Gamecollector wrote:

xjas wrote:

Gigabyte 8IPE1000-Gs which seem to be nice, solid & quick.

Until you plug somethin in a USB slot and get the burned southbridge. Maybe I'm unlicky but all my 8IPE1000 MBs (original, Pro's, -G's), have died this way.

WHAT were you plugging in? I've had two of those and never any trouble like that.
.
ASUS P4P and P4C mobos are notorious for USB faults.
Problem was found to be Asus not complying with Intel's recommendations in the chipset design guide.
.

Gamecollector wrote:

Until you plug somethin in a USB slot and get the burned southbridge. Maybe I'm unlicky but all my 8IPE1000 MBs (original, Pro's, -G's), have died this way.

Were you using the "floating" (un-grounded) case-front USB ports? I am aware of that problem.
I ran an earth wire back to the main chassis & I'm carful about hotplugging w/those. I'll report back if something bad happens, but it's been totally fine so far.

^ Same here, always run an earth wire to ground to chassis any 'floating on plastic' USB front ports.

These are all related.

PCBONEZ wrote:

Problem was found to be Asus not complying with Intel's recommendations in the chipset design guide..

xjas wrote:

Were you using the "floating" (un-grounded) case-front USB ports? I am aware of that problem.

TELVM wrote:

^ Same here, always run an earth wire to ground to chassis any 'floating on plastic' USB front ports.

.
Intel called for an EMI filter with ESD protection to be installed on the motherboard at the USB header or on USB near the chipset .
That is a small inexpensive circuit that only requires 5 or so SMD parts to pull off. (I will attach an example circuit though it's not specifically for USB.)
For a time motherboard manufacturers were 'cheaping-down' by not installing these parts.
[I pick on Asus for this because they didn't limit this 'cheaping down' to their low end models. Asus cheaped down their high end models this way too.]
Intel was not to happy about their chipsets being called defective when the real cause was cheap-skate implementations by mobo manufacturers but they beefed up the ESD protection internally in the chipsets anyway to cover their own arse. That helps but it's no where near as good as using discrete components.

In the example circuit the resistor may be an inductor instead. The caps are somewhere around 10pF (tiny SMD parts).

Then there is that case manufacturers cheap down on their front USB ports.
That ground wire you people (and I) add to those ports is actually required by the USB specification.
It should already be there.

If you have a cheaped down mobo -and- a case with a cheaped down USB ports there is a pretty good chance you'll kill the USB and/or your chipset.

.
----
While I'm on USB compliance.....
(You will see A LOT of misinformation about this one.)
I've come across a lot of USB gadgets that plug into USB headers that skimp and only use 8 pins.
That was 'legal' for USB1.x but to be truly compliant with USB2.x the 9th pin is required. It's not optional.
USB cables that are truly USB2.x compliant have a shield and a shield drain wire. The 9th pin is to properly ground the shield.
.
----
I've seen a lot of people blow out their USB ports by overloading them - usually with 2.5" SATA drive gadgets or chargers.
USB1.x and 2.x ports have a current limit of 500 mA (1/2 amp). - That's the 'official' number.
People got the idea with older 2.5" IDE drives that one port was enough for an external USB HDD but very few 2.5" Sata drives (or larger IDE drives) use less than 500 mA.
Yes - For USB powered USB HDDs you need one of those cables that uses two USB ports to not overload your ports.
For chargers - the people that build and sell these don't seem to know what a current limit is. Check the specs before you use one.
.

PCBONEZ wrote:

ASUS P4P and P4C mobos are notorious for USB faults.
Problem was found to be Asus not complying with Intel's recommendations in the chipset design guide.

Then again, my P4 rig has been running on P4P800-E and P4C800-E boards exclusively for 11 years, and this issue never happened.

I think we discussed it here not so long ago, and the conclusions were that (a) it's mostly an actual Intel chipset issue, and not really specific to any vendor, and (b), that cases with proper grounding of the front USB ports will not experience the issue.

I have been using the same case in my rig all this time (Thermaltake Lanmoto), so it could be a factor.

Anyways, what I like about the P4P800-E, is that it is one of first boards that feature 8-channel onboard audio, with a full set of jacks, if you don't feel like getting an audio card. Then again, the PCI cross-talk is quite bad, at least with the unit I had. I dropped the onboard audio for an Audigy 2 ZS.

dr_st wrote:

I think we discussed it here not so long ago, and the conclusions were that (a) it's mostly an actual Intel chipset issue, and not really specific to any vendor, and (b), that cases with proper grounding of the front USB ports will not experience the issue.

I suggest further reading then.
(a) Yes it affected more than one vender, Asus was just notorious for it. It was not a chipset issue it was a compliance (to design guidelines) issue.
If the mobo manufacturers had not ignored Intel's guidelines for ESD protection there would have never been an issue at all.
(b) Grounding the ports only helps when the static is only in the ground wires in the cable. While that's usually where it is, static can also get into the power wires of a cable and that ground is not going to help you at all in that case.
.

If you look at a properly done case front USB port there is more than just a ground.
There will be some small SMD parts. That's an example of one of the filters I've been talking about.

dr_st wrote:

Then again, my P4 rig has been running on P4P800-E and P4C800-E boards exclusively for 11 years, and this issue never happened.

What may factor is Intel altered the ICH5 to make it less susceptible. My P4P800SE rev2.00 is another that seems unaffected.

But it's worth being clear to the OP that this popular southbridge does have a heads-up on it. Wiki's entry is as good as any to have something to bookmark.
https://en.wikipedia.org/wiki/I/O_Controller_Hub#ICH5

Do wonder if any of the susceptible boards are still around. I'd expect the southbridges that can be fried, have been.

PCBONEZ wrote:

If the mobo manufacturers had not ignored Intel's guidelines for ESD protection there would have never been an issue at all.

It can be considered a design mistake. Motherboards makers had been cutting corners since forever, and relying on them and case manufacturers to provide ESD protection for something that is so easily accessed by the end user (and bound to be hotplugged) was a very unwise decision, and in the end they were the ones that took the blame for that and had to enhance the ESD protection of the chip.

alexanrs wrote:

PCBONEZ wrote:

If the mobo manufacturers had not ignored Intel's guidelines for ESD protection there would have never been an issue at all.

It can be considered a design mistake.

How is it a "design mistake" by the manufacturer (Intel) when they expressly told the buyers (mobo mfrs) in writing how to use the part and they don't do it that way.
.
This is like you putting vegetable oil in your engine rather than the manufacturer recommended oil and then trying to blame the manufacturer for a "design mistake" when your engine blows up.
.

Well, not necessarily.

There are what is called industry standards and practices. If you violate what is the common way to do things, and how people expect things to be done, then yes, you take at least partial blame, even if you explained it carefully enough.

For example, did other contemporary chips require the same level of ESD protection to not fry out?

A better analogy might be if you suddenly manufactured a car that does run on vegetable oil, wrote in the manual, and then blamed everyone who tried using regular machine oil as a force of habit, because they didn't read your manual.

Well, yes necessarily. Your argument is BS.

Intel specified how their part was to be used.
It make no difference what anyone else is doing or what is standard when you are given clear directions for THAT part and you don't follow them.

In your analogy a court would find the buyer as the responsible party. That's because that is what makes sense.

Guys, who cares? I really doubt the OP wanted an exhaustive argument as to who would have been at fault in court for a problem only slightly likely to pop up.

If you have a board that works then great. A few extra precautions will mitigate this particular problem 'well enough' if not completely. I would guess earthing the USB shield would take care of about 95% of the cases where this issue can occur. Or just ditch the front-mounted ports all together and use the back panel ones.

And if the southbridge pops - just get another! It's a P4 board, you can find them at the side of the road regularly these days.

FYI I got my stash of 8IPE1000-Gs (at least five or six boards) in a massive university lab dump-off. All machines that had been bought new and were in service for years in a dry, stuffy office environment with polyester carpets you could charge a cat on. Every single board fitted in a case with a cheap plastic bracket for the front USB ports and no earthing cable. Not a single popped southbridge in the lot.

PCBONEZ wrote:

How is it a "design mistake" by the manufacturer (Intel) when they expressly told the buyers (mobo mfrs) in writing how to use the part and they don't do it that way.
This is like you putting vegetable oil in your engine rather than the manufacturer recommended oil and then trying to blame the manufacturer for a "design mistake" when your engine blows up.

Oh, but the mistake is that Intel RELIED on motherboard makers to actually listen, with catastrophic consequences. No other chipset (Intel or not) needed extra protection, and they knew the likes of PCChips (or ASUS, or a bunch of other makers that cut corners during that time) would use their chipsets.
It is like going for a walk in a shady neighbourhood alone at night while carrying a bunch of your savings with a giant sign saying "DON'T MUG ME". You are not doing anything wrong in principle - you have the right to come and go as you please - but you knew the risks involved, you had a something to loose and, in the end, despite not being wrong the decision to go there carrying money was a mistake.

alexanrs wrote:

Oh, but the mistake is that Intel RELIED on motherboard makers to actually listen,

They were contractually obligated to 'listen' by way of their license agreement with Intel.
It was not 'extra' protection and EMI/ESD protection on ports IS the industry standard.

I'm not going to continue a conversation with someone who is evidently one of those people that tries to sue restaurants because the coffee is hot.
.

xjas wrote:

Guys, who cares? I really doubt the OP wanted an exhaustive argument as to who would have been at fault in court for a problem o […]
Show full quote

Guys, who cares? I really doubt the OP wanted an exhaustive argument as to who would have been at fault in court for a problem only slightly likely to pop up.

If you have a board that works then great. A few extra precautions will mitigate this particular problem 'well enough' if not completely. I would guess earthing the USB shield would take care of about 95% of the cases where this issue can occur. Or just ditch the front-mounted ports all together and use the back panel ones.

And if the southbridge pops - just get another! It's a P4 board, you can find them at the side of the road regularly these days.

FYI I got my stash of 8IPE1000-Gs (at least five or six boards) in a massive university lab dump-off. All machines that had been bought new and were in service for years in a dry, stuffy office environment with polyester carpets you could charge a cat on. Every single board fitted in a case with a cheap plastic bracket for the front USB ports and no earthing cable. Not a single popped southbridge in the lot.

Nice post and I tend to agree here.

To me this would be worth considering when using one of the affected boards, I couldn't care less who was to blame 😜

PCBONEZ wrote:

Well, yes necessarily. Your argument is BS. […]
Show full quote

Well, yes necessarily. Your argument is BS.

Intel specified how their part was to be used.
It make no difference what anyone else is doing or what is standard when you are given clear directions for THAT part and you don't follow them.

In your analogy a court would find the buyer as the responsible party. That's because that is what makes sense.

Yeah, I got it. I am full of BS. 😀

I see you know a lot about engineering, fair enough. You obviously don't know jack about law and how things work in court. I wish to you that you never ever have to deal with courts or the legal system. You may find yourself unpleasantly surprised.

Go troll somewhere else. 😀

Yes, I'm 100% sure all defected MBs already fried in the past 10+ years. So - now we must worry aboud bad caps and slots/jumpers corrosion.

P.S. About the CPU pins trouble. I made a tool from the upper part of a socket478. Have trimmed the 25x6 pins part so I can strain legs one-by-one or just check 1 side of the CPU.

Gamecollector wrote:

Yes, I'm 100% sure all defected MBs already fried in the past 10+ years.

Not that I can fix THAT, but I still see them coming in for repair.
I actually have two of the Asus ICH5 problem children in my "in box" right now.
.
ATM I'm winding down to fully retire (from this) but for the last decade or so I've repaired 500-600 boards a year.
When you see that may boards it's easier to see when a problem is wide-spread vs an isolated incident.
.