Dfi P35

Quote:

---

Originally Posted by Revv23

DFI May have done it right but thus far what you say about digital PWM isn't true, it runs hotter, is less stable, is hard too cool, and easy to burn out. This holds true for abit's, dfi's, you name it... digital PWM so far is not for extreme clocking. not to mention it seems to only be avialable on expensive boards so i would assume that it adds cost without reason.

Like i said in an earlier post, one of the most stable PWM circuits out there (and coolest running) is on abit's AW9D, which happens to be a 4-phase analouge solution.

---

You have to understand the technical benefits of each solution. Analogue is less accurate than digital, is considered less reliable due to the fact that is uses caps, and makes the socket area much more cluttered which is bad for those who wish to use very large heat sinks or phase-change/dry ice/ln2.

However, digital PWM is limited by it's phases, meaning if a digital solution has however many phases to deliver however many amps to the cpu, it will deliver up to that much and absolutely no more. An analogue solution is much more flexible and can generally extend far beyond the power requirements it was originally designed for. Again though, when pushing an analogue solution like this to it's max, you're running the risk of blowing a capacitor or suffering serious voltage fluctuations.

Digital in the end I would say is better because of it's accuracy and space saving. As for it's thermal characteristics, it should get better, especially if board designers decide to go even more overkill with the phases.

Quote:

---

Originally Posted by EnJoY

You have to understand the technical benefits of each solution. Analogue is less accurate than digital, is considered less reliable due to the fact that is uses caps, and makes the socket area much more cluttered which is bad for those who wish to use very large heat sinks or phase-change/dry ice/ln2.

However, digital PWM is limited by it's phases, meaning if a digital solution has however many phases to deliver however many amps to the cpu, it will deliver up to that much and absolutely no more. An analogue solution is much more flexible and can generally extend far beyond the power requirements it was originally designed for. Again though, when pushing an analogue solution like this to it's max, you're running the risk of blowing a capacitor or suffering serious voltage fluctuations.

Digital in the end I would say is better because of it's accuracy and space saving. As for it's thermal characteristics, it should get better, especially if board designers decide to go even more overkill with the phases.

---

What about the power loss that happens in every PWM phase? I read that low power/underclocking guys like boards with as little phases as possible because every phase has an efficiency of (lets say) 80%. So every additional phase increases the current that can be delivered but also decreases the overall efficiency of the PWM. Is this also true for the digital PWMs?

Well it's a parallel circuit, not series, so you wouldn't be multiplying the efficiencies. With more phases, you'd be splitting the load, so the efficiency would be the same, no? That said, when you push more load through a phase, I'd imagine it has an efficiency curve. So by using more phases, depending on where it falls on the curve, it could be more efficient, no?

Quote:

---

Originally Posted by mibo

What about the power loss that happens in every PWM phase? I read that low power/underclocking guys like boards with as little phases as possible because every phase has an efficiency of (lets say) 80%. So every additional phase increases the current that can be delivered but also decreases the overall efficiency of the PWM. Is this also true for the digital PWMs?

---

That I cannot comment on because I honestly do not know. However, generally the loss of efficiency and overlooked in multi-phase setups due to the fact that it is generally overclockers whom are utilizing them, and the overclockers don't care about low-power.

If I could guess though, I would say that digital PWM offers greater efficiency and far less loss.

the US product page for it is now up

http://us.dfi.com.tw/Product/xx_prod...YPE=LP&SITE=US

Looking at the PCI-E Slots I guess my Raid Card wont work unless its at X4 and I take a performance hit due to NO X8 SIgnal????

Some of you guys with better insight into the P35 could you please add some comments to my question?

i hope these start showing up in shops soon.

I just downloaded the manual from the web page. I can tell you that the whole "Transpiper heatsink" is going to be massive mistake on their part. I've installed many a complicated heatsink in my time but I can tell they are going to have massive RMA's from noobs screwing up. And especially if you take a look at the way you have to mount the "transpiper" in order for it to be outside the case.

They Use a copper plate that goes on top of the CPU (WTF?). This plate extends from the CPU to the PWM heatsink, and the then the heatsink mounts to that copper plate. And to use that copper plate with the CPU heatsink you automatically cant use the stock cooler because you have to mount a base plate under the CPu to use heatsinks that are screw on type.

In other words, it's a PITA

does this heatsink come pre mounted or is it put on by the user? I hope it isnt premounted so can just put on HR-05's etc. I am going to go get the manual myself and have a look. Thanks for the info.

It doesn't come pre mounted. Download the manual and you'll see:

DFI P35 Manual

I would just mount a fan on it instead of using the transpiper heatsink as I suspect I wouldn't have much room for it in case I get this mobo, depends on the price though.

WTF are they thinking here :confused:
anyone got an explanation why they have a copper plate there

http://img405.imageshack.us/img405/3...board01nw6.png

http://img520.imageshack.us/img520/9...board01ib1.png

http://img529.imageshack.us/img529/3...board01kl0.png

My guess is so that the "transpiper" also dissipates CPU heat? But then again, couldn't it also transfer NB/SB heat tot eh CPU? The design is just flawed IMO. They should have just kept it simple IMO.

but that means you also have to have two layers of thermal paste between heatsink and CPU

would that contact also affect the temps :confused:

if this thing drops in a week or two, I will buy it.

Any later will have to get something else...

I am very interested in the onboard teaming more than anything really...

but if I buy this, the first thing I do is throw out the heat pipe system... or atleast dismantle part of it.

Quote:

---

Originally Posted by dinos22

WTF are they thinking here :confused:
anyone got an explanation why they have a copper plate there

http://img405.imageshack.us/img405/3...board01nw6.png

http://img520.imageshack.us/img520/9...board01ib1.png

http://img529.imageshack.us/img529/3...board01kl0.png

---

Uhhhh, this thing wouldn't block say a Thermalright Ultra 120 eXtreme would it?? :confused:

I have been wondering as well how that copper plate is going to affect mounting a high end heatsink. Surely it is going to be thick and be raised right off the IHS of the cpu. I am not sure if this will hinder attaching a heatsink or create more pressure which could be a good thing. A layer of thermal paste will be needed between the cpu and plate as well as the plate and heatsink. If the surface condition of the copper plate is flawed that could also affect temps. We really need some real world tests done with this new board design.

yeah, when I order... Abit IP35 Pro... That's just messy!

Ryan

I think DFI have seriously fu*k up on this heatpipe rubbish, I wonder how much this scrap metal adds to the retail price :rolleyes:

It looks like the 'copper' ?? plate is used more to cool the digital power chips than anything else which makes me think they are going to be HOT :mad: :down:

I wonder when they are going to put the motherboard thats Oskar and Andre used on the DFI website.

it is clearly a DFI Infinity, while this is DFI Lanparty UT...

Quote:

---

Originally Posted by sergers

I wonder when they are going to put the motherboard thats Oskar and Andre used on the DFI website.

it is clearly a DFI Infinity, while this is DFI Lanparty UT...

---

no it is the lan party board, they used an infinity heatsink for the bench run as they stated in the posts.
all this heatpipe stuff is just going in a drawer when i get it.

Yip I am with jimmyz on this one. My heat pipe assembly is not even going to make it out the box the motherboard comes with. Just hope release date of this P35 is before the first X38 chip-set motherboard is released by Asus or Gigibyte, but with DFI you never know.

With respect to the heat pipe and the offending copper plate - that is only if the intention is to mount the transpiper out side the case. Page 41 of the user manual. If you are going to be mounting the transpiper inside then the copper plate will not be needed.

Judging from the pictures what after market cooling looks to fit on both the north and south bridge chips if we were to remove the heatpiping from the board entirely. Will an HR-05 fit on the North with an HR-05 SLI on the South?

Ive never used an air cooler on an Intel chip...would the retention clips still work ok, despite the extra mm or so added to the CPU height?

It looks like a dumb idea, but DFI will have seriously THOUGHT about this, so it either works or... something bad got into the aircon system and no-one was thinking straight for a few days.

Quote:

---

Originally Posted by K404

Ive never used an air cooler on an Intel chip...would the retention clips still work ok, despite the extra mm or so added to the CPU height?

It looks like a dumb idea, but DFI will have seriously THOUGHT about this, so it either works or... something bad got into the aircon system and no-one was thinking straight for a few days.

---

the manual states not to use retention clip style heatsinks.