External SoundCard Repair - **FIXED!**

[MrJoeyJoeJoe]

Hey guys!
I've been trying to fix my brother's Liquid Saffire 56 Soundcard and it's symptoms are no power on. Power gets to the powerboard from the switch but is not getting to any of the other boards - I think! - otherwise the lights would come on.




I've checked the first fuse for continuity and it's good. But then I've seen these two things;






The Green one says "WME CHJ471K"
And the Blue one says "WMR 70D471K"

I've searched Internet and can't find anything to do with those references. I'm pretty sure they're fuses no?
I checked continuity (in circuit) across the green one and that's good but the blue one hasn't got continuity!
Could this be the problem?
Are they fuses?

Thanks a lot,
Joe!

[SeanB]

Green is a NTC inrush limiter, and the blue one is a class Y interference supression capacitor. NTC will have a resistance of around 47ohms, and the other is a 470pF capacitor, like the 2 on the other side of the yellow common mode choke.

It is not your problem, look on the 2 heatsinks and there will be on the one a diode ( top pic the one on the left) and on the other either a power transistor or a TOPswitch device. Photo of the underside of the board will also help.

Please be careful, as this board will have high voltage on it, even after you have disconnected the power and left it alone for a while, and when in operation roughly half the board is at mains voltage.

[MrJoeyJoeJoe]

Thanks a lot for your speedy reply!
Damn I was hoping the blue capacitor was actually one of these resetable fuses!;
http://d1gsvnjtkwr6dd.cloudfront.net/large/PA-PTC-1100_LRG.jpg

As you may have guessed I'm a bit of an electronics Noob!
I did however make sure that scary 450v cap was pretty much drained before touching though! Thanks!
Here are some more pics.






So you think I should check what could be a TOPswitch on the left heatsink?

[MrJoeyJoeJoe]

I can't see any leaking caps, suspicious or burnt out components and since SeanB mentioned the TOPSwitch I've googled around.
I've seen a troubleshooting guide here;
http://wenku.baidu.com/view/1a46d019227916888486d7a5

And it mentions that if the System has no power, it could be the TOPSwitches shutdown latch has tripped...

So my new question is; Does anyone know how to reset a TOPSwitch!?
(Is it even possible?)

[David_AVD]

Did the unit suddenly stop during operation, or did it fail to power-up after a power interruption?

If the latter, I'd replace C5 (small electro cap on primary side) and also check all of the high value resistors on the primary side.

[janoc]

Are you actually sure the problem is in the power supply in the first place? If you have a short downstream, the PSU may not start - most have some sort of protection circuitry (no, it may not blow the fuse on the mains side!).

Unplug the PSU and check whether or not you have a power rail shorted to ground, by chance. Then check the PSU separately - it may actually start if you unplug it from the rest of the system! If not, you may have to put some load on the rail where the feedback optocoupler is getting its power from. Also make sure to check whether or not it has a soft power on signal - you may have to ground some pin on that output connector or bring some voltage to it for it to start fully - then you should see some standby voltage on the connector already (usually 3.3 or 5V), indicating that the supply is most likely good.

If the PSU seems dead, make sure to discharge the large mains cap again and then check the electrolytic caps. They don't need to be leaking to be bad! You need an ESR meter for that, though - capacitance meter is not enough, it won't show you if the caps are dried out.

Other common problem could be blown switching transistor - check Q1 for shorts. But that is likely not the issue in your case, because if the transistors blew, it would have likely been quite a bang and also the fuse would have blown.

If all that checks out, you are getting voltage on the secondary of the transformer but you aren't getting correct voltage at the output when loaded (when unloaded the voltages could be off because the feedback loop is not yet regulating properly) -> check the voltage regulator(s), diodes, optocoupler, caps on the secondary side.

If you are not getting any voltages on the secondary side and everything above checked out, then you may need to try to reverse engineer the schematics of the board (or find it online). Then the problem could be something more complicated involving the controller IC or some other parts on the high voltage side - but that is not something I would recommend a newbie to poke fingers into.

[MrJoeyJoeJoe]

Thanks David_AVD.
The unit just wouldn't turn on one day.
I'm thinking of buying an ESR... That way I could test the C5 cap in circuit no?

[MrJoeyJoeJoe]

Thanks for your reply janoc.
I presumed it was the PSU as the full 230v are getting to it, but no power appears to be getting to the dashboard lights.

I would love to know a way to trace how far the power's getting to, but can only see GND marked on the other boards, not on the PSU so I don't know what contacts to check with the meter and am worried of shorting something out.

I want to test the optocoupler and have googled how, but it's out of circuit. Is it worth me removing it to test?
Also thinking of removing the Q1 TOPSwitch to test...


BTW I'm thinking of buying an ESR - would that help?

[SeanB]

Not a TOPswitch but a high voltage MOSFET. The controller is the little 6 pin chip by your middle finger, and the resistors by the finger are the start up resistors, and are the most likely failure, along with the small electrolytic capacitor there being faulty. Resistors are possible to check in circuit using a DMM, simply measure the resistance, they should be lower than the marked value somewhat because of the surrounding components, but likely one or more has gone open circuit. If so replace all 6 with the same value.

this PSU will run unplugged ( some whil not) so simply put it back in the chassis and leave the output plug disconnected, but put all screws and plastic sheeting back on the board, and apply power. You should see some voltage on the output pins if it is working,  At least a 5V rail and some 12-18V rails both positive and negative with respect to the chassis, as the groundd of the output is connected to the chassis when it is screwed into the case.

[Rasz]

Thanks David_AVD.
The unit just wouldn't turn on one day.
I'm thinking of buying an ESR... That way I could test the C5 cap in circuit no?

dont, cap costs $0.1, meter $30-100, unless you are planning to make a living fixing electronics for couple of years

as SeanB told you power it up unplugged and start measuring voltages, but you can check for shorts before as its very simple procedure - you dont need to desolder anything when short hunting. Start with diode and transistor (big 3 legged parts on radiators).

[janoc]

Thanks for your reply janoc.
I presumed it was the PSU as the full 230v are getting to it, but no power appears to be getting to the dashboard lights.

I would love to know a way to trace how far the power's getting to, but can only see GND marked on the other boards, not on the PSU so I don't know what contacts to check with the meter and am worried of shorting something out.

There is a connector with a ribbon cable going out of it visible on your power supply board. That's your output. The connector seems to be marked, the individual power rails are likely marked on it. Put the negative (black) lead of your meter to ground (the mass of copper going all over the secondary part of the board, I think it is connected to the second pin from the left of the connector in the photo where you are holding the bottom of the PCB. And then check the voltages on the other pins of that connector with the red lead of your meter.

I want to test the optocoupler and have googled how, but it's out of circuit. Is it worth me removing it to test?

Don't bother with the opto until you have checked the rest. Optocouplers rarely fail and you could just break stuff that is not broken yet if the problem is elsewhere. That optocoupler is part of the feedback network and the power supply must be powering up at least for it to make sense to start checking the opto. It is a LED and a phototransistor and the LED needs power from somewhere - which is the output of the PSU. If there is no power, the opto can't work and has no effect anyway, so the problem will likely be elsewhere.

Also thinking of removing the Q1 TOPSwitch to test...

I am not sure what (or why) you call that as "TOPSwitch". It is a normal transistor, most likely a MOSFET switching the power to the transformer. If you want to quickly check it, you don't need to remove it - turn everything off, make sure the big capacitor is discharged (!!) and check the transistor for shorts. If it has the pins shorted together (0 ohms on the meter) regardless of the polarity (don't forget to swap the leads), then it is certainly blown. If you get different (non zero) readings, then it is likely OK, but to be sure you would have to desolder it and check it outside of the board, otherwise you could be measuring the resistance of the transformer windings, for example.

BTW I'm thinking of buying an ESR - would that help?

ESR meter is very useful for checking electrolytic caps in these power supplies. Especially cheap poor quality caps will dry out over time thanks to the high ripple currents and temperature and their serial resistance goes up. Then you end up with bulging/leaky caps and non working power supply. The ESR meter helps to diagnose such caps that may not be bulging/leaking yet but are about to fail already.

Dave has a video on this:


Now whether or not to buy one - it depends on whether or not you are planning to do more such repairs in the future. If it is a one off and it is an older device you are repairing, you can just blindly replace the electrolytic caps as a matter of course. It is very possible the caps are bad. However, a $30 meter from eBay is not such a huge expense neither.

[MrJoeyJoeJoe]

Thanks a lot for your replies and suggestions guys, I'm currently away from home but will try them out on Monday!
I'm very much an anti-planned-obsolescence kind of guy (in fact me and my brother did a music video about that!  ) so I'm always trying to fix things for the sheer 'screw you consumerism' factor, just like Dave in the EEVblog video!

I've bought an ESR meter now because it's something I could be doing with more and more!
And Solder Wick! What a discovery! I had never heard of it until seeing Janoc's post! (I have a solder suction pump thingy) so I'll definitely be getting some of that!

I'll let you know how I get on!

[janoc]

I've bought an ESR meter now because it's something I could be doing with more and more!
And Solder Wick! What a discovery! I had never heard of it until seeing Janoc's post! (I have a solder suction pump thingy) so I'll definitely be getting some of that!

I'll let you know how I get on!

Hehe, someone got bitten by the equipment bug Check other Dave's repair videos, there is a ton of stuff to learn from.

Good luck with the repair, but please stay safe. If you are a newbie, a switching mains power supply is a fairly dangerous item to work on. Well, it is dangerous for non-newbies too, but at least they know who or what to blame when they get their brains fried.

Remember - one hand always in  your pocket when poking around a live device so that you don't accidentally close a circuit through your body by touching the wrong things. I am sure you don't have an isolation transformer which one should use in such cases, so at least wear shoes with a thick rubber sole, in case.

[MrJoeyJoeJoe]

Sorry for the delay!
I've finally got round to checking the outputs of the ribbon cable. I plugged the board back in, screwed it down, attached earth cable and turned it on.
I followed Janoc's instructions, putting my meter's negative probe on pin 2 which is Ground and found it's definitely not outputting what it should be!
I have written what each pin reads on the photo - most of them apart from pin 1 fluctuate constantly.

Not sure what this could mean, but it's definitely a problem with a powerboard component. Could it be a cap?

[SeanB]

Stuck in startup mode, so very good. Says power switch and the little IC is still good, and that the secondary side is not totally shorted. Time to change all the secondary side electrolytics, and the little one on the primary side, and it should work properly. I would check the diode attached to pin 5 though, to see if it is not shorted. A simple diode test across it will show if it is shorted or not in circuit, and if short both ways either the diode or the capacitor is shorted.

[MrJoeyJoeJoe]

Thanks a lot! I will check that diode you mentioned, but just thinking I may as well wait for my ESR to arrive to check the caps rather than replacing them all.
Sounds like we're getting close to the problem! Cheers!

[janoc]

Thanks a lot! I will check that diode you mentioned, but just thinking I may as well wait for my ESR to arrive to check the caps rather than replacing them all.
Sounds like we're getting close to the problem! Cheers!

That's perhaps a good idea - if nothing else, it will be an educational experience. You won't learn to use an ESR meter on good caps only

And yeah, as SeanB said - that you are getting some voltages on the output is very good news. That means that most of the deal is still working ok.

On the other hand, the fluctuating voltages are not right - it most likely means that the PSU is attempting to start up but fails. That could be because the caps are bad and the excessive ripple is throwing things off or simply because there is not enough load on the output and the feedback loop doesn't work right.

Check the caps, replace them if needed and if it still doesn't work, more troubleshooting can be done.

[MrJoeyJoeJoe]

Hey guys!
My ESR finally came!
I've taken some readings and most of the caps on the power board seem fine apart from 3. Two of them give a reading as if they were resistors? And the third has a really high amount of microfarads. (It's supposed to hold 470uf but is registering at 2722uf!?)



I'm gonna replace those 3 - ordering them now!

Cheers,
Joe!

[janoc]

Well, if that meter shows them as resistors that's odd. Is that one of those cheap component testers? Something like this: http://thumbs.ebaystatic.com/images/g/GBQAAOSwpDdVbU3S/s-l225.jpg

That's not really very good as ESR meter, btw. and measuring in-circuit will give you nonsense results with this. Instead of sending a low level high frequency signal through the cap, like a normal ESR meter will do, these testers try to charge and discharge the cap and do a lot of other things trying to identify the connected part. That is going to be affected by the other components around, you can't measure in-circuit with it.

The 470uF cap could be ok. The capacity measurement of these universal testers is very approximate for large capacitors. Moreover, there could be easily other caps in parallel with it.

If you want to be sure, you will have to desolder the caps and measure them out of circuit.

[Monadnock]

Have you metered the output diodes?

[MrJoeyJoeJoe]

Well, if that meter shows them as resistors that's odd. Is that one of those cheap component testers? Something like this: http://thumbs.ebaystatic.com/images/g/GBQAAOSwpDdVbU3S/s-l225.jpg

That's not really very good as ESR meter, btw. and measuring in-circuit will give you nonsense results with this. Instead of sending a low level high frequency signal through the cap, like a normal ESR meter will do, these testers try to charge and discharge the cap and do a lot of other things trying to identify the connected part. That is going to be affected by the other components around, you can't measure in-circuit with it.

The 470uF cap could be ok. The capacity measurement of these universal testers is very approximate for large capacitors. Moreover, there could be easily other caps in parallel with it.

If you want to be sure, you will have to desolder the caps and measure them out of circuit.

Haha, the meter I bought is EXACTLY the one you posted! One of those cheap ones, yes.
If you could recommend me one that can be used with components in circuit like in the EEVBlog, that would be awesome.
You're right about it not working in circuit - I tried the meter out on a broken tv first and it read 3 bad caps. But when I removed them only 1 of the 3 was actually bad.

I guess it's time to start removing components then!

   

[MrJoeyJoeJoe]

Have you metered the output diodes?

Are the output diodes those two big ones to the left of the transformer?

[Monadnock]

Have you metered the output diodes?

Are the output diodes those two big ones to the left of the transformer?

Yes, those are 2 of them, looks like D6 and D9. There is also the big TO-220 one on the heatsink, but since you got voltage on that output it is OK.

I would look at D6 since you didn't seem to get any voltage on that output.

[EPTech]

Hi There,

I agree with David. If there are some voltages present, that means the control IC is showing some startup behavior but not quite getting there. This is often caused by undervoltage on it's own supply, that comes from an auxiliary winding of the transformer on the primary side in this case. I would also replace C5.

The ESR of a capacitor is best measured outside of a circuit. The meter uses an AC voltage to measure and that will also pass through the rectifier diode (partially) and the transformer's secondary coil. So you are not just measuring a capacitor.

[janoc]

Haha, the meter I bought is EXACTLY the one you posted! One of those cheap ones, yes.
If you could recommend me one that can be used with components in circuit like in the EEVBlog, that would be awesome.

That's what I thought, I have recognized the display screenshots.

Buy something that is actually marketed as ESR meter and not a component tester. These little testers are useful for quickly identifying/checking components, but don't rely on them for getting an accurate measurement.

An actual ESR meter looks somewhat like this:

http://www.ebay.com/itm/AnaTek-Blue-ESR-Low-Ohms-Meter-Complete-Kit-for-Assembly-BESR-Kit-/181171316406?hash=item2a2ea6e6b6:g:gV4AAOSw34FVDeYt

or this

http://www.ebay.com/itm/Portable-Digital-Capacitor-ESR-Tester-Internal-Resistance-Meter-Test-In-Circuit-/322034893012?hash=item4afac680d4:g:oNQAAOSwoudW4WGt

Dave has an early video on ESR meters too, explaining how it works and how to use one. I think the Anatek one is based on the same design Dave's one is using.

The other option is to get a proper LCR meter, those often include ESR measurement for capacitors, but they tend to be (a lot) more expensive and a massive overkill if you only want to check caps for repair.

You're right about it not working in circuit - I tried the meter out on a broken tv first and it read 3 bad caps. But when I removed them only 1 of the 3 was actually bad.

I guess it's time to start removing components then!

Yep. That little tester will get confused by other components in parallel with the component you are measuring, so that is to be expected.