This thread is getting too large, and too difficult to find useful information. Thank you to everyone who has put forth efforts into the investigation. At this time I feel we can safely say that there isn't any investigating left to do, so without further ado, please direct yourselves to this great help-thread with information about what the problem is, and how you can circumvent it:
http://forum.notebookreview.com/showthread.php?t=446193
====obsolete information====
EDIT Time, I've removed all the old information as its mostly speculative, and replaced it with useful factual information.
The following is from a post made by our user Khaledseif on the Dell Forum. I feel this update he posted is so well-written that it deserves to be included here.
ThrottleStop, brought to you by Unclewebb (developer of RealTemp) has been designed to combat the throttle "feature" Dell has implemented into the design of the XPS 1645 i7's BIOS. You WILL need a 130watt Dell branded adapter unless you understand how to limit your CPU through Throttlestop to keep it from blowing up your 90watt adapter (seriously).
This program will LITERALLY transform the performance of your i7!, it's amazing! So don't be a stingy leech and send a PM to Unclewebb to see how you can send him a donation. With this application he is literally giving you a huge performance increase in games and other applications, so $10 bucks or so is the least you can do. The link is hidden below the screenshot, and by downloading it and using it you agree to do so at your own risk.
![]()
http://www.fileden.com/files/2008/3/3/1794507/ThrottleStop.zip
-
old information from first post, (aka archive, skip over it)
UPDATE TIME, ALL PREVIOUS INFORMATION PRESERVED IN TINY-TYPE IN CASE WE NEED IT AGAIN!
More users have received their 130watt adapters and it has been discovered that different laptop configurations are benefiting less then others from the larger AC Adapter. I think it is the consensus that ordering a 130watt adapter is still helping out (it did with me in most cases, only some performance issues remain), however with the new reports and benchmarks and testing that has been provided there does seem to be another issue (related or unrelated) that is causing the 1645 to throttle it's performance down for some (currently unknown) reason, possibley thermal or power draw related.
At this time, I personally dont feel there is any reason to be worried about purchasing this laptop. I've really enjoyed having it and as long as you have it several notches below max brightness on the RGB screen it runs like a dreamboat sailing on a moonbeam (you read that right, lolol)
====================Old Stuff===========================
Thank you, Max420 for the hard work you put into researching and troubleshooting your way to sweet, sweet (almost)victory
This thread is for creating a central location for posting information and findings related to the Studio XPS 1645's performance being forced to throttle down due to the AC adapter not being powerful enough to provide adequet juice to the system.
ATTENTION: This seems to be resolved. Your Dell Studio XPS 1645 has shipped with an AC Adapter that does not provide enough power to your system causing it to choke and throttle itself during high-load tasks!
You are urged to contact Dell Support and report this issue, and link them to this thread, and request a replacement 130Watt adapter at no charge. Dell may resist doing so, so kindly link them to this thread to show them that not only have they already sent out some 130watt adapters to people, but that it is FIXING the issue.
Please follow the link below to see a screenshot of full proof of how our XPS performs on 90watt vs 130watt adapters as well as battery power.
>>> http://forum.notebookreview.com/show...&postcount=132 <<<
The Issue:
When connected to AC power, the 1645 will have degraded performance while in games. Users will experience slow downs to unplayable framerates and stuttering. People have done reports and investigations indicating that the CPU and/or GPU are failing to run at their full clock speeds. They are being throttled down as if by temperature or power requirements, however if you unplug your AC adapter and run off battery power you will experience no slow downs at all as long as you have battery saving modes disabled. (Please repost here to centralize the information!)
Summary: On battery power, the laptop does not throttle down (unless you put it into battery saving modes). On AC Adapter power, the laptop throttles down as if it is trying to save power, or reduce temperatures. It's essentially backwards.
Current owners' ideas of what the cause could be:
*Under-powered AC Adapter (Ships with a 90 Watt).
*Power Management Issues (Bug in the software, aka software patch needed).
*Screen drawing more power at higher brightness settings and causing the issue.
*Battery charge drawing enough power to cause the issue.
(Grayed-out issues reflect ideas that seem to have been ruled out already).
The results are in, we have confirmed that the 130watt adapter resolves all issues with stuttering, slowdowns, and crackling sound.
Mods, feel free to edit information into this post as you see fit.
Please post any information (copy/pasted or otherwise) that you have discovered or posted in the Owner's Lounge. This thread will serve as evidence towards our issue when we report to Dell with problems.Last edited by a moderator: May 8, 2015 -
Format for test results. Credits to user Synthesia:
-
-
Link to thread to Dell Community by khaledseif:
http://en.community.dell.com/forums/p/19306277/19596373.aspx#19596373 -
-
-
-
-
-
hey guy, sorry for this huge post, but it will save people a trip browsing through the OwnLounge.
-
i received the cord about ~5min ago via Purolator. It does fit the laptop physically, that being said, I have to go to the jobsite im running to drop off some materials, I'll be back home in about 3 hours to do the tests with the new cord.
and i guess i do have to send the old cord back
edit: the cord they sent me is PA-13 Family
Model No: PA-1131-02D2
DP/N :X9366
INPUT: 100-240v~ 2.5a 50-60hZ
output: 19.5v 6.7a 130w
lol, FAIL. seems I'll be waiting until next week to test the cord.
-
Hey guys, long time lurker, first time poster.
I'm one of the 'lucky' few who ordered in late sept, only to have it cancelled without being notified, and didn't receive it till yesterday.
My general impression is a consensus with the power consumption idea.
Running multi-core enabled games such as Crysis, TF2, L4D and L4D2 with time-demos I use to bench my desktop, I see a 25-50(!)% reduction in fps on fullbright battery vs plug. Lowering the brightness to its lowest setting makes it run more or less the same, around 5-10% less with plug. External monitors (VGA and HDMI to DVI) run the same as battery.
Also, interestingly, when running on fullbright with plug, the thing barely charges at all.
Also, I noticed general lag installing software and loading maps when the plug is in. A quick counting of a (second load for both) map load in TF2 showed that the map took about 12-13 seconds longer to load with the plug in.
I'm not entirely sure if this means the cpu is underpowered (electronically), or if its software throttling, and I eagerly await the results of the 130w test. -
Good to see you received it! Can't wait to see the results.
-
This cord is available at Radioshack
-
LOL would you look at that. It seems there is a real difference between cords. Didn't the Dell rep know that plug is different than the XPS's??
-
the cord that was on your 90watt should fit that just fine.
-
Yeah but don't you think only two of the three plugs connected can't supply 130W?
-
@Zlog
RBGLED and brightness: STILL an issue. I don't know if it happens with the WLED, but any RBGLED user can confirm this, I would say:
- Download Prime95 (choose Windows 64-bit)
- Download Real Temp and open the program called i7 Turbo
- Run In-Place large FFTs
- Adjust the brightness
Make sure that your performance options are all set to 100% CPU in the power options.
Results:
Full brightness + Adapter = Oscillation from 8x to 11x
Full brightness + Battery = 12x constant
Lowest Brightness + Adapter = 12x constant
Lowest brightness + Battery = 12x constant
You can also see the difference if you set the computer to full brightness and run the test with the adapter in. Remove the adapter and it will jump to 12x. Plug in the adapter and it will fall down to 8x-11x.
~Ibrahim~ -
-
-
the cord for the slim 90W has an angle to it that makes plugging it into the thick 130W brick not possible, unless i shave down plastic on either unit, and im not willing to do that.
ya the shack has it, but i live kind of far from an electronics store, and i don't really have teh time to leave, im @ home reading prints, pricing work etc. ill grab one sunday if i get some time, or remember. ill up a pic of what i mean, after i get off the phone with XPS Support. -
So I never got my call from the executive technical support team. No matter, I just spend another hour explaining the problem again, lol.
Anyways, now a SSR (Senior Support Representative) who is part of the Global Escalation Team is looking at the problem: he has seen the problem also before and will give me a call Monday....that should hopefully come, lol.
~Ibrahim~ -
yeah i just got home and saw that angle on the adapter plug and was like "oh nm"
I might go buy a universal 130 this weekend and see if it works.
then i'll return it >.> and make dell send me one if it does.
I ungrayed the RGB issue, even though i cannot duplicate it on mine. -
Just gonna add the summing up of the theoretical individual components' maximum power usage from the other post:
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 = 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. -
We cant say for sure about the wattage numbers adding up like that. as TDP doesnt reflect actuall draw, as been said before. In practice it could be more could be less we just dont know. but we can say by looking at it, its very sketchy
-
power is power i would think. ground is probably for noise cancellation / safety if something goes wrong over time. as long as the laptop gets its supplied steady voltage it shouldnt care. i would make sure if the dc voltage is the same though, even that might not matter though
-
Here's a neat one for you guys. I just noticed on the bottom of the laptop it says "for use with the pa-12 adapter" which is an old Inspiron series 65w dealie. The model I received was a pa-3e adapter.
Maybe dell doesn't even know what this thing needs? -
I just finished a chat session with a supervisor who was very interested in our findings, he asked me if i have the same dropping under power saver and dell recommended power plans. I ran the test on them and here is general observation.
Under power saver, working on AC is a little better, under DELL recommended or maximum performance working on battery is better. -
Sorry about the luck with the X-Fi drivers.
@PsYcNeT
Mine says "FOR USE WITH ADAPTER PA-12/PA-3E". Does yours not say that?
~Ibrahim~ -
So besides gaming problems and performance issues, has anyone else found that the USB devices on their machines resets or malfunctions with the XPS 1645? My G5 Laser mouse would reset it's DPI and lose the ability to change into custom DPI resolutions until it was reconnectioned, and my external storage disconnects and freezes while connected. It's very intermittent but seems likely related to power or performance problems. The system randomly gllitches or has corrupt information while doing high computational tasks like parallel mp3 encoding (8 mp3 encooders at once). Oddly enough, I noticed stranged things with it running at 60x with two threads and only 40-70x with 4 threads, and in a program called 7-zip it was faster with less threads than more. I was baffled by it, but all this makes sense. Gaming has been problematic to say the least.
-
Gah I just flipped it over again to look. I was at work and obviously my thourough reading skills are the pits :/
Either way, someone man up and test a 130w! We need to know! -
Games have been fine, as long as I play them on battery, with brightness on 0, or on an external. -
GUYS....GUYS.....GUYS....
I figured it out!!!!!!!!!!!!!!!!!!!!!!!!
Disable SpeedStep. That's it. So simple. Just disable EIST. I disabled it through Real Temp, but you could disable it through the BIOS as well.
That fixes the problem AND it allows us to finally reach the 13x multi we were supposed to be getting on four cores maxed out!
I'm testing 3DMark '06 right now, but preliminary results: HURRAY!
~Ibrahim~
EDIT: Never mind! Look a few posts down. :*( -
What is the part number for the XPS 1645 130-watt power supply/cable? I want to know so I can ask Dell to send a completely new power cable and supply. Hopefully it's idetical otherwise to the one it shipped with. edit: I want the slim 130-watt power supply part number.
-
Damn it....well, it fixed the Prime95 issue with the brightness, but I still only got 6.3K in 3DMark '06.
~Ibrahim~ -
ikjadoon said: ↑Damn it....well, it fixed the Prime95 issue with the brightness, but I still only got 6.3K in 3DMark '06.
~Ibrahim~Click to expand... -
eblock12 said: ↑Yeah I had tried disabling PowerPlay too, it was one of the first things. It kinda "fixes" the issue in that it locks the GPU into a "medium" clock rate and it doesn't draw enough power anymore to hit the issue. You need to stress the CPU and GPU at the same time. However, it may be an acceptible workaround for you if you're running a less intensive game.Click to expand...
~Ibrahim~ -
ikjadoon said: ↑I was talking about SpeedStep/EIST, not PowerPlay. PowerPlay = GPU ; SpeedStep/EIST = CPU.
~Ibrahim~Click to expand...
The 130w slim was the one I linked to -
Where can I disable PowerPlay and SpeedStep?
On another note i found something strange that may be related to the power supply problem.
I was playing Assasin Creed without the AC adaptor on max performances and full brightness. About an hour later, as expected Windows told me that I must connect the laptop to a power source because I was running out of battery. At this time the Windows battery indicator was at 10% so I plugged in the AC adaptor (didn't see any lag on the game..) but 10 minute later the game stopped and my computer wanted to shut down. I double clicked on the battery icon on the task bar and the level was at 3% and showing (3% - Plugged In - Not charging). I tried rebooting the laptop but I got a message just after the Dell logo telling me that my laptop could not charge the battery because the battery was not recognized.
After a complete power cycle I was able to boot the laptop and now my battery is charging. Strange.... -
Consistent with what we have found so far. At full load the charger is not supplying enough power to the machine and not able to charge the battery.
I found that the charging rate rate was 5mW while playing. -
@Ibrahim
RGBLED is not the cause of the problem as several people with WLED including me are reporting the problem.
RGBLED consumes more power and this is the reason the performance drop is more severe for the people with RGBLED. -
@Siphen
Yeah, I've found that disabling it in the BIOS causes the CPU to be locked to 7x on the i7-720QM. But disabling it in Real Temp allows the CPU to run at 13x, full brightness, adapter plugged in whereas it on makes the CPU bounce between 8-11x.
@khaledseif
Right...I think, Zlog, it might be best, then, to change the first post to brightness and remove the words 'RBG'.
@Tonkadan
PowerPlay can be disabled/enabled in CCC. SpeedStep can be disabled/enabled in the BIOS or in Real Temp. The difference, at least for me, is noted above in my response to Siphen.
~Ibrahim~ -
ikjadoon said: ↑@khaledseif
Right...I think, Zlog, it might be best, then, to change the first post to brightness and remove the words 'RBG'.
@Tonkadan
PowerPlay can be disabled/enabled in CCC. SpeedStep can be disabled/enabled in the BIOS or in Real Temp. The difference, at least for me, is noted above in my response to Siphen.
~Ibrahim~Click to expand...
Are you stressing the CPU only or both the GPU and the CPU. I am using Crysis benchmark on window and prime95 to stress both.
Disabling PowerPlay lets the GPU to work at a max of 300 vs the max 675 which will decrease power consumption. It is the case when you use power saver plan or DELL recommended on battery.
I found that the performance for the different plans as follows
Maximum brightness, Crysis GPU benchmark on high in window mode, Prime95 to stress the CPU, RivaTuner to monitor the GPU, RealTemp to monitor the clocks, and BatteryMon to display the discharging rate. 100%CPU and 100%GPU load.
1. Power saver
on battery
GPU 300 CPU at 7x
on charger
GPU 300 CPU at 12x
Observation
on charger performs better
my analysis
The power saved from underclocking the GPU helps in improving the power for the CPU.
2. DELL recommended
on battery
GPU 300 CPU at 11-12x
on charger
GPU 675 CPU at 7-10x
with GPU switching from time to time when CPU is at 7x due to lack of CPU orders
Observation
on battery is slightly better
My analysis
Battery provide enough power while the charger sometimes doesn't provide enough power which lead to the downclock of the CPU. the sudden downclock of the CPU lead the GPU to sense a decrease of activity and downclocks to 300.
3. Maximum Performance with PowerPlay changed to maximum performance under battery in the advanced power options
on battery
GPU 675 CPU at 11-12x
on charger
GPU 675 CPU at 7x
Observation
on battery is 20-30% better
My analysis
Battery provide enough power while the charger doesn't provide enough power which lead to the downclock of the CPU.
My conclusion is either the charger or BIOS, with the charger being my first suspect as it simply doesn't charge in Dell recommended and High Performance, and to the reporting that there is no problem on running on an external monitor -
Ikjadoon, are you still getting general stuttering performance with speedstep off and can yo uconfirm that the CPU DOES reach 13x multiplier? Some one below your post indicated that it gets the CPU locked at 1.6ghz.
-
-
khaledseif said: ↑Envy 15 consumes up to 109.5 watts Link
Studio 1557 with ATI4570 consumes upto 90.2W Link
Dv6t comes is 720p and a more efficient video card, yet equipped with a 120W charger
All the i7 laptops in the market now comes with 120+ watts charger
XPS 1640 +ATI 4670+ P9700 consumes up to 85.4W LinkClick to expand...
I have no substantiation for even thinking this, but I've been afraid that the charging circuit won't be able to use as much current as we need it to! I'm really looking forward to the 130W test.
S-XPS 1645 AC Power Throttle Issue Investigation
Discussion in 'Dell XPS and Studio XPS' started by Zlog, Nov 26, 2009.