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

[singletrackminded]

Sorry I'm not the first to be asking this, but I've spent days reading this thread, so much information to take in, and I just can't make a decision on what to get.

I have a mid-2015 MacBook Pro 11,5 with a bootrom version of 428.60.3.0.0, my priority is stability and low power consumption (I like to work around the house & out and about). I'm not particularly bothered about speed as long as it doesn't perform any worse than the stock SSD. To give you an idea of usage, I'm a web developer.

My current stock SSD is 500GB, would be looking at 1TB - more space can't hurt, mostly need to replace this SSD as I think it's on it's way out.

I was going to pull the trigger on a Sabrent Rocket at first, but the more I read, the less I know what to go for.

 

[robinasu]

Damn, Amazon just delivered the P2 UK/CA version. It's going back. Anybody know where I can get the good version? I like the low power consumption.

Ok, so I compared the P2 circuit board to some articles and noticed the articles said the non-TLC version had fewer chips on the PCB. Mine had all positions filled and the label says UK/CA. I was curious what the deal is, so I installed the SSD and did a macOS clean install. For what its worth, BlackMagic says ~1800MB/s write an ~1700MB/s read. My original Apple SSD was ~400MB/s write and ~1500MB/s read.

Any thoughts guys? Should I return or keep?

 

[Hillionaire]

Has anyone tried using WD green SN350 1TB on mac book air late 2015?

 

[herb2k]

Hi All,

Short update, I have Monterey 12.2.1 on an MBP 12,1 with Bootrom 428.60.3.0.0.

I have been testing whether or not Lilu+NVMeFix and SsdPmEnabler are still required - I think the use case has changed now with the latest OS.

Battery runtime:
6.5 to 7 hours (web browsing, youtube) with no KEXTs installed.
7 to 7.5 hours with the KEXTs installed (web browsing, youtube)

Temperatures:
With no KEXTs installed: Idle SSD Temperature while doing the above - anywhere from 37 to 39C
With KEXTs installed: Idle SSD Temperature while doing the above - typically ~35-37C

Bit of history - I have been active in this thread since mid-2020 and have experimented with the hardware in my signature (as well as a couple of different adapters and SSDs) on my machine, starting from Catalina, through the BigSur betas (as they came out in 2020) to Monterey now and I have to say that the runtime and heat characteristics are slightly improved under Monterey. In the past, I would struggle to achieve 5hrs of runtime on my battery without NVMeFix/SsdPmEnabler. The SSD would also typically heat up to and run at a consistent 42C while idle.

From this, I could conculde that it would seem Apple has done some optimizations under the hood with regards to the NVMe drivers in the BootRom/OS. Full ASPM and L0s/L1 states are not implemented but power consumption has improved.

As always, YMMV - but for anyone who does not want to go through the process of loading the KEXTs or OpenCore, this may help you decide if you want to go through the trouble at all now for what are now somewhat diminishing gains.

 

[macpro_mid2014]

Hi All,

Short update, I have Monterey 12.2.1 on an MBP 12,1 with Bootrom 428.60.3.0.0.

I have been testing whether or not Lilu+NVMeFix and SsdPmEnabler are still required - I think the use case has changed now with the latest OS.

Battery runtime:
6.5 to 7 hours (web browsing, youtube) with no KEXTs installed.
7 to 7.5 hours with the KEXTs installed (web browsing, youtube)

Temperatures:
With no KEXTs installed: Idle SSD Temperature while doing the above - anywhere from 37 to 39C
With KEXTs installed: Idle SSD Temperature while doing the above - typically ~35-37C

Bit of history - I have been active in this thread since mid-2020 and have experimented with the hardware in my signature (as well as a couple of different adapters and SSDs) on my machine, starting from Catalina, through the BigSur betas (as they came out in 2020) to Monterey now and I have to say that the runtime and heat characteristics are slightly improved under Monterey. In the past, I would struggle to achieve 5hrs of runtime on my battery without NVMeFix/SsdPmEnabler. The SSD would also typically heat up to and run at a consistent 42C while idle.

From this, I could conculde that it would seem Apple has done some optimizations under the hood with regards to the NVMe drivers in the BootRom/OS. Full ASPM and L0s/L1 states are not implemented but power consumption has improved.

As always, YMMV - but for anyone who does not want to go through the process of loading the KEXTs or OpenCore, this may help you decide if you want to go through the trouble at all now for what are now somewhat diminishing gains.

Thank you for your update.

With the Kexts installed, can you still get the delta update or does it download the full update (~12GB)?

I'm asking because when you root patch the system in OCLP, it messes up with the volume seal and you need to download the full update always. OCLP installs Lilu+NVMeFix automatically for me but, because I don't have any unsupported devices, I always get rid of all Kext OCLP wants to install.

Thanks again.

 

[User9]

if i recall correctly, Anandtech have deep review for that drive,

Just to recap, in term of speed, consumption and reliability, it has no real benefit compared to adapter solution. It just benefitted from direct fit socket.

But direct fit socket have big weakness when you need troubleshooting,

with adapter solution, you can always plug the NVME drive to any NVME enclosure, while drive like OWC aura gonna need another custom enclosure, which normally cost

Hi All,

Short update, I have Monterey 12.2.1 on an MBP 12,1 with Bootrom 428.60.3.0.0.

I have been testing whether or not Lilu+NVMeFix and SsdPmEnabler are still required - I think the use case has changed now with the latest OS.

Battery runtime:
6.5 to 7 hours (web browsing, youtube) with no KEXTs installed.
7 to 7.5 hours with the KEXTs installed (web browsing, youtube)

Temperatures:
With no KEXTs installed: Idle SSD Temperature while doing the above - anywhere from 37 to 39C
With KEXTs installed: Idle SSD Temperature while doing the above - typically ~35-37C

Bit of history - I have been active in this thread since mid-2020 and have experimented with the hardware in my signature (as well as a couple of different adapters and SSDs) on my machine, starting from Catalina, through the BigSur betas (as they came out in 2020) to Monterey now and I have to say that the runtime and heat characteristics are slightly improved under Monterey. In the past, I would struggle to achieve 5hrs of runtime on my battery without NVMeFix/SsdPmEnabler. The SSD would also typically heat up to and run at a consistent 42C while idle.

From this, I could conculde that it would seem Apple has done some optimizations under the hood with regards to the NVMe drivers in the BootRom/OS. Full ASPM and L0s/L1 states are not implemented but power consumption has improved.

As always, YMMV - but for anyone who does not want to go through the process of loading the KEXTs or OpenCore, this may help you decide if you want to go through the trouble at all now for what are now somewhat diminishing gains.

doesn’t look like kexts are worth it anymore if it is indeed small difference

 

[olfo]

Ok, so I compared the P2 circuit board to some articles and noticed the articles said the non-TLC version had fewer chips on the PCB. Mine had all positions filled and the label says UK/CA. I was curious what the deal is, so I installed the SSD and did a macOS clean install. For what its worth, BlackMagic says ~1800MB/s write an ~1700MB/s read. My original Apple SSD was ~400MB/s write and ~1500MB/s read.

Any thoughts guys? Should I return or keep?

I suggest you do a benchmark with much data (>135 GB) written continously. That should reveal the weak spots of the SSD.
This SSD has been discussed here before several times. If you do a search for my posts, you will probably get more benchmarks and opinions to compare, as I have commented occasionally on the Crucial P2 since May 2021. It's probably not too bad, especially for the momentary price.
Sorry for my short and lazy response, and good luck.

 

[otosan]

F

Thank you for your update.

With the Kexts installed, can you still get the delta update or does it download the full update (~12GB)?

I'm asking because when you root patch the system in OCLP, it messes up with the volume seal and you need to download the full update always. OCLP installs Lilu+NVMeFix automatically for me but, because I don't have any unsupported devices, I always get rid of all Kext OCLP wants to install.

Thanks again.

From what i tried, Delta is possible even with .kext installed, only around 1-2GB

i even deleted some kext on my machine (all related to RAID, since somehow it can randomly crash my macbook - traced it from crash log)

 

[vince22]

Hi All,

Short update, I have Monterey 12.2.1 on an MBP 12,1 with Bootrom 428.60.3.0.0.

I have been testing whether or not Lilu+NVMeFix and SsdPmEnabler are still required - I think the use case has changed now with the latest OS.

Battery runtime:
6.5 to 7 hours (web browsing, youtube) with no KEXTs installed.
7 to 7.5 hours with the KEXTs installed (web browsing, youtube)

Temperatures:
With no KEXTs installed: Idle SSD Temperature while doing the above - anywhere from 37 to 39C
With KEXTs installed: Idle SSD Temperature while doing the above - typically ~35-37C

Bit of history - I have been active in this thread since mid-2020 and have experimented with the hardware in my signature (as well as a couple of different adapters and SSDs) on my machine, starting from Catalina, through the BigSur betas (as they came out in 2020) to Monterey now and I have to say that the runtime and heat characteristics are slightly improved under Monterey. In the past, I would struggle to achieve 5hrs of runtime on my battery without NVMeFix/SsdPmEnabler. The SSD would also typically heat up to and run at a consistent 42C while idle.

From this, I could conculde that it would seem Apple has done some optimizations under the hood with regards to the NVMe drivers in the BootRom/OS. Full ASPM and L0s/L1 states are not implemented but power consumption has improved.

As always, YMMV - but for anyone who does not want to go through the process of loading the KEXTs or OpenCore, this may help you decide if you want to go through the trouble at all now for what are now somewhat diminishing gains.

thanks, very useful info, I've switched to Monterey 12.3 OCLP 0.43n with my 2015 13" MBP, Hynix P31, SIP enabled, Universal control enabled, Lilu,Nvmefix,ssdpmenabler kext loaded. Comparing with default Monterey 12.3 SIP disabled, boot arguments loaded for Lilu, Nvmefix, ssdpmenabler kext loaded, with identical settings for both system, power idling consumption ( doing nothing) 50% screen brightness, no apps running in the background, tested at least 1hr. OCLP 0.43n Monterey have a better battery power management in results.

Default Monterey 12.3 SIP disabled.

OCLP 0.43n Monterey 12.3 SIP enabled

 

[herb2k]

Thank you for your update.

With the Kexts installed, can you still get the delta update or does it download the full update (~12GB)?

I'm asking because when you root patch the system in OCLP, it messes up with the volume seal and you need to download the full update always. OCLP installs Lilu+NVMeFix automatically for me but, because I don't have any unsupported devices, I always get rid of all Kext OCLP wants to install.

Thanks again.

To be honest - I have not noticed. I believe my machine was still getting the delta updates as they generally downloaded and installed quick... but I will keep an eye out next time.

Note - I am using OpenCore (configured it manually with the bare minimum required to load the KEXTs), not OCLP... as I wasn't able to get that to work at all on my 13" MBP at the time.

 

[maily.spamy]

Being known to work doesn't necessarily mean it will work the same way for everyone if at all. These adapters are a hit and miss, so you might have to get a couple or so to get one that works right.

In my case, like I mentioned in my post and one of my comments, I lucked out with the one adapter I bought for cheap, with the A2000.



There are now at least 5 people who have installed the Kingston A2000 (regardless of adapter) successfully without issues so far. It's the cheaper option among the SSDs and has the least amount of issues if at all. I've installed it a few months ago and have not experienced any sleep or hibernation issues. Power consumption is decent even without the kext, and those of us who have installed it, afaik, have opted not to use any kext at all since the power saving isn't really that significant.

I also have a 2014, so, the A2000 might be your best bet. Read from my post onward (and a few posts back) and you'll see a couple of others who have used the A2000, or do a search for it.

Hi thanks for info, but I anyway bought the SN750 and the cheapest adapter and it works.
This is the working adapter: adapter
And here a speed test:

I thought that the speed will be higher, but it is anyway much more than the original drive.
It was everything plug and play and sleeping (lid closing) works.
On the machine is installed the newest last compatible macos big sur.
Thanks and regards
standal

 

[ahm12]

This Thread was a big help so here is some info that might come in handy for some. I have never used macOS, first time. I bought 2015 MacBook Air 13' from eBay.
I am using sinitech small adaptor and bought WD SN570 since that one was newer and found some review which said 100mw better at idle compared to sn550, eventhough overall power consumption is higher but idle is better than sn550.

installation went fine write speed 1400, read speed 1497. I was hoping higher read speeds but I guess 1497 is fine. I am not sure what's the limiting factor here laptop itself or sinitech adaptor.

Laptop is going in and coming out of sleep just fine. Can't comment on battery life because I never used the laptop with the original 128Gb SSD. got the laptop and literally did a clean install right away.

 

[keepsec]

thanks, very useful info, I've switched to Monterey 12.3 OCLP 0.43n with my 2015 13" MBP, Hynix P31, SIP enabled, Universal control enabled, Lilu,Nvmefix,ssdpmenabler kext loaded. Comparing with default Monterey 12.3 SIP disabled, boot arguments loaded for Lilu, Nvmefix, ssdpmenabler kext loaded, with identical settings for both system, power idling consumption ( doing nothing) 50% screen brightness, no apps running in the background, tested at least 1hr. OCLP 0.43n Monterey have a better battery power management in results.

Default Monterey 12.3 SIP disabled.

View attachment 1961272

OCLP 0.43n Monterey 12.3 SIP enabled

View attachment 1961274

Can you please explain which boot arguments do you use for Lilu, NVMeFix with default Monterey without OCLP? I have a P31 (2TB) but it can't go under 28mA with all 3 kext loaded?

 

[ahm12]

Can you please explain which boot arguments do you use for Lilu, NVMeFix with default Monterey without OCLP? I have a P31 (2TB) but it can't go under 28mA with all 3 kext loaded?

sorry I thought you were asking me, can't figure out to delete

 

[herb2k]

Can you please explain which boot arguments do you use for Lilu, NVMeFix with default Monterey without OCLP? I have a P31 (2TB) but it can't go under 28mA with all 3 kext loaded?

You will need to use keepsyms=1 as boot arguments.

 

[User9]

thanks, very useful info, I've switched to Monterey 12.3 OCLP 0.43n with my 2015 13" MBP, Hynix P31, SIP enabled, Universal control enabled, Lilu,Nvmefix,ssdpmenabler kext loaded. Comparing with default Monterey 12.3 SIP disabled, boot arguments loaded for Lilu, Nvmefix, ssdpmenabler kext loaded, with identical settings for both system, power idling consumption ( doing nothing) 50% screen brightness, no apps running in the background, tested at least 1hr. OCLP 0.43n Monterey have a better battery power management in results.

Default Monterey 12.3 SIP disabled.

View attachment 1961272

OCLP 0.43n Monterey 12.3 SIP enabled

View attachment 1961274

there is something strange about latest Monterey update as without kexts the battery seemed to be just as good; anyone know if Apple enabled some power saving feature in non-apple ssd?

 

[keepsec]

Can you please explain which boot varguments do you use for Lilu, NVMeFix with default Monterey without OCLP? I have a P31 (2TB) but it can't go under 28mA with all 3 kext loaded?

I have that boot argument otherwise I have I have kernal panic.

I just can't understand why that some people can get as low as 1mA with their P31.

I have a MBP 15 2015.

 

[User9]

I have that boot argument otherwise I have I have kernal panic.

I just can't understand why that some people can get as low as 1mA with their P31.

I have a MBP 15 2015.

due to 13" fewer pci-e lanes therefore lower speed and power, also 1TB consumes less than 2TB. It could also be false reading as the lowest I've seen is with Intel 660P 2TB 5mA. Also note the fan is 0 rpm which may not be right.

 

[olfo]

An interesting observation about the SSD wearout. My 2TB Crucial P2 has reached today 2% wearout as of SSDReporter / smartmontools. In 9 months, 21 TB have been written. While at first this seems a lot, it can be easily calculated that it will reach it’s full 600 TBW in 23 years.
And if 20 TB are really 2% of the SSD’s lifespan, then 100% equal 1000 TB, not 600. Anyway, this makes me feel relaxed about wearout.

The only thing that worries me is the “launchd” process that is sometimes on top of the list in Activity Monitor’s hard drive usage. Today it shows about 900 GB written in 4 days.

 

[robinasu]

thanks, very useful info, I've switched to Monterey 12.3 OCLP 0.43n with my 2015 13" MBP, Hynix P31, SIP enabled, Universal control enabled, Lilu,Nvmefix,ssdpmenabler kext loaded. Comparing with default Monterey 12.3 SIP disabled, boot arguments loaded for Lilu, Nvmefix, ssdpmenabler kext loaded, with identical settings for both system, power idling consumption ( doing nothing) 50% screen brightness, no apps running in the background, tested at least 1hr. OCLP 0.43n Monterey have a better battery power management in results.

Default Monterey 12.3 SIP disabled.

View attachment 1961272

OCLP 0.43n Monterey 12.3 SIP enabled

View attachment 1961274

Hi, what monitoring application is this? I would also like to monitor SSD current in real time.

 

[keepsec]

Hi, what monitoring application is this? I would also like to monitor SSD current in real time.

It is iStat Menus from Bjango.

 

[robinasu]

Quick update.. bought the P2 2TB the other day. Has the UK/CA stamp. Been benchmarking it on my Macbook Retina 2015 (USB installed Monterey; no SSD hacks) and this is a wild guess, but it seems like the new QLC firmware maybe improved things compared to previous forum feedback as BlackMagic test seems consistent for multiple tests in a row, but randomly performance drops significantly. Pure speculation at this point. Using iStat, the SSD idle current is ~0.155A and with sustained SSD activity it increases to about ~0.55A average with peaks to 1A depending on the tests. Clean up actions after big writes keep current at ~0.5A for a while. I did not see a noticeable difference between write/read current. Let me know if you all want me to do any tests. If these rough numbers are correct, the P2 power numbers are now in the range of the Phision E12 based SSDs. I will try the Rocket.

Also, write speeds are much faster than reads. Seems similar to the E12 based SSDs. When I take out the P2 I will take some pictures of it.

With this said, works fine for general use cases. Faster than original Apple SSD.

 

[olfo]

Quick update.. bought the P2 2TB the other day. Has the UK/CA stamp. Been benchmarking it on my Macbook Retina 2015 (USB installed Monterey; no SSD hacks) and this is a wild guess, but it seems like the new QLC firmware maybe improved things compared to previous forum feedback as BlackMagic test seems consistent for multiple tests in a row, but randomly performance drops significantly. Pure speculation at this point. Using iStat, the SSD idle current is ~0.155A and with sustained SSD activity it increases to about ~0.55A average with peaks to 1A depending on the tests. Clean up actions after big writes keep current at ~0.5A for a while. I did not see a noticeable difference between write/read current. Let me know if you all want me to do any tests. If these rough numbers are correct, the P2 power numbers are now in the range of the Phision E12 based SSDs. I will try the Rocket.

Also, write speeds are much faster than reads. Seems similar to the E12 based SSDs. When I take out the P2 I will take some pictures of it.

With this said, works fine for general use cases. Faster than original Apple SSD.

Always good to see new tests of the mysterious variants of this SSD.
As for the benchmark – AJA system test would show performance drops graphically over time. You can set it to test a 64GB file continously. I'd be curious to see this done with your SSD.

Upgrading 2013/2014 Macbook Pro SSD to M.2 NVMe

AJA System Test set to 64 GB, and set to “run continously”. Click on the “graph” icon to show the data rate graph. This is my 2 TB Crucial P2: Look into System Report under “Extensions”. Not necessarily – one or the other, or both. As you think what works best for you. Not really sure what...

forums.macrumors.com

Also AmorphousDiskmark can test all kinds of file sizes. You could start with the default setting and see how it changes when you run a larger file. Compare with mine here if you like.
The idle power draw can be reduced with SsdPmEnabler.
Examples:

Confirmed working Mac models

Contribute to kvic-z/SsdPmEnabler development by creating an account on GitHub.

github.com

 

[xanderx007]

An interesting observation about the SSD wearout. My 2TB Crucial P2 has reached today 2% wearout as of SSDReporter / smartmontools. In 9 months, 21 TB have been written. While at first this seems a lot, it can be easily calculated that it will reach it’s full 600 TBW in 23 years.
And if 20 TB are really 2% of the SSD’s lifespan, then 100% equal 1000 TB, not 600. Anyway, this makes me feel relaxed about wearout.
View attachment 1961908
The only thing that worries me is the “launchd” process that is sometimes on top of the list in Activity Monitor’s hard drive usage. Today it shows about 900 GB written in 4 days.

SSD wear depends on how heavy you use your drive. Since my work is design/graphics heavy, disk access, read/write is understandably heavy.Compared to yours, you have 2% wear in 9 months, my A2000 has 1% wear in 4 months, so the wear between the two is roughly similar. Both the P2 and the A2000 has 600TBW, so they would basically last the same.

 

[User9]

SSD wear depends on how heavy you use your drive. Since my work is design/graphics heavy, disk access, read/write is understandably heavy.Compared to yours, you have 2% wear in 9 months, my A2000 has 1% wear in 4 months, so the wear between the two is roughly similar. Both the P2 and the A2000 has 600TBW, so they would basically last the same.

just curious what usually happens at the end of 600 TBW does the drive just flat out write data with errors?