[XFastest] Puts different boards through VRM heat test, which will burn first?

[Raja@ASUS]

Not enough capacitance on either side of the fet could allow for out of spec operation dependent on the implementation method and BIOS settings. As was stated earlier though, the capacitance issue is more of a general issue and was not meant to be the "end all be all" statement nor to point to it being the largest of the issues or the cause of the fets entering cascade, only that due to the implementation methods used it is an issue.

You monitor each fet separately? I highly doubt that.

Do you even know how the monitoring functions of the drivers and fets work? If one fet enters a cascade scenario there is almost no method of reliably saving it from damage, let alone outright failure.

I think I know my way around a VRM solution well enough to make a statement. And I certainly don't think you are in a position to be telling me how things work nor do I need educating on cascade scenarios

[sin0822]

I think everyone knows me and Raja don't always see eye to eye, but what he said about switching frequency is very true and that is the reason i quoted him.

Lol good discussion, no body has mentioned digital PWm algorithms like ATA? which can turn off all pulses at a load release or increase the frequency of the pulses when there is a huge load step. I would think turn off all the pulses at the same time would greatly help with overshoot. That is an example of a Digital PWm technology.

This has nothing to do with the VRM heat thingy, I was just talking about how it would be hard to compare VRMs based solely on temperature alone, a lot of boards would off the bat run at higher switching frequencies opening room for people to say no, they have to be equal.

[Raja@ASUS]

I would not have responded to be honest Steve but post 8 was very misinformed.

[STEvil]

I think I know my way around a VRM solution well enough to make a statement. And I certainly don't think you are in a position to be telling me how things work nor do I need educating on cascade scenarios

So you are monitoring each fet and driver separately, or are you not?

If you want to banter on the less than two digits of temperature change due to switching efficiency on an average system (read as stock to mild overclock), then have fun.

[Raja@ASUS]

So you are monitoring each fet and driver separately, or are you not?

If you want to banter on the less than two digits of temperature change due to switching efficiency on an average system (read as stock to mild overclock), then have fun.

That information is publically available in the datasheet, which I am sure you will have read by now but probably misunderstood. Take a look and then think about how a cascade event would occur with the fault protection in place (that's how we have it implemented).

[Kayin]

After my experiences with ASUS boards (11 of 12 dead in six years, including two workstation boards that arrived DOA and they refused to even talk about support) I'm rather unsure that ASUS has an idea as to what to do. But, if ASUS reps speak that way to people on a forum, then I'm pretty sure I'm not interested in their products even if I felt like giving them a second chance.

[sin0822]

So you are monitoring each fet and driver separately, or are you not?

If you want to banter on the less than two digits of temperature change due to switching efficiency on an average system (read as stock to mild overclock), then have fun.

I think I just now realized what you were trying to say, that the minor PWM adjustments available wouldn't cause too much temperature change, right? With my point i was trying to make it was that it is hard to get different boards on the same level, when there are so many aspects to the equalized testing that readers could complain about. Switching frequency is widely known to increase temperatures, but yes that amount could be small or it could be big. On most boards now, at least the Digital PWM ones, you are able to greatly influence the switching frequency across a large range and that might pose to be an issue when comparing boards, especially analog vs digital.

[STEvil]

Sin - yes.

Raja - So you did what everyone else did, a thermal probe on the PCB or against the VRM heatsink and a current divider for sensing, and called it good. I noticed you caught some flak over that in the ROG forums..

edit

Wait, is there even a sensor? http://www.xtremesystems.org/forums/...=1#post5024295

[StAndrew]

After my experiences with ASUS boards (11 of 12 dead in six years, including two workstation boards that arrived DOA and they refused to even talk about support) I'm rather unsure that ASUS has an idea as to what to do. But, if ASUS reps speak that way to people on a forum, then I'm pretty sure I'm not interested in their products even if I felt like giving them a second chance.

11 out of 12 . Sounds some user error to me .

[Kayin]

Hmm. Let's see, so two DOA, so I'm down to nine. A8N-SLI Deluxe 1, died with mosfet failure, stock clocks. RMA replacement for the board blew up at the first PCI-E slot, killing my 6600GT in that slot and my RAM as well. Replacement two went in my little brother's system, lived at stock clocks until it too crisped its VRM area, taking PSU, CPU and RAM. Six to go, so let's see. P5ND32-SLI Deluxe simply never could overclock at all. Also didn't like cards in slot 1 after a month or two. A8N32-SLi Deluxe was the same, started throwing Memtest errors on sticks that passed in other boards. Its replacement was the same. For all of those boards the stock cooling wasn't touched. So no improper remounting or anything. Two P5Bs decided to have boot issues, flashing didn't help. The P7P55 I had when I first got my i7 decided it wouldn't run with my Ballistix which are still memtest fine now. And the one that's living has a dead PCI-E slot.

I think that somebody ought to go back to the drawing board. Especially when you buy a $400 WS board that arrives DOA, and they refuse to assist you. I have plenty of other boards that haven't died, and they happen to be other brands. So I don't think it's user error-I'm more inclined to believe it's substandard parts. It's one reason I moved to workstation/server boards-they're simply built better than desktop equipment. I'm quite a big fan of Intel boards-all the ones I had were unkillable for the most part-as in it took a direct lightning strike on the house to kill my BadAxe2. I've used a LOT of Gigabyte and MSI, and barring mechanical damage I've lost two MSI total and no Gigabyte. I don't really want to spend my tie chasing motherboard bugs, so I moved to manufacturers that give me less trouble. That's my experience, YMMV. It's more the support I've seen than the failures even-seeing a rep be rude on a forum where they're supposed to be here to help just reinforces the idea that I have better places to spend my money.

My i7 is on a Big Bang Fuzion, my media center is a Zotac, my workstation is an Intel board, and my server is an Arima. And they're all fine. My wife's board is a Crosshair IV and like I said, it has PCI-E 1 dead or flaky. She'd rather run it like it is than fight ASUS RMA. I don't blame her. But, as stated earlier, this is an opinion, and you're entitled to yours, but I feel that I have based mine on sufficient evidence.

[Raja@ASUS]

Sin - yes.

Raja - So you did what everyone else did, a thermal probe on the PCB or against the VRM heatsink and a current divider for sensing, and called it good. I noticed you caught some flak over that in the ROG forums..

edit

Wait, is there even a sensor? http://www.xtremesystems.org/forums/...=1#post5024295

That question is referring to an external temp monitor or linkage to the Super IO that reports temps to an end user - has nothing to do with the VRM sensing mechanism. The board will throttle past a certain point irrespective of end-user settings (the way it's implemented and the thresholds that are observed by the thermistor or current balancing are the critical factors). The rest is down to what kind of features each vendor's buck controllers support for the failsafe mechanisms to work and how they have gone about insuring that the hard-set/programmable thresholds are adequate (we did our homework on this platform).

One must be careful when using various online sources to glean information from, while some are accurate at the time of writing, they need updates to reflect the capabilities of modern digital controllers.

[STEvil]

So I was right - you tossed a thermal probe in the mix and called it good, and are relying on the manufacturer to have done their part.

Its par for the course, good enough to cover your butts from parts failure from your supplier, just like everyone else.

Doesnt seem to fit the bill for this statement, though:

And seeing this thread is about X79, we had OTP and OCP implemented properly on our boards from the get-go, so you cannot push them "out of spec" to the point of failure.

[Raja@ASUS]

So I was right - you tossed a thermal probe in the mix and called it good, and are relying on the manufacturer to have done their part.

Its par for the course, good enough to cover your butts from parts failure from your supplier, just like everyone else.

Doesnt seem to fit the bill for this statement, though:

No, you are assuming. How do you know where the temp limit is set or what the OCP thresholds or vendor specific conditions are? You don't. Don't assume. Any more than this also implies that Chil have put no thought behind their controller and its safety features, which is completely absurd. Use the built-in features correctly and the solution works fine.

[STEvil]

I dont recall telling you your products are not currently working fine, but as you've stated "Any more than this also implies that Chil have put no thought behind their controller and its safety features, which is completely absurd. Use the built-in features correctly and the solution works fine." means you're just relying on what they said was good enough to cover warranty claims.

[Raja@ASUS]

I dont recall telling you your products are not currently working fine, but as you've stated "Any more than this also implies that Chil have put no thought behind their controller and its safety features, which is completely absurd. Use the built-in features correctly and the solution works fine." means you're just relying on what they said was good enough to cover warranty claims.

This latest post is full of further assumptions on your part. It means nothing of the sort; the component values and thresholds are set after months of rigorous testing. I am currently at HQ and I know what kind of work we put into the VRM and BIOS for these boards. You have convinced yourself and that is fine, the sad part is that you are creating misinformation for others so I have to respond.

[STEvil]

Every manufacturer uses pretty much the same monitoring scheme, what makes you think you're special for using different values on OCP or OTP?

[Raja@ASUS]

Every manufacturer uses pretty much the same monitoring scheme, what makes you think you're special for using different values on OCP or OTP?

Because we have the boards here and we test them. The implementation does not have to be "special", it just has to be right for the task at hand.

That means the right FETs, inductors and placement of thermistor/s together with ensuring the current and temp trip thresholds are effective in worst case scenarios. All of this takes time and manpower - R&D time, together with investment in equipment. Get it wrong and you encounter issues in the field. That's how things evolve.

The point here is that Chil provide the failsafe mechanisms, it is up to the vendor to invest time into researching optimal layouts and programming those failsafes.

[STEvil]

And everyone else did the same testing as you to validate the parts for their boards and the boards for resale, some got lucky or unlucky, or even had to backpedal (nVidia 570 and 590) for errors.

You did more than the rest, i'll give you that, but that doesnt mean you've swayed me to believe you've built a VRM as well as you could have.

[Raja@ASUS]

And everyone else did the same testing as you to validate the parts for their boards and the boards for resale, some got lucky or unlucky, or even had to backpedal (nVidia 570 and 590) for errors.

You did more than the rest, i'll give you that, but that doesnt mean you've swayed me to believe you've built a VRM as well as you could have.

The Chil controller is pretty much the best you can get right now as an off-the-shelf solution, we teamed that with good FETs and inductors so the design is solid.

[STEvil]

nevermind the single piece heatsink and not monitoring each fet and driver individually.

Congrats on getting me away from general VRM design and specifically into your design btw, walked right into that. Maybe I am more aggravated than I had realized over the lack of action on the Z7S-WS fiasco.

[Raja@ASUS]

nevermind the single piece heatsink and not monitoring each fet and driver individually.

Congrats on getting me away from general VRM design and specifically into your design btw, walked right into that. Maybe I am more aggravated than I had realized over the lack of action on the Z7S-WS fiasco.

You obviously are disgruntled looking at the posts. No hard feeling on my side to be honest, I don't expect anyone to know everything that goes on behind the scenes. The team we have here right now is the best we have ever had, past mistakes withstanding. The initial assumption was that very little work goes into this stuff at all, and that is not true.

[STEvil]

Very little work at all? No, not all.

[Raja@ASUS]

I dont recall telling you your products are not currently working fine, but as you've stated "Any more than this also implies that Chil have put no thought behind their controller and its safety features, which is completely absurd. Use the built-in features correctly and the solution works fine." means you're just relying on what they said was good enough to cover warranty claims.

Very little work at all? No, not all.

That part there - relying on what they said, it implies we don't test anything and just take the stock solution and put it on the board. Anyway, let's move on.

[PatRaceTin]

interesting test

[CrazyNutz]

Very little work at all? No, not all.

My p5k-e has served me well. However there was (is) a huge problem. All the bios voltage settings were wacky, i.e. if you set the lowest voltage for pll, fsb term, etc. it will actually run at the highest possible voltage (Checked with DMM)

I't degraded a couple of e8400's before I finally found out what was going on Definitely left a sour taste in my mouth.


I still use the board and it works fine, but I did have to use the help of my DMM to get the voltages in range.