As stated in this long thread, there have been two versions of Sintech "short" black adapter.
You may have the old one which doesn't handle properly reset and hibernation. When and where did you buy it ?
Got a tip for us?
Let us know
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Upgrading 2013-2015 Macbook Pro SSD to M.2 NVMe
- Thread starter maxthackray
- WikiPost WikiPost
- Start date
- Sort by reaction score
- Status
As stated in this long thread, there have been two versions of Sintech "short" black adapter.
You may have the old one which doesn't handle properly reset and hibernation. When and where did you buy it ?
If you want to compare properly battery life, you need to do proper comparisons.
Battery life depends on :
- software installed and running in background
- wifi usage (which highly depends on distance to the router)
- display luminosity settings
- battery Full Charge capacity (which is, of course, not always the battery Design capacity)
etc.
So it's very difficult to find trustable battery life tests...
That said, I'm quite upset you don't trust the long work I have done and the consumption tables...
Do you have tested more than 25 models of AHCI and NVMe SSD ? I did.
Here is a screenshot taken with a MyDigital SSD inside a rMBP 13" 2013 :
Here is a graph of the total power consumption of the whole macbook pro in idle with this SSD :
Here is the total power consumption during a read operation :
Now, the same captures with an Apple AHCI 128 GB SSD :
idle :
and during a read operation :
The tested computer was the exact same rMBP 13".
Of course 13 hours while doing nothing doesn't mean that there's a 13h battery life in real usage.
But all of those tests mean that the Phison E7 based NVMe SSDs are not only very cheap and affordable, but also they consume less power than Apple AHCI drives... which were only 2013-based Samsung (or toshiba) SSDs sold to Apple according to their specifications.
Actual NVMe drives are cheaper and, except QLC based drives, bette. that's a simple fact.
As of macOS 10.14, macOS still doesn't support APST on any regular NVMe drive.
Nor does it enable standby mode on AHCI SSDs.
As Apple has moved to their own NVMe controller on 2016-2017 12" Macbook and rMBP, which was even replaced by the T2 chip in later mac, I highly doubt Apple will make any future effort on this.
Take note that lot of people complained of a lower battery life but :
- this was due to the upgrade and data copy itself : the first hours / days, the system reindex the data
- lots of people here did upgrade their OS which usually lower the battery life
If you compare things that are identical (same SSD capacity, same os, same indexed datas), from my experience battery life can be better with some NVMe models (mostly Phison based models) and worse with some models (especially P1 / 660p QLC drives which consume a lot of power to write their data to QLC mode after SLC cache)
Hibernation is enabled by default, but wake from hibernation isn't supported on 2013-2014 macs with the "stock" un patched Bootrom.
We have developed here a method to mod the BootRom so that hibernation can be enabled again on 2013-2014 macs with NVMe drive (it involves replacing the NVMe DXE driver with an uncompressed one).
But you have a 2015 model : and battery drain in sleep mode is not an issue with a 2015 model.
That said, with a NVMe drive you can lower the hibernation delay : I did it on my own 2015 rMBP that I'm using for 2 years now with a NVMe drive. I've setted standbydelay to 180s on battery, so my mac goes to hibernation nearly immediatly. This is not an issue because wake from hibernation is very fast with a NVMe drive.
I'm happy to go halves with you. I'm based in the UK, but will pay extra for the postage. This eBay ad seems to be the best value, as it includes the programmer. Let me know if you're interested.
i bought the same thing. do we need to buy directly from sintech link below ?
http://eshop.sintech.cn/ngff-m2-pcie-ssd-card-as-2013-2014-2015-macbook-ssd-p-1143.html
Last edited:
Man, you did a great job. It's just unclear to me why the average power consumption is calulated like this. Why it consists of 80% idle power, 10% for write and 10% for read, is it based on some measurments how disk is used during regular usage ?
Plus different drives may consume different power under different workloads (like writing big file to disk vs background spotlight index writes)
It would be great if you tested the actual battery life for different drives in same conditions (fixed brightness, same wifi signal level etc.) maybe some web-surfing test or movie playback.
Also what do you think about Kingston A2000 with Silicon Motion SM2263ENG controller. At price 60$ per 512Gb it's going to be the budget king. Is there any chance you will test it ?
There are a lot of factors at play apart from what gilles_pollysoft claims. There are hardware functionalities in the drives that are not supported by macOS such as hibernation, PCI-e active state power management, etc.
Although the drives are more efficient than original AHCI drives designed by Apple it doesn't mean that macOS is capable of making use of those efficiency features.
One of those is CPU architectures that are evolving to be more efficient. NVME drives cause the CPU to work harder but in smaller bursts since the drive is extremely fast compared to AHCI. CPUs with newer architecture will definitely be more efficient and therefore have less impact on battery life. If you noticed, the new mac models like 2017-2019 have significantly smaller battery capacities than the 2013-2015 models yet they are faster and maintain the same battery life guaranteed by Apple and there's a reason for this, its not only software related.
I did a test of both OEM and NVME drives on my 2014 without anything extreme. Just running a bare macOS mojave with basic browsing, some reading and writing and I've lost about 20% of battery compared to the OEM drive, therefore there is something else probably hardware related that is causing the CPU and NVME drive to utilize more power than the designed specs even though they're technically more efficient drives. There is most definitely some compatibility issues behind the scenes even though there is full NVME support for 2015 models.
I would recommend an additional test with the OWC aura pro x2 on a 2015 model to see if OWC has designed their drives to be on par with apple's in terms of efficiency. THere have been reports that people actually get better battery life with the Aura n or Aura Pro x2...
Thanks
Regarding the 80% idle 10% read 10% write : it's not a definitive choice.
I thought it was a good way to show "average" power consumption, of course it can't be representative of all of the different usage people do have with their computer.
A video editor will have an usage with lot of long reads/writes, a ProTools user or a Web developper will have more random access to little files, someone watching a Netflix streaming will have another usage with lots of Wifi but nearly no hard drive writing...
The choice was based on lots of testings on both computers in my office (~10 macs) and from what I've seen with some of my customers, and I based it on observed average reads / write proportions.
It may be more 90% idle, 5% read, 5% write, it is only here to give a good idea.
I'm open to any suggestion.
Well that would be a lot of work : I estimate that I would need to do at least 3 full charge / discharge tests with each model to have a significant and accurate average battery life test.
It would take at least 20 hours of tests for each model of SSD.
I have 29 models to test, to which I would like to add the OWC Aura "N" (a 2x lanes Phison E8 based NVMe SSD) and the Aura Pro X2 (which seems far better than the Aura Pro X which was an horrible raid of 2 PCIe Sata controllers..).
At least 600 hours of test, which obviously I don't have at all to spend :-/
Also the battery of the test model may suffer from 100x full charge / discharge cycles, so it may have a different capacity between the beginning and the end of the test !
Quite a headache...
Mainly, I don't have this time to spend..
I have already spend an enormous time here, in this forum regarding NVMe upgrades :
- I tested and find out the differences between revisions of Sintech / Chenyang adapters
- I tested nearly every Mac models (every MacPro from 4.1 to 6.1, every MacBook Air from 2013 to 2017, every MacBook Pro retina 13" and 15" from late 2013 to mid 2015)...
- I have tested nearly every NVMe SSD model on the market..
- I have tested and found out what caused the hibernation issue, and spent a lot of evenings and saturday to test and find out the DXE NVMe BootRom mod solution...
- I also did some tests with a MacBook Pro 13" 2017 2 TB and I killed a motherboard..
All this was time consuming, hopefully there were a lot of help, support and feedback from other forum members here.
We could ask various users to get their battery life results, but it has proven difficult to get reliable, reproductible datas.
I think the test of power consumption is at least a reliable test on what power is really pumped by the SSD, at least I have not found better idea than measuring the SSD power consumption...
I did it with iStat Menus, and with writing a 100 GB file to the disk, let the disk rest a little (mainly to mesure the time for some SSD to write their SLC cache to TLC or QLC data), and then read back the datas.
It can be reproduced by anyone..
I have it and it seems a good drive, definitely not the faster but very cheap.
I just have tested it last week and did not have time to add it to the tables right now.
Here are the results for a 500 GB drive (I always do tests on 500 GB drives on all SSD brands, in order to have comparable results, 1 TB are faster)
:
- no 4K bloc size support
- idle power consumption = 0,59W (181mA x 3,3V)
- write speed on a 100 GB file = 639MB/s
- write power consumption = 3.2W
- read speed on a 100 GB file = 1642 MB/s (PCIe 2.0, not tested on PCIe 3.0)
- read power consumption = 2.6W
Last edited:
So thoughts on this version?
I've searched but can't really find reference to a "bad" short black Sintech adapter
[doublepost=1568160260][/doublepost]
my system information is the same
I had no idea that there might be a "bad" short Sintech adapter.
but it could be coincidence but changing autopoweroffdelay to 43200 (from 28800), (even though autopoweroff 0)
has so far (one day in) avoided the previous "? folder kernel panic" first thing in the morning and another in the evening when I get back from work...
Hi Everyone!
Thanks for enabling me in performing the long needed upgrade to my '14 13" MBP replacing my 128GB Apple SSD to an 480GB Adata 8200 XPG NVMe drive.
Anyways, I've partitioned my drive in preparation to use boot camp and when I get to the portion of selecting a disk to install Windows on, within the Windows installer, I am unable to see the drive/ partition.
Please note that this is through Boot Camp and it may be just glitchy, but I've never had a drive not show when installing Windows on an NVMe or installing Windows using Boot Camp.
I am installing Windows 10 LTSC. Any ideas?
Thanks for enabling me in performing the long needed upgrade to my '14 13" MBP replacing my 128GB Apple SSD to an 480GB Adata 8200 XPG NVMe drive.
Anyways, I've partitioned my drive in preparation to use boot camp and when I get to the portion of selecting a disk to install Windows on, within the Windows installer, I am unable to see the drive/ partition.
Please note that this is through Boot Camp and it may be just glitchy, but I've never had a drive not show when installing Windows on an NVMe or installing Windows using Boot Camp.
I am installing Windows 10 LTSC. Any ideas?
Hello,
Anyone in France would have a J6100 SPI-ROM adapter which I can borrow for a few days? I pay the postal service expenses.
I just need to flash a new EFI ROM.
I already have the CH341A chip/stick.
Maybe @gilles_polysoft ?
Big thanks!
Anyone in France would have a J6100 SPI-ROM adapter which I can borrow for a few days? I pay the postal service expenses.
I just need to flash a new EFI ROM.
I already have the CH341A chip/stick.
Maybe @gilles_polysoft ?
Big thanks!
I didnt have any kernel panics, no errors Apple hardware test, I’m noob to this mod, just want to have best hibernation battery saving settings. Thanks
Gilles,
First, thank you very much for that that time and effort that you've put into the research, even more for sharing it all with everyone here.
I think I wouldn't be alone in wanting to show my appreciation in some fashion. I think besides paypal, there is some way that I (we) could sent you a tiny token of our appreciation. A few bucks to show our appreciation for your work and for sharing it with the community.
If you have some link or other donation form, please post it here so we can contribute. Even if it's just the equiv of a beer of coffee, it'll add up.
Appreciated,
Paul
First, thank you very much for that that time and effort that you've put into the research, even more for sharing it all with everyone here.
I think I wouldn't be alone in wanting to show my appreciation in some fashion. I think besides paypal, there is some way that I (we) could sent you a tiny token of our appreciation. A few bucks to show our appreciation for your work and for sharing it with the community.
If you have some link or other donation form, please post it here so we can contribute. Even if it's just the equiv of a beer of coffee, it'll add up.
Appreciated,
Paul
Clearly the issue was running the Catalina beta ♂️
Everything running smoothly on Mojave.
Apologies for reporting this so late!
I'm not sure if my adapter is the same as yours (it is also a short Sintech) but I was having the ? folder and frequent crashes until I wrapped my adapter in kapton tape. While my adapter did come with a bit of tape (or something) covering the exposed wires, it wasn't perfect. I've definitely been having overheating issues so it could also be that. Anyhow, the system seems relatively stable now (occasional crash on hibernate but not too often). Not perfect, but since my Apple SSD had failed this is the best I can do on an old machine (2013 15"). I didn't take any measures to avoid hiberation issues (yet).
You might try the kapton tape and see if it helps?
thank you for the response.
I have the same issues the 2TB Sabrent + 13" 2015 MBP when using my previous adapter, which was part of a rock solid 1TB Adata SX8200 upgrade for my 2013 13" MBP, so I'm not sure it's adapter related, but I might order another directly from Sintech.
My MBP 15" (2015, no dGPU) is doing well with the long Sintech adapter and an Intel 760p 2TB SSD for several months now under Mojave. I changed to an Apple SSD temporarily to update the firmware to the latest version.
I plan to keep this MacBook on Mojave because I still use some 32bit apps the are expensive or impossible to replace. From your experience this seems to be the right decision.
Can others confirm that Catalina doesn't play nice with our homemade SSD upgrades? Just want to make sure in case I would change my decision to stay on Mojave.
Hello to all from Japan.
Really appreciated for all of here sharing valuable informations.
Being upgraded my early 2015 13" MBP with 2TB SiliconPower 34A80 bought at Amazon.com and with "Black short adapter" which been bought at Aliexpress (at quite cheap price) safely.
it been working very well with prompt recovery from sleep mode and restart.
Surprising to feel almost no stress to install all step at all.
Very easy.
Hope somebody will follow me and become happy with value for the money.
Regards, JJ.