4+1 vs. 8+2(8+1) Phase Power Design

[Entropicity]

Dr. MOS is, in short, a technology where a pair of mosfets (high & low) and a mosfet driver are integrated into one circuit. Instead of three (driver, mosfet low, mosfet up) you have one. They have nothing to do with phase count, but MSI says they run cooler and output less ripple. MSI is not the only company that uses integrated mosfets; Gigabyte has them too, for example on GA770T-UD3P.

Are you sure about that? I've never heard or seen this model. Could you provide me a link to it?

[G.Foyle]

Here you go: http://www.gigabyte.com.tw/Products/...ProductID=3096
I used it in my Regor and Callisto review.



Here's the PWM area with integrated mosfets:

[Entropicity]

Dude, I believe its just marketing of normal mosfets. Maybe better components but nevertheless it aint driver mosfets.

Here are some shots of Driver Mosfets. Notice they have combined the 3 components into a single chip for each phase?

[G.Foyle]

You are right, these are not driver mosfets. My bad
But I'm sure I have seen such fets somewehere not on MSI board...

[Entropicity]

You are right, these are not driver mosfets. My bad
But I'm sure I have seen such fets somewehere not on MSI board...

No worries man.

Another company that has used non traditional power circuits is DFI. I believe they were one of the first to use digital pwm mainstream and it seems EVGA has started to do so too.

Just curious how DrMOS and Digital PWM stack up to traditional mosfets since Asus and Gigabyte seem to swear by them (in huge numbers) in their top range boards.

[erek]

ASUS M4A89GTD PRO/USB3

is that a true 8+2 Power Design?






here's some data on the design:

[EPU (ASP0902) PWM + PEM (ASP0910) PWM + uP6282AD Driver]

ASUS M4A89GTD Pro/USB3 Intel CPU from the mainboard, the power supply have been used primarily by a combination of EPU and the PEM CPU up 8Phase, memory support 2Phase configuration has 8 +2 Phase Power Supply. Also, Phase 2, the MOSFET and 5 per uP6282AD MOSFET driver is a combination. EPU processor clocks and voltages depending on the load, lower power consumption, ASUS EPU-6 engine that can be reduced (CPU, memory, chipset, HDD, VGA, paenkeonteurol management) is the key.

[erek]

they say..


ASUS new improved ASP0902 series PWM chip, with PEM form 8 +2 Phase Power Supply System.

[haylui]

to me, 8+2 phase design is to evenly shared the load over a parallel channel
so that each MOSFET only take up the load half of what 4+1 phase design is taking;
hence, this would lower the operating temperature of MOSFET while user pushes the voltage when overclocking.

[chew*]

afa8ik there is not true 8 phase PWM......they are all 4+4+2 or whatever....thats how it works.......ati once made an extremely well laid out true 6 phase back in rd 480 580 days.....

[[XC] Oj101]

Quality phases mean more than quantity

Actually there is NO NEED for 164 phases that companies put in nowadays mobos.... FOUR (4) are enough with nowadays materials and maybe SIX (6) with the upcoming Gulftowns....
The 1024 phases is a "marketing trick" nowadays....
You will buy MY mobo because I put 2048 phases in my motherboard....

Someone please calculate the Wattage of a P4 Gallatin at 2.23VCore at 4600MHz with ONLY FOUR phases and tell us what you see.....

The Legend Speaketh (TM)

[BadNizze]

afa8ik there is not true 8 phase PWM......they are all 4+4+2 or whatever....thats how it works.......ati once made an extremely well laid out true 6 phase back in rd 480 580 days.....

This has gotten me to read thru old Grouper/Manta/Sturegon threads, but I could only find 4 phase PWM´s. Wich board hade the 6 phase PWM?

[DedEmbryonicCe1]

There seems to be a lot of confusion about ASUS 8+2 power circuitry.
Both Crosshair 3 and M4A79T-Deluxe use L6740 PWM controller made by ST Micro.
It's a 4+1 phase controller, which means power regulation is done in 4 phases for core voltage and in 1 phase for NB voltage.

That is the same controller used by Biostar for the TA790GXE 128M and presumably all the other 4+1 boards they released at the same time. The TA890FXE is now on the L6717 controller. My experience with the L6740 from my TA790GXE 128M was that it was best to disable the active phase switching when overclocking. With phase switching enabled I could crash/shutdown the board stopping a Prime95 run whereas the same frequencies and voltages were perfectly stable in the same scenario with phase switching disabled via the BIOS.

[Smartidiot89]

afa8ik there is not true 8 phase PWM......they are all 4+4+2 or whatever....thats how it works.......ati once made an extremely well laid out true 6 phase back in rd 480 580 days.....

Feels like I am nagging, but how about the ASUS Crosshair IV Formula? Is that also a 4+1 + 4+1 phase design?

And what about those Intel board with 24 and even 32 phases are those 6 and even 8 pairs of 4-phase designs? Seems kinda stupid to me... Not disbelieving you, I am sure this kind of marketing works great.

[haylui]

Feels like I am nagging, but how about the ASUS Crosshair IV Formula? Is that also a 4+1 + 4+1 phase design?

And what about those Intel board with 24 and even 32 phases are those 6 and even 8 pairs of 4-phase designs? Seems kinda stupid to me... Not disbelieving you, I am sure this kind of marketing works great.

IMO, if they make the power phase parallel, it helps when you up the voltage to a certain level since the current were shared by more MOSFET, those MOSFET definitely would run cooler than lesser phase

[BeepBeep2]

ASUS/GIGABYTE's 8+2 phase is indeed a 4+1/4+1 phase design. It was an AM2 socket design contstraint and the board is only able to have a true 4+1 phase design. That didn't stop manufacturers from doubling up though.

As for my board's 6+2 design and BIOSTAR's 6+2 design I'm not exactly 100% sure on how they achieved that.

[chew*]

This has gotten me to read thru old Grouper/Manta/Sturegon threads, but I could only find 4 phase PWM´s. Wich board hade the 6 phase PWM?

some PWMS chokes and caps count does not equal phase

[chew*]

ASUS/GIGABYTE's 8+2 phase is indeed a 4+1/4+1 phase design. It was an AM2 socket design contstraint and the board is only able to have a true 4+1 phase design. That didn't stop manufacturers from doubling up though.

As for my board's 6+2 design and BIOSTAR's 6+2 design I'm not exactly 100% sure on how they achieved that.

thats not 6 phase...........

[BeepBeep2]

thats not 6 phase...........

*scratches chin*

[DedEmbryonicCe1]

The Biostar FX board is 4+2. Although the stock cooling solution is possibly worse than what I had modded onto my TA790GXE 128M 4+1 board the FX power supply is still much stronger in practice. The 4+1 board I have can handle quads fine but not a X6 at high voltage. The TA890GXE is 4+1 too but judging from other users here it is definitely improved over the previous generation of 4+1 boards Biostar had.

[TheBlueChanell]

I think you are all reading way to far into this. Everyone's been on the phase kick since the MSI with 4+1 kept blowing up. The key is to keep the PWM's cool. I've pushed my GD70 pretty hard but I've always had active cooling over the NB/PWM area and I've not had any issues. I plan on pushing it harder when I get around to picking up a Thuban before replacing it with a 890FX board. There was an article, I believe on bit-tech, a while back regarding "Phase" design. 4+1, 4+4+1 etc.

Less is more can apply here as long as the quality of components is high, and to those worrying about your boards being 4+4+2/1 whatever and not true 8+2 dont worry and just enjoy your board and the overclock that comes with it. If this thread was never made a majority of you would of never known the difference and/or care.

[BeepBeep2]

I think you are all reading way to far into this. Everyone's been on the phase kick since the MSI with 4+1 kept blowing up. The key is to keep the PWM's cool. I've pushed my GD70 pretty hard but I've always had active cooling over the NB/PWM area and I've not had any issues. I plan on pushing it harder when I get around to picking up a Thuban before replacing it with a 890FX board. There was an article, I believe on bit-tech, a while back regarding "Phase" design. 4+1, 4+4+1 etc.

Less is more can apply here as long as the quality of components is high, and to those worrying about your boards being 4+4+2/1 whatever and not true 8+2 dont worry and just enjoy your board and the overclock that comes with it. If this thread was never made a majority of you would of never known the difference and/or care.

6-core?

All I was saying in the other thread is that MSI should NOT need a fan over the NB/PWM area. They design those heatsinks for a reason. It doesn't say in the manual to buy a fan separately and stick it over them. It's supposed to stay cool. If ASRock, ASUS, Gigabyte, ECS, DFI can do it then why can't MSI?

[TheBlueChanell]

6-core?

All I was saying in the other thread is that MSI should NOT need a fan over the NB/PWM area. They design those heatsinks for a reason. It doesn't say in the manual to buy a fan separately and stick it over them. It's supposed to stay cool. If ASRock, ASUS, Gigabyte, ECS, DFI can do it then why can't MSI?

If my memory serve's me well DFI had issue's with pwm's blowing up on the M2R boards. The one's that were stable, were as rock solid as you can get.

On the MSI if you run the CPU stock, or push light voltage I don't think you'd need a fan but if you're pushing heavy clocks and volts you should probably throw some cooling on there just like any other component that you're overvolting and clocking. Especially on air. It just seems like good practice to me but idk :/ ymmv.

Edit: Forget to tell you what I was running. It's a 965BE 140w, if you read my earlier post more carefully you'll see I mention moving to a Thuban. I'm been doing some other stuff so it's not high on my priority list atm.

[crazydiamond]

i beat the hell out my 790fx-gd70 & 965be - over 1.54v 24/7 & constantly over 1.6v benching, never w/ active cooling on pwm's - and stable as a rock. they were always hot as hell too because they were hiddin out of the case airflow behind my 120mm rad. the gd70 was prolly the best board ive owned regardless if it has 0+1 phases

[[XC] Oj101]

Anyone care to work out the wattage on an FX-55 at 2v/3.92GHz max (albeit idle) and 1.68v/3.45GHz 100% load running F@H 24/7? I get 183w, which was stable for MONTHS without the proper cooling on the PWMs - and that was only four phase. It really isn't the NUMBER of phases but the quality. Now take George's example of a P4 3.4GHz EE Gallatin core at 4600MHz and 2.23v and I get 288w, again done on 4 phase back then.

What's the TDP on an X4 965BE C2 at 4GHz 1.54v? And an X6 1090T at the same speed? I don't understand why the number of cores affects anything if the TDP is similar. Please someone enlighten me

[zanzabar]

What's the TDP on an X4 965BE C2 at 4GHz 1.54v? And an X6 1090T at the same speed? I don't understand why the number of cores affects anything if the TDP is similar. Please someone enlighten me

the thulban parts have a better design than the denab, the l3 and IMC (or the cpu-NB on the cpu settings) make up most of the TDP and amd specs teh TDP for a batch of parts not the individual. so i would guess that they have an over estiment on the 965 c3 and that the 1090 comes close to 125W, but at least with my chip amd gave it a VID of 1.3V but it can do stock at 1.18V so it is lower watt than amd is saying it is and they over volted it. for amd TDP is a measure of the cooling needed not the power consumption so to really tell u need a board that can read the amps going to the cpu (i know that the gd70 can in the bios it would be nice to see some1 tap that setting.)