PDA

View Full Version : Overheating SB3 CPU ?



Adam_F
2008-10-26, 07:34
My SB3 recently developed a fault in that it now invariably crashes after a brief period of use (display goes blank and no sound). It is now out of warranty, and as this is now the third hardware fault I have had with this unit, I am inclined to bin it rather than pay for a further repair. However, I thought I would make some attempt to diagnose the problem first.

When this SB3 was repaired previously, the symptoms were exactly the same. On that occasion, the wireless card was found to be the cause and it was replaced. Based on a guess that I had the same issue again I replaced the wireless card myself, but this did not solve the problem. Removing the wireless card completely did not stop the crashes either, but the SB3 did run for considerably longer before doing so without the card in situ. Then I noticed that the CPU (IP3023) was running very hot. Holding a temperature probe against the CPU (i.e. not the best thermal contact), the probe read 100 degrees celcius after a few minutes.

Is this normal (doubtful), or could this point to the root cause (a problem with the power circuitry feeding the CPU for example). Can anyone suggest where to look next ?

Thanks, Adam.

Phil Leigh
2008-10-26, 08:35
My SB3 recently developed a fault in that it now invariably crashes after a brief period of use (display goes blank and no sound). It is now out of warranty, and as this is now the third hardware fault I have had with this unit, I am inclined to bin it rather than pay for a further repair. However, I thought I would make some attempt to diagnose the problem first.

When this SB3 was repaired previously, the symptoms were exactly the same. On that occasion, the wireless card was found to be the cause and it was replaced. Based on a guess that I had the same issue again I replaced the wireless card myself, but this did not solve the problem. Removing the wireless card completely did not stop the crashes either, but the SB3 did run for considerably longer before doing so without the card in situ. Then I noticed that the CPU (IP3023) was running very hot. Holding a temperature probe against the CPU (i.e. not the best thermal contact), the probe read 100 degrees celcius after a few minutes.

Is this normal (doubtful), or could this point to the root cause (a problem with the power circuitry feeding the CPU for example). Can anyone suggest where to look next ?

Thanks, Adam.


100 degrees on a CPU? - that cannot be right. Must be a faulty m-board or PSU in the SB3. Can you check the current being drawn by the PSU with a multimeter?

Adam_F
2008-10-26, 14:28
100 degrees on a CPU? - that cannot be right. Must be a faulty m-board or PSU in the SB3. Can you check the current being drawn by the PSU with a multimeter?

My thinking exactly - 50 or 60 degrees I could understand, but 100 must be well outside the device spec.

I measured the voltage being supplied by the external PSU at 5.17V off load, which seems reasonable. The SB3 is drawing 1.45A, with the wireless card still removed.

Adam.

pfarrell
2008-10-26, 14:34
100 degrees F? or C?

100 F is nothing. 100 C is bad.

My laptop just hit 81 C, but I was streaming video, which is a big load. Normal is 45C idling in my living room, which is about 70 degrees F.

Adam_F
2008-10-26, 14:38
Definately 100 C, which as you say is bad. Ambient temperature in here is about 60 F at present.

Also just spotted this:

http://www.audiocircle.com/circles/index.php?topic=24771.msg219713#msg219713

Mine is drawing 1.45A, which does not look good by comparison.

Adam.

pfarrell
2008-10-26, 14:45
Adam_F wrote:
> Definately 100 C.

That is not good


> Also just spotted this:
> http://www.audiocircle.com/circles/v...c.php?p=219713
> Mine is drawing 1.45A, which does not look good by comparison.

You may want to use tinyurl to fix that url, it didn't post cleanly.

--
Pat Farrell
http://www.pfarrell.com/

norman12
2008-10-26, 14:47
My thinking exactly - 50 or 60 degrees I could understand, but 100 must be well outside the device spec.

I measured the voltage being supplied by the external PSU at 5.17V off load, which seems reasonable. The SB3 is drawing 1.45A, with the wireless card still removed.

Adam.

I've measured 64 degrees C on the cpu with a type K T/C, what type of sensor/probe is being used to measure the case temperature. Also, 7.5 watts seems a little high, I measure 6 watts from the wall. What file format is the squeezebox decoding?

Adam_F
2008-10-26, 15:03
I've measured 64 degrees C on the cpu with a type K T/C, what type of sensor/probe is being used to measure the case temperature. Also, 7.5 watts seems a little high, I measure 6 watts from the wall. What file format is the squeezebox decoding?

The temperature was measured with the probe that came with my multimeter - not sure exactly what type it is. Even if the reading is a bit off, the CPU still gets hot enough to burn skin :/

The current draw of 1.45A was measured with the SB3 idling.

Adam.

norman12
2008-10-26, 15:22
The temperature was measured with the probe that came with my multimeter - not sure exactly what type it is. Even if the reading is a bit off, the CPU still gets hot enough to burn skin :/

The current draw of 1.45A was measured with the SB3 idling.

Adam.

Adam,
Try dunking the probe into a kettle as it boils from cold - just need to make sure we're not getting erroneous readings.

That 1.45A does seem high, has anyone else measured this?

norman12
2008-10-26, 15:29
Don't worry found a thread that stated the max current draw seen was 0.93A which I believe to be correct.

norman12
2008-10-26, 15:54
Ok,
The cpu should be receiving 1.2v from the onboard regulators, not sure where to probe at the moment (easy on the SBR - the test pads are plain to see). I'll check for you tomorrow.

Adam_F
2008-10-26, 15:57
Ok,
The cpu should be receiving 1.2v from the onboard regulators, not sure where to probe at the moment (easy on the SBR - the test pads are plain to see). I'll check for you tomorrow.

Many thanks.

Just tried the probe in a freshly boiled kettle thing, and got a reading of 98C, so it isn't far off.

Adam.

Adam_F
2008-10-29, 13:56
Not sure if this is relevant, but Q3 is running at close to 150 degrees C and a small area of board near the top of Q3 looks slightly burnt.

sfraser
2008-10-30, 06:19
100 degrees F? or C?

100 F is nothing. 100 C is bad.

My laptop just hit 81 C, but I was streaming video, which is a big load. Normal is 45C idling in my living room, which is about 70 degrees F.

Pat, good rule of thumb for translating C to F is this: "double and add 30" therefore 45C is about 120F, not 75F.

Cheers,

MrSinatra
2008-10-31, 12:25
wow, scary.

have u tried contacting support? i would think they may be interested in getting the unit back to see what went wrong.

Adam_F
2008-11-02, 02:11
wow, scary.

have u tried contacting support? i would think they may be interested in getting the unit back to see what went wrong.

I have been in touch with them and this has ended with EMEA support sending me a replacement SB3.

Partly out of idle curiosity, and partly due to some enforced idleness caused by a bout of 'flu, I decided to track down the cause of the problem. It appears that the switch mode regulator (NJU7630) that generates the supply for the CPU has failed, as the output of it was stuck at the supply voltage (5V) regardless of the input. The end result was that Q3 was passing about 2.7V to the CPU board, hence the hot CPU and the very hot Q3 due to the excess current it was passing.

This is probably not something I can repair directly, as the regulator part is not available retail and even if it was, it hard to say what other damage may have been caused.

Adam.