Anyone who has fiddled a bit with USB devices recognizes the infamous Windows pop up message "This device can perform faster". This message is a result of two USB communication speeds, known as USB1.1 and USB2.0. Great! But heck, it pops up on two UBS2.0 DELL latitude D610 laptops and one USB2.0 workstation. On top of that on a LINUX UBUNTU installation in a different format. Time for a bit of digging where "This device can perform faster" Windows message comes from.
But my laptop has USB 2.0 ports!
One USB 2.0 Hi-Speed card reader produced the "This device can perform faster" pop up message from day 1. A quick test with various other USB 2.0 mass storage devices (USB stick, USB hard drive) and USB printers DO NOT produce "This device can perform faster" pop up message. So the USB drivers and USB hardware seem to be very much OK.
| This device can perform faster - USB 2.0 Hell
In a hurry? Go directly to USB 2.0 Hi-Speed findings
Another slow USB 1.1 device
Next a Tiptel 150 USB connected VOIP phone caused the same Windows XP behavior, right from start. Although the consequence here is less, the phone works ok, the slower USB1.1 speed is enough for calling over the internet. The USB 2.0 card reader however operates very slow, copying 2Gb memory cards took ages.
Now what is wrong here? The Dell Latitude D610 has USB 2.0 ports as the other mass storage devices already indicate. The generic USB 2.0 capable drivers from windows XP pro operate as they should. It seems like there's an incompatibility problem between USB devices on a lower level than USB driver software. So at this stage installing different USB drivers doesn't seem to make any sense.
|Cheap All-in-One Cardreader advertised as USB2.0||Tiptel 150 USB VOIP phone|
The D610 Latitude is equipped with a Intel 82801 FB/FBM USB2 Enhanced Host Controller. Very common USB controller hardware found in many laptops and desktops. Through "Properties" of "My Computer" the "Manage" menu item shows if a Windows system has USB2.0 Hi-Speed capabilities:
The "Enhanced" part in the name indicated USB 2.0 specifications.
Opening the USB 2.0 Hi-Speed card reader reveals a single chip - the AU6371 USB 2.0 Single LUN Multiple Flash Card Reader Controller from Alcor Micro Corporation located in Taiwan. Surprisingly the controller chip range is mentioned at the website, however the AU6371 is not available anymore. The newer AU6372 - the 9-in-1 USB flash card reader chip - now has the spotlight and is released in October 2003. The AU6371 most likely is released before 2003. Also the Intel 82801 USB2 Enhanced Host Controller stems from this era. The USB 2.0 Hi-Speed card reader is bought in Dubai in 2008, so the AU6371 is still in stock and sold as recent USB hardware.
The Alcor Micro website doesn't host any drivers for the USB controller chip so generic Windows XP drivers should do the job. But it doesn't, the card reader operates at slow USB 1.1 speeds.
Opening the Tiptel 150 VOIP phone reveals another brainy controller chip. The CY7C64713 is a EZ-USB FX1™ USB Microcontroller Full-speed USB Peripheral Controller made by Cypress Semiconductor Corporation. A quick glance through the chip specification results in an interesting statement: "FX1 operates at one of the three rates defined in the USB Specification Revision 2.0, dated April 27, 2000: Full speed, with a signaling bit rate of 12 Mbps". Well that's basically USB 1.1 speed! That's one problem solved - the VOIP phone is not designed for USB 2.0 Hi-Speed use. The Windows pop-up message "This device can perform faster" is therefor wrong and misleading; The Tiptel 150 USB VOIP phone can not operate at high USB speeds.
It get's even crazier in Windows XP USB High Speed HELL!
Now with the Tiptel 150 VOIP phone, clearly identified as USB 1.1 because it's USB controller (CY7C64713) can't go any faster, let's plug it into a USB 1.1 port on an ole DELL Optiplex with Windows XP Pro (SP1). Heck! The USB1.1 VOIP phone is AGAIN classified as High Speed USB2.0. At least Windows XP gets the controller right, it's a non-Hi-Speed USB Hub - know as the Intel 82371 AB/EB PCI to USB universal host controller. What a mess - no wonder the USB driver (re-)installation orgy found on the Internet!
Note: SP1 has no USB2.0 support by default. USB 2.0 Hi-Speed drivers have to be installed manually, starting with the USB host controller.
AU6371 USB 2.0 Controller chip malfunction ?
Now for the USB 2.0 Hi-Speed card reader - somehow USB 2.0 High Speed - capability identification goes sour between the Intel 82801 USB2 Enhanced Host Controller and the AU6371 USB 2.0 Controller chip. The Intel 82801 already proved it can operate at USB 2.0 speeds, so what about the AU6371 Controller chip? A quick test in a Pentium 4 - USB2.0 workstation with Windows XP Home produces the same message, the USB 2.0 Hi-Speed card reader is detected as a USB1.1 device. The Pentium 4 work station has the same chipset, the Intel 82801 USB2 Enhanced Host Controller.
At the same time a different 6 in 1 USB 2.0 card reader connects without a problem. This card reader has a Integrated Circuit Solution, Inc. (ICSI) IC1110 USB 2.0 Flash Card Reader controller inside. (ICSI is also known as Integrated Silicon Solutions, Inc - ISSI)
With these strange findings, let's see how a LINUX install copes with the USB 2.0 Hi-Speed card reader on a DELL Latitude D610 laptop.
The UBUNTU website serves a neat LINUX install resulting in a bootable UBUNTU 9.04 CD within 2,5 hours using NERO Burning ROM. This free LINUX version is installed as a multi-boot partition next to a Windows XP SP3 install - all straight out of the UBUNTU box. With UBUNTU up and running, the USB 2.0 test with the USB 2.0 Hi-Speed card reader can commence. And it connects smoothly, silently popping up as a USB disk device. UNBUNTU doesnt show if it's USB2.0 directly. So a 256 Mb Memory Card from a Sony T1 is installed for a quick copy test. 71 Mb of pictures are copied at a speed of approximatly 1012 KByte/sec. Hmmmm, sounds like that's USB 1.1 again. To know for sure, /var/log/messages was opened and it contained the following log entries:
- usb 4-1 new full speed USB device using uhci_hcd and address 3
- usb 4-1 not running at top speed; connect to a high speed usb
Looks similar to the Windows XP Pro USB mess. The log entries with "uhci_hcd" and "full speed" is similar to USB 1.1. It should mention "ehci-hcd" and "high speed". So with a totally different OS on the same hardware, the USB 2.0 Hi-Speed card reader still connects as a slow USB1.1 device.
Connecting a LaCie 160 Gb USB2.0 disk with UBUNTU 9.04 running on a DELL D610 Latitude laptop immediately results in a Hi-Speed USB2.0 link. No need for manual "sudo modprobe ehci_hcd" commands because UBUNTU 9.04 automatically loads the right USB module. Now /var/log/messages looks like this;
- USB 1-5 new high speed USB device using ehci_hcd and address 3
- USB 1-5 configuration #1 chosen from 1 choice
- scsi 3 : SCSI emulation for USB Mass Storage devices
- scsi 3:0:0:0: Direct-Access FUJITSU MHW2160BH PQ: 0 ANSI: 2 CCS
NOTE:: Trying to manually load the ehci_hcd module using modprobe results in "FATAL: module ehci_hcd not found". With the above results, this seems to be normal.
Chances are bleak that the USB 2.0 Hi-Speed Memory Card Reader will ever perform at High USB Speeds.
So far it seems to be a hardware issue with the Memory Card Reader where Windows XP Pro adds even more confusion.
Now back to Windows XP Pro again. A much repeated remedy to get USB 2.0 devices operate at USB High speed is to uninstall all USB devices in the Windows Device Manager and restart. The reasoning behind it is; Service Pack 1 has no USB 2.0 support. So for all those who migrated their XP installation from SP1 to SP2, already registered USB 2.0 devices continue to operate on USB 1.1 speeds. Sound legit. So lets give that one a try with SP3 installed. So open the Computer Management console (My Computer - right click - Manage), select the device manager and start uninstalling all USB host controllers. Pretty simple - they're gone in a few minutes. Restart and Windows XP Pro detects the USB hardware right away.
Now it got really really really weird with re-installed USB device drivers. Windows XP Pro SP3 reports the USB 2.0 Hi-Speed card reader as a High Speed USB 2.0 device and the Host controller as USB 1.1. With SP2 it was the other way around. So what's the verdict with the USB 1.1 VOIP phone? Wow!! it's also reported as a USB2.0 device. What a total mess - SP3 should be an improvement?
The list is long of devices that do work at High Speed WITHOUT re-installing USB drivers. Too give a few examples of much used USB Hi-Speed devices:
- LaCie 160 Gb USB disk (Design By Porsche) with JM20339 USB2.0 to SATA Bridge
- NoName 120 Gb USB disk with M110B USB 2.0 To IDE Bridge (PATA)
- HP Deskjet F300 series with USB2.0 interface with unknown USB controller
And those USB 2.0 devices work flawlessly from day 1 on USB 2.0 speeds with:
- DELL Latitude D610 with XP Pro SP2
- DELL Latitude D610 with XP Pro SP3
- DELL Optiplex Gx110 with Xp Pro SP1
- NoName P4 with XP Home
- DELL Latitude D610 with UBUNTU 9.04 (Linux variant)
All have been tested with the above mentioned USB2.0 memory card reader and all OS configurations switch to USB1.1. Only Windows XP add's more confusion to the already hard to diagnose USB 2.0 Hi-Speed mess. The correct USB message here would be: "USB non-Hi-Speed device connected to USB Hi-Speed Hub".
Looking closer at the USB 2.0 High Speed - capability identification shows a rather simple mechanism and is known in the USB industry as "chirping". At connect the USB device receives "USB RESET" command from the host controller. If the USB device has high speed capabilities it answers the USB RESET by raising one of the communication lines (D-) to +5Volts. In turn the host controller "chirps" (raising D- and D+ lines alternating) indicating it's own USB 2.0 High speed capabilities. From here communication continues at USB 2.0 speeds.
Now for the USB 2.0 Hi-speed card reader, where does the USB Hi-speed identification go sour? A few possibilities, assuming the host controller initiates the USB reset;
- The USB device doesn't respond according to USB specification, because;
- The USB reset is not received correctly (voltage & timing)
- The USB reset response is incorrect (voltage & timing)
- The USB device does respond within USB specifications, but
- the host controller doesn't interpret the response as correct (timing) and stays at USB1.1.
- the USB device doesn't see the "chirping" correctly and stays at USB1.1 speed.
Most likely - in this case, it's scenario 1b where the timing of the D- signal is off mark and the USB host controller continues at USB 1.1 (Full speed iso Hi-Speed).
In this case changing Windows driver software doesn't help for one bit. The problem lies at chip level, not at the Windows driver level. For those with USB devices with on-board firmware in (flash) memory, there is a chance changing USB device driver software might help. Controller chip dependent USB2.0 detection software, following signal and timing specification in the graph above, requires good programming usually implemented on small real-time OS kernels. This is specialist embedded software programming where timing is crucial. For PC based host controllers the right driver does the trick. And as shown in this case, compatible USB 2.0 controller chipsets can make a big difference.
Not surprisingly, there's USB2.0 High Speed certification. Those USB integrators / manufacturers who pass the USB-IF Compliance Testing Program may carry the USB Hi-speed certificate on their products. The cheap USB2.0 card reader mentioned above did not have the USB Hi-speed certificate - at least not on the product package. But as usual, logo's can be copied easily and can appear on products not certified at all. For those USB devices where high speed communication is crucial, such as a memory card reader, it might safe a bit of time and frustration to have a device with a USB 2.0 compliant USB controller inside.
Now, just for the heck of it, did the AU6371 chip from Alcor Micro Corporation pass the USB-IF Compliance Testing Program? Yes it did, the AU6371 chip is listed in the the USB Hi-Speed compliance database, however with revision number A41. The USB Hi-Speed compliance entry is created at 6 august 2006. According to the Alcor website, the AU6371 is released in October 2003. It could be that early releases of the AU6371 controller didn't function properly and are still available in the "wild".
The UBUNTU linux installation lists the Memory Card Reader with serial number Multi_Flash_Reader_058F001111B1-0;0 with firmware version 1.0. The latter version sounds very virgin...
Tired of getting the Windows USB messages?
With SP3, it's possible to get rid of the infamous USB messages. Apparently some-one at Microsoft agreed the slow USB speed reminders are not always so much fun to look at.
Don't try to find this "feature" in Windows XP versions prior to SP3.
Based on the findings above, the USB 2.0 Hi-Speed stuff that really sticks out;
|1) Not all USB 2.0 devices are really USB2.0 compliant,
in this case the AU6371.
2) Windows USB messages are misleading, specially
in the case of 1)
3) Windows SP1, SP2 and SP3 have different USB
Hi-Speed messaging behaviors
4) UBUNTU 9.04 on the same hardware also
logs misleading USB info in case of 1).
|High Power WIFI adapter - USB connected
No wonder the USB driver (re)-installation orgy found on the internet. When USB messages can not be trusted in combination with the absence of decent USB error reporting it takes a lot of cross-PC / OS testing to make any sense of this USB 2.0 Hi-Speed Hell.
To get your system communicating at USB2.0 hi-speed the following can be done:
|- Check USB 2.0 hardware in your computer
- Check USB 2.0 drivers for your OS (Windows, Linux)
- Check USB2.0 device if it communicate at USB2.0 Hi-speed
- Check USB2.0 certification of the USB2.0 device
As in the above example, USB devices branded as USB2.0 can fail to communicate at USB2.0.
This is decided within 10 ms (mili-seconds) using the USB 2.0 detection mechanism.
A new USB standard : USB 3.0 SuperSpeed
In 2010 USB 3.0 will make a debut, read how to use USB 3.0 with your laptop or notebook.