As I reported, I had issues with my 17" MacBook Pro and its flaky SATA cable.
The folks over at OWC figured this one out: it’s a badly shielded cable (or perhaps a low quality one). See OWC Offers Fix For 2011 17" MBP SATA Problems.
OWC will be offering a shielding kit for this Apple design defect. Hopefully Apple with fix it under warranty. This seems to be an issue only with the 17" model.
I don’t yet have the shielding kit, but I tried Reynolds Wrap aluminum foil. As a friend of mine might say: “You’re shitting me, right?”.
No, I’m not shitting you.
In short, if you can cook, you can fix your MacBook Pro 17" while you prepare dinner. Just don’t short out your Macbook Pro with stray foil.
The aluminum foil speed upgrade for your 17" Macbook Pro
Shielded here means wrapped in aluminum foil. How you wrap the cable influences how well the shielding works. The OWC shielding kit should be better, I presume.
Unshielded: 50 MB/sec (and highly erratic speed)
Cable shielded: 296 MB/sec
Cable + drive shielded: 482 MB/sec (but exact wrapping makes results vary)
Results varied wildly depending on how the cable was wrapped meaning I couldn’t really make sense of the best way to wrap it (without spending hours experimenting, which I did not).
Cable wrap and drive wrap results and photos
Shown below is the 17" 2011 MacBook Pro drive bay with the drive removed. The SATA cable has been wrapped in aluminum foil. This approach yielded only partial improvement, with read speeds around 300MB/sec, but also dropping sometimes to 70MB/sec. I used the DiskTester run-sequential command to test the speed.
Maybe I didn’t wrap the cable all that well, or the foil is not heavy enough (I did not have the “heavy duty” foil, just regular thickness.
Each time I wrapped it the results varied. When I re-wrapped the drive, I did not get as high a performance as I had the first time. So maybe a RF leak requires just the right wrap-job.
Even more interesting, I was getting absolutely horrible numbers (50MB/sec or so) with no wrapping. Later, I got much higher numbers after wrapping/unwrapping again. I have no explanation other than perhaps RF interference could vary for other unknown reasons.
Holding a strong neodymium magnet near the drive bay, especially towards the optical bay side, but not actually touching it, would immediately force the MacBook Pro into sleep mode. On the 15" MacBook Pro, I was able to force sleep on the left side of the unit, well away from the drive bay. The magnet had no effect near the drive bay. But the guys at OWC tell me that this might just be the way the MBP knows when the lid is closed.
Wind:~ lloyd$ disktester rs -t 4G -i 20 owc6G DiskTester 2.4.1b3 64-bit, diglloydTools 2.0.1b3, 2010-07-19 0822 ----------------- Averages for "owc6G" (4GB/4MB, 20 iterations) ---------------- Iteration Write MB/sec Read MB/sec 1 477 326.4 2 477 336.7 3 473 272.8 4 470 318.1 5 479 377.8 6 469 297.1 7 469 342.8 8 466 341.1 9 482 330.9 10 471 304.9 11 468 306.8 12 460 302.8 13 481 74.7 14 480 342.5 15 471 324.2 16 468 375.4 17 478 81.7 18 480 82.1 19 480 360.1 20 470 430.4 Slowest 460 74.7 Fastest 482 430 Average 474 296 Median 472 325 Range 21.2 356
With the OWC Mercury Extreme Pro 6G SSD installed, just the cable wrapped (sits underneath the drive).