**The Official Studio XPS 1645 Intel Core i7 "Owners Lounge"

I second this suggestion. If we get numerous people to try these options then we can rule out what is causing the problem or not (adapter, BIOS, etc.)

 

I've always disliked 3D Mark benchmarks because of that reason. I've always had large variances of 10-15% on any computer I've ever ran it on. It always runs better the second time too.

Its really just an estimate. Nothing is more solid then a benchmark loop in an actual game ran several times minimum to get an average like the crysis bench that was ran.

 

Could also rule out most of the software stack by reproducing the problem on Linux or something.

 

Hello

I just recieved my new 1645 yesterday, but have been lurking this thread since it was first created.

Thought i might tip in with some testing of my own, so here goes:

Specs:
Windows 7 x64 Pro
Core I7 720qm
4gb DDR3 ram
16'' RGBled
ATI mobility 4670*
500gb sata

*I have installed latest ATI drivers (8.671-091104a-091527C-ATI) using the MobilityModder from Driverheaven. All the other drivers I've just had windows handle.

Performance settings:
Power plan - High performance
Powerplay - Maximum performance

1. Test on 100% charged battery system WITH the 90W AC Adapter @ Full Brightness, Wifi+Bluetooth On, etc. (Max Performance)
6396
sm2 2240
sm3 2773
cpu 2898

2. Test on 100% charged battery without the adapter @ Full Brightness, Wifi+Bluetooth On, etc. (Max Performance)*
7098
sm2 2539
sm3 3106
cpu 2937

3. Test on <90% charged battery without the adapter @ Full Brightness, Wifi+Bluetooth On, etc. (Max Performance)
7026
sm2 2489
sm3 3092
cpu 2930

4. Test on <90% charged battery WITH the 90W Adapter @ Full Brightness, Wifi+Bluetooth On, etc. (Max Performance)
6280
sm2 2215
sm3 2749
cpu 2695

So it seems I'm experiencing somewhat similar numbers as everyone else, atleast I'm seeing the same pattern with battery vs. adapter.

I have also been running some wPrime tests aswell, using CoreTemp for monitoring meanwhile. While running wPrime 32M with only one thread, it doesn't seem turbo boost is kicking in. Not according to CoreTemp atleast. Fault with CoreTemp, or shouldn't turbo boost be kicking in at all in a situation like that ?

On a side node. I'm experiencing sporadic flickering with the screen, when brightness is below 70ish %. Seems like the brightness switches between two different brightness levels very fast, if that makes sense? It's not very apparant, but now that I'm aware of it, it's very annoying. Anyone else experienced something similar ?

All things aside though, I'm very happy with my laptop so far It's quite an upgrade from my previous 5 year old laptop.

 

the more reports im seeing, the less worried I am and the more convinced I am that the laptop is just not setting itself to the proper performance setting when it is plugged in to the AC adapter. the fix is probably going to be correcting the line of code that tells the BIOS what power/performance state it should be using when on AC power.

No big deal IMO. Thats something that will likely be fixed very quick as long as we all report it to Dell. I'll check mine out and submit a tech ticket when i get home tonight.

 

It's a 99% power supply problem - while playing game on max Brightness - got lags, changed to minimum and the game is running perfectly (Left 4 Dead 2) on my RGMLED LCS screen.
Back again to max brightness and game is slowing down every 1 min.

1% is for graphic drivers, cos I'm not sure if they got anything to do with LCD

Will call Dell Helpdesk tomorrow (UK) and ask for power supply replacement.

 

I called earlier and they have a case open for this issue and they're troubleshooting. Supposed to hear back from them in a few days. But you can call in and add more details to the case, maybe if more people call it'll be a higher priority. They were interested in what games I tried for reproducing. I'm pretty sure it affects any modern 3D game though.

 

Seems like it indeed IS the 90W power adapter indeed not being able to handle the power demand for this laptop. And this would make sense when adding up what the individual components need (on their theoretical max.):

Core i7 processor TDP = 45W
ATI 4670 Graphics = 35W

These 2 components together are already 80W.

Add to that the WLED screen which will use a few W too, even more on full brightness, and with an RGBLED screen added some more on top of that.
The SATA HDD spinning at 7200rpm is also likely to use a couple of Watts of power - moreso than the SSD drives.
Add to that the other components, the odd peripheral USB mouse and keyboards and already with these you'll be quite a bit over 90W I'm guessing.

That's not to mention the battery charging requirements...

So yea I think we should point this out to Dell and file as many complaints as possible so they change the adapter to at least 130W as a standard on all new (and not yet shipped) orders.

 

Btw we're still waiting for some people to post results with a 130W adapter

 

Yeah but TDP isn't really intended to be used to calculate the maximum power use, it tells you how much heat needs to be dissipated (probably with a margin added on).

For example, the upgraded i7, the 820QM which runs at a higher clock speed, ALSO has a TDP of 45. Yet running at a higher clock speed means it uses more power. So they can't both have a maximum of 45 watts. In reality the power use will be much lower than TDP.

 

Ok even then when adding up all the theoretical maximum wattage of the individual components + the battery charging its quite possible to be over 90W..
And even if it is ~90W a 90W adapter is really not sufficient since there will be slight variances in adapters themselves...

 

Seriously, the next person who uses TDP to indicate how much power the laptop is drawing loses the game....

....forever.

 

The Thermal Design Power (TDP) is the average maximum power a processor can dissipate while running commercially available software. TDP is primarily used as a guideline for manufacturers of thermal solutions (heatsinks/fans, etc) which tells them how much heat their solution should dissipate.

It is NOT a measurement of how much power the CPU consumes. To know the actual power draw, we need the voltage which we have (up to 1.4v) and what amperage the CPU is running at. figure this out and we'll have the actual wattage (which may or may not be around 45watts)

EDIT: What I'm trying to say, and i dont think i did it well above, is that your CPU even when gaming is not running constantly maxed out. It will occasionally spike to 45watts of TDP, but it isn't goign to stay there unless you are running a stress test or rendering graphics.

You guys seem pretty confident that a larger adapter is going to provide more power to the unit, and I just dont think that is how it works.

 

do realize the heat generated has to come from the electricity supplied. and as far as i know there isnt a difference between electrical watts and thermal watts.

adding up the wattage tdp even if it doesnt heat that much. What it shows is that even if it doesnt reach the max, it seems uncomfortably close. especially when you add in screen/fans/ and other sources of power draw. Correlate that to when adjusting the brightness of the display affects the performance greatly, it would seem the system is at the edge if not a bit past. especially those tested using an external monitor showing zero slow down. All evidence that supports its the ac adapter even more so.

IMO we cant say 100% that its the adapter until we test a bigger one, and/or measure power draw on the bigger one.

 

The CPU isn't constantly sucking 45 watts of power. Thats the point I'm trying to make.

Also, you have to remember that the battery is going to supply the same amount of power. I think that if there is any issue with hardware, its going to be more then just the adapter. If the system is so close to its max, then the battery would never charge. The battery clearly charges, and clearly powers it well enough. Its a power management issue. I'm sure of it.

But we'll see. some one has a 130 on the way. i'll accept being wrong if the 130 solves the issue, but i'll be one heck of a surprised computer geek

Also, as much as I hate wiki, http://en.wikipedia.org/wiki/Thermal_Design_Power

Please please understand that TDP is really referring to what a CPU cooler needs to be able to manage in a worst case scenario, and doesn't directly represent power draw or consumption (although it is a good indicator)

 

well from the test that khal did. it seems the battery stops charging under load and also doesnt discharge underload. and begins charging when not under load. Thats the problem where seeing is load.

i agree with you though. we cant draw conclusions off of tdp alone. what i think it does is it doesnt disprove the power adapter idea. i agree as well, we think we know the solution the only way to back it up 100% is to use the solution such as the bigger adapter once someone gets it.

 

i should have it this week.

 

I was under the impression the benchmarking IS taxing everthing to the max to see its limits....Is this not true? I want to know its limits.
If all you do on it is watch fish swim around in a screen saver then ya, you will never have to worry about its limits and/or current draw.
It seems you guys are getting close to a solution. thanks for doing that.
sorry i cant help (still waiting)
Am i wrong about this?
I apologize if so,

 

Well, I am another person who ordered an xps 16 i7 in september only to have dell send out a delay notice the day it was supposed to be deliverd (oct 27th) and then again in november they cancelled the order, then they recreated the order and set a new date of december 14th. I then immidiatly canceled and went to best buy and got the envy 15.

Although the envy 15 has its own issues I have found that the lack of onboard dvd drive is no problem at all and being able to use 2 ssd drives or one ssd and one mechanical drive is a great feature. I am very happy with the display. it is leaps and bounds better than the mac book pro I was using.

I am not trying to convince anyone that the hp is better it has its own problems but I will say that the constant delays was really deplorable.

 

stress/load testing is the only way to max out your computer's load to test temps, or stability.

benchmarking tests you computer's ability to do certain tasks and assigns a rating/number to the result. Benchmarking will not really max your whole system, but rather whatever you're testing (CPU, GPU, and so on)

If you run a game benchmark its going to stress the CPU mostly, thats why 3D Mark has a seperate set of CPU tests. but it doesn't really max both out at the same time.

 

i agree that yes, it cpu won't always be at max power....but im not always having issues either. I'm only having in some sections, of many games.

if nothing else, the 130w should charge my battery faster. ^^

 

ok, thanks.
I did find this, does it fit in somewhere?
Power for the LCD display is usually done seperately. Generally, LCD backlights require high voltage to operate. A component that might be called a power inverter will step the voltage up to what is needed for the backlight. Sometimes this inverter is a discrete and replacable component, and sometimes just a chip on your mainboard. It is also involved in the brightness control.
From here: Fixing Laptop Power Problems

 

I thought id add, i see on my desktop crysis will max out a single core just while playing. seeing how turboboost is to essentially overclock the cores to keep it at max power, i would assume the cpu would be working harder then conventional core2duos etc. if that makes sense? I dunno just some ideas for you guys that our testing.. anyone have TF2? that will max multiple cores with multicore enabled.

tenax i dunno if that applies so much i think wled is different technology. but there will be a lot of things like that to convert the power adapter voltage to cpu voltage, memory voltage etc etc..

 

In order to further torture the system i ran prime95 with 8 threads and Crysis benchmark at the same time.

Under Battery, the processor worked at 9-11x, with the GPU not switching to 2D clock.

Under Charger, the processor almost stayed at 7X, with the GPU starving for CPU orders.

 

is there anyway you can monitor the actual voltages at the cpu? using cpuz or something. see if theres a difference?

nice test btw

edit what if you do different brightness settings?

 

I think this might be out of date though, the displays that are backlit by cold cathode lamps (CCL) needs high voltage of course. But since they are using LED backlighting now, high voltage isn't needed.

 

A few misnomers have been floating around here:

i7-820QM and the i7-720QM can have the same TDP because of binning. I don't have time to explain that, just google it. It's just a better wafer that is more efficient, thus it can be clocked higher and have the same TDP.

Battery versus adapter giving "same" power: I think this is the meat of the issue. What we're saying is that when it is attached to the adapter, it can only eat 90W in x amount of time. But it can use as much power from the battery, possibly at a faster/higher rate.

I don't have time for all these things, unfortunately, and just got sick today. It kind of blows. These tests require at least an hour's worth of time that I just don't have right now. No call from the "executive technical support team". TBH, I'm just going to wait for that call. I have all the evidence I need: it is not MY burden to do all these tests, damn it, it's DELL's! Not trying to discourage you guys, but my time has been cut short pretty bad now: the whole day has gone by and I have maybe done a 1/3 of the work that I had allotted out for the day? Which means what I don't finish today, I get to do tomorrow. While I'm still sick. Ugh...

I know about benchmark caching, but I'd LOVE to see someone explain the 40% performance increase with the battery with the reason being caching, bad benchmark, or anything else.

These things are completely repeatable. I know it sounds freaky to someone who doesn't see the numbers first-hand, but I don't have time for a scientific break down. 5000 to 70000. No caching will do that, lol.

Also, forget Core Temp. It is not as useful as Real Temp. If someone is getting ambitious, Real Temp can turn off ALL clock modulations so that the CPU is fixed @ 1.6GHz, including Turbo Boost, EIST, and C1E.

~Ibrahim~

 

Unless they disable cache ala Celeron, they're not the same proc.

 

Yes, when you are playing a game, and see it degrade into a slideshow everytime you increase the display brightness, and the sound starts crackling after awhile, and then you can repeat it over and over again, it is very freaky.

Like there is lag that is independent of the framerate counters too. Even though the counter says 15 fps or something, you may actually see only 1 frame every second on the display.

 

this is what i'm going to do. ive been basing my problems from the gaming issue, and i tend to think it will be solved with the adapter, since the battery gives great performance, the adapter doesn't.

so, i am going to run 3d mark, with ATI power play on max performance, and my power plan on max performance. then, when i receive the adapter, i will run 3dmark with everything the same, and compare. if this doesnt solve my gaming problem, and 3d mark diff, then i will investigate other options.

if the 130w adapter solves my issues, id say case closed for me, and others should try.

btw, countless times today in CoD MW2 at the level where you are in suburban america, the crackling/slowdown started, and when i turned the brightness down, or switched to battery, it played silky smooth.

EDIT: also, to account for any possible diff regarding temperature ( one factor that controls the turbo mode ) , i will be running my tests outside, where it is currently 10 degress C

 

Crysis benchmark and prime95 at full load and using realtemp to record the log, clearly the processor switch from an average of 11x to a 7x

 

Hello,
I've been lurking and following these posts since the order thread. I'm not sure if this will help but I've done some real world cpu testing. I have the xps 1645 i7 720Q 4g ram 500g hd and the rgbled screen. After reading all of the posts about the power supply issue ect... I decided to run some test through Lightwave. Lightwave is a commercial 3D package and I've rendered some test scenes through it to give the cpu a work out. The program uses all 4 cores hyperthreaded to 8 (max load while rendering) the display was set to 100% brightness. in all tests with or without the charger connected battery at 100% or below all results were almost the same with only negligable differences going either way = dam fast no matter what.
The only comment I would have is that the rgbled display sucks in non color managed programs (spider3elite calibrated) it is way over saturated. With color managed programs it is the best display I have ever used (3D modeller and semi-pro photographer).
Still waiting on my Win 7 upgrade from dell, should be here on Friday.

Thanks,
Ed

 

It is very fast indeed. No matter what you throw in, it is very hard to let the system work on max load. Personally i had to run Crysis and prime 95 to reach full load on both the CPU and GPU.

 

I would love to know whether the 130w adapater fixed all the problems or not... It's great that you really like your laptop!

 

shoot i just beat borderlands... i was wanting to play it on my 1645 but i accidentally played too much today

 

Yeah stressing the CPU isn't enough, need to get the GPU to spin up as well with some hardware rendering. I totally agree about the RGBLED display. Maybe now that they're pushing these displays on consumers we'll see better support from Windows and other apps.

I read on some blog that the window manager in Win7 was supposed to support high-range floating-point rendering but it got cut, maybe it'll be in Win8. But it would be nice to have for these wide-gamut displays.

 

Well I just got done playing a ton of left 4 dead 2. Runs like a dream on all settings maxed (i turned AF down to tripple buffering though because i just done see an improvement personally).

I can pretty much conclude that the laptop is NOT throttleing when it is on battery power, and it IS throttleing when its on AC power. We've established this.

However, my AC adapter isn't even that warm. I mean its toasty, but it isn't scorching hot like I'd expect it to be after running it at capacity like some people are theorizing.

I'm going to tinker around with some windows power profiles and see if i cant force all throttleing off completly and see if that improves. I"ll let you know.

In the meantime, I'm still very confident that it is just a software bug and not a real big issue that will be hard for dell to fix.

 

I would really be surprised if it turns out to be a software issue, in this review Test-Dell-Studio-1557 (if you scroll all the way down) you can see that the Studio 1557 with i7 is using up to 90,2 Watt. A studio 1645 will use as much or more. so if you push it to it's limits it looks like the AC adapter is going to have problems.

 

Well has anyone tried 130W yet?

I am thinking about getting the Studio XPS myself, but now I'm not so sure if I should get it now or wait until some of this is resolved.

 

No one has tested this yet. I have my 130w adaptor but my laptop hasn’t gotten to me yet and probably won't until Monday. Someone else is waiting for theirs and probably won't get it until Monday as well

 

I've called Dell today and they told me that there is no 130W power supply for that model.
They will call me back tonight and try to resolve the problem by running some diagnostics on my laptop.

 

I'm going to use my 130W M1710 xps ps, same plug & voltage etc, but like Siphen I dont have the 16 xps yet. Been running my Latitude on the brick for a few days now just as a check and the only difference is that I dont need to put a magazine under it any more to stop it melting the carpet - the brick on the xps was very hot, the 65w on the latitude was very hot (I'm playing lotro on it as the 1710 is dead, 3dmark06 a massive 630) , this is far more sensible heatwise.

 

And that is completely expected they would say that. Dell makes all their current laptop adaptors interchangeable, it’s a cost savings thing for manufacturing. As long as the 130W fits into the Studio XPS, the BIOS should be able to tell it’s a 130W adaptor without issue. Dell is just telling you that "we don’t have a 130W adaptor on our parts page to send you, blah, blah, blah." This in all honesty is correct. If you look at the parts page for this model computer, you will only find a 65 auto/air adaptor (which in all honest will not provide enough power to this system). They don’t even list the 90W adaptor that gets shipped with this laptop. what they also probably do not tell you is that the 90W adaptor shipped with the studio is the EXACT SAME 90W adaptor that they ship with the E series Dell laptops, the exact same E series laptops have a 130W adaptor (the one linked to earlier), and the 130W adaptor says its supported on the E series laptops.

 

Here is the link for that 130W adapter by dell:

http://accessories.us.dell.com/sna/...s&l=en&s=dhs&cs=19&sku=330-1829&mfgpid=198276

 

U sure it works with 1645?? There is no mention of the XPS on the item description

 

I can not confirm that, you need to check with dell. That is the adapter that has been in previous discussions here on what we should have instead of the 90W, once again this has not been confirmed.

 

It's not even confirmed that the adapter is the problem yet.

I'm going to start a thread for this issue, so that we can all discuss it and link dell reps to it. I'll edit it in soon.

Edit: Heres the thread
http://forum.notebookreview.com/showthread.php?t=437800

 

I'll be receiving mine early next week. Maybe in some strange way the ones in my country or in Europe at all don't have this problem. We'll know by then.

 

lol, sometimes you have to tell them what you need. ^-^

 

This may not be a preferred fix, but for those of us that use the iGo power adapters, I think most of those adapters are greater than 90...

As long as you are using the correct tip for the laptop with the adapter, it should be fine, correct?

They have a 130 and a 160 iGo adapter.