*OFFICIAL* Alienware 13 R3 Owner's Lounge

Discussion in '2015+ Alienware 13 / 15 / 17' started by katalin_2003, Oct 24, 2016.

  1. propeldragon

    propeldragon Notebook Evangelist

    Reputations:
    122
    Messages:
    536
    Likes Received:
    365
    Trophy Points:
    76
    I have a nvme for my os and m.2 for games. So you're good
     
  2. c69k

    c69k Notebook Deity

    Reputations:
    615
    Messages:
    1,192
    Likes Received:
    894
    Trophy Points:
    131
    This guy completely nailed it with his 'BF5 graphics settings on FPS impact' review.
    TAA high vs TAA low has, unfortunately, minimal impact on FPS, but high impact on picture quality.

     
    Last edited: Nov 18, 2018
  3. propeldragon

    propeldragon Notebook Evangelist

    Reputations:
    122
    Messages:
    536
    Likes Received:
    365
    Trophy Points:
    76
    Those guys are great.
     
  4. forbyefor_gold_eagle

    forbyefor_gold_eagle Notebook Guru

    Reputations:
    36
    Messages:
    65
    Likes Received:
    58
    Trophy Points:
    26
    So Thought I'd give an update after I had issues with my machine and a bit of rambling.

    Last I had posted I had a motherboard failure. Actually 3 total and had to send my 13r3 into Dell AW to be serviced. I got it back about 2 months ago and have since been using it without issue. So far I've owned 2 Samsung VRAMs, 1 Micron and now Hynix. I'm liking the board I got now( more on that later)
    As far as the cause of my troubles, I was never able to pin the cause down with absolute surety but after considering all the possibilities I believe it was the thermal pads I had used. The troubles began when I installed Panasonic graphite thermal pads to replace the stock pads. And while the boards were replaced and subsequently failed, it was the only mod neither I nor the tech had considered as the source of the problem. So needless to say I can't say for sure it was the cause but, I haven't had a problem since. So if you decide to experiment with those, due so knowing there may be more risks and issues than commonly thought.
    I'll say ( and the more experienced members of this forum will probably roll their eyes at this point) but in all cases with these laptops you're best sticking with the stock or stock type pads to avoid issues related to contact fit. At this stage I tried TG Minus 8, fujipoly, Artic, and all of them lack the ability to compress as much as may be needed for these machines. Currently I have TG Minus8 0.5mm on all with decent pad contact. Changing the size was necessary to get the die contact just right for Conductonaut which I insist on.

    Note to 13r3 users to clear up some misinformation out there:
    • the stock thermal pads which top the VRAM and the VRMs are 0.8mm NOT 1.0mm,
    • I measured these both on my relatively new machine and the ones that came on the repaired board (with a brand new heatsink) on the uncompressed part of the pads with Mitutoyo Digimatic calipers accurate to .0001inches. I'm positive of this.
    • the others are 0.5mm.
    What this means is that to even get to the point of where the stock pads begin to contact the heatsink, if you use 1.0mm pads there, the pad must be compressed 0.2mm first. So compressing a fujipoly 17W/mk pad will require something more than 150psi iirc to get to the point of ideal contact then. I don't know what amount the heatsink can compress to but I guarantee you it won't get there. At this point I've spent more than a grocery order on thermal pads and come full circle to "it's best to stick with stock type pads"--LOL, I guess some lessons cost more than others.

    I was finally able to get Wildlands to run at 1080p and the issue was stupid. If Nvidia GPU and Max power were selected in the Nvidia CP, I could not change into 1080p from 1440 in the game. It would window the game within the game itself and mis-align the curser and controls onscreen. Switching to Auto-select and adaptive power and then back after changing res worked perfectly. Stupid.

    So The Hynix board I have now runs a little hotter on the GPU and colder on the CPU than the Samsung I had originally. However, OC'ing this little fella has shown much better results.
    In GR Wildlands (1080p, Very High settings with SMAA +FXAA), which typically tolerates little to no memory OC, from my original stable 175MHz on the clock and 125MHz on the memory I've been able to play hours at time with 200MHz core and 175MHz memery offsets with no instability nor crashes. Clockspeed maintains very close to over 1900MHz at these offsets and 60fps or higher for most of the gameplay. The frame rate does have an occasional dip to 55-58 fps but not too often. What's weird is the dips don't seem to coincide with temps, because frame rate is highest when GPU temp is 80-83c, though the temp does go as high as 83 and then cools down to 79-80c. Sometimes at these settings and temps it'll be running at 70 - 90 fps, depending what part of the game I'm at. (My Benchmark OC setting of +175MHz core and +525 Memory didn't even hiccup on this board where as with my original Samsung board was on the bleeding edge of stability.)
    I use Innovation Cooling's Contact and Test kit to check my CPU/ GPU contact, and I'm glad I did. It took a fair amount of careful bending and fitting and retesting until I could get the optimum fit and thermal pad thickness. Temps got a few degrees better on average and it's peace of mind anyway.
    The pics show initial contact and then initial with final to the right.
    [​IMG]

    [​IMG]

    [​IMG]
    In addition, I performed some mods as well I do believe they have an observable effect though because I can't do before/ after and different machines I can't prove how much at the highest temps and loads.
    • I made what can best be described as PGS heat pipes by putting PGS tape on the bottom of the heatsink blocks and heatpipes and attached them to the fins on both CPU and GPU sides.
    • Replaced the thermal pads with Thermal Grizzly Minus8 0.5mm all over.
    • SSD received PGS from the controller to the copper sheet below and two 1.0mm Artic pads over the length
    • PCH has a PGS heat pipe attached to the heatpipes on the heat sink and two 1.0mm Artic pads over the top.
    • It took some templating and planning but I think you can get the idea from the pics
    • You'll note in the graph (below) how long it actually takes for the temps to rise to operating temperature, this is where I think an observable effect is evident.
    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]

    [​IMG]
    Here's a graph showing performance over about 15 minutes of Wildlands gameplay.
    Settings:
    • 1080p Very High, SMAA + FXAA, Adaptive VSync
    • +200MHz Core, +175 MHz memory
    • Conductonaut GPU/ CPU
    • Thermal GRizzly Minus8 pads 0.5mm all over
    • Above heat mods
    • (The extreme high surges where frequency is above 2000MHz is where I paused for a break or map check.)
    [​IMG]

    [​IMG]
     
    Last edited: Nov 26, 2018
  5. cn555ic

    cn555ic Notebook Deity

    Reputations:
    137
    Messages:
    884
    Likes Received:
    441
    Trophy Points:
    76
    Glad you got it all sorted

    I read your post above and did you say you used Panasonic Graphite pads on the heatsink to replace the stock pads on the computers that had MB failure?? You do know that those pads are conductive correct !
     
  6. forbyefor_gold_eagle

    forbyefor_gold_eagle Notebook Guru

    Reputations:
    36
    Messages:
    65
    Likes Received:
    58
    Trophy Points:
    26
    Thx. Yeah I was aware they are said (by consumers and tech enthusiasts) to be conductive. For the record Panasonic's specs only state so on one of their brochures, that I could find --for the product that I used. And it says these are 'conductive' and 'insulation should be used if in direct contact with components that are normally insulated' --note it doesn't specify if it is electrically or thermally that they are referring to-meaning the specs are ambiguous, though I assumed it was electrically. All other info on these suggests they are appropriate for the application including their drawings on how it should / could be used. Though, knowing this, I tested them for conductivity/ continuity with a very good DMM and could not get a reading. And took care to prevent that they touch legs or anything on the board. Anyway, like I said, no proof that was the cause, though I suspect it might have been and I'm not willing to give it a shot again, haha. Dumb me. When in doubt err on the side of caution, I guess. At least now I lived the stupid, so none of you have to.
     
  7. propeldragon

    propeldragon Notebook Evangelist

    Reputations:
    122
    Messages:
    536
    Likes Received:
    365
    Trophy Points:
    76
    Definitely post a full hwinfo64 screenshot. Playing BFV for hour and a half or something like that! What were your cpu temps like? Kinda hard to see.
     
    Last edited: Nov 22, 2018
    forbyefor_gold_eagle likes this.
  8. Marksman30k

    Marksman30k Notebook Deity

    Reputations:
    2,080
    Messages:
    1,072
    Likes Received:
    179
    Trophy Points:
    81
    Pyrolytic graphite is conductive although I would imagine in a more specific direction/form, as with most forms of graphite. It's the whole reason why we used to be able to pencil mod overclock back in the day.
     
  9. forbyefor_gold_eagle

    forbyefor_gold_eagle Notebook Guru

    Reputations:
    36
    Messages:
    65
    Likes Received:
    58
    Trophy Points:
    26
    Sorry I didn't save a screenshot of the HWInfo64 at that time. This graph is made from HWInfo64 recording for about 1hr 17 min. The whole graph of the previously posted graph is below. The GPU freq. is on the right secondary axis, all other lines reference the left primary axis.
    I don't save screenshots of the read out since it looks kind of disingenuous to me since you have no idea whether or not I Iet the machine idle for an hour of that time, so I make graphs.
    • CPU temps average at 68 and change for all cores put together for that 1.25 hrs of Wildlands at 1080p very high settings.
    • Cores 0, 1, 2 average all the same temps, min and maxes
    • Core 3 averages 2 degress celcius lower on average--I guess the contact on this one is a little better as you can see from the pics above
    • GPU frequency-(discluding the initial 0 and 2100MHz readings when the game is starting up, since that's not representative of the 1060's performance under load)--average 1858.54MHz for the 1.25 hrs. This is actual performance under load, in a demanding game. I haven't tried the undervolt OC method on this board yet--I tried that on the original Samsung and was not impressed, it ran hotter and lower frame rate, which should be expected, I guess. So far my limited experience tells me the dynamic response of boost can squeeze better results out of the GPU with an offset at least in a laptop where you don't have case volume and bigger sinks/ fans and more power on your side.
    [​IMG]

    Yeah, having been aware of that, like I said IF it was the cause, well it was a stupid move. I don't know for sure. Not trying to alarm anyone, just being too honest I suppose in hopes it can help anyone who may be as bull headed as myself. ------>Can I join the club of famous guys here who have bricked (several) machines for a dumb move? Haha. Anyway Not sure why Panasonic gives all indications the product is appropriate for this type of application if it is not, it's vol. resistivity on one of the spec pages is slightly out where it would be considered a good conductor-about in between being a good conductor and a good insulator, but closer to the conductor side. Meh, water under the bridge now. She run's good. Thanks for all the input. I'm going to have to look up pencil mod overclock, never heard of that!
     
    Last edited: Nov 26, 2018
    c69k, propeldragon and Vasudev like this.
  10. c69k

    c69k Notebook Deity

    Reputations:
    615
    Messages:
    1,192
    Likes Received:
    894
    Trophy Points:
    131
    Hi all, I installed XG Nvidia latest driver and it DOES overclock and undervolt better than default driver.

    Default Nvidia driver: core: 1810 MHz, 875 mV, memory +200
    XG Nvidia for laptop: core: 1797 MHZ, 850 mV, memory +200

    My Nvidia 1060 is pretty hungry compared to what most 13 R3 users report, so I believe that majority of gamers would be able to tweak it with much better results.

    I see more fluid gameplay and it feels like latency is much lower than with default driver. When I move mouse, I definitely feel reaction on the screen is faster. I tested for 3 days already, not going back to default drivers, unless some problems appear.

    I have framecap with all games @ 60 fps as I hate lag with V-sync. I definitely noticed less horizontal tearing, especially in games like No mans sky and Everspace.

    https://www.reddit.com/r/XtremeG/

    Thanx @6.|THE|1|BOSS|.9 you made my day!!!
     
    Last edited: Nov 23, 2018
Loading...

Share This Page