.  It is possible that there was also damage to the PWM chip.  This is the four pin device on a heatsink next to the transformer.  You can check if it’s pins are shorted or not.  If there are short circuits between pins then it is dead.  If there are no shorts, it doesn’t mean that it isn’t broken though.

Hi Tim, thank you. I have now ordered two of them. Can I destroy them by changing the AC voltage from 230V to 115V on the 230V grid?

Hi Max.  Those are both NTC thermistors (not capacitors).  They can be replaced with part # SCK-054.

They act as 5 ohm resistors at room temperature.  As they warm up, their resistance drops.  This prevents a large amount of current inrush at power up.

In a pinch, you could try replacing them with jumpers, but I would only do that if you are powering the board with a dim-bulb (as described earlier in this message thread).  In that case, the incandescent light bulb will limit inrush current.  

Hello, im from Germany and need your Help.

These two capacitors are defective. do you know the data for these?

TH16001 and TH16002.

 

Thank You

Max

Hi, sorry for not coming back sooner.

looks like the 4 resistors are acting as a voltage divider off the 320v rail. The 19v zener is then just working in the normal way. If any of the resistors in the chain is out you wont get 19v. I thin d16041 is there to stop feedback once the pwm chip is providing its own supply.

if you got it working ok as you described, then i guess thats fine. If not, then check the resistors.

 

Kind regards,

 

Andy

Userlevel 3
Badge +3

So I gave up on trying to solve the problem with the existing circuit.  I ended up replacing the 240k resistor at R16082 with a resistor divider network.  I used 15k+4.7k+100k, with one leg of the 100k tied to the 59k resistor, and one leg of the 15k tied to GND.  The junction between the 4.7k and 100k resistor should then end at about 18V (ie approx (15+4.7)/(15+4.7+100+59)*sqrt(2)*115).  I tied that to the VCC input of the PWM.  I used ¼ watt resistors for all three of the resistors.  The lower resistance at R16082 (~120k instead of 240k) drops the voltage at the junction between R16079 and R16082 to 109V.  The current flowing through the resistor divider is 0.9 milli-amp when the PWM draws no power, and <1.1mA when it acts as a short circuit.  As a result, power dissipation for the largest resistor (100k) in the resistor network is under 1/10 watt (making the ¼ watt resistors overkill)

The circuit works fine now.  I didn’t time it, but it may take a couple of seconds longer to boot up.  There is a 100uF capacitor on the PWM VCC, and that needs to charge up before the PWM VCC line exceeds the 16V it needs to start oscillating.  Otherwise, there is no issues with the circuit.

Maybe this only works so well since the PWM chip is tolerant of voltage fluctuations on its input (it needs to start up with >16V, but then will operate with any voltage from 8-24V).  The original circuit may have provided a more stable VCC, but it’s just not needed given there is only one chip drawing less than a milliamp.

 

Userlevel 3
Badge +3

I had a typo:  I mislabelled the zener.  It is D16035 on the board (not D16041).

Userlevel 3
Badge +3

Sure.  The attached is how I think the circuit is organized.  I don’t quite ‘get’ how the 19V zener is providing 19V in this, but working devices do in fact read 15-18V when running.

D16040 reads with 0.68 forward voltage drop.  I wonder if there is a reverse voltage avalanche breakdown property that is allowing some current to pass to the cathode of D16041.

 

Hi, can you sketch the circuit around these diodes out for me?

I would imagine if the other one gives .68v, then it is probably a general purpose silicone diode. 

As I say, sketch it out and I'll have a think.

 

Kind Regards,

Andy

Userlevel 3
Badge +3

I’m a bit stumped on a ZP120 and am looking for help.  The zp120 had a bad varistor, between VPRI and VMID), with lots of scorched bits when it caught fire.  It took out the PWM chip, the fuse and a board trace.  I’ve replaced those but still can’t get it working properly.

The attached image is from the underside of the zp120 board, below the large capacitors.

I have circled two small circuits in green.  The one on the left looks like a bleeder resistor that is intended to discharge the large 1000uf capacitor that sits between VPRI and VMID (about 150V).  The 68k and 240k resistors provide 308k of resistance across the capacitor.  I thought the small “p1.c” diode is some kind of snubber diode, but I’m not so sure any more, based on what’s going on in the other circuit...

The similar circuit one on the right seems to bleed the 150V between VMID and PGND.  That circuit also has the addition of D16035, which sits between the bottom legs of R16082 and D16041.  This additional chip is a 19V zener diode in a SOT-23 package.  (MMBZ5249B).  Pin 3 of this zener diode (which is presumably supposed to be at 19v when the circuit is working property) is the voltage that drives the PWM chip.

My board reads 0V at that junction.  I have replaced the zener with no luck.  When I provide 16V from an external source, the PWM operates as it’s supposed to.  The PWM chip needs 15V to get started, but operates ok with as little as 9V… if I remove my external 16V, the circuit continues to operate (and somehow the circuit now internally generates about 12.5V at the same junction of R16082 and D16041).

I can’t figure out how this circuit works.  Specifically, the junction between R16079 amd R16082 reads about 110V when powered up (this seems right given the resistor divider created by R16082 and R16079).  I understand that the zener is supposed to clamp that to 19V, but what is the P1C diode (D16041) doing?  If it is a regular diode then it wouldn’t pass any voltage to the zener.  If it isn’t a vanilla diode, then what is it?  I’m guessing D16041 isn’t working but I don’t know what to replace it with.  It does not read a voltage drop in either direction when tested (the other P1C reads 0.68V).  I can’t find a reference to that marking anywhere.

 

 

Fantastic, thanks Tim(?). 
not often a cheap repair with Sonos!

i’ll go and find a bit of wire now!

 

Kindest regards,

Andy

Userlevel 3
Badge +3

Hi, can anybody tell me the vale of resistor R15550, located under the WiFi module on the ZP120?

This component is missing on my board ( knocked off and lost by previous “repairer”). Oddly, the player still works, but I would like to replace it.

 

You’re in luck!  That’s a 0 ohm jumper, so no parts search required...

 

Hi, can anybody tell me the vale of resistor R15550, located under the WiFi module on the ZP120?

This component is missing on my board ( knocked off and lost by previous “repairer”). Oddly, the player still works, but I would like to replace it.

 

Userlevel 2
Badge +4

Here’s what it looks like to me:  Sonos may have done away with their “recycle” program, but now they’re using “forced obsolescence”, and shutting down older units to force people into upgrading. 

 

Upgrade to what? The Amp isn’t available to ship until at least Sept. 3. My ZP-120 continues to work fine. 

That's some sketchy paranoia right there.

I’m in the same boat as most of you (though I haven’t opened up my unit to see what’s going on - I wouldn’t know how to fix it anyway).  My unit refuses to power on, and it happened I’m sure “coincidentally” right after I updated my app.  

Here’s what it looks like to me:  Sonos may have done away with their “recycle” program, but now they’re using “forced obsolescence”, and shutting down older units to force people into upgrading. 

My unit - a ZP120 - had been used so infrequently that it should have lasted 40 years.  And now, all of the sudden, and right as I updated my app, it goes bad for no reason?   And Sonos is telling me that I can’t get another ZP-120 and I have to get one of the newer models.  

That’s some sketchy business practices right there.  

Userlevel 3
Badge +3

That is likely not beyond repair!

It’s hard to tell from the picture, but there are 6 components under that area and likely at least one of them is bad.

There are four large flat diodes.  These can be tested with a multimeter to see what the forward voltage drop is across each.  They should conduct power in one direction only.

The other two components are on the other side of the board.  They have 3 legs each, and are power mosfets.  (like this: http://www.datasheetcafe.com/k3797-datasheet-mosfet/).  You can test like a diode between each pair of legs.  You should not see any short circuits.

The mosfets are part of the circuit that provides 36V to the amplifier.  That 36V is not actually required for the device to power up properly.  The fact that you don’t have power means that there is another problem.  It could be as simple as the fuse being blown. That is on the other side of the circuit board, and is a small short cylinder about 1cm tall  (usually brown).

Do you have a multimeter to test, and have you ever done electronic soldering before?

Hi Im not tech enough to open and start fixing mine.

Could there be anyone that could fix mine?

Ive talked to Sonos, but no help there. It online and apparently “working” but no sound.

Cables and speakers working fine.

PM me - maybe I can help.

 

Looks like a lightning has toasted something?

Hi Im not tech enough to open and start fixing mine.

Could there be anyone that could fix mine?

Ive talked to Sonos, but no help there. It online and apparently “working” but no sound.

Cables and speakers working fine.

PM me - maybe I can help.

 

Userlevel 3
Badge +3

Hi Im not tech enough to open and start fixing mine.

Could there be anyone that could fix mine?

Ive talked to Sonos, but no help there. It online and apparently “working” but no sound.

Cables and speakers working fine.

PM me - maybe I can help.

Hi Im not tech enough to open and start fixing mine.

Could there be anyone that could fix mine?

Ive talked to Sonos, but no help there. It online and apparently “working” but no sound.

Cables and speakers working fine.

Userlevel 1
Badge +3

Hey everybody! It’s been a while.

I just found something that is contrary to something i have said over and over in these forums. I have a AMP (1st gen version) that had the 36V transformer winding return shorted to the 5V winding return. I shorted them like i suggested in many posts in this thread.

The amp functions fine.

What i noticed was that in the time delta between the 36V coming on and the sound starting there was quite a bit of noise. Normally i use these things in noisy environments and i don’t hear that noise at all. But now in my quite home office that noise really bugs me. I removed the short and not only does the unit still function, but the noise is gone.

I am not sure how this thing works without the 36V and the low voltage biases returns tied together, but it does on mine. 

 

Just my 2 cents.

 

Man how much does recycle mode suck?!?! LOL

Userlevel 3
Badge +3

Thanks, will check it out!

Userlevel 3
Badge +3

In the meantime, are there test points where I can check for this 36VDC and/or other voltages to see if my amp section is at least attempting to engage when I hit play?

I’ve forgotten where the switching on occurs in the unit.  However, I did have a power board lying around and took some measurements.  You should find 36V between the labelled pins.

 

Reply