[IN TICKET] Piece of mainboard partially melted

[Matt_Hartley]

@Lord_Emloy

Just contacted Support probably 20 minutes ago but just had the same issue happen to me. I was in batch 5 or 6 cant remember. Was charging on the upper left side under medium/high duty use

Thanks for sharing this with us. Contacting support for resolution is the correct step to getting this sorted.

[Eric_Putney]

Batch 4 DIY 11th gen here (1165-G7), found this thread when I was researching tb docks and just popped mine open to check. There’s no sign of melting/overheating. Do keep us posted though.

[ztaira]

I also charge on the upper left side, have a mainboard from batch 5, and have a melted plastic cover. I noticed it because I was looking to reset the mainboard after my laptop died, but wasn’t seeing any lights on the left hand side.

Now that I see the damage, I’m guessing it might be the cause of the laptop/left USBs not working.

Is it still worth sending this in for analysis?

Edit: edited “because the USBs aren’t working” to correctly order cause and effect. Also, I contacted support ^-^

[MJ1]

@ztaira

Now that I see the damage, I’m guessing it’s because the left USBs aren’t working.

Both of the USB ports on that side don’t work?
If so, the overheating of that chip could be the reason those ports stopped working.

You have a 2nd Gen Framework? Regardless of the generation, I’d suggest contacting support.

[Adrian_Joachim]

@ztaira

Now that I see the damage, I’m guessing it might be the cause of the laptop/left USBs not working.

The overheated component does look a lot like a mosfet and it is a bit far away from the cpu to part of the vrm so if that is part of the power routing to the ports on that side it would explain the usb ports not working. That does however not mean it would be the initial cause of the problem, mosfets tend to die from issues down/upstream a lot more than just by themselves.

Very curious what support finds out. Unfortunately normal users can’t get the actual schematic and board-view, cause then you could exactly tell what component that was and what it was connected to.

[MJ1]

@Adrian_Joachim

The overheated component does look a lot like a mosfet

I believe it’s one of these EMZB08P03 MOSFETs.

Power MOSFETs-Products-Excelliance MOS 杰力科技

The number on the chips can be read here [IN TICKET] Piece of mainboard partially melted - #6 by Yunlee, and the logo is clear in other pictures.

[carboncountyutah]

Hi guys, firing up an older thread but, we’ve got 1 of 24 of our framework laptops doing this same thing. Mosfets show up to 302F/150C while plugged into a 20v 45w PD power supply. If I can get the computer to boot without flashing codes, it boots with zero USB port functionality and no errors in the device manager. Tried installing latest bios release and it then failed to boot. Anyone having luck getting these boards replaced being out of warranty?

[Azure]

Mosfets can get pretty hot, but if you’re seeing parts of the mainboard melting, you should absolutely contact support so that they can look over the hardware and help you with a possible replacement.

[wkos]

Yesterday my left rear port stopped charging, opening up show the same melted spot. Reported to support today.

Mine is a batch 10. Order date april 1 2022. DIY i7-1165G7

[suliblian]

And a question to all who had such a board. Did you charge mainly through the left rear port? The melting locations point to specific MOSFETs. Those for the right ports are on the backside and have no plastic cover to indicate too much heat, but they could slowly cook the memory in slot 0.

Now there are 18Edit: 12 of the BnnP03ZB08P03 type (‘Z’ is manufacturer? Third line is a time/batch marker) on the board that may fail later on.

Frameworkers: Does Framework still exchange boards with this issue, even when they are out of warranty?`

[MJ1]

@suliblian

(‘Z’ is manufacturer? Third line is a time marker)

The manufacturer, Excelliance MOS, is represented by the stylized “M” logo.

[IN TICKET] Piece of mainboard partially melted - #90 by MJ1
Follow the link for datasheet

@suliblian

Now there are 18 of the BnnP03 type (‘Z’ is manufacturer? Third line is a time marker) on the board that may fail later on.

Users have posted quite a few pictures, and every one shows overheating in only one mosfet, the one located right next to the LED. There is no reason to believe that all of those chips are inherently bad or destined to fail. Rather, it looks like there was an issue that overloaded just the one chip in that particular spot.

[suliblian]

@MJ1

The manufacturer, Excelliance MOS, is represented by the stylized “M” logo.

Follow the link for datasheet

Ah, thanks! I had thought the B08/B09/B12/B20/… were all the same only with different ratings, and read one datasheet only.

I had to correct their number, it’s only twelve B08P03, six per side (of laptop, as well as of board).

@MJ1

Users have posted quite a few pictures, and every one shows overheating in only one mosfet, the one located right next to the LED. There is no reason to believe that all of those chips are inherently bad or destined to fail. Rather, it looks like there was an issue that overloaded just the one chip in that particular spot.

I beg to differ. Here is a mainboard for sale where also the second chip I marked in the pic had been too hot. The seller has likely drawn a high current (to conclude from 64GB RAM included in the offer), possibly while charging the battery, through a left side USB port. When it failed, he tried the other left port, until it overheated, too. Is that correct, @Reese_Borel?

In this thread, where the port is mentioned, is is the left rear port, hence always the same chip. For the other ports and chips, “absence of evidence is not evidence of absence”.

The backside B08P03 chips were found too hot, too, pointing to a right-hand port.

@Usernames

Charged only with the FW adapter and once briefly with an Apple 61W USB-C PD adapter (pictured) which wouldn’t charge the system for some reason.
…
No external GPU but I was stress testing the system.

The one on the underside of the board was cooking my RAM a fair bit, it burnt my finger touching a module when I was trying to figure out the heat source.

@Usernames: Can you confirm/debunk that this happened when the power went through a right side port?

FW seeks the root cause upstream, the diagnosis went from FW to mainboard manufacturer to chip manufacturer.

@TheTwistgibber

We’ve shipped four affected boards to our manufacturing partner in Taiwan. The chips are being decapped and analyzed.

@TheTwistgibber

After a review of additional boards by our manufacturing partner, the results were still inconclusive and no root cause was identified. The only finding was that the failure was due to component stress, but again, no root cause. The failure rate is extremely low based on shipped units vs failures in the field but we continue to evaluate any returned units to see if a definitive cause can be found. X-ray evaluation of the MOSFET unfortunately does not point to a cause of failure.

To me the most likely the root cause is a (sub)batch of bad chips. Those chips would have come on the same reel, placed on the same board together, and die successively from too high currents, causing a variety of different symptoms: Black Screen Issue, intermittent contact with USB storage, thunderbolt not working, etc.

@TheTwistgibber: Has the real cause been found?

[Kyle_Reis]

@suliblian

@TheTwistgibber: Has the real cause been found?

FYI he no longer works at Framework.

[suliblian]

Thanks. I’ll take it to a new post, someone from FW should see it there.

[suliblian]

Does Framework replace the out-of-warranty mainboards that had the overheating/plastic-melting ZB08P03 chips? These MOSFETs were on mainbards of the batches 5 and 6, iirc. Some reports mention or suggest a high CPU load while charging the battery- If that is not the regular use of the working mainboards and they carry the same fault, the defect may manifest much later after a period of such high-load situations.

And to satisfy our curiosity: has the root cause ever been found?

[amoun]

@suliblian

Does Framework replace the out-of-warranty mainboards that had the overheating/plastic-melting ZB08P03 chips?

Wouldn’t a direct ask of Framework via support@frame.work be more appropriate rather than a public forum query ??

[Brian_Gregory]

@suliblian

ZB08P03

Best case: it was a batch of fake chips (something lower spec relabelled).
Worst case:

[amoun]

Why a new topic? Framework have seen this and answered . . . to the degree they could

[MJ1]

@suliblian

Does Framework replace the out-of-warranty mainboards that had the overheating/plastic-melting ZB08P03 chips?

@suliblian

And to satisfy our curiosity: has the root cause ever been found?

While Framework staff does sometimes answer questions here, this is a primarily a community forum. And Framework has said that this forum is not a channel for obtaining official support from them / direct questions to Framework. So I would suggest contacting support if you want a response from Framework staff.

You can of course link them to this thread & also post the response you get here if you wish.

[suliblian]

@amoun

Framework have seen this and answered . . . to the degree they could

A guess, or have you been initiated into the secret and do know?

FW did no tell us of a result of their investigation here, in the most likely place and I could not find it in other threads. The latest info is from TheTwistgibber on Dec 28, 2022(!), who has left the company.

@TheTwistgibber

We continue to investigate the returns and have been shipping them to our HQ for review. If we do identify a root cause, rest assured we will let the Community know.

Now, were FW and upstream companies unable to identify the cause, or was it simply forgotten, maybe because the man who had it in mind has left FW?

The question about out of warranty replacements has not been answered. At the time of the early replacements mentioned, FW was seeking samples and asked for them. Most recently it was asked on May 8, still not answered.

@carboncountyutah

Anyone having luck getting these boards replaced being out of warranty?

@amoun

Wouldn’t a direct ask of Framework via support@frame.work be more appropriate rather than a public forum query ??

Support is seemingly always under a heavy workload. I did not want to add to the pile.

Also, this thread here is mostly technical and reports of the issue. I thought it better to place a question about policy in its own thread. Could have asked about the investigation result here, though, that’s right.