Rebuttal to Dell's reply regarding 630i issues

by Gregory Donner
September 5, 2008
(updated: October 21, 2008)

This is in response to Dell's XPS 630i Responses post to My630i.com on August 20th, 2008. Dell's responses are in blue; my replies are immediately below them.


(1) Customer statement - No LightFX 2.0 dancing lightshow software: LightFX 2.0 is not the same as the older LightFX 1.0. LightFX 2.0 was not designed to have a dancing lightshow. It was only designed to change the colored LEDs using profiles. We have removed the misleading references to dancing lightshow LEDs on our sales pages. If you are unhappy with this feature set, Dell will provide a full refund [expired October 15, 2008]. To do so, please contact DELL-Chris_M via a private message on the Dell Community Forum.

How--and why--does false advertising become a "feature set" at Dell? Dell failed, and a refund neither corrects, nor makes amends to the customer for the false advertising of this lightshow software. The sales page made reference to this for far too long for it to have been an 'oversight' by Dell's marketing department. I doubt that Dell would ever have removed the false advertising had their customers not brought it to light.

It seems to me that Dell has a wonderful opportunity to make amends and prove just how well their Master I/O Board's ESA architecture works by adapting their existing LightFX software for the 630i. According to Dell's own LightFX development page, it should be a piece of cake...or are there bigger issues with the 630i that Dell is trying to hide?


(2) Customer statement - Dell modified the Nvidia 650i SLI chipset/motherboard to restrict it: Dell's products are unique and manufacturers build the motherboards to our specifications. The retail motherboard specifications frequently do not apply to the Dell proprietary product. There are many reasons Dell does this, but our goal is provide a reliable product to our customers and to provide to you the ability of having two graphics cards. When building the XPS 630, Nvidia gave us two 650i chipset motherboard choices:
* 16,1,1 = 1 PCI Express x16 slot (16 lanes), 1 PCI Express x8 slot (1 lane), 1 PCI Express x1 slot (1 lane)
* 8,8,1,1 = 2 PCI Express x16 slot (8 lanes), 1 PCI Express x8 slot (1 lane), 1 PCI Express x1 slot (1 lane)
We chose the second one to enable you to use either Nvidia SLI or ATI CrossFireX. The XPS 630i Owner's Manual clearly states the PCI Express lane configuration is locked in as follows:
- PCI Express x16 (SLOT1 and SLOT4) = connector size 164 pins, connector data width maximum 8 PCI Express lanes
- PCI Express x8 (SLOT3) = connector size 98 pins, connector data width maximum 1 PCI Express lane
- PCI Express x1 (SLOT3) = connector size 36 pins, connector data width maximum 1 PCI Express lane
http://support.dell.com/support/edocs/systems/xps630i/en/OM/HTML/specs.htm#wp1104354

If you are unhappy with this feature set, Dell will provide a full refund [expired October 15, 2008]. To do so, please contact DELL-Chris_M via a private message on the Dell Community Forum.

Dell failed. The sales page made reference to this for far too long for it to have been an 'oversight' by Dell's marketing department. What you advertise and sell to your customers must be accurate and in no way misleading. This was not the case with the 630i, and a refund does not resolve the false/misleading advertising of the 630i's technical specifications that customers understood and expected when they purchased the system.


(3) Customer statement - The motherboard is not ESA-certified: Nvidia's specifications clearly identify the 650i SLI chipset as ESA supported. Additionally, the Nvidia ESA Certified Products page specifically identifies the XPS 630i.

This is in direct conflict with what Dell Engineeer Patrick Dubois stated on June 12, 2008, via an interview on crave.cnet.com. He made it very clear that it is the Master I/O Board that holds the ESA architecture--not the 650i SLI motherboard:

"The ESA architecture is actually a separate control board. It's a separate control board that controls fans, lighting, all of the value-added features that we added in our boxes. It's a controller with firmware, its own microcontroller that controls all of those things. But it's physically a separate board, if you open the 630 or a 730, you'll see a separate board, which is actually an ESA board and it holds the ESA architecture."

What part of the statement above identifies the motherboard as being part of the 630i's ESA architecture?

Nvidia modified their specifications page only after being approached by Dell to change it, and after users pointed out the obvious problems. I wonder why this was? The Nvidia ESA Certified Products page specifically identifies the XPS 630i system (meaning only the Master I/O Board) as ESA-certified; not the motherboard.


(4) Customer statement - No SLI-Ready memory support: Nvidia does not officially support EPP with the 650i chipset, but it does support overclocking to EPP speeds. Again, this is a difference between the retail specifications and our proprietary specifications. The XPS 630i will support 2GB of 800MHz Corsair Dominator Dual-Channel DDR2 SDRAM overclockable to 1066MHz. 4GB of SLI memory does not function 100% reliably because of a limitation with the 650i chipset. Because of this, Dell does not ship the XPS 630i with overclocking enabled, but customers may turn on that feature. If you are unhappy with this feature set, Dell will provide a full refund [expired October 15, 2008]. To do so, please contact DELL-Chris_M via a private message on the Dell Community Forum.

Dell failed. Regardless of whether or not Dell was aware of the 650i chipset limitation with 4GB of SLI RAM before they began selling 630i systems does not excuse them from providing their customers with a proper fix for this issue. A refund is not a fix. Furthermore, why did Dell take so long before informing their customers of this issue, and why hasn't Dell disclosed the nature of the potential problems stemming from this bug? If they are aware of the limitation, they must surely also be aware of the problems it causes.


(5) Customer statement - CrossFireX compatibility restrictions: The Dell motherboard supports ATI CrossFireX using both our drivers and the retail drivers. But, to function properly, the Dell drivers must be completely uninstalled before loading the retail drivers.

CrossFireX configurations only work when the graphics card is one of the following two that Dell has specially modified by way of flashing the card's BIOS before shipment to the customer:

Users cannot use any other after-market ATI graphics card in this configuration, forcing them to buy their graphics cards at higher mark-up prices from Dell. The exception to this are single graphics card users. It is unknown what (or if) there are any compatibility issues Dell's proprietary drivers have on either a single or CrossFireX configuration.


(6) Customer statement – Constant hard disk drive LED activity: The operating system performs regular maintenance on the drives at periodic intervals: indexing service, constant scandisk, rolling defrag, AV, memory paging, write cache commit, polling of the HDD and optical drives, antivirus software, etc. These functions are occurring on all desktops. However, the hard drive LED on the XPS 630i is brighter than on any other Dell desktop PC, so the normal activity is more noticeable. We are investigating whether we can add a setting into the PC BIOS to turn the HDD LED off. Would you be interested in this?

I can accomplish the same thing with a pair of wire cutters or a piece of duct tape. How would this provide useful--and expected--information about hard disk drive activity to the customer? I expect a hard disk drive LED to accurately and properly inform me of disk activity like it is supposed to. My old Dimension 2400 and 4700 did this just fine. The LED needs replacement, and Dell is responsible to fix it at their expense--the 'official' fix makes no improvement whatsoever.


Note: Item seven was apparently due to a bug in HWMonitor and has been removed. The two negative temperatures were removed from HWMonitor as of v1.11.


(8) Customer statement - Missing temperature sensor after BIOS 1.0.3: BIOS 1.0.3 and later does hide a temperature sensor reading reported by NVMonitor. This reading is used only by the BIOS and does not need to be monitored by the user or NVMonitor. There is no overheating condition. The processor and chipset have built in overtemp protection and will shut down the system if there is a temperature problem.

Why is this sensor being used by the BIOS, and how is it providing accurate data to the BIOS if it is reporting such abnormal temperatures? If this is only used by the BIOS, why was it ever used by Nvidia in the first place to report the CPU temperature as a feature of their ESA software? Did Nvidia somehow overlook that 'small' detail? Are overheating conditions required before the user is allowed to monitor sensors? Nvidia GPUs can also shut themselves down if they nearing overheating conditions (around 120°C), but they still have an accurate--and user-accessible--temperature reading.


(9) Customer statement – Confusing/misleading advertising of optional audio cards: The sales websites either list the XtremeMusic or the XtremeGamer sound cards. Online configuration listings are not always the same between countries. The XtremeMusic and XtremeGamer have the same X-Fi audio processor. These are both X-Fi audio cards of the same quality.

Apparently they are (unlike the XtremeAudio card), but you won't know this unless you spend some significant time doing online research--it took me quite a while to find out what the differences were back when I owned an XtremeGamer card. Obviously the confusion has not helped Dell's customers.

Incidentally, where are the missing responses to Section F - "Shoddy technical support from XPS Desktop staff", and Section G - "Contradictions, hypocrisy, and other failures" of the Issues page? Does Dell have any plans to address them?