Posted at 10:12 by Richard Swinburne with 7 comments
Asus has responded with a detailed report as to why its M3A-H/HDMI went up in flames and smoke. After publishing the first episode following the freshening smell of burnt electronics festered in our labs, Asus was keen to take the board back to Taiwan for further testing and let us know how it got on.
Here's its conclusion, with Asus' words and pictures (all we have done is tidy the English a little):
Customer feedback to the damaged M3A-H/HDMI was a burnt out MOSFET.
Detailed information on the cause is in the statement below.
Backstory: The board was running fine with an FX62 for 9 months, then they dropped in a 9600 Phenom B2, without removing the motherboard or changing any other hardware and it firstly killed the Enermax PSU, they then replaced the PSU and the motherboard ignited within about 10 seconds.
Analysis:
1．Check the impedance of MOSFET Rgs. The test result is below. The normal Rgs is around 8K OHM and there is just VDD_NB high side MOSFET damage.
Phase 1 phase2 phase3 phase4 phase VDD_NB
High side Rgs (OHM) 8.16K 8.28K 8.2K 8.07K 12.7
Low side Rgs (OHM) 8.22K 8.25K 8.25K 8.16K 8.15K
2. Then we removed the VDD_NB high side MOSFET and re-check the impedance of PCB Rgs. The result is 8.05K OHM. This tells us that just MOSFET is damaged and the Driver function is OK.
3. Next we changed the MOSFET with good one and turn on the system. Then we checked the MOSFET gate waveform. According to the waveform all the function of VRM is still good.
CH1:VDD_NB High side gate
CH2:VDD_NB Low side gate
CH3:VDD_NB Phase node
VID=1V DC Load 0A
VID=1V DC Load 20A
Conclusion:
There is just VDD_NB high side MOSFET damage and it maybe caused by an abornal voltage or current spike. To determine the root cause we need the failed MOSFETanalysis report to help us find out. We have passed the failed part to vendor to do analysis so after receive the report we will update you again.
Here's its conclusion, with Asus' words and pictures (all we have done is tidy the English a little):
Customer feedback to the damaged M3A-H/HDMI was a burnt out MOSFET.
Detailed information on the cause is in the statement below.
Backstory: The board was running fine with an FX62 for 9 months, then they dropped in a 9600 Phenom B2, without removing the motherboard or changing any other hardware and it firstly killed the Enermax PSU, they then replaced the PSU and the motherboard ignited within about 10 seconds.
Analysis:
1．Check the impedance of MOSFET Rgs. The test result is below. The normal Rgs is around 8K OHM and there is just VDD_NB high side MOSFET damage.
Phase 1 phase2 phase3 phase4 phase VDD_NB
High side Rgs (OHM) 8.16K 8.28K 8.2K 8.07K 12.7
Low side Rgs (OHM) 8.22K 8.25K 8.25K 8.16K 8.15K
2. Then we removed the VDD_NB high side MOSFET and re-check the impedance of PCB Rgs. The result is 8.05K OHM. This tells us that just MOSFET is damaged and the Driver function is OK.
3. Next we changed the MOSFET with good one and turn on the system. Then we checked the MOSFET gate waveform. According to the waveform all the function of VRM is still good.
CH1:VDD_NB High side gate
CH2:VDD_NB Low side gate
CH3:VDD_NB Phase node
VID=1V DC Load 0A
VID=1V DC Load 20A
Conclusion:
There is just VDD_NB high side MOSFET damage and it maybe caused by an abornal voltage or current spike. To determine the root cause we need the failed MOSFETanalysis report to help us find out. We have passed the failed part to vendor to do analysis so after receive the report we will update you again.
Comments (7)
Discuss in the forumsSeconded, site and downloads are atrocious from ASUS
Third-ed.
Great hardware, but horrible support.
New Xonar D1 Firmware please!
It's quite interesting to see the kind of checks they're doing on the board.
For most enthusiasts, once you see a part of the board melted or burnt-out, the initial reaction is to treat the whole board as being fried.
For someone who knows their electronics that'd seem kinda silly, but seeing ASUS describe the result as :
Makes me wish I had the tools and know-how to test and repair components like this myself instead of going through RMAs. :p
Let's take a hypothetical tour
The board comes back to the manufacturer for RMA.
Without the association of Bit-Tech name you just throw it in a recycle pile.
BUT this one is high visiblity so we send it to "the lab".
Now these are not minimum wage individuals - these are highly schooled and trained people with a bazillion $ worth of state of the art equipment.
One 8 hour day for this engineer is probably in the $50 to $100 per hour so you have $400 to $800 invested in one day testing this board by the engineer. That assumes only one engineer. In my experience it is seldom 1 engineer doing the testing.
It is so difficult to spend the $$ to test that it is no longer economically feasible for a manu. to test hardly any of the RMAs.
In a perfect world a chain of RMAs for a device with the same defect should alert the people that they have a problem with component XX. But it seldom happens.
I don't know what the internal cost of this board was to Asus but it was probably less than a good evening meal out.
One can only hope that by this exercise Asus will improve either the component or the design.
It's sad to see so many $$ parts scrapped out because of one 4 cent part.
I do hardware and software for a living and have for 30 years.
Pulling a card and replacing it is cheaper than fixing.
john
I would bet that most RMA's don't involve fire, only a defective flash or similar.