The problem is that the cable is unshielded and not capable of running 6Gb. If it's working for you, it's because you are very lucky with the cable and ambient noise, or the negotiated data rate is 3Gb.
Since the data rate is negotiated by the PHY, you have to tell the PHY to not negotiate the highest possible rate. This is a PITA, but it's possible by writing a 2h value to Offset 2Ch: PxSCTL – Port x Serial ATA Control (SCR2: SControl) register at 07:04. This is possible, but requires writing a KEXT which does a method override on the IOAHCIBlockStorage.kext, and has a better matching priority than that KEXT.
This will basically lock you at 3Gb, but then your system won't be crashing all the time.
Your method sounds promising. Do you mind guiding me thru the method of changing the KEXT?
Not too familiar with that.
It seems like ppl at Crucial has the problem fixed with a firmware update.
Not sure if it is related though.
My MBP negotiated 3 Gbps with an intel ssd 510, but still experienced the beach ball problem.
I thought resetting the SMC was a good idea, and since then the link it at 6 Gbps - with the same amount of beach balls.
I found this strange. The original apple drive works fine and runs at 3 Gbps, but why did the 510 SSD not manage this?
And yes, I really regret resetting the SMC, but was desperate for a solution. My hope now is for the 10.6.8 update (due in a few days), and possibly the tin-foil method. Even getting a stable 3 Gbps would be acceptable.
I manage to get beachball free on 3gbps. Strange.
As for 10.6.8, do not be too hopeful on the final release. I have tried the beta 10.6.8 10K537, beachballing came back after two hours.
As for tin foil method, I have strange result.
Tin foil the Sata3 cable - Can't boot from cold start - from shutdown (getting mach_kernel read error - status 6)
*manage to get it to boot by using the Install CD and point it to Startup from the HDD. Not a single beachballs.
* removing the tin foil will allow the MBP to boot again. Beachballs will occur
From the test results, the sata cable is highly sensitive to interference if runs on 6gbps.
You may need a shielded SATA cable, some are, some aren't, and almost impossible to tell by looking at them.
IMO, the main problem here is simply due to the way Apple/MACs basically are, using products not found in the apple tree are at your own risk. The item may work fine if you are a techie-type expert on the internals of Apple products, but otherwise their PCs are not anything goes hardware-wise as Windows PCs are. Apple products are on the opposite end of the "open source" concept of PC software and peripherals. As stated in a post at the beginning of this thread, Apple does not support products that are not "approved" by them. While that may be changing due to the anger of their customers, we have yet to see products like SSDs that state "Compatible with Apple..." on the box. Apple might sue them for doing that.
I have a 15" 2011 MBP with an intel 510 and was suffering from the intermittent 30 second freeze/beachballs described in this thread. So far, the only solution I have found is described over at the crucial forums. Basically you ping the drive to avoid LPM (power management) from locking up the drive.
Packaged up in script format:
I don't think these LPM freeze bugs are anything new (you can google and see people having them under windows, it just so happens that they can more easily disable LPM.)
Not super thrilled about the nature of this workaround, but I have not beachballed since applying it.
"Basically you ping the drive to avoid LPM (power management) from locking up the drive."
That makes sense. The 510 uses more power than other Intel SSDs due to the SATA 6Gb/s interface, and it's power requirements are above that supplied when LPM is active. Whatever that threshold is, this and other drives need more than it allows. The Apple approved SSDs are fine at that level. It must be very low, since the 510's idle power usage is low.
It looks like there hasn't been much activity on this thread for over a month.
Is that because the problem has been solved somehow?
After reading for about 4 hours: from other Macbook threads where SATA-3 / 6gbps problems are discussed, this problem appears to happen with *all* 6gbps drives (intel 510, OCZ Vertex3, Kingston HyperX, Crucial's drives, etc). It also appears to have some dependency on when the macbook was manufactured, with the trend appearing to be that the later in the year, the better the chances the drive will work (no beachballs/hangs), but no guarantee (there are no reports I can find of an August 2011 machine having a problem, but certainly macbooks made in June have been reported as having the problem). It's not clear whether the cable was changed by Apple at some point, but it is clear that Apple silently changed the SATA configuration on the current Macbooks: from 6gbps main / 3gbps optibay to 6gbps for both. And that also appears to affect the problem (but still not completely solve it!)
Hmm. Here's where I am: i've just ordered a macbook pro 15", it'll arrive around Sept. 15. I want a 6gbps SSD to try, and I want to try it in the first 14 days because if it doesn't work then I believe I have the option to still return my macbook.
Here's the question: do any of you believe the intel 510 is a better option than say the OCZ Vertex3 MIOPS or the Kingston HyperX?
thanks for your advice.
No, I received my new MacBook Pro and my Intel SSD 510 about a month ago and can confirm that the problem persists.
What’s more, I did some investigation on the fix described by Lance French, regarding LPM and it appears to me that this is nonsense, that is, from what I’ve seen it’s not clear to me how that could work. I’m weary to install binaries that just float around in the web, so I’ve downloaded the original utility on which that fix is based, hdapm, which can be obtained here: http://mckinlay.net.nz/hdapm
From what I understand after reading the originial discussion of that fix and the discussion on the “scripted” version, hdapm was only modified so that it sends APM commands continuously, rather than once. That said, when I try to use hdapm as described, the result is as follows:
$ sudo hdapm disk0 max
disk0: INTEL SSDSC2MH120A2
Setting APM level to 0xfe: FAILED: APM not supported
Now, doing that continuously rather than once shouldn’t have much of an effect either, or am I missing out on something?
Any other reports on the tinfoil approach?