USB External SLOW

[dnrsslr]

I have a Coolmax HD-338-U2 USB external enclosure with a Maxtor DiamondMax 21, Ultra ATA/100 hard drive inside. The drive seemed to work fine for awhile. Lately, it has been VERY SLOW sometimes. An Acronis image said it would take like 7 hours! But, other times, it seemed like Acronis was able to make the scheduled backup in a few hours.

Now, when I connect it to USB 2.0 ports, a message comes up that says it is a 2.0 device that has been connected to a port that is not 2.0. I have tried it on three different computers with the same result.

Any suggestions?

 

[WyomingKnott]

Try swapping the easiest things first. Have you tried a new cable?

 

[dnrsslr]

I tried a different cable but it did the same thing. However, I may have found the problem; I read in the little manual that came with the Coolmax enclosure, I am supposed to format as FAT16 or FAT32. The drive looks like it is formatted NTFS. Would it make be logical that it would work on NTFS for awhile and then start causing problems? I plan to take it out of the enclosure and re-format it to see if that helps.

I'm thinking about buying a new enclosure. Is this common that an enclosure does not allow NTFS?

 

[WyomingKnott]

NTFS is a perfectly good and happy filesystem. Some older OSes won't work with it, so you have to use FATx, but I have never heard of NTFS being slower even if it has the slight overhead of journaling ( http://en.wikipedia.org/wiki/NTFS#NTFS_Log ).

To the best of my knowledge... No, make that "No USB enclosure cares what filesystem is on the drive." It only provides a connection from the drive interface to a USB interface; the motherboard does the reading. (The exception is external RAID enclosures that do hardware RAID inside the enclosure, but that's not what we are talking about).

================

Reformatting is a decent idea as long as you don't care about the data that is on it now. You might want to do one of three things based on the possibility that the disk is getting too old to use.

1) Go to Maxtor and download a utility to check the disk. See if it is showing signs of failure.

2) Attach the drive directly to the motherboard's EIDE port (if it has one) and see if the same behaviour persists, in which case you likely have a drive problem.

3) Just buy a newer drive.


4) Run a disk-speed test utility. USB is a pretty limiting interface; if you are getting full USB speeds then it's not the disk. The following is blatantly copied from Wikipedia, at this address: http://en.wikipedia.org/wiki/Usb#Transfer_speeds_in_practice

Transfer speeds in practice

As of 2004[update], the actual throughput of USB 2.0 high bandwidth attained with a hard drive tested on a Mac was about 18 MiB/s, 30% of the maximum theoretical bulk data transfer rate of 60 MiB/s (480 Mbit/s). On Windows, the highest speed observed was 33 MiB/s, or 55% of the theoretical max. The drive could reach 58 MiB/s on Firewire, so the drive's speed was not a limiting factor.[56]

According to a USB-IF chairman, "at least 10 to 15 percent of the stated peak 60 MB/s (480 Mbit/s) of Hi-Speed USB goes to overhead — the communication protocol between the card and the peripheral. Overhead is a component of all connectivity standards."[57] Tables illustrating the transfer limits are shown in Chapter 5 of the USB spec.

Typical high bandwidth USB devices operate at lower data rates, often about 3 MiB/s overall, sometimes up to 10–20 MiB/s[citation needed]. A Full Speed device (for example PIC18f2550) easily reaches a little over 1MiB/s over a bulk endpoint[citati

 

[dnrsslr]

Thanks for the information. I plan to attach the drive directly to the motherboard but I have not had a chance to do that yet.

Not to be argumentative but, if the enclosure doesn't care about the format why does the manual state the following? "For new HDD: before connecting with USB IDE Box, please partition and format your HDD into FAT16 or FAT32. If your HDD partition is NTFS, please let your HDD be FAT16 or FAT32 before connection." The enclosure is almost four years old so that was before Vista...

Do you have a disk-speed test utility you would recommend? Should I just Google to fine one?

 

[dissembled]

Had this problem myself.

Before you go formatting the drive, go to Device Manager and check to make sure you don't have any little yellow exclamation marks on your USB devices. If you do, update the drivers on them.

 

[WyomingKnott]

Not to be argumentative but, if the enclosure doesn't care about the format why does the manual state the following? "For new HDD: before connecting with USB IDE Box, please partition and format your HDD into FAT16 or FAT32. If your HDD partition is NTFS, please let your HDD be FAT16 or FAT32 before connection." The enclosure is almost four years old so that was before Vista...

Do you have a disk-speed test utility you would recommend? Should I just Google to fine one?

First question: Either A) I am confused and ignorant, or B) the device's manufacturer was confused and ignorant at the time, or C) NT-based OSes could not use NTFS on a USB drive. I consider B to be slightly, but only slightly, more likely than A. Just because I can't imagine a reason doesn't mean there isn't one; I have trouble imagining Victoria Falls, but I believe that they do exist.

Second question: I do very, very brief testing and use the Atto tool: http://www.attotech.com/products/product.php?sku=Disk_Benchmark . It's just a quick glance; if you look at articles from Tom's you'll see transfer rates over the whole surface of the drive, IO/Sec, and more relevant information. But a quick look is good enough to tell you if you are getting 480Kb/s or 80 Kb/s.

 

[dnrsslr]

I installed the drive directly to the motherboard. The Acronis backup image took less than 10 minutes! Here are two disk programs I ran:

Acronis Drive Monitor Disk overview says
- "Disk health: 30% health of disk is below the warning level at 70."
- "Disk temperature: 120 F. Temperature of the disk is above the warning level 107 F." I wonder about this because the drive is now in a wide open case so it shouldn't be hot...
- "Reallocated Sectors count S.M.A.R.T. attribute reported bad block on the drive. Increasing number of bad blocks may be an indicator of imminent drive failure...."

I also have HD Tune. I shows Reallocated Sector Count of 99 (threshold 36) but says Status "OK". It also has Spin Retry Count of 100 (threshold 97) also Status "OK".

What does this mean?

 

[WyomingKnott]

It means
A) Copy everything to another disk. Oh - it's your backup disk?
B) Buy a new disk

That drive may survive for another two years or stop working tomorrow. It is unhappy. SMART attributes tend to overstate how bad things are, but why take a chance. Other people on the forum will have better ideas about SMART criteria than I do. Download the disk-maker's diagnostic tool and run it.

-----------------------------------------------------------------------

I suffer from curiosity. If I had a spare drive and your issue, I would amuse myself by putting a known good drive into the enclosure, formatting it NTFS, testing, formatting it FAT32, and testing again. I would simply want to find out if the manufacturer was right. But that's what I do FOR FUN. Also, the fact that the backup ran that much faster directly connected to the mobo may be within the range of acceptable: USB 2.0 is slow compared to the direct disk interfaces.

So I still can't say, from what you've told me, whether the speed in the enclosure is actually reasonable, whether the speed problem is really due to the format used (despite the fact that I have never seen an external enclosure which cared; that doesn't mean that there isn't one) and would pick up with FAT32 format, whether the drive was getting just barely enough power through USB and retrying over and over and over. At a minumum, we would need benchmarks of the drive in the USB enclosure as-is and, if the speed is significantly below USB rated speeds, another drive formatted NTFS and that other drive reformatted FAT32.

Edit: I think we've reached the end of what I can do for you that would be useful. I'll lurk to see if you post benchmarks, because I like to understand things, but if you need more you might start a new thread with the info from the original post plus the SMART info and speed with direct-connect. New thread so a competent reader won't skip it thinking that I'm handling it.

 

[dnrsslr]

The disk is attached directly to the motherboard with IDE cable. While copying files, I ran HD Tune Benchmark. The maximum Transfer Rate was 72.8 MB/sec, Minimum was 0.4 and Average was 59.9 MB/sec. Access tme was 15.7 ms. Burst Rate was 65.3 MB/sec and CPU usage was -1.0%. What does this mean? Will these numbers be different since I ran the test at the same time I was copying files?

While copying files, it seemed to be cruising faster at first then slowed down. Or, maybe I just imagined that...

 

[WyomingKnott]

The disk is attached directly to the motherboard with IDE cable. While copying files, I ran HD Tune Benchmark. The maximum Transfer Rate was 72.8 MB/sec, Minimum was 0.4 and Average was 59.9 MB/sec. Access tme was 15.7 ms. Burst Rate was 65.3 MB/sec and CPU usage was -1.0%. What does this mean? Will these numbers be different since I ran the test at the same time I was copying files?

While copying files, it seemed to be cruising faster at first then slowed down. Or, maybe I just imagined that...

-1.0% CPU usage is a statistical fluke.
What the rest of it means is, that for an IDE disk, the transfer rates are reasonable.
Yes, the rates will almost certainly be higher if you measure them while the disk isn't busy with something else. Unless that benchmark package can claim exclusive use of the disk, which would certainly have stalled your copy.

As to it seeming to slow down, I have three possible interpretations and no way to tell which is correct. First, you started the copy before the benchmark; competition for bandwidth then slowed it down. Second, the disk has a failing area and your copy started taking space in that area. Third, anything else: some cylinders are faster than others, having to grab fragmented space can be slower than writing to contiguous free space, lots of things that I don't know.

I'd emphasize two important points here: Get the manufacturer's diagnostics and run them on the disk, and I can't really answer the original question without benchmarking in the USB box.

 

[dnrsslr]

The Maxtor software I downloaded didn't seem to do much diagnostics... Maybe I didn't download the right thing or didn't know how to run it...

With the drive back in the enclosure, it was back to being unreasonably slow and shows up as USB 1.0. I don't have a spare IDE hard drive at home to test in the enclosure. But, maybe I can come up with one somewhere to test...

But, I do have a spare SATA drive. So, I'm buying a SATA (internal connection) enclosure to connect USB and that should work.

Thanks so much for all your help and the information!!!