Wiki Consolidated eMMC Thread (MCU repair) (Black Center Screen)

[Krash]

Periodically my slacker streaming will stop mid song and immediately start playing the next track. Sometimes several times in a row. Soundls like a really bad DJ remix. Doesn’t feel like a mobile data dropout so I suspect another pending eMMC failure. My first was before they put in upsized replacements.

 

[M0rpH3x]

Periodically my slacker streaming will stop mid song and immediately start playing the next track. Sometimes several times in a row. Soundls like a really bad DJ remix. Doesn’t feel like a mobile data dropout so I suspect another pending eMMC failure. My first was before they put in upsized replacements.

My eMMC is superfresh and sometimes this happens to me. I really think it's truly a connectivity issue. Just think that, even if you have 4 bars of connectivity, many tings can happen that would cut the connection. (It could also be a software bug).

 

[FlatSix911]

Periodically my slacker streaming will stop mid song and immediately start playing the next track. Sometimes several times in a row. Soundls like a really bad DJ remix. Doesn’t feel like a mobile data dropout so I suspect another pending eMMC failure. My first was before they put in upsized replacements.

I don't think the track skipping is related to the eMMC. I upgraded to MCU2 and the skip still occasionally happens.

 

[tldickerson]

I too get the skipping with my upgraded MCU2 in my July 2016 P90D. Seams to almost be getting worse everyday.

 

[The Duke]

Son says the app has cell tower change problems, mistakes it as an end of file and goes to the next song.
Spotify has a work around so try them.

 

[BigNick]

Son says the app has cell tower change problems, mistakes it as an end of file and goes to the next song.
Spotify has a work around so try them.

Interesting. I'll have to pull up something like ipchicken.com during a drive (assuming the browser will handle such an intense load) and refresh when a skip occurs to see if the IP has changed (which might be an indication of changing cell towers?)

 

[rxlawdude]

Interesting. I'll have to pull up something like ipchicken.com during a drive (assuming the browser will handle such an intense load) and refresh when a skip occurs to see if the IP has changed (which might be an indication of changing cell towers?)

I'm not sure your IP address changes on handoff between towers.

 

[Krash]

I'm not sure your IP address changes on handoff between towers.

Your IP address should stay the same. And I believe the tower infrastructure is obfuscated so that even if you had root and ran a traceroute, I think it would not show. Also the songs cache and you can see the bar as they do so. I haven't noticed whether skipped songs have finished caching and if not, if they skip at the exact location in which the cache sticks, which would be helpful information to know. Usually I'm driving so it is hard to tell.

 

[flipb]

Just curious how many owners of early (not updated) MCUs have had the eMMC recall performed. I received the email in February advising that I'm part of the recall and that I'll be contacted when parts are available for my eMMC/daughterboard upgrade. Haven't heard anything since. Been patiently waiting; haven't contacted Tesla. In fact my MCU seems to be on its best behavior lately, not much flakiness as I've sometimes experienced in nearly 5 years of ownership.

I'm planning to ask for the discounted 4G upgrade when they schedule my eMMC recall work. Just curious how far along they are with performing the recalls and if they're going in any particular order.

 

[MP3Mike]

Just curious how far along they are with performing the recalls and if they're going in any particular order.

Not very far, just over 10%:

If they keep up the current rate it will take a little over 3 more years to get through everyone.

As far as the order, my understanding is that they are prioritizing people that have had the eMMC fail. Followed by people where the stats shows it is getting close to failure, etc...

 

[Krash]

I haven’t been contacted. I had a replacement on my 2017 MCU1 before the eMMC upgrade, so that may put me at the back of the list.

I also bought FSD. I am curious as to whether there are any eMMC asymptomatic FSD owners being contacted to do the eMMC upgrade proactively.

 

[gbdesai]

Just curious how many owners of early (not updated) MCUs have had the eMMC recall performed. I received the email in February advising that I'm part of the recall and that I'll be contacted when parts are available for my eMMC/daughterboard upgrade. Haven't heard anything since. Been patiently waiting; haven't contacted Tesla. In fact my MCU seems to be on its best behavior lately, not much flakiness as I've sometimes experienced in nearly 5 years of ownership.

I'm planning to ask for the discounted 4G upgrade when they schedule my eMMC recall work. Just curious how far along they are with performing the recalls and if they're going in any particular order.

I've had my third MCU die today. Original lasted 3.5 years, refurb they put in 2.2 years, next refurb lasted 1.3 years. This is getting stupid.

 

[FlatSix911]

I've had my third MCU die today. Original lasted 3.5 years, refurb they put in 2.2 years, next refurb lasted 1.3 years. This is getting stupid.

This will only get worse as the memory demands grow in the future... best to bite the bullet and get the MCU2 upgrade.

 

[gbdesai]

Last time they told me my 2013 can't take the MCU2. Is that not true?

 

[tga]

This will only get worse as the memory demands grow in the future... best to bite the bullet and get the MCU2 upgrade.

Not sure I agree with that, now that they are replacing the Tegra boards with new, higher quality, larger (so more room for wear leveling) eMMC parts. A refurb MCU1 that has been in a car for over a year may very well have the original, crappy, 8GB, well-worn eMMC.

 

[FlatSix911]

Not sure I agree with that, now that they are replacing the Tegra boards with new, higher quality, larger (so more room for wear leveling) eMMC parts.
A refurb MCU1 that has been in a car for over a year may very well have the original, crappy, 8GB, well-worn eMMC.

This is still happening...

I've had my third MCU die today. Original lasted 3.5 years, refurb they put in 2.2 years, next refurb lasted 1.3 years.
This is getting stupid.

 

[immolated]

Periodically my slacker streaming will stop mid song and immediately start playing the next track. Sometimes several times in a row. Soundls like a really bad DJ remix. Doesn’t feel like a mobile data dropout so I suspect another pending eMMC failure. My first was before they put in upsized replacements.

Same problem, my streaming music randomly skips to the next song even will full bars LTE.

 

[tga]

This is still happening...

I think you missed my point. In all likelyhood, this was an old 8GB part that failed, not the new 64GB one. @gbdesai had a refurb installed 1.3 years ago. That was before Tesla started using the improved 64GB eMMC. Plus, a refurbished (not new) MCU would be even older and more likely to have the bad part, and one that had a lot of wear to boot.

 

[CB198]

I’ve been experiencing the same glitchy audio with streaming music as well. I recently had to power down the car to replace my onboard charger. Funny thing, my MCU works flawlessly for 24 hours after electrically disconnecting the whole car. I disconnected the negative 12V terminal on the battery and disconnected the Emergency HV disconnect cable under the cabin filter housing (MSP85D ‘15). I was dreading doing it. Today I was driving and the MCU restarted several times in a span of an hour. It was bad enough that I dug everything out of the frunk (be sure to let the driver window down beforehand). It’s a huge pain in the ass.
There has to be an easier way to power down the MCU so that it clears the logs and resets the eMMC (mine has been replaced via desoldering).
Thought about making an extension cable for the 12V negative and the emergency disconnect so that I could access them by simply opening the frunk lid. But I don’t know how best to set up the connectors and cables so that it won’t disconnect while driving. The car is a 1000X better after disconnecting power.

Is there a way to accomplish the same thing via SSH? If only there was a way to isolate what is overloading the Tegra resources, and kill it. There are a lot of dead-end traffic seen in syslog because I’m running 2019.16.1.1 (no more openvpn and other resources). But I think there is an over accumulation of junk that is hogging resources. Because disconnecting the car fixes all these issues for a short period of time. The screen becomes extremely responsive, no delays. Everything downloads faster. No random restarts. Anyone else with similar issues, give this a shot and see if it helps. I expect it will.

 

[Kaveman]

CB198​

There are Fuses to pull power for the MCU and the IC. I had the 2 stop communicating, pulled the fuses, waited a bit, plugged the fuses back in, and back in action. Plenty of info on which fuses, but depends on your year car as to which ones.
Kaveman