Upgrading 2013-2015 Macbook Pro SSD to M.2 NVMe

[mdgm]

An older Mac won’t be able to utilise the full speed of the SSD, but it still should be much faster than the SSD that shipped from the factory.

The mid-2015 MBP is seen as the best model to support replaceable SSDs for this reason.

 

[herb2k]

I expect LiLu/NVMeFix will be updated eventually, but no guarantees... that said, on BigSur with the Barracuda 510, I get close to the same runtime as I did with NVMeFix installed (so it's not as necessary on BigSur).

For those of you on Big Sur and who want to try enabling NVMeFix again - I did a bit of research and found the Kernel Panics were due to the Lilu kext having a compatibility issue on Big Sur.

You need to add "keepsyms=1" to boot-args, after loading in the kext, before your first boot.

To do this:
Download and install Lilu.kext and NVMeFix.kext
Reboot in recovery mode (CMD + R)
Open up terminal

Type in:

nvram boot-args="keepsyms=1"

Reboot and you should enter macOS normally.

 

[herb2k]

For those of you on Big Sur and who want to try enabling NVMeFix again - I did a bit of research and found the Kernel Panics were due to the Lilu kext having a compatibility issue on Big Sur.

You need to add "keepsyms=1" to boot-args, after loading in the kext, before your first boot.

To do this:
Download and install Lilu.kext and NVMeFix.kext
Reboot in recovery mode (CMD + R)
Open up terminal

Type in:

nvram boot-args="keepsyms=1"

Reboot and you should enter macOS normally.

So far, (On Big Sur) with only @kivc's kext, I was at around 0.05A lowest on my Seagate 510.

Now with NVMeFix, I've seen it drop as low as 0.02A

 

[hopefulPA62]

I recently upgraded the SSD on my early 2015 13" macbook pro with a 500 gb SN550 nvme ssd and for some reason ive gotten kernel panic errors twice within the past week after closing my laptop and putting it to sleep for several hours. The first time it happened, it occured after a few hours of the laptop being put to sleep and today it happened after the laptop was put to sleep overnight. I did update my macbook with Big Sur prior to installing the new SSD and im currently running 425.0.0.0.0 firmare. Im not sure if anyone else has had this issue or if anyone can help guide me on to resolve this?

 

[developeralioz]

Our WD SN550 1TB + 2015 13-inch MBP has been flawless so far for almost three months. While not always under demanding workload, it does include hours of Zoom calls almost every work day, and often up to hour long of video compression by Zoom after those calls. During the video compression, I was told CPU temp shoots up to 100C, and SSD temp could reach as high as 50C.

So it saddens me to see a good SSD like SX8200 Pro has to live in a hell as hot as an enclosure. Your issues are very likely caused by poor craftsmanship of installation and/or the adaptor and perhaps coupled with brutal handling of the laptop on a daily basis. That could result in poor contacts and short circuits. Usually could be resolved by re-seating the SSD, applying Kapton tape (won't hurt to overdose) on suspected joints of possible short circuit, tightening the SSD's screw and/or inserting a thin rubber washer between the SSD and the screw.


vince22 is right. It should work. Based on current feedback, ssdpmEnabler is popular among 2015 models both 13 & 15 inches.

I recall bursts of complaints from you right about the time I started participating in this thread. My key take-away was _never_ putting a Samsung stick inside Macbook's. lol. Seriously though I was impressed by your observation regarding the iGPU never dipping below 500MHz.

There is rumour saying that ssdpmEnabler has the butterfly effect on iGPU which starts hopping up and down around 500MHz just like with a AppleSSD. This does add further power saving! It's a "rumor" because nobody understands why there is such an effect, and what's the exact behaviour the iGPU is expected to be. Also I only have (limited) access to 2015 13-inch MBP + WD SN550 1TB and verify the observation.

So do let us know how it performs on your Samsung SSD + 13-inch combo.


Perhaps installing ssdpmEnabler (disclosure: I'm its creator)?

My iGPU was not going below 500 mHz. Installing "ssdpmEnabler" the problem got a little better. The iGPU now runs at 100 mHz when the system is idle. But after installing "ssdpmEnabler" I started to hear noise from my SSD. It's kind of like an electric sound. Looks like coil whine. There was no such problem before installing this kext.

 

[oryan_dunn]

@kvic I'm testing out SsdPmEnabler on a fresh install of Big Sur. Since enabling it, I've noticed there is no screen on time shown in the battery preferences screen. Is this a known issue? I'm going to try disabling it to see if screen on usage comes back.
Edit: I used this nearly all day, but no screen usage shown.

 

[herb2k]

@kvic I'm testing out SsdPmEnabler on a fresh install of Big Sur. Since enabling it, I've noticed there is no screen on time shown in the battery preferences screen. Is this a known issue? I'm going to try disabling it to see if screen on usage comes back.

I'm using SsdPmEnabler and screentime works on my BigSur install.

If this is a totally fresh install, it might need some time to gather statistics before any data shows up.

 

[oryan_dunn]

I'm using SsdPmEnabler and screentime works on my BigSur install.

If this is a totally fresh install, it might need some time to gather statistics before any data shows up.

It had screen time data before I installed SsdPmEnabler. I'll keep digging to see if I can find anything.

 

[Stratus Fear]

So far, (On Big Sur) with only @kivc's kext, I was at around 0.05A lowest on my Seagate 510.

Now with NVMeFix, I've seen it drop as low as 0.02A

View attachment 1706789

Thanks for the keepsyms=1 tip... This worked well for me. Whereas NVMeFix alone never used to make any difference, and then I couldn't even use it on Big Sur because I wasn't aware of the aforementioned boot argument, I just stopped using it after SsdPmEnabler came out. But after seeing this, I decided to go back and try it again.

I'm on a 2014 15" with the ADATA SX8200 Pro. With nothing, SSD idles around 0.17A, with SsdPmEnabler it will idle at 0.1A. But with the combination of both SsdPmEnabler and NVMeFix, it bottoms all the way out to 0.00A with the occasional blip between 0.01A-0.03A! This is excellent.

 

[kvic]

@herb2k Congrats on your entry into sub 50mA club. You earned it.

@hopefulPA62 Try to look into Console for traces of "kernel panic" logs. Google for help. If you suspect your installation of 3rd party SSDs cause the "kernel panic", check out my suggestions to tzevoon in the previous page.

@developeralioz
SSDs won't have coil whine. If you hear what you claim, very likely due to some aging components inside your laptop.

Great to hear you also saw iGPU down clocking! So I'm not alone.

@oryan_dunn
Sorry, no idea about "Screen On Usage" issue.

I recall you have Seagate Barracuda 510 1TB. How is the idle current before and after? Good to hear back from you, btw.

@Stratus Fear
0.00A is bogus. Consider it as unknown.

To check your idle current with both kexts. Connect your laptop to AC power. Let it idle for a minute and check in iStat Menus. It will show leveling off at 0.01A (or 0.03A) if it's able to.

Thanks for updating me on your machine spec, I'll add your data to the User Guide.

 

[quaresma]

NVMeFix updated:

v1.0.5​

• Fixed quirks enabling per controller
• Fixed initialisation on 10.15+

 

[blackfang]

For those of you on Big Sur and who want to try enabling NVMeFix again - I did a bit of research and found the Kernel Panics were due to the Lilu kext having a compatibility issue on Big Sur.

You need to add "keepsyms=1" to boot-args, after loading in the kext, before your first boot.

To do this:
Download and install Lilu.kext and NVMeFix.kext
Reboot in recovery mode (CMD + R)
Open up terminal

Type in:

nvram boot-args="keepsyms=1"

Reboot and you should enter macOS normally.

How do you install Lilu and NVMEFIX? I tried to open the kexts on Big sur and It asks me to select application. I tried with show package contents and selected manually open it but i get error "zsh: exec format error:"
terminal log below:

"Last login: Tue Jan 5 13:50:00 on ttys000


/Users/X/Downloads/Lilu-1/Lilu.kext/Contents/MacOS/Lilu ; exit;


flaviusbei@Flaviuss-MBP ~ % /Users/X/Downloads/Lilu-1/Lilu.kext/Contents/MacOS/Lilu ; exit;


zsh: exec format error: /Users/X/Downloads/Lilu-1/Lilu.kext/Contents/MacOS/Lilu


Saving session...


...copying shared history...


...saving history...truncating history files...


...completed."

Did it worked?



[Process completed]

 

[developeralioz]

Lilu and NvmeFix have been updated. Does it work at BigSur now?

 

[oryan_dunn]

@oryan_dunn
Sorry, no idea about "Screen On Usage" issue.

I recall you have Seagate Barracuda 510 1TB. How is the idle current before and after? Good to hear back from you, btw.

Thanks, so I've been able to hang on to my sisters laptop for a bit to test. I put in that Seagate Barracuda 510 1TB (likely Phison E12s, how do I verify?). In general, I'm not sure there is a really good way to test power usage, other than to either spot check current readings, or just to use the laptop normally and note battery life. If a current reading of 0.00A is bogus, how can we be sure the rest of the readings are accurate?

It would be nice if there were a standardized way to test, so all these comparisons would be more meaningful.

I've been running with SsdPmEnabler enabled for a couple days. I seem to be able to get 8hrs or so of use, and the current readings for lowest are either 0.00A or 0.05A, usually sitting at 0.00A. I'll likely disable SsdPmEnabler and use in a similar manner for a couple days to compare.

Also, I was able to get screen on time to show again after a reboot. Not sure why that was not showing. Hopefully it's a one time issue.

 

[kvic]

I'm afraid I can't show you 0.05A equal to 0.05A. I did show people 0.00A not equal to 0A in one of my previous posts. I haven't re-tested since then but I doubt there is any change. I'm happy to be proved wrong.

@vince22 demonstrated to people two pages back how to test if the battery does last longer though I think he omitted a paragraph or two. I think others could come up with a controlled test that suit themselves to do the verification.

So Barracuda 510 1TB is basically same as 500GB on idle power consumption. Thanks for the update and I'll add it to my table.

I still consider the silver chips are Phison E12s but I also think people's bias against E12s is with little ground. I'm with everyone.. spiritually.. in their quest to find the truth.

 

[BoPl]

@Earl Urley Thanks for the update. To be conservative. I'll take 0.11A (/w NvmeFix) as your stock idle power for Inland Premium 1TB because I think >0.4A is likely not a proper observation for stock value.

@TiBook550 Thanks for posting. Looks like Crucial P2 is good with idle power at stock and responds very well to ssdpmEnabler. Currently P2 tops my chart at 90% reduction in idle power.

Thank you to all who have tried and/or provided feedback on ssdpmEnabler. From those of you who had included sufficient details, I've included your data under the "Confirmed working models" section of the User Guide.

To those of you who tried but your combinations of Mac and SSD models not listed on the table, some of us here would really like to hear from you.

At the moment, we're particularly interested in 2014 13-inch MBP owners who could share success stories with ssdpmEnabler. So that we would be able to isolate a kernel panic issue to this Mac model or specific SSDs. Possibly also come up with a fix.

EDIT:

@le9508 Thanks for the feedback. Updated the table to include your Mac & SSD.

Gents, if my guess of your Mac's model IDs are incorrect, pls let me know.

Hi kvic.
Just to remind my post last week:
I'm on rMBP 13" mid 2014 upgraded to Big Sur 11.1 (BootRom is 429.0.0.0.0) and installed new 1TB Corsair Force MP510 SSD with old version Phison E12. Have idles at 0.17A.
Today I tried to install your SsdPmEnabler.kext twice and twice my Mac didn't boot after that...

Today I replaced my 1TB Corsair MP510 with 1TB Samsung 970 Evo Plus with clean Big Sur 11.1 on and tried to install the SsdPmEnabler.kext. Same story - my Mac didn't boot back after that.

2021-01-05 17:27:39.416225+0300 localhost kernelmanagerd[100]: Validate approval for /Library/Extensions/SsdPmEnabler.kext in auxKC: not approved

2021-01-05 17:27:40.536083+0300 localhost kernelmanagerd[100]: failed to create bundle from path, unable to include: /Library/Extensions/SsdPmEnabler.kext

panic(cpu 0 caller 0xffffff8017183d06): nvme: "Fatal error occurred. CSTS=0xffffffff US[1]=0x0 US[0]=0x1b VID=0xffff DID=0xffff
. FW Revision=2B2QEXM7\n"@/AppleInternal/BuildRoot/Library/Caches/com.apple.xbs/Sources/IONVMeFamily/IONVMeFamily-557.60.1/Common/IONVMeController.cpp:5472
Backtrace (CPU 0), Frame : Return Address
0xffffffb09b2ab960 : 0xffffff8014ab9aed
0xffffffb09b2ab9b0 : 0xffffff8014bfc6e3
0xffffffb09b2ab9f0 : 0xffffff8014becd1a
0xffffffb09b2aba40 : 0xffffff8014a5ea2f
0xffffffb09b2aba60 : 0xffffff8014ab938d
0xffffffb09b2abb80 : 0xffffff8014ab9678
0xffffffb09b2abbf0 : 0xffffff80152be3ca
0xffffffb09b2abc60 : 0xffffff8017183d06
0xffffffb09b2abc80 : 0xffffff8017168427
0xffffffb09b2abde0 : 0xffffff801521c795
0xffffffb09b2abe50 : 0xffffff801521c696
0xffffffb09b2abe80 : 0xffffff8014afe6b5
0xffffffb09b2abef0 : 0xffffff8014aff5c4
0xffffffb09b2abfa0 : 0xffffff8014a5e13e

Kernel Extensions in backtrace:
com.apple.iokit.IONVMeFamily(2.1)[008D1169-B6FD-39ED-9033-C67AC409031C]@0xffffff8017161000->0xffffff801718afff

dependency: com.apple.driver.AppleEFINVRAM(2.1)[E4DF1D8B-0DA9-33A8-A7B6-17DCA1678FC0]@0xffffff8015e59000->0xffffff8015e62fff

dependency: com.apple.driver.AppleMobileFileIntegrity(1.0.5)[9BA70327-883F-3E4F-960A-14203BB191FD]@0xffffff8016000000->0xffffff8016014fff

dependency: com.apple.iokit.IOPCIFamily(2.9)[D0F1BFB8-C28E-3E70-891B-73B783189394]@0xffffff801742c000->0xffffff8017453fff

dependency: com.apple.iokit.IOReportFamily(47)[3A404946-7B6B-371A-A7B5-96E25F6F449D]@0xffffff8017462000->0xffffff8017464fff

dependency: com.apple.iokit.IOStorageFamily(2.1)[AA8F8B9D-BA6E-3EBB-8195-4792042ADF71]@0xffffff801754c000->0xffffff801755dfff

Process name corresponding to current thread: kernel_task

Mac OS version:
20C69

 

[mork4]

Hello everybody, merry christmas!
I've just got an early christmas present, Sabrent Rocket 1TB with phison e12s controller.
My current setup was Sabrent Rocket 512 GB with Phison E12 and macOS Catalina + NVMeFix installed - everything worked as charm and battery was awesome.
So I decided to make an upgrade and also install new macOS Big Sur 11.1.
After clean install of macOS Big Sur (Apple SSD), my bootroom got upgraded to 425.0.0.0.0 (Macbook Pro 15 Mid 2015). Then I've tried to install Lilu with NVMeFix kexts via Hackintool (SIP disabled), but have no luck . My system got stuck on restart loop and only solution was do PRAM reset (enable SIP) and then system boots normal.
So my question is:
Is there anybody that have successfully installed Lilu and NVMeFix on macOS Big Sur 11.1?
Thank you!

For those of you on Big Sur and who want to try enabling NVMeFix again - I did a bit of research and found the Kernel Panics were due to the Lilu kext having a compatibility issue on Big Sur.

You need to add "keepsyms=1" to boot-args, after loading in the kext, before your first boot.

To do this:
Download and install Lilu.kext and NVMeFix.kext
Reboot in recovery mode (CMD + R)
Open up terminal

Type in:

nvram boot-args="keepsyms=1"

Reboot and you should enter macOS normally.

Hi,
thanks for your insights, it works.
Just in case to revert that boot-args back, what is a default value for keepsyms ?

So now, I've got installed kvic's SsdPmEnabler and Lilu with NvmeFix. Power consumption goes down to 0.00A (0.04A before with just SsdPmEnabler) thats great!

But question is:

Is this stable ?
Are we allowed to have installed that many 3rd party kexts, which in fact do the "same" thing ?
How come, that SsdPmEnabler works even without Lilu, but not handling power states correctly or most efficiently ?

In the ideal world shouldn't be there something like SsdPmEnabler behaving like NvmeFix (efficiently handling power states) and working without Lilu?

Thanks.

 

[chrickers]

Is it safe and worth to install any of these kext on Catalina with the new Bootrom?

 

[herb2k]

Just in case to revert that boot-args back, what is a default value for keepsyms ?

From my research, keepsyms isn't set by default, so to clear it you could simply clear the boot-args of your nvram with a command like this:

sudo nvram boot-args=""

 

[herb2k]

Lilu and NvmeFix have been updated. Does it work at BigSur now?

On Big Sur, you still need to do the steps I outlined above or you will get a Kernel Panic.

#8104

 

[herb2k]

I'm afraid I can't show you 0.05A equal to 0.05A. I did show people 0.00A not equal to 0A in one of my previous posts. I haven't re-tested since then but I doubt there is any change. I'm happy to be proved wrong.

@vince22 demonstrated to people two pages back how to test if the battery does last longer though I think he omitted a paragraph or two. I think others could come up with a controlled test that suit themselves to do the verification.

So Barracuda 510 1TB is basically same as 500GB on idle power consumption. Thanks for the update and I'll add it to my table.

I still consider the silver chips are Phison E12s but I also think people's bias against E12s is with little ground. I'm with everyone.. spiritually.. in their quest to find the truth.

@kvic - I pulled the fw details in an earlier post #7902 - the Seagate 510 is a standard E12.

Agree the bias against the E12s has little ground - most of the negatives of that chipset appear to be simply the manufacturers including less DRAM on the module as a result...great reason to be upset at the manufacturers, but not to hate on the chipset

 

[herb2k]

FWIW - After getting Lilu+NVMeFix + SsdPmEnabler running, during those times of "0.00A", the iStats graphs show a very small mA draw on the 3.3V SSD line (less than <0.005A), instead of 0.00A.. this may be from an iStats update too.

 

[kvic]

Hi kvic.
Just to remind my post last week:
I'm on rMBP 13" mid 2014 upgraded to Big Sur 11.1 (BootRom is 429.0.0.0.0) and installed new 1TB Corsair Force MP510 SSD with old version Phison E12. Have idles at 0.17A.
Today I tried to install your SsdPmEnabler.kext twice and twice my Mac didn't boot after that...

Today I replaced my 1TB Corsair MP510 with 1TB Samsung 970 Evo Plus with clean Big Sur 11.1 on and tried to install the SsdPmEnabler.kext. Same story - my Mac didn't boot back after that.

Hi BoPI,

1. Without ssdpmEnabler, does Samsung 970 EVO Plus boot correctly?
2. With ssdpmEnabler installed, is the panic error from Samsung 970 EVO plus same as that from Corsair 510 last week?

But question is:

Is this stable ?
Are we allowed to have installed that many 3rd party kexts, which in fact do the "same" thing ?
How come, that SsdPmEnabler works even without Lilu, but not handling power states correctly or most efficiently ?

In the ideal world shouldn't be there something like SsdPmEnabler behaving like NvmeFix (efficiently handling power states) and working without Lilu?

ssdpmEnabler, and NvmeFix (+Lilu) perform different functions. Also see my response to you in post #8052 in case you missed.

If you want to measure idle current, pls follow instructions in post #8111.

@kvic - I pulled the fw details in an earlier post #7902 - the Seagate 510 is a standard E12.

Yeah, I saw & read your post back then. But...it is only as accurate as the program's author. I'm not saying it's incorrect or correct. I just think the same silver package coming in two versions E12 and E12s doesn't make sense with how the semiconductor industry works..

 

[hopefulPA62]

@hopefulPA62 Try to look into Console for traces of "kernel panic" logs. Google for help. If you suspect your installation of 3rd party SSDs cause the "kernel panic", check out my suggestions to tzevoon in the previous page.

Thank you for the advice kvic! I just ordered some Kapton tape and will try it out when it comes in on Thursday. Do you have any advice on where to exactly put the kapton tape on the adapter?

 

[oryan_dunn]

FWIW - After getting Lilu+NVMeFix + SsdPmEnabler running, during those times of "0.00A", the iStats graphs show a very small mA draw on the 3.3V SSD line (less than <0.005A), instead of 0.00A.. this may be from an iStats update too.

View attachment 1707592

How did you switch units from A to mA?