USB 2.0 problem is highly device dependent

[dcouzin]

I'm one of the many T42 (and T41 and T40) owners not achieving USB 2.0 speed. It is a fascinating problem because (1) it is still (http://forum.thinkpads.com/viewtopic.php?t=29057) unclear whether it is caused by software or hardware and (2) the vicitims never give up.
Since my T42 was under warranty and Lenovo was eager to replace the system board, I took that course. It didn't help at all. Still no high speed USB. I restored the system to "factory" state. Still no high speed USB. So it could only be the BIOS. I returned that to earlier states. Still no high speed USB.
I think Lenovo did replace the board, because things were actually worse than before. With the unsigned drivers, my USB hard drive wasn't recognized. With the USB2 (as opposed to USB 2.0) drivers the hard drive blinked on and off. Only with the signed USB 2.0 drivers could I use the USB hard drive, but slowly.
I supplied 5 VDC externally to the USB hard drive, and this didn't help.
My evidence was then based on two devices. A two year old Lexar Jump Drive (JDA256-00-501) and a one year old Western Digital Silver Passport (WD400UQ17-RTE). The Jump Drive did achieve high speed with my wife's Mac -- a little humiliating. Both devices did achieve high speed once a month ago with my T42 after I cleaned the sockets with air, perhaps introducing static. So I knew the devices were USB 2.0 capable, and I had a new system board on the T42.
The plot thickens. I bought a new memory stick -- Transcend TS2GJF168. And I bought the cheapest USB 2.0 cardbus I could find. NO brand -- it boasts "Design in Taiwan" on the rear. It uses an external 5 VDC supply.
Here are some benchmarks:

Using cardbus:
WD HD: 130 Mbit/s
Transcend: 65 Mbit/s
Lexar: 78 Mbit/s

Using T42 port:
WD HD: 7.8 Mbit/s
Transcend: 60 Mbit/s
Lexar: 7.7 Mbit/s

USB 1.1 transfer speed limit is 12 Mbit/s. So the cardbus achieves USB 2.0 speeds with all three devices, while the sad T42 port achieves USB 2.0 speed with just the Transcend memory stick. This shows that there's more to our T42 problem than USB 2.0 vs. no USB 2.0. There is a device dependency which might provide clues to the problem.
For the time being, I'm happy carrying my new Transcend stick when out with the T42 and using the big barnacle cardbus when at home backing up on the external WD drive. But I do want to get to the bottom of it.

[t41user]

i'm not one of the many super smart and savy people on this forum. nonetheless, as far as I can tell, the problem seems to be a design flaw that includes engineering matters related to the motherboard and the interfaces.

I do not see why persistence of the problem after replacing the motherboard (and restoring to factory image) should count as evidence that the problem is not with the motherboard/hardware design. If the problem is indeed a design flaw with the engineering of the motherboard and interfaces, a new motherboard could be expected to repeat the problem, even if the problem only occurs or is detected intermittently. Am I missing something? Replacing the motherboard does not change the design of the motherboard or the hardware integration of the motherboard and interfaces. So what's the mystery? Unfortunate design flaw, I thought the consensus was.

[agarza]

dcouzin, you much time your T42 has been with you? I own a T42p and use a lot my external E7K100 HDD via USB 2.0 and I can't even imagine working again with USB 1.1

Previous to this laptop I've been usin a T30, A22 and my 1st laptop a Compaq Presario 1685 which died suddenly . I never use preinstalled configurations shipped to the machines, I don't know if Power Manager would have anything to do in damaging the USB ports, but I hope I keep trouble free.

[dcouzin]

t41user, I agree that it looks like a design flaw relating to the motherboard and the interfaces, but against this hypothesis consider: (1) Why aren't the many thousands of other owners with the same motherboards having the problem? Are they having it and stupidly not knowing it? (2) Why is Lenovo shelling out money on replacing the flawed board with new flawed boards rather than revising the board? (3) Even if the motherboard is flawed or "screwy" there seem to be software workarounds such as described in http://forum.thinkpads.com/viewtopic.php?t=29057. Lenovo could then write new drivers instead of revising the board.

Where was the mentioned "consensus" was reached? Not in the 29057 strand.

I hoped by giving benchmark information to encourage other victims to find other USB 2.0 devices which function at high speed even on the flawed T42 ports. Transcend's Jet Flash 168 stick does. Are there perhaps external hard drives which do too?

[ed_vickery]

I sent my laptop back to IBM due to a seperate physical hardware problem and when they returned it the USB ran at 2.0 speed. The only clue to how they fixed it was an updated BIOS note on the paperwork that came back with the fixed machine.

I do know that my problem of USB 1.1 speed began when I reinstalled Windows and used a non IBM Windows CD. THis might have been the cause for losing my USB 2.0 speed. At the same time my Access IBM key stopped functioning.

[dcouzin]

Ed, can you share which BIOS version you now have, and can you save its settings to a file? Thanks.

[dcouzin]

Absolute evidence of hardware fault on my T42.
The upper USB port functions differently from the lower USB port!
When the WD hard drive is attached to the lower USB port it works, but at USB 1.1 speed. When it is attached to the upper port, it is unstable -- turning off and on -- its drive letter appearing and disappearing from Windows Explorer. Sometimes it stabilizes, generally with USB 1.1 speed, but very rarely with USB 2.0 speed. (The external hard drive has its own external 5 VDC power supply.)

[coreman]

Well I've got to jump into this thread since I originated the other thread...I've not replaced my motherboard yet but I am still convinced that it isn't solely a hardware problem. It could be a combination of the hardware and driver...I have also seen where my connections are variable, sometimes I get USB 1.1 speeds which later (and with different device) seem to run at USB 2.0 without doing the "driver dance" in between.

When I absolutely need the USB 2 however I do the "driver dance" and get the ports temporarily running at 2.0 speeds.

But it is encouraging that the USB 2 adapter worked fine for you so I am going to go that route, I just don't like changing motherboards out for a problem that I can workaround.

DCOUZIN, how did you measure your speeds? Time and known file size?

[dcouzin]

coreman, I hoped you'd comment, because your software hypothesis still tugs me. Whatever might be bad in the system board, since USB 2.0 works some of the time there must be a software workaround. But this might entail Lenovo writing new drivers.

My cheap cardbus has NEC chips and it works. Lenovo wants to replace my system board again. Are they crazy for waste? Or do they recycle their badly rebuilt boards just to run out our warranties? You might buy a USB 2.0 + Firewire cardbus so as to feel less defeated by the T42's failure.

I measure speeds using always the same .zip file of approximately 256 MB. I always write to the external drives. I assume the .zip file has a non-tangled structure. If not, the reported speeds are a little slow. File transfer isn't like clockwork. There are interfering events in a computer (and I confess I'm always online). But for this big transfer, times are consistent to within +/- 5%. (I once wrote code to measure computer speed and it suffered from large errors unless allowed to spend more time at the task.) When reporting transfer speed in Mbits/sec, I follow industry convention that Mbit refers to 1,000,000 bits, not 1,048,576 bits.

Thank you for launching http://forum.thinkpads.com/viewtopic.php?t=29057. "Driver dance" will someday be the name of a poem or a movie.

[t41user]

In my infinite ignorance, the fact that USB 2.0 works some of the time seems perfectly compatible with (if not a prediction of) the hypothesis that the problem is a hardware problem. The fact that people can get things to work by using different drivers also seems perfectly compatible with (if not a prediction of) the hypothesis that the problem is a hardware problem. Could people who disagree please explain more in the friendly spirit of communication and helping others learn?

[w0qj]

guys, what utility program can we use to check USB speed, v1.1 (slower) or USB v2.0 (faster)?

[Nick Y]

Absolute evidence of hardware fault on my T42.
The upper USB port functions differently from the lower USB port!

I recently bought a USB device (Pinnacle 500-USB with Studio 10 s/w) to convert my old VHS tapes to DVD. I found that I had problems with USB and checked the on-line forum and FAQs. It appears that many computers appear to have a problem between two USB ports and the recommendation was to use the one closest to the motherboard. Since adopting this approach I have not had any problems -I use the lower USB port on my T43. What causes the difference was not explained from what I recall.

[coreman]

In my infinite ignorance, the fact that USB 2.0 works some of the time seems perfectly compatible with (if not a prediction of) the hypothesis that the problem is a hardware problem. The fact that people can get things to work by using different drivers also seems perfectly compatible with (if not a prediction of) the hypothesis that the problem is a hardware problem. Could people who disagree please explain more in the friendly spirit of communication and helping others learn?

I'll take a stab at this...my T30 USB port went "intermittent", meaning that sometimes it would work and sometimes it wouldn't but "not working" meant exactly that, it wouldn't work at all.

Motherboard was replaced and (2) years later, still working fine. USB ports are a particular weak area of motherboards...

When USB 2.0 ports throttle down to 1.1, the hardware is still working...but it is the driver that controls the hardware and how it performs. As I understand, the drivers are written such that any software failures cause the ports to throttle down rather than quit working.

In my case, my throttle down problem was exactly timed to my upgrade to Win XP SP2 and IBM/Lenovo updates (detailed in other thread) so I know it isn't hardware problem.

Secondly I can always, 100% of the time, get the ports working doing the "driver dance" albeit temporarily (meaning it works for a couple of tasks).

Finally even if Win XP pops up with the ubiguitous "This device can perform faster if..." message, many times I will actually connect/transfer at ~ USB 2.0 speeds (similar to description of dcouzin)...what I haven't investigated is whether this could be USB device related (and it very well could be).

This sort of tells me the fault may be related to the amount of power required by the USB device, if it calls for too much...the port throttle downs...this is just a guess on my part (at this time) but since the drivers control how the ports operate I still think it is a poorly written driver for this product that is majority cause of problem.

As for the reason motherboards are replaced so often, I personally know an IBM hardware engineer who is familiar with the RMA procedure and he told me "90% of the motherboards that are replaced are fine, after going full QC inspection they are used for new RMA replacements". This is faster/cheaper than extensive troubleshooting of MBOs to find customer fault.

But this leaves, at least on the surface, ~10% of faults that are true MBO failures...as in my T30 example above.

After being in the industry for many years and having a lot of experience with notebook PC makers here in Taiwan, I can safely say that they produce very high quality products that simply don't fail in the numbers that seem to be reported. But because of warranties...it is just easier to dispose of a problem by replacing MBO and simply "moving on" to next customer complaint.

[mtroxel]

I've read these threads with interest as my T42 USB's worked flawlessly for the first year. For the last 4 months I've had erratic USB problems. But in my case, I would connect on either port, sometimes get the USB 1.1 warning, sometimes not. Even if I connected at 2.0 speeds, I would often hear the disconnect sound, then it would immediately reconnect with the 1.1 warning bubble. And when I saw the bubble, it was truely the slow 1.1 speeds. It's not hard to tell.

I did the hardware diagnostics on the ports and one of the two would fail, despite the fact it worked....sorta.

Mine is not SP2 related. Despite about 3 complete rebuilds, I have never been without SP2.

Motherboard was replaced and (2) years later, still working fine. USB ports are a particular weak area of motherboards...

DHL picked up my ThinkP Monday night at 7:30, it was back in my hands Wed AM at 10:30 with a new mobo (kudos for ThinkPad service). So far, no problems. And I backed up my hard drive image to an external hard drive, all at USB 2.0 speeds right after I got it back.

As for the reason motherboards are replaced so often, I personally know an IBM hardware engineer who is familiar with the RMA procedure and he told me "90% of the motherboards that are replaced are fine, after going full QC inspection they are used for new RMA replacements". This is faster/cheaper than extensive troubleshooting of MBOs to find customer fault.

From what I've seen so far, and the fact that a diagnotic showed a failure on mine, I am sure mine was bad. And I hope no one at Lenovo is reading this but I'll admit, there are times when I have a flash drive pugged in and I move the computer on my lap and an awful lot of leverage type weight gets put on the USB port as the flash drive rests against my leg before I catch it...quickly.

I also wonder, is there a chance some static charge from a flash drive zaps directly into the mobo?

[dcouzin]

mtroxel, please let us know which hardware diagnostic faulted your USB ports. PC-Doctor 5 doesn't fault mine, even though I have strong evidence of hardware caused problem. It would really help if we who struggle with this problem found a clear diagnostic.

[mtroxel]

ThinkPad forum has always helped me so I spent some time making sure I could answer your question. But….it looks like I’ve got some explaining to do.

It was through PC Doctor, and not even the one from the recovery partition but just through the Windows version. If you go there you see two USB devices. It was my second one that failed. Now I have a brand new mobo with two USB ports that work great……and USB device 2 failed again. Twice.

It’s the fingerprint processor. I always have it disabled. It’s about as reliable as the cable guy and I don’t want to waste power, so I always kill that, the IR port and the serial port too. When I enable the fingerprint processor, USB 2 passes.

So…..nevermind. Sorry guys.

[apblank]

I'm just posting here to add my experience to the thread, in the hopes of helping someone else out.

I have a T42 running Windows XP sp2 issued by my company which I use for work. I have a work issued Western Digital "My Book Essential Edition" (WD 2500JD) External USB 2.0 drive which I use for system backup. This drive has its own external power source.

I always got USB 2.0 throughput, and could back up my entire drive in reasonable time (2-3 hours). After a recent office move, I went to run a backup and I got both the popup warning about "this device could work better, etc." and USB 1.1 throughput, with an estimated 18 hours for complete backup.

Nothing had changed so I assumed hardware failure from the move (and from reading some of these threads). Then I realized one thing had changed. The old office had wired network, and the new office did not, so I was using my built in 802.11g wireless for networking.

Simply turning off the wireless radio in the Thinkpad, and unplugging and then replugging in the external hard drive got rid of the popup warning and gave me USB 2.0 speed (2 hours for a complete backup).

I have since repeated this test several times just to double check my results. With the wireless on, using this particular external drive, I get USB 1.1 only. With the wireless off, this drive gets USB 2.0. I have no other devices to try this with, so that's as much info as I could give.

It leaves me curious why the one device would conflict with the other, but I have a work around, so I'm happy. Good luck to others exploring this problem on their thinkpads.

Andy

[richk]

I am jumping into the middle of this and I am sorry if I am suggesting something that has already been suggested, but I work on lots of machines and it has been my experience that this kind of problem occurs when the motherboard doesn't put out enough current to drive the device. It is generally fixed if you use an external power supply for the device or run off of a powered hub. If the device has to little current, it won't respond fast enough and the motherboard USB controller assumes it is a slow device. If the draw is to high, you can also burn out the motherboard or the device.

[mtroxel]

That wasn't my problem. It happened with:

• 1. flash drives
  2. external hard drives with their own power supply

Both instances work fine now that I have a new motherboard.

[dcouzin]

Responding to apblank's post. I now have a new motherboard, and full USB 2.0 capability, and so can't check his fascinating example. What flaky USB 2.0 the T42 must have, whether working or not.
Incidentally, 2-3 hours is a long time for a backup, unless it's of a 200 GB hard drive. SelfImage freeware backs up at about 1.5 GB/min when imaging a whole partition to a whole partition.

[mtroxel]

SelfImage freeware backs up at about 1.5 GB/min when imaging a whole partition to a whole partition.

dcouzin I've used Acronis for years but know of that SelfImage. Have you used it much? I don't think has the ability to boot to a recovery disk and reimage that way does it?

[DaveMackenzie]

This USB 2.0 Problem annoys me quite a bit, i think its a combination of software and hardware conflicts, if i do a clean install of xp with the the drivers, i get roughly 1 minute of usb 2 speeds then it reverts back to usb 1, however, i think ive found one of the problems, if i use my IBM Rapid Restore external hdd, it runs for about 5 minutes then reverts back, but if i use my archos mini jukebox gx20 mp3 player, it stays the same, now the difference is, the archos has its own power supply, so i did further testing and found that devices that use the usb port for power aswell as data feedback crash and revert to usb 1,whilst anything with its own battery tends to stay unpeaved by the error

[dcouzin]

To mtroxel. I've made heavy use SelfImage 1.2 during the past two months. I always run it from a bootable CD called "BartPE". BartPE is a Windows CD made from your own operating system, hence quite legal. It has no trouble with NTFS volumes and runs USB at 2.0 speed, etc. SelfImage is available as a beta plugin for BartPE. It has always worked for me. SelfImage can image a whole drive to a whole drive, or a partition to a partion, obviously uncompressed, or either one to a compressed file and vice versa. If your partition is more than 1/3 full, it's quicker to work uncompressed with SelfImage. Imaging carried out from an external platform -- no bells and whistles -- is not difficult to program. So freeware should be perfectly fine.

[mtroxel]

Interesting. I carry Barts with me all the time for troubleshooting. It's particulary good if you want to go into an existing registry and edit from "outside" the operating system.

I gave up on Access Connections. I just could not solve my problem and the only advantage for me is that AC tends to drop connections less often than W Zero Config. It does make things more complicated though and I can't get around this new thing where it disables network sharing for every new connection.