Overclock.net › Forums › Intel › Intel Motherboards › GIGABYTE Z77X Owners/Discussion/Information/Support Thread and Club (UD5H, UD3H, D3H, and ALL)
New Posts  All Forums:Forum Nav:

GIGABYTE Z77X Owners/Discussion/Information/Support Thread and Club (UD5H, UD3H, D3H, and ALL) - Page 268

post #2671 of 6656
Quote:
Originally Posted by Sin0822 View Post

A word to the wise watching these forums, asus, asrock, msi, evga all hate GB for their Z77 pricing, thus companies with the resources(there really are only 2 who can afford full time forum trolls) I already caught 1 troll who worked for one of those companies posting in my threads on more than 1 forum, i am not saying that every new member who comes here with 2 posts and 0 rep is a troll, they aren't. They are here so that we can help them as this is the most active thread for these boards on the internet.

Ok, since I happen to have only a handful of posts on my back, zero rep and I got that "but you probably already knew that" from barkeater, I feel compelled to address this comment. Sorry, I won't be brief, since being brief before has led, I believe, to a misunderstanding.

Once it was witches, then there were commies, now it's trolls. Every age has its obsessions. :-)
I have no way to convince you, but let me state this in a clearer way: I have been positively impressed by Gigabyte's outstanding board quality. I am writing this from a Ga-8iexp-something (I don't even remember since it's been so long...) that I bought 10 years ago. Yes I am still using a Pentium 4 8-O and not even a fast one. Apart from some minor problems that were fixed in the next revision (mine is rev 1.0), the board has never let me down. My cat even peed on it, I had to take it out of the case, wash it with alcohol, scrape away some (electrolyte?) that seeped from the capacitors near the CPU (solid state capacitors were prolly off limits back then, I guess) and it's still rock solid.

That is the reason I want to buy Gigabyte again (and I have bought Gigabyte for my nieces and cousin's computers).
That is the reason I am here.

I looked up Asrock, Asus and MSI, because I am not a fanboy. I believe Asrock should fire (or at least crucify for half an afternoon) the PR guy who let a picture of a bent board be used for its publicity. I mean, what is the message you give about board quality when you show a board that bends like cardboard? If that's the level of attention to details of your company, then I am not going to be your customer.
MSI didn't feel just right for my needs, tough luck.
Asus almost got me with their "super power saving that will make polar bears shove the snow from your front yard and pandas write you love letters with haikus". But then their power consumption - according to all the comparative reviews I've seen so far - are way higher than Gigabyte's (who knows, maybe because of the ceramic capacitors... :-) ). Moreover, they did away with DVI on the boards that I was interested in, so...

...so here I am. Still thinking "D***n, another rev 1.0 board, what could be wrong this time? something I don't give a **** about, like the distorted audio on the 8iexp (and something else I do not even remember), or something I do care about, like stability at low power consumption?"
And then I read opinions about a bios that seems a bit too buggy (early versions of the uefi not working with logitech and microsoft mice? random freezings, the 100.1 bclk fix, people who has to step up their voltages more than they wished to in order to achieve stability...), probably because of an early release, or because Intel supposedly had the manufacturers wait until they perfected (so to speak, thinking what they did with the TIM) their Ivy Bridge act. Then I stumbled upon the page mentioning the missing capacitors and I think: mmmm, what would happen if the caps were indeed required and gigabyte's engineers were overconfident in their "our copper is thicker than the others"? Is it possible for the cpu to receive the wrong voltage and then shut down (-> random freezes)?

I know it sounds like "the earthquake happened only after Urugh ate the sacred berries, so let's throw Urugh into the volcano to appease the angry gods". But I believe it's quite normal trying to see the source of (all?) problems in what has been done differently from the past - or from other manufacturers (including Intel itself, if I am not mistaken).

I read fasty's and yours replies to that issue (I apologize I did that after barkeater pointed out it had already been discussed but I have little or no trust in forum search engines so I tend not to use them much - I am a Usenet kind of guy). Basically the reason for the lack of caps is that their VRM is the best out there and they can do away with the caps under the cpu. You pointed out that Asrock has no digital PWM suppliers, but that made me wonder: is ASUS's all digital power supply so inferior to gigabyte's (I could not find the chip used by their P8Z77 boards) that they HAVE TO USE CAPS (pun intended)?

It also strikes me as odd the lack of curiosity for a solution that differentiate these boards from the competition. Just saying "GB's VRM is the best bar none, so if you bring the matter up you must be a troll, or a mudslinger paid by the competition - hence we should not discuss it anymore" is a bit too fanboyish for my tastes.
Nobody is interested in the details - and possible drawbacks - of such a feat? Or is that the ultimate perfection in voltage reg design?

I believe Gigabyte engineers are human, hence fallible: after all, there is a rev 1.2 of the GA 8iexp, so, at least in the past, they felt like they did something that could be done better . From what I've read thus far the capacitors can be done away with if you have a speedy voltage regulator and low ohmic losses. Murphy rules at hand, there are lots of things that could go wrong: overestimating the thickness of the copper stripes ("what do you mean UD4? Isn't this a UD5 board?" :-) ), an engineering error ("Oh, the capacitors are not those cute curly symbols?"), a bug in the simulation program ("What do you mean 'in meters'? Didn't we use feet?", a rerouting that had to be done on the fly ("What does it mean 'they're like microwaves, you can't run a stripe around the board twice'?"), some other problem that had to be fixed by bringing certain parameters out of specs ("How, c'mon, don't tell me your PCI card can't hold 101.10 MHz! That's ridiculous!"). Literally tons of things could go wrong. And then maybe they didn't.

"What if" is the key to discovery and to a better understanding.
Hence, let me restate my question: "somebody knows what would happen if we took away the ceramic caps from under the cpu of a z77 board that has to use them (like ASUS's P8Z77)?"
Fanboys please abstain.
Edited by Sredni Vashtar - 7/15/12 at 5:05pm
post #2672 of 6656
Quote:
Originally Posted by Sin0822 View Post

Wait so your only issue is that when you restart your motherboard is shuts itself all the way down and then turns on?
How is that so annoying to you? When you are OCed the systems have to double check themselves, GB has built in fault check points throughout the operation of the system when it is OCed. Have you tried the behavior when the system is not OCed?.

There was something that I forgot pointing out but I will now.
As I said that after CPU PPL Overvolt is disabled, only then i can enable the PC from it's sleep mode, from the mouse/Keyboard. But when CPU PLL is disabled, the system BSOD's randomly which obviously means that it's not stable while being OC'ed and with it being disabled. So now I gotta choose between either having the comfort of using the sleep mode properly or having my system stable, lol I just can't have both now?

I'll put this in as easy words as I can. I hope you get it this time. If I restart the PC, it would shutdown for a few seconds and then turn on and pop into windows. When I put the PC to sleep, it would shutdown (Now instead of just staying there), it would turn back on again and then finally turn back off and then go into sleep mode.

Yes it does piss me off because I don't like my Rig turning on and off time and again, apart from the fact that it's not good for the hardware itself. And I'm not sure that why did you mention those check points? I mean for what reason? I'm not doubting the systems stability when OC'ed, with LLC set to extreme, the voltage comes as close as when entered in the bios, it's that accurate which is a really good thing. But it acts the same way even when it's not even OC'ed.
Quote:
Originally Posted by pushaz View Post

well,i cant see what you doing with your pc,and that "user related" was overal,not for you,you taking it to personally,nobody called you n00b
if you think that Asus, or MSI dont have problems - you are very mistaken,i spent 2 years in one forum,where people talking about Asus P8P67 boards problems,and still after 2 years have same problems,they not updated bios for 4 months,even with very well know bugs,while GB did more than 4 times during last 30 days,that says something and i must say that with every bios update my board just keeps getting better
good software takes time,just as with any other piece of pc hardware,you always need better drivers for vga,for other things,same with bios,i dont see problem here,im yet to see ANY board who release they mobo with perfect bios,especially that UEFI not long ago introduced
Sin0822 - can you look at my post 2644 at 265 page,and what you think what problem might be?
i dont have any friends around with 1155 socket cpu...

Every new Beta Bios that gigabyte comes out with doesn't guarantee any way of being better than the previous one. Your statement is really flawed here. And I even said that not every other company has a perfect end-product, I'm not sure if you even read what I said previously.
post #2673 of 6656
Quote:
Originally Posted by Sin0822 View Post

Quote:
Originally Posted by Ali Man View Post

Your argument of user-related issues applies to noobs, I'm not one. The point is that it had been working alright at one time and now it's not. It's as simple as that and I didn't take out the mobo or damage it in any other way, physically.
Yes I agree, no company is perfect, but we all know that gigabyte hands down has the best build quality in their mobo's, but without good software to make use of that hardware is useless. It's not rocket science to understand man.
The only reason why I updated it was to get better OC'ing potential out of my memory sticks. They're 2133Mhz G.Skill's 4x2GB where they perfectly boot-up at 2200Mhz but never at 2400Mhz and no matter how lose the timings are. But then i realized that it was the sticks themselves, which was an obvious conclusion.
Yes I'am the same guy who had even posted before about the sleeping issues and no one even cared to address my problem. However, you were right till a certain extent that after I turn off the 'CPU PLL overvolt' option in the bios, the sleep function works but turning it on from the keyboard/mouse, but now another problem is also related to it.
First let me tell you that I did update me PC specs a while back, but I'm not sure if you guys can view it. Well no, I'm not running an SB processor, but rather a 3570K.
Check out this link for my overall build:
http://www.pakgamers.com/forums/f139/ivy-bridge-build-130677/
I'm running it at 4.5Ghz @ 1.200V (LLC @ Extreme). This is the lowest Vcore that I can get 4.5Ghz stable at and the hottest core spikes till 65C max.
Now coming to the problem, when I shutdown the PC, it plainly shuts down which is good. But when i restart the PC or even put it in sleep mode, it would first shutdown for around 2 seconds and then turn back on and perform the desired operation. If you see what I'm trying to say, this thing really does get to be a pain. I've tried almost all bioses available except for the F6 one as that's the first bios, so I was just reluctant to go too back.
Wait so your only issue is that when you restart your motherboard is shuts itself all the way down and then turns on?
How is that so annoying to you? When you are OCed the systems have to double check themselves, GB has built in fault check points throughout the operation of the system when it is OCed. Have you tried the behavior when the system is not OCed?.

GB boards have done this for ages. My EP45-UD3P used to do this 2-3 times on really big OCs... (4.5GHz+) and once when below 4.5GHz.
Beast of Burden
(19 items)
 
Cool story, bro!
(16 items)
 
 
CPUMotherboardGraphicsRAM
i7 3770k Gigabyte Z77X-UD5H Gigabyte HD 7970 Reference Ripjaws Z 
Hard DriveHard DriveOptical DriveCooling
Crucial M4 Caviar Black ASUS DVD XSPC Raystorm 
CoolingCoolingCoolingOS
EK FC-7970 w/ Backplate Copper/Acetal RX360 w/ 3x Yate Loon High Speed in Push XSPC Dual Bay Res w/ D5 Pump Windows 7 
MonitorMonitorKeyboardPower
ASUS VE226H ASUS VE226H Cooler Master Quickfire Rapid Corsair HX850 
CaseMouseMouse Pad
HAF 932 Advanced Razer Deathadder QcK Mini 
CPUMotherboardGraphicsRAM
Q9550@ 3.5GHz w/ 1.191v Gigabyte GA-EP45-UD3P Sapphire Reference 5850 @ 775/1100 G.SKILL 4GB (2 x 2GB) DDR2 1066 
Hard DriveHard DriveHard DriveOptical Drive
Crucial C300 Spinpoint F3 Caviar Black ASUS DVD-RW 
CoolingOSMonitorKeyboard
Arctic Cooling Freezer 7 Pro Windows 7 Home Premium 64-bit Samsung 23" Saitek Eclipse III 
PowerCaseMouse
Corsair 430W NZXT Source 210 Wolf King 
  hide details  
Reply
Beast of Burden
(19 items)
 
Cool story, bro!
(16 items)
 
 
CPUMotherboardGraphicsRAM
i7 3770k Gigabyte Z77X-UD5H Gigabyte HD 7970 Reference Ripjaws Z 
Hard DriveHard DriveOptical DriveCooling
Crucial M4 Caviar Black ASUS DVD XSPC Raystorm 
CoolingCoolingCoolingOS
EK FC-7970 w/ Backplate Copper/Acetal RX360 w/ 3x Yate Loon High Speed in Push XSPC Dual Bay Res w/ D5 Pump Windows 7 
MonitorMonitorKeyboardPower
ASUS VE226H ASUS VE226H Cooler Master Quickfire Rapid Corsair HX850 
CaseMouseMouse Pad
HAF 932 Advanced Razer Deathadder QcK Mini 
CPUMotherboardGraphicsRAM
Q9550@ 3.5GHz w/ 1.191v Gigabyte GA-EP45-UD3P Sapphire Reference 5850 @ 775/1100 G.SKILL 4GB (2 x 2GB) DDR2 1066 
Hard DriveHard DriveHard DriveOptical Drive
Crucial C300 Spinpoint F3 Caviar Black ASUS DVD-RW 
CoolingOSMonitorKeyboard
Arctic Cooling Freezer 7 Pro Windows 7 Home Premium 64-bit Samsung 23" Saitek Eclipse III 
PowerCaseMouse
Corsair 430W NZXT Source 210 Wolf King 
  hide details  
Reply
post #2674 of 6656
Thread Starter 
Quote:
Originally Posted by Sredni Vashtar View Post

Warning: Spoiler! (Click to show)
Ok, since I happen to have only a handful of posts on my back, zero rep and I got that "but you probably already knew that" from barkeater, I feel compelled to address this comment. Sorry, I won't be brief, since being brief before has led, I believe, to a misunderstanding.

Once it was witches, then there were commies, now it's trolls. Every age has its obsessions. :-)
I have no way to convince you, but let me state this in a clearer way: I have been positively impressed by Gigabyte's outstanding board quality. I am writing this from a Ga-8iexp-something (I don't even remember since it's been so long...) that I bought 10 years ago. Yes I am still using a Pentium 4 8-O and not even a fast one. Apart from some minor problems that were fixed in the next revision (mine is rev 1.0), the board has never let me down. My cat even peed on it, I had to take it out of the case, wash it with alcohol, scrape away some (electrolyte?) that seeped from the capacitors near the CPU (solid state capacitors were prolly off limits back then, I guess) and it's still rock solid.
That is the reason I want to buy Gigabyte again (and I have bought Gigabyte for my nieces and cousin's computers).
That is the reason I am here.
I looked up Asrock, Asus and MSI, because I am not a fanboy. I believe Asrock should fire (or at least crucify for half an afternoon) the PR guy who let a picture of a bent board be used for its publicity. I mean, what is the message you give about board quality when you show a board that bends like cardboard? If that's the level of attention to details of your company, then I am not going to be your customer.
MSI didn't feel just right for my needs, tough luck.
Asus almost got me with their "super power saving that will make polar bears shove the snow from your front yard and pandas write you love letters with haikus". But then their power consumption - according to all the comparative reviews I've seen so far - are way higher than Gigabyte's (who knows, maybe because of the ceramic capacitors... :-) ). Moreover, they did away with DVI on the boards that I was interested in, so...
...so here I am. Still thinking "D***n, another rev 1.0 board, what could be wrong this time? something I don't give a **** about, like the distorted audio on the 8iexp (and someting else I do not even remember), or something I do care about, like stability at low power consumption?"
And then I read opinions about a bios that seems a bit too buggy (early versions of the uefi not working with logitech and microsoft mice? random freezings, the 100.1 bclk fix, people who has to step up their voltages more than they wished to in order to achieve stability...), probably because of an early release, or because Intel supposedly had the manufacturers wait until they perfected (so to speak, thinking what they did with the TIM) their Ivy Bridge act. Then I stumbled upon the page mentioning the missing capacitors and I think: mmmm, what would happen if the caps were indeed required and gigabyte's engineers were overconfindent in their "our copper is thicker than the others"? Is it possible for the cpu to receive the wrong voltage and then shut down (-> random freezes)?
I know it sounds like "the earthquake happened only after Urugh ate the sacred berries, so let's throw Urugh into the volcano to appease the angry gods". But I believe it's quite normal trying to see the source of (all?) problems in what has been done differently from the past - or from other manufacturers (including Intel itself, if I am not mistaken).
I read fasty's and yours replies to that issue (I apologize I did that after barkeater pointed out it had already been discussed but I have little or no trust in forum search engines so I tend not to use them much - I am a Usenet kind of guy). Basically the reason for the lack of caps is that their VRM is the best out there and they can do away with the caps under the cpu. You pointed out that Asrock has no digital PWM suppliers, but that made me wonder: is ASUS's all digital power supply so inferior to gigabyte's (I could not find the chip used by their P8Z77 boards) that they HAVE TO USE CAPS (pun intended)?
It also strikes me as odd the lack of curiosity for a solution that differitiate these boards from the competition. Just saying "GB's VRM is the best bar none, so if you bring the matter up you must be a troll, or a mudslinger paid by the competion - hence we should not discuss it anymore" is a bit too fanboish for my tastes.
Nobody is interested in the details - and possible drawbacks - of such a feat? Or is that the ultimate perfection in voltage reg design?
I believe Gigabyte engineers are human, hence fallible: after all, there is a rev 1.2 of the GA 8iexp, so they flet like they did something that could be done better in the past. From what I've read thus far the capacitors can be done away with if you have a speedy voltage regulator and low ohmic losses. Murphy rules at hand, there are lots of things that could go wrong: overestimating the thickness of the copper stripes ("what do you meand UD4? Isn't this a UD5 board?" :-) ), an engineering error ("Oh, the capacitors are not those cute curly symbols?"), a bug in the simulation program ("What do you mean 'in meters'? Didn't we use feet?", a rerouting that had to be done on the fly ("What does it mean 'they're like microwaves, you can't run a stripe around the board twice'?"), some other problem that had to be fixed by bringing certain parameters out of specs ("How, c'mon, don't tell me your PCI card can't hold 101.10 MHz! That's ridicolous!"). Literally tons of things could go wrong. And then maybe they didn't.
"What if" is the key to discovery and to a better understanding.
Hence, let me restate my question: "somebody knows what would happen if we took away the cercamic caps from under the cpu of a z77 board that has to use them (like ASUS's P8Z77)?"
Fanboys please abstain.
No GB and ASUS currently use the same supplier, but that has little to nothing to do with the need for extra capacitors. If you look inside the socket there are high frequency capacitors in there, the point is the put them as close to the CPU as possible, so you start inside the socket and then to the back, however there is an optimum amount to balance out everything. Saying that capacitors are missing in the back must have been written by a new ASUS hire who has no knowledge of electrical engineering, probably some media asus hired recently, as that is what they have started to do(take people who used to review motherboards and quit, and offer them jobs in marketing).

I don't work for GB, all the help i provide here is out of my own free will, and what i say doesn't represent GB as a company. GB and other manufacturers do their best to minimize variation in the PCB so they can use the same PCB for two or more boards. That means that some of the layout considerations in terms of the MLCCs(the ceramic capacitors) are different board to board.

There are many reasons that capacitor count varies on different boards, that is because almost all boards vary in what MOSFETs, Inductors, capacitors, as well as settings and transient controls they use. Switching frequency for example is one factor that differs board to board and directly has an impact on ripple, and it has an impact on how much output capacitance is present(higher switching frequency=lower output capacitors used). Even though all the ceramic capacitors look the same their specifications can differ too.

Let's take one example of a way to decrease ripple other than using capacitors(switching frequency is a big one, but here is one that is much more complicated), I will use the ASRock Z77 Extreme4 Gen 3 and the GIGABYTE Z77X-UD3H as examples. I could use the UD5H but i am using two boards in the same price segment. The ASRock on the surface looks like it has 8 phases for the CPU VCore, the UD3H uses 6 phases. However the UD3H uses more true phases. That sounds odd right? How can an 8 phase board use less true phases than a 6 phase board? That is because the PWM the ASRock uses is a 6+1 phase PWM and the GB uses a 6+2 phase PWM, on neither PWM can the +1 or +2 be added to the second set(on some PWMs this is possible such as CHL8328, but not on these two).

So the question remains how does ASRock get 8 phases from 6 phases? Phases are all about being symmetric, so you cannot double a few of the true phases while still using the others. No, you have to take 4 phases and double them and then disable 2 of the true phases. That means that they are only using 4 true phases. Now all manufacturers do this, GB did it a lot with X58 and Z68 and X79, ASUS does it with their Maximus 5 Gene, Maximus 4 Gene as well as many other boards even though they use a PWM that can do 8 phases on a single rail.

The reason being that the +1 or +2 allows the manufacturer to use 1 PWM to drive CPU VCore and either iGPU or VTT(and IMC) which requires their own phases on a separate rail(s), then use a second cheaper PWM to power the VTT or iGPU. If they used 1 PWM for all VCore then they woudl have to use 2 other PWMs or 1 more expensive PWM. So it is most cost effective to use 1 PWM for VCore and iGPU is most common, that is why PWMs come in 4+1, 6+1, 6+2, and 7+1 dual output models. There are some 8 phase and 6 phase only singe outputs, and one CHil 7+1 can also do 8 flat, however the intersil and the IR that asrock and GB use cannot. I am getting to my point, so please hang on.

Now why dose this make a difference? The reason is beucase the more true phases you use, the more of an interleaving affect you get. The interleaving affect uses the fact that each phase is out of phase and symmetrical with all the other phases(it is on when the others are off) in the end you get combined ripple frequency of n(times) the amount of a single phase. So if you have a true 4 phase/virtual 8 phase such as Extreme4 then your combined ripple frequency is 4x that of any individual phase. On the UD3H you have 6 true phases and thus the combined ripple frequency is 6x that of any single phase. This interleaving affect increases ripple frequency which decreases the amplitude of the ripple(think of a graph with amplitude being the length of the peak), when you decrease the amplitude of the ripple then you also decrease the ripple current. That reduction of ripple through increased ripple frequency lessens the demand on the output filter to get rid of all the ripple, and thus you need to use more output capacitors when you have less true phases. However while this affect is MUCH greater from 2-4 and 4-6, from 6-8 this affect is a bit less.
649
This graph shows you the decrease and it is straight from the Intersil design guide for the ISL6366 which is the VRD12 certified analog PWM. The ISL6367 is the VRD 12/12.5 certified analog PWM that asrock uses on the extreme4 same as the ISL6366 except in some minor SVID things. The GB UD3H uses an IR3567 which is a VRD 12/12.5 certified digital PWM.

Now that isn't the reason GB doesn't fill up those bottom caps, the truth is they just aren't needed, there are SO many factors that adding extra capacitors can actually hurt the performance. GIGABYTE uses different aluminum capacitors, they use different chokes(inductors) and they use different PCB and MOSFETs and PWM and traces and EVERYTHING is different. Balance is the key, and almost all modern VRMs use a mix of high cost aluminum polymers and cheap ceramics to attain what the manufacturer desires. On the UD3H perhaps more of the caps would be used, but on the UD5H the use of more inductors could also help reduce the ripple as less current is flowing through each as could the difference in the specs of the MOSFETs. The UD5H actually has 1 more output capacitor than the UD3H perhaps because it is needed, thus even though the interleaving affect is the same, perhaps the spread of the current requires and extra cap, everything is different on every board.

I don't mean to call anyone a noob, i just get frustrated when people report issues but are in no way willing to explain further or accept that it could possibly be their fault. Everyone makes mistakes and everyone is a noob in some sense. I am a noob when it comes to SSDs past 2011, but i know my stuff when it comes to motherboards, however many other people don't. I try to help them and share my knowledge, but to do that i need as much information as possible.

I treat all of you equally, no one differently, however since i made this thread I have seen a HUGE influx of new users, now that makes me happy to see users are coming here for me to help them, but it also makes me suspicious, as it is very easy to take advantage. I just want to let people know issues reported in this thread doesn't mean everyone will have them.

Please if you have an issue elaborate, but don't thin i or others can read your mind, please give us all the statistics, exactly what happens, and exactly what parts and settings you are using. That is all I ask.

Also the random freezing issue, it shouldn't happen at all unless you have a fault memory modules, a bad OC, or you have a PCI add in card which is causing DPC latency spikes, which is now the number 1 cause of the BSOD. Now if that is you then GB is looking into the issue. Also CPu PLL OV is needed for OCing, in all previous BIOSes it is enabled by default, enabling it here shouldn't hurt anything with ivy bridge. i will try it out tonight Ali Man, and investigate what you are reporting.
X99 Main Rig
(10 items)
 
  
CPUMotherboardGraphicsRAM
Intel 5960X Extreme Edition @ 4.5GHz Always Changing VisonTek R9 290 G.Skill Ripjaws 4 16GB (4x4GB) DDR4 @ 3200MHz 
Hard DriveHard DriveHard DriveCooling
Samsung 128GB M.2 PCI-E 4x SSD Apotop 256GB SSD 1.82TB NAS Noctua NH-D15 with both fans 
OSPower
Win7 Pro Enermax 1000W 
  hide details  
Reply
X99 Main Rig
(10 items)
 
  
CPUMotherboardGraphicsRAM
Intel 5960X Extreme Edition @ 4.5GHz Always Changing VisonTek R9 290 G.Skill Ripjaws 4 16GB (4x4GB) DDR4 @ 3200MHz 
Hard DriveHard DriveHard DriveCooling
Samsung 128GB M.2 PCI-E 4x SSD Apotop 256GB SSD 1.82TB NAS Noctua NH-D15 with both fans 
OSPower
Win7 Pro Enermax 1000W 
  hide details  
Reply
post #2675 of 6656
Quote:
Originally Posted by Sin0822 View Post

Warning: Spoiler! (Click to show)
Quote:
Originally Posted by Sredni Vashtar View Post

Warning: Spoiler! (Click to show)
Ok, since I happen to have only a handful of posts on my back, zero rep and I got that "but you probably already knew that" from barkeater, I feel compelled to address this comment. Sorry, I won't be brief, since being brief before has led, I believe, to a misunderstanding.

Once it was witches, then there were commies, now it's trolls. Every age has its obsessions. :-)
I have no way to convince you, but let me state this in a clearer way: I have been positively impressed by Gigabyte's outstanding board quality. I am writing this from a Ga-8iexp-something (I don't even remember since it's been so long...) that I bought 10 years ago. Yes I am still using a Pentium 4 8-O and not even a fast one. Apart from some minor problems that were fixed in the next revision (mine is rev 1.0), the board has never let me down. My cat even peed on it, I had to take it out of the case, wash it with alcohol, scrape away some (electrolyte?) that seeped from the capacitors near the CPU (solid state capacitors were prolly off limits back then, I guess) and it's still rock solid.
That is the reason I want to buy Gigabyte again (and I have bought Gigabyte for my nieces and cousin's computers).
That is the reason I am here.
I looked up Asrock, Asus and MSI, because I am not a fanboy. I believe Asrock should fire (or at least crucify for half an afternoon) the PR guy who let a picture of a bent board be used for its publicity. I mean, what is the message you give about board quality when you show a board that bends like cardboard? If that's the level of attention to details of your company, then I am not going to be your customer.
MSI didn't feel just right for my needs, tough luck.
Asus almost got me with their "super power saving that will make polar bears shove the snow from your front yard and pandas write you love letters with haikus". But then their power consumption - according to all the comparative reviews I've seen so far - are way higher than Gigabyte's (who knows, maybe because of the ceramic capacitors... :-) ). Moreover, they did away with DVI on the boards that I was interested in, so...
...so here I am. Still thinking "D***n, another rev 1.0 board, what could be wrong this time? something I don't give a **** about, like the distorted audio on the 8iexp (and someting else I do not even remember), or something I do care about, like stability at low power consumption?"
And then I read opinions about a bios that seems a bit too buggy (early versions of the uefi not working with logitech and microsoft mice? random freezings, the 100.1 bclk fix, people who has to step up their voltages more than they wished to in order to achieve stability...), probably because of an early release, or because Intel supposedly had the manufacturers wait until they perfected (so to speak, thinking what they did with the TIM) their Ivy Bridge act. Then I stumbled upon the page mentioning the missing capacitors and I think: mmmm, what would happen if the caps were indeed required and gigabyte's engineers were overconfindent in their "our copper is thicker than the others"? Is it possible for the cpu to receive the wrong voltage and then shut down (-> random freezes)?
I know it sounds like "the earthquake happened only after Urugh ate the sacred berries, so let's throw Urugh into the volcano to appease the angry gods". But I believe it's quite normal trying to see the source of (all?) problems in what has been done differently from the past - or from other manufacturers (including Intel itself, if I am not mistaken).
I read fasty's and yours replies to that issue (I apologize I did that after barkeater pointed out it had already been discussed but I have little or no trust in forum search engines so I tend not to use them much - I am a Usenet kind of guy). Basically the reason for the lack of caps is that their VRM is the best out there and they can do away with the caps under the cpu. You pointed out that Asrock has no digital PWM suppliers, but that made me wonder: is ASUS's all digital power supply so inferior to gigabyte's (I could not find the chip used by their P8Z77 boards) that they HAVE TO USE CAPS (pun intended)?
It also strikes me as odd the lack of curiosity for a solution that differitiate these boards from the competition. Just saying "GB's VRM is the best bar none, so if you bring the matter up you must be a troll, or a mudslinger paid by the competion - hence we should not discuss it anymore" is a bit too fanboish for my tastes.
Nobody is interested in the details - and possible drawbacks - of such a feat? Or is that the ultimate perfection in voltage reg design?
I believe Gigabyte engineers are human, hence fallible: after all, there is a rev 1.2 of the GA 8iexp, so they flet like they did something that could be done better in the past. From what I've read thus far the capacitors can be done away with if you have a speedy voltage regulator and low ohmic losses. Murphy rules at hand, there are lots of things that could go wrong: overestimating the thickness of the copper stripes ("what do you meand UD4? Isn't this a UD5 board?" :-) ), an engineering error ("Oh, the capacitors are not those cute curly symbols?"), a bug in the simulation program ("What do you mean 'in meters'? Didn't we use feet?", a rerouting that had to be done on the fly ("What does it mean 'they're like microwaves, you can't run a stripe around the board twice'?"), some other problem that had to be fixed by bringing certain parameters out of specs ("How, c'mon, don't tell me your PCI card can't hold 101.10 MHz! That's ridicolous!"). Literally tons of things could go wrong. And then maybe they didn't.
"What if" is the key to discovery and to a better understanding.
Hence, let me restate my question: "somebody knows what would happen if we took away the cercamic caps from under the cpu of a z77 board that has to use them (like ASUS's P8Z77)?"
Fanboys please abstain.
No GB and ASUS currently use the same supplier, but that has little to nothing to do with the need for extra capacitors. If you look inside the socket there are high frequency capacitors in there, the point is the put them as close to the CPU as possible, so you start inside the socket and then to the back, however there is an optimum amount to balance out everything. Saying that capacitors are missing in the back must have been written by a new ASUS hire who has no knowledge of electrical engineering, probably some media asus hired recently, as that is what they have started to do(take people who used to review motherboards and quit, and offer them jobs in marketing).

I don't work for GB, all the help i provide here is out of my own free will, and what i say doesn't represent GB as a company. GB and other manufacturers do their best to minimize variation in the PCB so they can use the same PCB for two or more boards. That means that some of the layout considerations in terms of the MLCCs(the ceramic capacitors) are different board to board.

There are many reasons that capacitor count varies on different boards, that is because almost all boards vary in what MOSFETs, Inductors, capacitors, as well as settings and transient controls they use. Switching frequency for example is one factor that differs board to board and directly has an impact on ripple, and it has an impact on how much output capacitance is present(higher switching frequency=lower output capacitors used). Even though all the ceramic capacitors look the same their specifications can differ too.

Let's take one example of a way to decrease ripple other than using capacitors(switching frequency is a big one, but here is one that is much more complicated), I will use the ASRock Z77 Extreme4 Gen 3 and the GIGABYTE Z77X-UD3H as examples. I could use the UD5H but i am using two boards in the same price segment. The ASRock on the surface looks like it has 8 phases for the CPU VCore, the UD3H uses 6 phases. However the UD3H uses more true phases. That sounds odd right? How can an 8 phase board use less true phases than a 6 phase board? That is because the PWM the ASRock uses is a 6+1 phase PWM and the GB uses a 6+2 phase PWM, on neither PWM can the +1 or +2 be added to the second set(on some PWMs this is possible such as CHL8328, but not on these two).

So the question remains how does ASRock get 8 phases from 6 phases? Phases are all about being symmetric, so you cannot double a few of the true phases while still using the others. No, you have to take 4 phases and double them and then disable 2 of the true phases. That means that they are only using 4 true phases. Now all manufacturers do this, GB did it a lot with X58 and Z68 and X79, ASUS does it with their Maximus 5 Gene, Maximus 4 Gene as well as many other boards even though they use a PWM that can do 8 phases on a single rail.

The reason being that the +1 or +2 allows the manufacturer to use 1 PWM to drive CPU VCore and either iGPU or VTT(and IMC) which requires their own phases on a separate rail(s), then use a second cheaper PWM to power the VTT or iGPU. If they used 1 PWM for all VCore then they woudl have to use 2 other PWMs or 1 more expensive PWM. So it is most cost effective to use 1 PWM for VCore and iGPU is most common, that is why PWMs come in 4+1, 6+1, 6+2, and 7+1 dual output models. There are some 8 phase and 6 phase only singe outputs, and one CHil 7+1 can also do 8 flat, however the intersil and the IR that asrock and GB use cannot. I am getting to my point, so please hang on.

Now why dose this make a difference? The reason is beucase the more true phases you use, the more of an interleaving affect you get. The interleaving affect uses the fact that each phase is out of phase and symmetrical with all the other phases(it is on when the others are off) in the end you get combined ripple frequency of n(times) the amount of a single phase. So if you have a true 4 phase/virtual 8 phase such as Extreme4 then your combined ripple frequency is 4x that of any individual phase. On the UD3H you have 6 true phases and thus the combined ripple frequency is 6x that of any single phase. This interleaving affect increases ripple frequency which decreases the amplitude of the ripple(think of a graph with amplitude being the length of the peak), when you decrease the amplitude of the ripple then you also decrease the ripple current. That reduction of ripple through increased ripple frequency lessens the demand on the output filter to get rid of all the ripple, and thus you need to use more output capacitors when you have less true phases. However while this affect is MUCH greater from 2-4 and 4-6, from 6-8 this affect is a bit less.
649
This graph shows you the decrease and it is straight from the Intersil design guide for the ISL6366 which is the VRD12 certified analog PWM. The ISL6367 is the VRD 12/12.5 certified analog PWM that asrock uses on the extreme4 same as the ISL6366 except in some minor SVID things. The GB UD3H uses an IR3567 which is a VRD 12/12.5 certified digital PWM.

Now that isn't the reason GB doesn't fill up those bottom caps, the truth is they just aren't needed, there are SO many factors that adding extra capacitors can actually hurt the performance. GIGABYTE uses different aluminum capacitors, they use different chokes(inductors) and they use different PCB and MOSFETs and PWM and traces and EVERYTHING is different. Balance is the key, and almost all modern VRMs use a mix of high cost aluminum polymers and cheap ceramics to attain what the manufacturer desires. On the UD3H perhaps more of the caps would be used, but on the UD5H the use of more inductors could also help reduce the ripple as less current is flowing through each as could the difference in the specs of the MOSFETs. The UD5H actually has 1 more output capacitor than the UD3H perhaps because it is needed, thus even though the interleaving affect is the same, perhaps the spread of the current requires and extra cap, everything is different on every board.

I don't mean to call anyone a noob, i just get frustrated when people report issues but are in no way willing to explain further or accept that it could possibly be their fault. Everyone makes mistakes and everyone is a noob in some sense. I am a noob when it comes to SSDs past 2011, but i know my stuff when it comes to motherboards, however many other people don't. I try to help them and share my knowledge, but to do that i need as much information as possible.

I treat all of you equally, no one differently, however since i made this thread I have seen a HUGE influx of new users, now that makes me happy to see users are coming here for me to help them, but it also makes me suspicious, as it is very easy to take advantage. I just want to let people know issues reported in this thread doesn't mean everyone will have them.

Please if you have an issue elaborate, but don't thin i or others can read your mind, please give us all the statistics, exactly what happens, and exactly what parts and settings you are using. That is all I ask.

Also the random freezing issue, it shouldn't happen at all unless you have a fault memory modules, a bad OC, or you have a PCI add in card which is causing DPC latency spikes, which is now the number 1 cause of the BSOD. Now if that is you then GB is looking into the issue. Also CPu PLL OV is needed for OCing, in all previous BIOSes it is enabled by default, enabling it here shouldn't hurt anything with ivy bridge. i will try it out tonight Ali Man, and investigate what you are reporting.
Excellent post biggrin.gif
I also agree that the UD5H is one of the best mobo in the world, ever thumb.gif

On the other hand I still have somehow a desire to the X79-UD5 and I haven't decide yet what to choose (and this is a 3770k ihs remove issue related too)
Excellent power consumption on load for the Z77X-UD5H (and idle too) and really nice idle power consumption for the X79-UD5:

581581

http://www.techpowerup.com/reviews/Biostar/TZ77XE4_review/8.html

But I still don't get how come the X79-UD5's vrm will consume so much power on load. Are the guys from techpowerup somehow wrong?
They do reviewed recently the Sniper M3 (I honestly didn't like their review) and the two UD5's gets the same positions.

http://www.techpowerup.com/reviews/Gigabyte/G1_Sniper_M3_review/8.html

What do you personally think of the X79-UD5?
Edited by arrow0309 - 7/15/12 at 10:37pm
Arrow I7
(15 items)
 
  
CPUMotherboardGraphicsRAM
i7 6850K 4.4 @3.7 cache Heatkiller IV Pro Asus X99M WS/SE TITAN X Pascal @ Heatkiller IV G.Skill F4-3200C14Q-32GTZSW 
Hard DriveHard DriveCoolingCooling
Intel 750 NVMe 400Gb, 2500Pro 180Gb Crucial MX300 750Gb LE, 2x Barracuda 3Tb Cooled by EKWB, Watercool, Alphacool, Koolance XSPC RX360 v3, 2x240 rads 
MonitorKeyboardPowerCase
Acer Predator XB271HU G.Skill Ripjaws KM780 RGB SuperFlower Titanium 1000W Phanteks Enthoo Mini XL 
Mouse
Logitech G602 
  hide details  
Reply
Arrow I7
(15 items)
 
  
CPUMotherboardGraphicsRAM
i7 6850K 4.4 @3.7 cache Heatkiller IV Pro Asus X99M WS/SE TITAN X Pascal @ Heatkiller IV G.Skill F4-3200C14Q-32GTZSW 
Hard DriveHard DriveCoolingCooling
Intel 750 NVMe 400Gb, 2500Pro 180Gb Crucial MX300 750Gb LE, 2x Barracuda 3Tb Cooled by EKWB, Watercool, Alphacool, Koolance XSPC RX360 v3, 2x240 rads 
MonitorKeyboardPowerCase
Acer Predator XB271HU G.Skill Ripjaws KM780 RGB SuperFlower Titanium 1000W Phanteks Enthoo Mini XL 
Mouse
Logitech G602 
  hide details  
Reply
post #2676 of 6656
You should write a book on motherboards man and I'll be the first one to buy it!

It's interesting seeing that the Extreme4 uses only 4-true phases, it also must be the same with the Z77 Extreme6 as even that has an 8-phase VRM but with just more beefier sinks. But what's more catching is that those so-called 8-phases still doesn't hinder the mobo's OC'ing capability.
post #2677 of 6656
Thread Starter 
well yes they over kill, but the extreme6 uses good quality MOSFETs, but the extreme4 has some D-PAk fets, and they cause issues such as hot running VRMs(that is why a sub $150 has heatsinks like those) i mean you see the UD3H? it doesn't even have heatsinks on half of its FETs and it does WRs lol.. The VCore doesn't hold steady like it does on the UD3H or UD5H when you compare to E4. Get a DMM and read the real vcore using LLC extreme, it actually raises the voltage under load, then try LLC Turbo, and it barley drops in voltage. CPUz creates a reading which isn't so accurate.

The issue is that every VRM is different that is what i was trying to get across. I wasn't trying to be a dick, but the whole thing about marketing which aims to be technical yet is just wrong, it really gets to me. The worst is when people say stuff and can't back it up.


I don't mean to get at people, sorry if i was a bit over the top, i was just trying to explain that it isn't so simple as missing capacitors. They aren't missing, they are just not in the design. Seriously take a DMM to it and compare to any other board, i doubt any other board will have steady voltage like that. They tuned the LLC to perfection, there is even 1 LLC level higher which u cannot use with normal release BIOS, it is only for LN2, as you get higher in voltage and voltage, like 1.9v loading 150A at 100% then LLC extreme will droop down to like 1.88v, thus LLC +6 is used and it will keep the vcore very steady.

listen Ali Man i will look into your issue, but are you sure if you flash back to BIOS F7(try both) that the issue isn't different?
X99 Main Rig
(10 items)
 
  
CPUMotherboardGraphicsRAM
Intel 5960X Extreme Edition @ 4.5GHz Always Changing VisonTek R9 290 G.Skill Ripjaws 4 16GB (4x4GB) DDR4 @ 3200MHz 
Hard DriveHard DriveHard DriveCooling
Samsung 128GB M.2 PCI-E 4x SSD Apotop 256GB SSD 1.82TB NAS Noctua NH-D15 with both fans 
OSPower
Win7 Pro Enermax 1000W 
  hide details  
Reply
X99 Main Rig
(10 items)
 
  
CPUMotherboardGraphicsRAM
Intel 5960X Extreme Edition @ 4.5GHz Always Changing VisonTek R9 290 G.Skill Ripjaws 4 16GB (4x4GB) DDR4 @ 3200MHz 
Hard DriveHard DriveHard DriveCooling
Samsung 128GB M.2 PCI-E 4x SSD Apotop 256GB SSD 1.82TB NAS Noctua NH-D15 with both fans 
OSPower
Win7 Pro Enermax 1000W 
  hide details  
Reply
post #2678 of 6656
Well that is really interesting to know about the E4. I guess that price really does matter. But I'm sure that the previous Z68 Extreme4 GEN3 had better mosfets than the current Z77 one. I also owned this mobo (before the UD5H) and had a great experience with it.

I always thought of the same thing that you pointed out lol. That they always used the UD3H for WRs and not the UD5H tongue.gif
I knew that there must have been something behind it. The only reason why I didn't go for the UD3H was for the same reason that it didn't have sinks covering half of its VRM zone!

Well you even said that before of there being two F7 bioses, but I just flashed the one available at the giga website. However I just flashed back to F6 and the problem still continues. I'm thinking that this may only have to do with my mobo instead of anyone elses.
Edited by Ali Man - 7/15/12 at 11:20pm
post #2679 of 6656
@Sin0822

Hope I wasn't so OT before, with my invasive X79-UD5 vrm's power issue, I really like this board and its vrm (even powerful) and the only reason I posted here is that I can't find a X79 Gigabyte motherboards thread here on the ocn.
Feel free non to answer my question if it's worth it

OT out
Arrow I7
(15 items)
 
  
CPUMotherboardGraphicsRAM
i7 6850K 4.4 @3.7 cache Heatkiller IV Pro Asus X99M WS/SE TITAN X Pascal @ Heatkiller IV G.Skill F4-3200C14Q-32GTZSW 
Hard DriveHard DriveCoolingCooling
Intel 750 NVMe 400Gb, 2500Pro 180Gb Crucial MX300 750Gb LE, 2x Barracuda 3Tb Cooled by EKWB, Watercool, Alphacool, Koolance XSPC RX360 v3, 2x240 rads 
MonitorKeyboardPowerCase
Acer Predator XB271HU G.Skill Ripjaws KM780 RGB SuperFlower Titanium 1000W Phanteks Enthoo Mini XL 
Mouse
Logitech G602 
  hide details  
Reply
Arrow I7
(15 items)
 
  
CPUMotherboardGraphicsRAM
i7 6850K 4.4 @3.7 cache Heatkiller IV Pro Asus X99M WS/SE TITAN X Pascal @ Heatkiller IV G.Skill F4-3200C14Q-32GTZSW 
Hard DriveHard DriveCoolingCooling
Intel 750 NVMe 400Gb, 2500Pro 180Gb Crucial MX300 750Gb LE, 2x Barracuda 3Tb Cooled by EKWB, Watercool, Alphacool, Koolance XSPC RX360 v3, 2x240 rads 
MonitorKeyboardPowerCase
Acer Predator XB271HU G.Skill Ripjaws KM780 RGB SuperFlower Titanium 1000W Phanteks Enthoo Mini XL 
Mouse
Logitech G602 
  hide details  
Reply
post #2680 of 6656
Any difference between Z77X-UD3H-WB WIFI and Z77X-Ud3H aside from the wireless connector? would it be worth the extra price?
AMD
(9 items)
 
Home PC
(9 items)
 
Current
(9 items)
 
CPUMotherboardGraphicsRAM
AMD 8320 Asus M5A99X EVO Rev 2.0 Sapphire 7970 Kingston HyperX Beast 2x4Gb 9CL 1600Mhz 
Hard DriveCoolingOSPower
Samsung 830 128 Gb ThermalRight Macho r2-h0 Windows 8.1 Corair 650W TX 
Case
Zalman Z11 Plus 
CPUMotherboardGraphicsRAM
3770k Gigabyte Z77x UD3H Intel HD 4000 Geil Black Dragon 2 x 4 GB 1600 Mhz 
Hard DriveCoolingOSPower
Samsung 830  Thermalright True Spirit 120 Windows 7 SP1 x64 Corsair 650TX v2 
Case
Corsair Carbide 400R 
CPUMotherboardGraphicsRAM
4790k Asus VII Ranger Gigabyte 290 8 Gb Corsair DDR3 
Hard DriveCoolingKeyboardPower
3 x Samsung 850 250gb SSD Swiftech H220 BlackWidow Ultimate EVGA 850 G2 
Case
Phanteks Enthoo Evolv ATX 
  hide details  
Reply
AMD
(9 items)
 
Home PC
(9 items)
 
Current
(9 items)
 
CPUMotherboardGraphicsRAM
AMD 8320 Asus M5A99X EVO Rev 2.0 Sapphire 7970 Kingston HyperX Beast 2x4Gb 9CL 1600Mhz 
Hard DriveCoolingOSPower
Samsung 830 128 Gb ThermalRight Macho r2-h0 Windows 8.1 Corair 650W TX 
Case
Zalman Z11 Plus 
CPUMotherboardGraphicsRAM
3770k Gigabyte Z77x UD3H Intel HD 4000 Geil Black Dragon 2 x 4 GB 1600 Mhz 
Hard DriveCoolingOSPower
Samsung 830  Thermalright True Spirit 120 Windows 7 SP1 x64 Corsair 650TX v2 
Case
Corsair Carbide 400R 
CPUMotherboardGraphicsRAM
4790k Asus VII Ranger Gigabyte 290 8 Gb Corsair DDR3 
Hard DriveCoolingKeyboardPower
3 x Samsung 850 250gb SSD Swiftech H220 BlackWidow Ultimate EVGA 850 G2 
Case
Phanteks Enthoo Evolv ATX 
  hide details  
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Intel Motherboards
Overclock.net › Forums › Intel › Intel Motherboards › GIGABYTE Z77X Owners/Discussion/Information/Support Thread and Club (UD5H, UD3H, D3H, and ALL)