ICH4/ICH5 post Sept change cause 865 high return rate in China

[lsman]

In Chinese:
http://market.intozgc.com/news...-47294-1833054237.html

also post in VIA's chinese site:
http://www.viatech.com.tw/big5/news/news.jsp

Note:
-Reports in China store that 865PE has high return rate (over 10%) on the big 3 mb (asus, MSI, GA)
- many are SB burned when using USB device.
-Rumor:
1.Intel change internal design of SB ICH4/ICH5 Sept that leave off some volt protection circuit AND do not inform mb co.
2. MB manf. use old design without any knowing.
3. user in China with cheap/ not up to standard USB device / case front port that do not ground property plug in and overheat the SB.
4. high return rate.

 

[lsman]

continue to read forum at asia....

some ppl plug in USB external hd, or USB connector to phone..... dead mb immediately (has to exchange mb) on their Gigabyte 8IG1000.

I wonder if any bad USB stuff go oversea. and how many will match with the ICH5...

 

[VirtualLarry]

I wonder if Intel chose to leave off the OC (over-current) protection signal on their USB host ports? I know that many people get confused or otherwise have problems, wiring up cases with front-panel USB connectors, some use that pin as a mechanical guide, others use it as not-connected, chassis-ground, and some actually do use it for overcurrent. I'm not an EE, so I don't have the details, but seems like a possible explaination.

 

[lsman]

(In chinese)
http://www.intozgc.com/news/10...7497-1254137069_1.html


Updated as MSI saying the problem lay on the design (lack of ESD protection) of front USB port of case....

Some cheap case do not have those design (intel ask for and there is suggestion but since case front port was made by some lower tech company they may simple ignore it, BTW, VIA has build in protection on chipset)
e.g. At least the port has to be on a PCB. (and with ESD protection build on the PCB) some case just use a direct wire connection (to skip cost)

So better check your case design for front port connection....


story about MSI test and re-test , show the mb will burn sb when a USB drive plug in the case with problem front USB port
[again in chinese]
http://market.intozgc.com/news...-47576-2233337976.html

 

[lsman]

So the problem does not lie on Post Sept production.
If you have those bad design case and use a USB 2.0 device... it will be a timing bomb.
But post-sept production ICH5 chip even raise the rate...

MSI suggest ppl use it will care (ground yourself before unplug, touching the medal part of usb connector when unplug,
do not buy cheap usb wire, product that do not have PLL label, and make sure the long time usb product like printer is off before unplug)

Because its not just intel problem (lack of instruction and design missing), its part of mb manf. problem (need to replace high amount of RMA mb, and need to add additional circuit ... but it is near the end of line product....) and most important the vast amount (and hard to find) case manf. and their supplier.

Who will pay the bill? or if there will be any recall? Time will tells.

 

[lsman]

Originally posted by: VirtualLarry
I wonder if Intel chose to leave off the OC (over-current) protection signal on their USB host ports? I know that many people get confused or otherwise have problems, wiring up cases with front-panel USB connectors, some use that pin as a mechanical guide, others use it as not-connected, chassis-ground, and some actually do use it for overcurrent. I'm not an EE, so I don't have the details, but seems like a possible explaination.

Exactly, the problem is the usb connector do not ground and then match up with some low quality USB device that also do not ground the outside (the outside metal of USB). and it burn the south bridge.

(in the second link above, MSI test and re-test and give this example)