What makes a good VRM

OPTIONAL PREFACE
I have some good news and some bad news.
Good news when I tried to power on the GTX 590 it didn't catch fire, make magic smoke or explode.
Bad news is that I didn't get any video so I need to fix the card. It's probably the PCI-e slot or I need to trip the PWR_GOOD pin on the controllers. I also tried to power the GPU using 2 separate PSU so that might have something to do with it too.

However since there's no entertainment article I have prepped for today other than the GTX 590. You're gonna get an education on VRMs.
END OF OPTIONAL PREFACE

First of all you need to understand how a VRM that converts 12V DC to a lower voltage works. Since this is rather complicated and better explained elsewhere you can just go read this. That will explain the basics of a low power single phase VRM.

So now that you've read that lets expand that and apply it to computer VRMs. First of all the fly wheel circuit uses a diode. This is really inefficient and massively limits the maximum current through put so in computer VRMs you will find instead of the diode what is called a low side MOSFET. This MOSFET is only on when the High side MOSFET(the component labeled switching transistor) is off or else you would get a short circuit. This low side MOSFET handles the bulk of the current that flows through your load(CPU/GPU core RAM chip...) so these MOSFETs are the most important when building a powerful VRM. Low side and high side MOSFETs typically have current handling capabilities between 20 and 60A at 125C° case temperature.

The article I linked shows the PWM being fed directly to the high side MOSFET. In computer VRMs the MOSFETs used have a rather large gate capacitance. Meaning that if you want to switch them on you need to let them charge up. If you want them to charge quickly you will need to supply a current greater than what PWM controllers can provide. So to supply the current a driver MOSFET is used. This MOSFET is typically capable of only handling currents smaller than 10A and can be switched on and off directly by the PWM signal coming out of the PWM controller. The driver MOSFET is not very important to a VRMs current handling capability but they are a key component of computer VRMs so you should know about them.

So far I have explained everything as far as a single phase is concerned. As you probably know computer VRMs for the CPU and GPU typically have 3 or more phases. So how does that work.

Well  each phase handles a chunk of the total current that your load requires. To do this the PWM controller generates as many PWM signals as there are phases. These signals are offset so that only one phase has it's high side MOSFET on at any given time. All the other phases have their low side MOSFET in the on state and the high side MOSFET is off. So if you have a 4 phase VRM you have 3 phases running in fly wheel mode and 1 phase charging. You can more or less gauge the current handling capability of a multi phase VRM by taking the current capability of the low side MOSFETs and multiplying it by the number of phases.

Now VRMs also include capacitors and many better VRMs will include more capacitors than cheaper VRMs. This is because you need capacitors to smooth out the voltage being produced by the VRM and the more capacitors you have the more capacity the VRM has and the less your voltage drops while your highside MOSFET is off. If you had very small capacitors and a very high current draw the capacitor could end up completely draining before the high side comes on resulting the voltage that your load is being provided reaching 0V. As we all know that is bad. Which is why high end VRMs have huge capacitor banks. Now capacitors also cause an efficiency loss and take a ton of space so just slapping 1F of capcitance on a VRM is not the best idea. However if you have VRM that has high ripple adding more caps can help.The other issue with capacitors is that some capacitors(electrolytes) have a maximum current that can be pulled from them, If you exceed this current the cap will fail.

The other way to lower voltage ripple is to increase how often you turn on the high side. This is dictated by the PWM controller's switching frequency. When you turn on the high side MOSFET your VRM output voltage starts to rise until the PWM signal turn it off again and your voltage starts to drop. The longer the wait between the on and off the longer the voltage will rise and drop increasing the minimum voltage and maximum voltage that your VRM outputs when trying to hit a set voltage. This is what ripple is. So if you cut down the amount of time you voltage spends dropping and rising by increasing the frequency of the PWM signal you decrease the ripple. This is why many overclocking centric boards have a VRM switching frequency option in the BIOS. The down side to this is that you need to charge your MOSFETs on and off more often and that lowers the VRM's efficiency. Which is why OCing GPUs like the Lightning are so damn power hungry.

The final way to lower voltage ripple is to add more phases. Because then you basically increase the switching frequency of the PWM because instead of cycling through X MOSFETs turning on and off in time Y you cycle through Z>X MOSFETS turning on and off. So you get more switching in time Y. Accompanied by that same efficiency loss as before.

So how many phases does your motherboard/GPU have? No more than 8.
8 is the largest number of phases that any PWM controller currently used in computers can produce. So how do we have VRMs with 10 12 14 16 20 24 and 32 phases? Doublers.
Doublers are are specially ICs that take one PWM signal and split it into 2. In the process they cut switching frequency in half but they do give you more phases so you do get the extra current capability and get lower operating temps but don't gain anything in terms of voltage ripple suppression. Another trick motherboard manufactures use  that I hate is putting stuff in parallel. There is a good way to do it where they put extra MOSFETS in one phase which basically creates a "super phase" if they are using high end MOSFETs but more often than not they double the number of inductors. This means that inexperienced buyers who count inductors to get phase counts can easily be fooled into thinking that a board has 8 phases when in reality it only has 4 but with each phase having 2 inductors. having 2 inductors on 1 phase is completely pointless. It does nothing that a single inductor couldn't do other than looking more complex than it is.

A perfect example of all of these is the MPOWER MAX motherboard I bought. Here are photos detailing it's VRM design that looks like a 20(20 low and high side MOSFETs and 20 inductors) is driven like a 10(10 driver MOSFETS) and is only fed by 5 PWM(5 doubler ICs and the PWM controller is an 8 phase IR running in 5 phase mode) signals before the doublers.














What about VRMs that are listed as having X+Y phases?
Those VRMs just mean that there are 2 different VRMs one with X phases making voltage A and one with Y phases making voltage B. Many PWM controllers offer this type of configuration natively but often you will see more than 1 PWM controller being used. It all depends on the manufacturer.

So what makes a good VRM?

First of all the VRM has to handle the load. This is very important when overclocking because if the overclocked current draw of your CPU/GPU exceeds what the low side MOSFETs can handle the MOSFETs will burn up. The same happens if you exceed the ripple current of the capacitor bank, You end up with a burnt capacitor. The first is common with cheap AMD and X79 motherboards and reference PCB Nvidia GPUs when pushing the voltage. I've only heard of the second once and that is on the EVGA E-power board when heavily over volting(1.7V) the GTX TITAN-X. You can calculate current capabilities by multiplying the low side MOSFET current rating by the number of low side MOSFETs but with the capacitors you just gotta trust the manufacturer(this is almost never an issue). For a 20% over volt and 20% overclock you will want a VRM with at least 44% more current capability than the stock current draw(~TDP / stock voltage). So for an FX 8350(stock 125W 4Ghz 1.35V) at 5Ghz at 1.525V you would want a VRM that can handle at least 131A. That's 10A more than the typical cheap 4 phase VRM and 31A more than the super cheap 4 phase VRM. Also running VRMs close to spec is bad for them so you'd want 10% head room or 144A.

Now that your VRM doesn't explode when you overclock you need a VRM with low voltage ripple. Voltage ripple basically causes your CPU/GPU to degrade at the rate of the voltage you set however it's maximum clock is tied to the minimum voltage that the ripple creates. So if you set 1.525V and have 25mv ripple you can only achieve an overclock as high as if you had flat line 1.5V while the chip is degrading at the rate that 1.525V causes. To get as little ripple as possible you want the highest number of PWM drive signals coming from the controller as possible at the highest switching frequency possible. So ideally you want an 8 phase controller running in 8 phase mode with a 1MHz switching frequency. The difference this makes is usually minimal but if you're overclocking something with a really high power draw it helps. I also suspect that the stock VRM of the R9 290X has really bad voltage ripple but until I get more equipment I can't test that.

Note:
Haswell and Haswell-e use the FIVR so only the current thing applies and you have to do the calculation differently. As of right now there is no motherboard that will fail from too much current if you're overclocking with air/water cooling. If you're on LN2 you know what you need.

Also I'd like to thank silicon lottery for sponsoring me and this blog. They bin intel i7 CPUs so if you want to buy a CPU that is guaranteed to not suck at overclocking go check them out.

Ladies and gentlemen I bring you the GTX 590 AHOC edition

It's huge 
It's impractical
It might not even work
But now it's finally compelete and ready to occupy the top 1% of GTX 590s

The E-power is fully attached I will be mounting the coolers and doing the final chekups tomorrow

Here are some pictures of the monster I've created.

Well it won't fit into any case I know of the the power plane leading to one of the cores seems to be shorted because I'm reading 0.6 ohms to GND. So worst case I'll have to rewire that plane. I did manage to attach pin 17 properly so if pin 16 really was GND this card will still work.
Good news there is most likely not a short on my GTX 590 because I just checked VCC to GND of my  HD 5850 and got 1.8 ohm. So assuming that the GF 110 core has around 1 to 1.1 ohms of resistance my GTX 590 where the VCC is hooked up to both cores should read 1.1/2 so 0.55ohm which would be in line with the reading I'm getting right now. So now I just have to fix one resistor which went missing and fix 2 VCC wires that I pulled when looking for shorts and the card will be fully operational tomorrow.

Sorry about the photo quality but I can't change that without getting a new phone or a camera.
Also I recently revised the support me page so please go check it out.


As always remember to checkout SiliconLottery if you are considering buying an i7 and overclocking it. They sell binned CPUs so you won't have to worry about getting a crap chip.

Bad news EDIT: Now kinda good

So I've been working on attaching the E-power to the GTX 590 and due to my lack of skill and equipment it looks like the card won't make it.
1. The VCC to GND reads 0.8ohms. IDK if this is wrong or right because that's what it read even before I started soldering and GND to GND is 0.5 ohms so maybe my multimeter just really sucks or the cores of the 590s are really low resistance or I have a short.

2. I managed to spill solder on the PCI-e gold fingers and in the process of trying to clean them I managed to completely destroy 2 of them. Hopefully they were GND pins. If they weren't this card is most likely not going to work.

The fact is I'm not really equipped to do these kinds of mods I have a 50W regulated and 150W unregulated soldering iron. The problem is that the 150W is intended for connecting metals not PCB work however low wattage irons like my 50W do not have a chance when trying to work on circuit boards that contain as much copper as GPU or motherboard. Also the PCB of the GTX 590 is a cramped mess with tiny voltage plains and no where near enough GND plain on the core side to work with. Hopefully the GTX 590 will make but right now I think the card is dead so if there is no weekend post it's because I'm digging it a grave.

Go checkout silicon lottery their my sponsor. They bin i7s so if you want an i7 with which will surely run a certain clock go check them out.

EDIT: So I just checked the pin against Wikipedia and it should be a GND pin. If it is then I'm fine. I will still finish of the E-power attachment however I will need to find new ways to mount the cooler since the wires are everywhere and the card still isn't fully grounded(18/30 wires).

Hide your motherboards and GPUs buildzoid got an E-power

LADIES AND GENTLEMEN I BRING YOU
POWER!

You have no idea how excited getting an EVGA post box made me today. I've been wanting to buy the E-power for 2 years and waiting to do put this on the GTX 590 for just as long. Now you may be thinking. Why did he only buy 1 E-power to power a GTX 590. The E-power is only rated at 400A on the EVGA spec sheet. The thing is that EVGA spec sheet is actually very conservative. The low side MOSFETs on this beast are IR 6725s these are absolute power houses rated at a continuous drain current of 170A at 25C. This derates to 60A at 125C however there are 14 of them. 14 60A MOSFETs that's a total of 840 amps at 125C. Now if you actually use the E-power properly you should be running it cooler. So in fact a single E-power can power an overclocked GTX 590 just fine. That is as long as you don't use LN2 because the E-power does come with a 900A OCP which would trip when using LN2 and 1.5V.
The GTX 590 isn't the only thing I plan to use this one of these. I also want to use it on this motherboard to build an MATX 5+Ghz FX 9590 computer. No I don't need to go see a doctor about having OC sickness. Why wouldn't you want to attach a 59 euro VRM to a 60 euro motherboard.

Also checkout that image layout Blogger's awkward interface is going to teach me how to HTML at this rate.

I would like to thank Silicon Lottery for being my sponsor. They sell pre binned i7s so if you want to avoid having to deal with the variance in CPU overclocking capabilities you can just buy a CPU from them.

A weekend of overclocking.

I didn't manage to get any full article finished for today so I'll just go over some things I've done overclocking wise.

Firstly HWbot.org has launched GPUPI. It basically SuperPi for the GPU. You can either do 1B or 32B to earn points. Currently the best card for the benchmark is the R9 290X. I did my own little session on 1B and 32B.
Here are my results:

The orange background makes no difference. However it runs best with the non transparent windows 7 theme and having as few windows open as possible does help. This is a 99.9% GPU centric benchmark so even those of you with slow CPUs can get competitive scores given that you're running a high end GPU(AMD is better). Just like SuperPi, GPUPI will complain if your OC is not stable and you get results that aren't correct invalidating your run. 1B finishes in about 20-24 seconds on an R9 290X so if your running air cooling it is entirely possible to start the benchmark at 35C and finish before the GPU starts experiencing heat related instability(for me that is about 70C at 1203/1634 +200mv). 32B on the other hand is a beast that takes 20 to 24 minutes on an R9 290X. Heat is a real problem and my GPU ended up running at 82C throughout the test so you can't use temperature buffering to run higher core clocks. However GPUPI does like VRAM clock more than the core clock so you can get a good scores by maxing the VRAM and running a slower core than for 1B.
Obviously this was my first session so I didn't test everything. I will take suggestions on what software tweaks I should try in the comments.

Other than that I also got started prepping my GTX 590 for the E-power that should arrive this week. So here are photos of the card after I gave it a bath to get rid of the mess that soldering AWG 12 wires with a 150W iron leaves:

Hey if you like reading this blog you should go check out my sponsor Silicon Lottery. They sell pre binned i7s so you don't have to worry about a chip that clocks badly.

A tour of my room

 So I just realized that I've never put up images of where I work from and my main computer on here. So here are some pictures of both.

A shot form the door

A shot from my chair

The keyboard is a Blackwidow Ultimate Stealth 2014 and the mouse is a 2012 Razer Naga Epic

This is what I like to call the CPU cannon since it's running a 5Ghz capable 3960X. However ever since the 5960X came out that name isn't as justified as it used to be.

Another shot to get a better view of the board GPU and cooling loop. Yes that PCI-e slot is missing the tab and I did that on purpose because when I install cards like my GTX 590 it only gets in the way.

This is the RAM and FAN test bench R3.0. It's currently prepped for doing another round of thermal paste testing. However the H100 is a pain to mount on AMD systems so until I grow a 3rd arm that's on hold.

And this is my water chiller. I haven't used it for any benching yet because I don't have a block compatible with the R7 260X(I do have a bunch of new zipties...) and I don't want to abuse my R9 290X until I get a different GPU for the CPU cannon.

And here is the before mentioned GTX 590 at my soldering work space. I've pulled of the controller ICs and inductors now I just need to get my hands on a pair of E-power 2.0 boards. The Lego is a magnifying glass holder.

And this is where I do all my photo shoots for reviews. I should probably get larger and whiter sheets of paper.

And here is the obligatory collection of boxes... I need more boxes.

You must've noticed but I've redone the blog's layout and color scheme. If you have any suggestions on how I should improve it you can just leave a comment saying so below.

As always I'd like to thank Silicon Lottery for supporting me. They sort intel i7 CPUs so that you can avoid the silicon lottery and buy a CPU that gets the clocks you want. They also sell the bad chips with a discount so if you're looking either well clocking i7s go check them out.

Thermal paste application methods

(YOU CAN SKIP THIS!)Well I got into an argument. I do that a lot. This one was about thermal paste application methods. I was given a video from Tek Syndicate as proof that I'm wrong. As you can see this video is not scientific in the slightest. The plexi glass and pressure is applied by hand. Plexi glass is flexible. And there is 0 real world testing. So just like I decided to test fans by measuring temps on a real CPU because that's actually useful. I decided to do the same for thermal  paste.

I'm using the stock 125W AMD heatsink because it has a simple and very pressure consistent mounting system minimizing the amount of systematic and human error in the testing. As always I'm also monitoring room temps with my crappy thermometer and using the default IBT for my load. The 750K is at 4Ghz using 1.35V(medium LLC) and the RAM is the G.skill ECO on XMP. My margin of error is 1C° because I couldn't keep the room at a steady 24C° it was swinging between 23C° and 25C°. The thermal paste I'm using is Arctic MX-2.
The test system                                                          The AMD stock 125W heatsink

 
















Results:
1) Pea

Temp margin: 29.38 (higher is better)
Notes: temp basically even swings between 24 and 25C°

2) X

Temp margin: 29.12 (higher is better)
Notes: This was tested mostly at 24C° with a few deviations up to 25C°














3) 5 pointer

Temp margin: 29.00 (higher is better)
Notes: For this one my ambient temps drifted heavily into the 25C area so while this looks like a bad score it's actually about on par with everything else I tested.

4) manual flat

Temp margin: 30.25 (higher is better)
Notes: This one did have a short spell of 25C° at the start but spent most of it's time between 23 and 24C°. So no this method is not terrible but it doesn't have any benefit that I can see and it takes for ever to do it right.

5) Line
Temp margin: 29.75 (higher is better)
Notes:This method went through a full range of temps from 23 to 25C°

6) Smaller pea
 
Temp margin: 29.38 (higher is better)
Notes: Spent most of it's time at 23C and still scores like methods that were tested at 25C. This method is the worst.
















7) No paste

Temp margin: 7.62  (higher is better)
Notes: This test is just to check that the paste actually works. However I must say I'm surprised I thought this would totally overheat but it actually managed to finish IBT at 25C ambient even with the overclock.











Conclusion:
The testing revealed exactly what I thought I would see. Given that you use enough paste you will not see much variation between application methods. I would however like to note that the AMD 125W stock cooler doesn't not have a very high mounting pressure and the AMD CPUs have massive dies compared to intel CPUs. This means that if there was an air bubble right in the middle of the IHS and heatsinks contact it would have very little impact on performance. I currently do not have a cooler with a spring loaded mounting system to test with and I don't have a haswell system but once I get those I will do this testing again because I believe there will be much more variance there. I also believe that large IHS like LGA 2011 will see a difference. One last thing to note is that the viscosity of the paste you use will change the results. IF you have a paste with high viscosity like MX2 it will easily yield to pressure and spread out very thinly on it's own. However if you have a less viscus paste even very high pressure will not spread it out fully. In those situations methods like the flats spread, 5 pointer and X should work better. Also I think 4Ghz was a little tame for the cooler I was using. If enough people complain I'll do a retest at a higher thermal load. You can also leave suggestions for more application methods in the ANONYMOUS and PUBLIC comments bellow.

The results of this testing mean that I'm still undecided about what method is the best but at least I know that none of the methods I use on my builds is bad.

Please go check out my sponsor Sillicon Lottery. They are the #1 reason that I am putting out more and better content and more consistently. They sort CPUs by their overclocking capability so that you can get around the silicon lottery and get a CPU that won't dash your hopes by need 1.35V at 4.5Ghz.

I got a new toy.

 I have mixed feelings about the design. On one had it has insane output current capabilty on the other it's a doubled 5 phase driving 2 parallel inductors and MOSFETs. So it looks like a 20 phase but is only switching in a 10 phase mode and is driven by 5 PWM signals. Luckily it was the cheapest OC centric motherboard I could get new for LGA 1150 otherwise I would've gone for a Gigabyte Z97X SOC-Force.

Also this toy ate into mu budget for fan reviews so those are now on delay.

I know it's a horrible thing to do at the end of such a short post. However I have a sponsor who will be supplying the CPU that will go into the this motherboard. So you should check them out so that they can continue to be this generous. Of course I'm talking about Silicon Lottery.

R7 260X VRM setup and full voltage modding.

I know I've already done 2 other post on volt modding the R7 260X. However none of them are complete and one of them is plain wrong and just like the saying goes 3rd time's the charm so here's a guide made after I actually went and tested everything. If you don't know anything about hardware volt modding do not try this. If you do something wrong your card will be bricked before your brain registers it(your brain takes 5about ms to process visual stimulus). I will make a post about volt modding basics some time soon but not now. Also I'm taking down the 2 old post because of the before mentioned reasons.
So lets run down the VRM setup of the R7 260X.

Here you have the important controller pinouts. For the FB Vcore a 22K ohm variable resistor from the pin to GND will get you a starting voltage of about 1.3V under load. For the FB aux a 22K ohm variable resistor from the pin to GND will give 1.05V under load. Cut the pins or the traces connected to the pins labeled ILIM to disable over current protection for that VRM.So if you cut ILIM AUX the AUX VRM from the first picture will no longer have a power limit.

The VRAM voltage controller is a super simple controller and does not have over current protection. All you need to do to get it under manual control is to hook it up to a 220K ohm variable resistor to get a starting voltage of about 1.585V-1.6V regardless of load.

I would like to thank techpowerup.com for allowing anyone to use their PCB pics.
Also check out my sponsor Silicon Lottery. They sort CPUs by overclocking capability. Right now you can get an i7 4790K, 5820K or 5960X in a variety of  frequencies from above average to extremely rare. For example a 5Ghz i7 4790K.