Issues with the 3.5" SATA to CompactFlash SSD Adapter

[ppgrainbow]

Hey there!

I had someone install the StarTech 3.5" SATA to CompactFlash SSD adapter in the 3.5" drive bay and so far, the computer is working good.

Windows Vista can detect what CompactFlash card is inside the 3.5" SATA to CF drive bay without any major problems. However, I'm running into issues after I insert the CF card in the 3.5" SATA to CF drive bay and turn the computer on:

1. The BIOS will recognise the CF card that is connected to the 3.5" SATA to CF drive bay as it is connected on SATA Port 4 on the ASUS M5A97 R2.0 motherboard. However, when Windows Vista attempts to boot up with the "© Microsoft Corporation" Segoe UI text and the green progress bar above it, the green light indicator on the SATA to CF drive goes off and the red hard drive LED stays on and the computer hangs indefinitely. In a interval of one minute, the hard drive LED will go off and the green light indicator on the SATA to CF drive goes on for a short period of time.

One temporary work around is that when I try to eject the CF card from the SATA to CF drive, the hard drive LED light would go off. I would leave it ejected for up to one minute and it would let the internal hard drive on SATA Port 1 continue to boot the Windows Vista operating system.

2. When there is data being read or written on the CF card, not only the green light indicator is blinking, but the hard drive LED light is blinking as well and it's not supposed to do that.

 

I have a couple of questions to ask to try to remedy this if possible:

1. Is there a way to fix the issue that will allow Windows Vista to boot successfully even if the CF card is inserted inside the internal SATA to CF drive?

2. In any event there is read and write activity on the CF card, is there a way to stop it from using the hard drive LED light?

The model number on the StarTech SATA to CompactFlash card is 35BAYCF2SAT. The model number can easily be searched on Google. The issues that I mentioned above are minor and do not affect the functionality and performance of my computer.

 

On a sidenote, in order to successfully recognise the CF card inside the the SATA to CF drive, I had to set the SATA ports to use AHCI mode and not IDE mode.

Edited September 13, 2014 by ppgrainbow

[jaclaz]

My guess is that the "hard drive led" is actually an "activity on SATA bus led", and if this is the case there is no way you can change it's behaviour.

As well if you set in the BIOS SATA/AHCI, that setting affects the way the controller is detected by the booting OS and it's "global" for all devices connected to the SATA bus, so last sidenote seems to me like perfectly "normal".

 

The "booting issue" sounds strange, it seems like the booting OS "interrogates" the device and cannot have the reply it expects.

 

Is the behaviour the same with?:

1. a CF card inserted (wiped, i.e. all 00's)
2. a CF card partitioned/formatted (i.e. with first sector a MBR)
3. a CF card only formatted (i.e. as "superfloppy" with first sector a PBR)

Can you make a boot log with the CF card (i.e. "hanging") and one without the CF card (i.e. booting normally)?

 

Additionally, can you try tracing the boot with xbootmgr/xperf?

http://helpdeskgeek.com/how-to/speed-up-boot-time-in-windows-vista/

 

I believe you need the WAIK or the SDK to get the tools.

 

jaclaz

[ppgrainbow]

My guess is that the "hard drive led" is actually an "activity on SATA bus led", and if this is the case there is no way you can change it's behaviour.

As well if you set in the BIOS SATA/AHCI, that setting affects the way the controller is detected by the booting OS and it's "global" for all devices connected to the SATA bus, so last sidenote seems to me like perfectly "normal".

 

The "booting issue" sounds strange, it seems like the booting OS "interrogates" the device and cannot have the reply it expects.

 

Is the behaviour the same with?:

1. a CF card inserted (wiped, i.e. all 00's)
2. a CF card partitioned/formatted (i.e. with first sector a MBR)
3. a CF card only formatted (i.e. as "superfloppy" with first sector a PBR)

Can you make a boot log with the CF card (i.e. "hanging") and one without the CF card (i.e. booting normally)?

 

Additionally, can you try tracing the boot with xbootmgr/xperf?

http://helpdeskgeek.com/how-to/speed-up-boot-time-in-windows-vista/

 

I believe you need the WAIK or the SDK to get the tools.

 

jaclaz

 

Sorry for not responding earlier, but this behaviour occours only when the CompactFlash card is inserted and it doesn't matter if the card is partitioned/formatted or not.

In a forum post by asicnewbie in early May 2008 complianed that the Addonics SATA to CompactFlash adapter doesn't work with the majority of the cards that the user owns and stated that the user's PC will either sit there for a very long time (and will eventually error out) when the user tries to acces the drive or that it simply hangs the Windows XP operating system: http://arstechnica.com/civis/viewtopic.php?f=11&t=127190

On a side note, I'm wondering if it's possible to obtain xbootmg/xperf without going through the hassle of reinstalling the Windows SDK or not? :\

At one point, I went into Safe Mode and found that the hanging error points to the crcdisk.sys which has a timestamp of 2008-01-19. So, I'm wondering if a outdated crcdisk.sys has to do with this booting issue or not.

 

I might try a smaller capacity CF card to see if it makes a difference or not soon.

[jaclaz]

On a side note, I'm wondering if it's possible to obtain xbootmg/xperf without going through the hassle of reinstalling the Windows SDK or not? :\

 

You can seemingly do a partial download/install:

http://windowsitpro.com/windows-server/q-where-can-i-download-xperf-and-xperfview-tools

 

jaclaz

[MagicAndre1981]

You need the Windows 7 SDK under Vista (the tools from the Windows 8 SDK no longer work under Vista ):

 

 

http://social.technet.microsoft.com/wiki/contents/articles/4847.install-the-windows-performance-toolkit-wpt/revision/2.aspx

 

 

During setup, you can select to only get the WPT (you can uncheck all other tools)

Edited September 17, 2014 by MagicAndre1981

[ppgrainbow]

You need the Windows 7 SDK under Vista (the tools from the Windows 8 SDK no longer work under Vista ):

 

 

http://social.technet.microsoft.com/wiki/contents/articles/4847.install-the-windows-performance-toolkit-wpt/revision/2.aspx

 

 

During setup, you can select to only get the WPT (you can uncheck all other tools)

 

Thanks for the help. I'll look into reinstalling the Windows 7.1 SDK. However, this requires me to uninstall the Visual Studio 2010 runtime first. Installing the Windows 7.1 SDK will reinstall the VS2010 runtine.

[ppgrainbow]

Okay, I would like to post a update to this forum post.

I used this command to perform the boot trace below:
 

xbootmgr -trace rebootCycle -noPrepReboot

 

Although, that it performed a boot cycle trace, it traced the output on the root directory of drive C. Therefore, when the tracing was finished, I decided to move it to the C:\Output\Boot sub-directory.

But when I tried to perform this command:
 

xbootmgr -trace boot -noPrepReboot -traceFlags BASE+CSWITCH+DRIVERS+POWER -resultPath C:\Output\Trace

 

I ended up getting this error message:
 

'-noPrepReboot': This option can only be used with shutdown and rebootCycle traces.

 

See 'xbootmgr.exe -help' for more information.

 

Nevertheless, I viewed the following files in the C:\Output directory:

1. First, looking at the boot_BASE+CSWITCH_1.etl and boot_BASE+CSWITCH_1_km_premerge.etl files.

Between around 10 to 315 seconds elapsed time (ET), no activity was being sampled. Between 315 to 520 seconds ET, activity usage started to spike with almost 2,500 hard faults at 320 seconds.

The shutdown_BASE+CSWITCH_1_km_premerge.etl file stated that it took 44 seconds to reboot the computer. Not bad.

2. And then it comes the C:\Output\Trace\boot_BASE+CSWITCH+DRIVERS+POWER_1.etl file. The analysis stated that it took about 80 seconds longer than boot_BASE+CSWITCH_1.etl/boot_BASE+CSWITCH_1_km_premerge.etl files

If you look in the Driver Delays section of this file, you'll notice that partmgr.sys took 310.78 seconds to load and CLASSPNP.SYS taking 1 ms shorter to load. Numerous delays between the 320 to 600 second ET mark indicate long delays with fltmgr.sys delaying for as much as 8 seconds in the 420 to 470 second ET mark.

Something sounds fishy here and the cause of this issue is difficult to decipher here. If you have the Windows 7.1 SDKs installed should I upload the files from the C:\Output\Trace directory onto my OneDrive account so that the issue can be looked into?

If so, please let me know.

Edited September 22, 2014 by ppgrainbow

[MagicAndre1981]

-noPrepReboot is only needed for shutdown or hibernation, but not for boot.

 

 

Run this command please, which adds wait analysis:

 

xbootmgr -trace boot -traceFlags BASE+CSWITCH+DRIVERS+POWER+LATENCY+DISK_IO_INIT+DISPATCHER+FILE_IO+FILE_IO_INIT+FILENAME -stackwalk profile+CSwitch+ReadyThread+DiskReadInit+DiskWriteInit -resultPath C:\Output

 

compress the generated ETL as 7z/RAR, upload it and send me the link via PM.

Edited September 23, 2014 by MagicAndre1981

[ppgrainbow]

-noPrepReboot is only needed for shutdown or hibernation, but not for boot.

 

 

Run this command please, which adds wait analysis:

 

xbootmgr -trace boot -traceFlags BASE+CSWITCH+DRIVERS+POWER+LATENCY+DISK_IO_INIT+DISPATCHER+FILE_IO+FILE_IO_INIT+FILENAME -stackwalk profile+CSwitch+ReadyThread+DiskReadInit+DiskWriteInit -resultPath C:\Output

 

compress the generated ETL as 7z/RAR, upload it and send me the link via PM.

 

I'm gonna be uploading it onto my OneDrive account and privately send you a note.

 

I've noticed that while Windows Vista was attempting to boot, I've noticed a long wait in the CLASSPNP.SYS and partmgr.sys driver delays of 620.01 seconds! Ouch! I always though that crcdisk.sys was the issue as it was determined in Safe Mode, but it appears that I was wrong.

 

The datestamps on the classpnp.sys file is 2009-04-11 and for the partmgr.sys, it's 2009-04-11 for one file and 2012-03-20 for the other two files.

Edited September 23, 2014 by ppgrainbow

[MagicAndre1981]

the long delay occurs by starting the ULTIMATE CF CARD ATA Device:

- <phase name="bootStart" startTime="53" endTime="625731" duration="625677">  <pnpObject name="IDE\DiskULTIMATE_CF_CARD________________________Ver7.02K\5+2efc85c4+0+3.0.0" type="Device" activity="Start" startTime="5287" endTime="625301" duration="620014" prePendTime="620014" description="Disk drive" friendlyName="ULTIMATE CF CARD ATA Device" /> 

this delays boot for 626 seconds. Is this the adapter?

 

In the trace I see the delay in CLASSPNP.SYS!ClassPnpStartDevice, so that's why you see the long classpnp.sys entry.

 

[jaclaz]

Still, is the delay actually in:

1. CLASSPNP.SYS
2. PARTMGR.SYS
3. DISK.SYS

(as hinted before the two latter ones may have a connection with partitioned/not partitioned/fixed/removable status of the media).

 

And personally I wouldn't rule out specific brand/model/*whatever* of the actual card, it seems to me like CF cards are one of the pieces of hardware where every make/model behaves slightly differently.

 

At least some time ago I converted a couple old machines (win2K) to use CF cards (but on IDE/PATA bus, and obviously using el-cheapo no-name adapters) and I remember having quite a few difficulties in finding "decently performing" CF cards (this might be completely unrelated, cannot really remember if the issues I had initially were connected to booting).

 

jaclaz

[ppgrainbow]

the long delay occurs by starting the ULTIMATE CF CARD ATA Device:

- <phase name="bootStart" startTime="53" endTime="625731" duration="625677">  <pnpObject name="IDE\DiskULTIMATE_CF_CARD________________________Ver7.02K\5+2efc85c4+0+3.0.0" type="Device" activity="Start" startTime="5287" endTime="625301" duration="620014" prePendTime="620014" description="Disk drive" friendlyName="ULTIMATE CF CARD ATA Device" /> 

this delays boot for 626 seconds. Is this the adapter?

 

In the trace I see the delay in CLASSPNP.SYS!ClassPnpStartDevice, so that's why you see the long classpnp.sys entry.

 

 

Yep. It's the adapter itself and not the CF card. This boot delay issue occurs only if the CF card is inserted inside the SATA-to-CF adapter.

Edited September 23, 2014 by ppgrainbow

[ppgrainbow]

Still, is the delay actually in:

1. CLASSPNP.SYS
2. PARTMGR.SYS
3. DISK.SYS

(as hinted before the two latter ones may have a connection with partitioned/not partitioned/fixed/removable status of the media).

 

And personally I wouldn't rule out specific brand/model/*whatever* of the actual card, it seems to me like CF cards are one of the pieces of hardware where every make/model behaves slightly differently.

 

At least some time ago I converted a couple old machines (win2K) to use CF cards (but on IDE/PATA bus, and obviously using el-cheapo no-name adapters) and I remember having quite a few difficulties in finding "decently performing" CF cards (this might be completely unrelated, cannot really remember if the issues I had initially were connected to booting).

 

jaclaz

 

Like I said before, the boot delay occurs in the classpnp.sys and it only happens when the CF card is inserted inside the SATA-to-CF adapter.

[MagicAndre1981]

The driver is fine. It is the device which is slow to start.

[ppgrainbow]

The driver is fine. It is the device which is slow to start.

 

I knew it.

 

Do you have any idea what can be done that can make Vista detect the CF card in the drive at boot without having to hang on startup?