S-XPS 1645 AC Power Throttle Issue Investigation

I would recommend to put in an explanation what Throttling is, how to detect it (i.e run Prime95/FurMark and watch for modulation, and lower multipliers, etc...).

 

Sager has 1 hour (perhaps less) battery life. If not for this one issue, I would have gotten the 17" version instead of my Dell Studio 17 (has same throttling problem/fix as XPS 16).

 

I agree, less than one hour of battery is a major downer... but very few other negatives other than that plain plastic chassis. I'm still considering the Sagar in light of the many Dell issues.

 

i heard dell found a fix to the problem and are going to release it in January, is this true? if not im just going to return my laptop

 

There is nothing confirmed. And of course Dell has an incentive to make people think a fix is around the corner so the 21 day return deadline passes.

 

it blows my mind how many people

A) Get told by Dell that there is a fix/has been fixed

and

B) Believe it.

 

As requested.

http://forum.notebookreview.com/showthread.php?p=5669984#post5669984

here pm me for suggestions etc. Hopefully this works a little better for newcomers etc.

 

Any Dell adapter with 19.5v should be fine. Dell rushed out the 1645 to be the first on the market and they knew the design was not good enough to handle the heat the i7 would put out and thats why you see the throttling. Except those same idiotic engineers crippled the M15x in a similar (albeit not as bad) way despite it having more than good enough cooling. I've gotten around their ridiculous bios limitations and have freed my M15x to perform as it should and it never gets hot either.

 

how did you get around the throttling issues on the M15x? Does the throttlestop program work for that BIOS as well?

 

I think the finger should probably be pointed at the accountants and lawyers. Why include a 130 watt adapter when a 90 watt adapter will be good enough for most users. There are hundreds of happy laptop owners on the Dell website giving their crippled laptops 4 and 5 star ratings. Most users don't realize what they're missing out on.

I'm sure the lawyers shook their heads in agreement so they wouldn't have to deal with any exploding battery issues. Crippling these beasts probably seemed like a good idea at the time.

Any guy in India that tells you this problem is all solved isn't being 100% honest. If this problem was solved there would be a new bios available and there would be a massive recall of the under powered 90 watt adapters. That is the only way to fix this problem. ThorttleStop is a nice band aid just in case they never get around to properly fixing this issue.

The M15x uses a second type of clock modulation throttling that ThrottleStop isn't able to deal with at the moment. I'm working with 5150Joker and hopefully I will be able to come up with something specifically for the Alienware guys. The present solution is a little complex so I'm very interested in coming up with a much simpler fix.

 

What we need to do is find a way to get some mainstream media attention. Fox news, CNN, or some other big station to do a news segment on how this big corporation is taking advantage of customer's trust.

 

So does this problem infest the Studio 1747 as well?

 

How about Help me Howard, he is affiliated with a major news company and has helped many in similar situations.

I located a link below that allows you to contact Howard.

http://weblogs.wpix.com/news/helpmehoward/

 

I finally decided to get a more demanding game. So I bought Crysis. I'm definately thinking this is a power issue and not heat. With the rgbled at full brightness I'm getting major clock modulation (12.5%). Temperatures are 61 (CPU) and 63 (GPU). That is with 1024x768, medium quality. When I reduce brightness down to a quarter, I don't see any throttling and actually see turbo boost kick in occasionally. Temperatures are 72 (CPU) and 77 (GPU). I'll push up the resolution and quality later to see what the system can do.

 

So i called dell tech support to request a 130w adapter explaining that my system required more then the 90w adapter shipped with my system can provide.. girl explained that the system only needs 65-90w and that power throttling doesnt exist. i explained that it does and in order to fuly use my xps i need a 130w adapter.. again she explained the system comes with 90w and thats all it needs so i fully explained the issue at which point she placed me on hold and the phone somehow disconnected.. i love these guys

 

HSeldon,
The guys here have done excellent work to show that this is a power related issue, and not temperature. I have used a Kill-A-Watt power meter and some games to verify power consumption. The throttling kicks in when the system starts using ~90W (usually because the GPU starts working hard for 3d), the throttling kicks it back to about 60W and framerate suffers significantly. There may be cases where heat related throttling may be taking place, but this is remote compared to the power issue.

Also, keep in mind, this throttling is dictated by the BIOS, therefore just getting a larger supply does not help. You need a software fix such as unclewebb's Throttlestop, or a bios fix from Dell that has yet to released (holding breath and turning blue).

 

12.5% modulation is a bad modulation. the cpu is crawling when its at that. around 4% of its potential lol.

If ya guys could eventually transfer your discussions on over to the new thread. As information/experiences like this is exactly what needs to be in that thread, to help spread/educate others.

 

Winks,
She is either playing dumb or truely does not know about the issue. Dell engineering is aware of it, but you never know how well they inform there tech support of every issue out there.

They sent me a 130W after complaining about it. Just be persistent. Personally, I got better results by emailing the tech info on the issue after talking to him on the phone (send them links to here or dell forums). He initially was not going to send it to me, insisting that the 130W was not the proper solution. Technically this is correct, as you need at least a BIOS or software fix to see improvement. I think it is still open to debate if there are more hardware changes needed internally for system robustness, and if the 130W is even large enough.

Typical power consumption is around 55W with battery at full power, just surfing the web or word processing does not tax this system. This rises quickly if doing gaming, or any CPU intensive tasks (compiling software, photo filters/manipulation, file compression/decompression)

 

That's what I was getting at. I've seen some comments about Dell throttling the system to control heat but seems like the system can handle quite a bit of heat. When temperatures are in the 70's, the fan kicks into high speed occassionally. Hopefully the 130W (or greater) and BIOS update will take care of this. Otherwise I'll be buying a 130W and using Throttlestop.

 

For any user that doesn't fully understand how big a kick in the balls a clock modulation setting of 12.5% is in the middle of a game, here's a benchmark you can run.

http://www.techpowerup.com/downloads/366/Super_PI_Mod_v1.5.html

Download Super PI mod and run a quick 1M or 2M benchmark for a baseline with the ThrottleStop clock modulation locked to 100%. Now run the same benchmark with clock modulation set to 12.5% and with your multiplier locked to 7.0. That's exactly how slow your XPS laptop can get during the middle of a game.

Do you see any difference in your benchmark scores? No wonder games start to chug when your CPU starts performing like a Pentium II or worse.

If anyone has access to an antique computer then run the same benchmark on it to put this into perspective. Maybe some public humiliation will encourage Dell to take ownership of this problem.

In a way the Dell girl is right. These computers only draw 65 to 90 watts. When the bios is throttling them to death, that's all they can draw. Maybe we also need someone to post a screen shot of a Kill-a-Watt meter while running Prime95, Furmark and ThrottleStop. Then when a Dell rep tells you nonsense like 65 to 90 watts, you can show them the screen shot and tell them that you would like to talk to someone a little more knowledgeable about this subject.

 

i keep seeing this over and over, and it's a joke. if a larger adapter doesn't "help" then how come when i unplug my laptop, or use the 130W adapter, i have no lag, or audio crackling in games? not only does it help, but the 130W solves my problem.

 

hi guys, I was giving throttlestop a try. And as advised, I've set the multiplier at 13, which was awesome. However, since I'm using 90W, I decided to stop it and restarted my laptop. Upon restarting, somehow the program i7 turbo shows that my multiplier is still hitting 13. Shouldn't it have gone down to 8?

 

Well it doesn't solve my problem or the problem of numerous others who have posted here. The problem for us is only solved when we use a 130/150w adapter and the throttlestop program. So congrats to you if you believe just a 130w adapter on its own completely solves your problem but that doesn't make everyone else a liar does it? I mean what do you want? A medal?

 

Max,
Did you update the BIOS? The last BIOS update (mid/early december) fixed the sound crackling for me, without changing the power supply.

All the testing I have done is with the 90W supply, I just received the 130W a couple days ago, and been pretty busy on family activities the last couple days and haven't done any benchmarks with it. My understanding from other's testing is the 130W is still limited to ~90W consumption by the BIOS, maybe I misunderstood what they said, either way, I'll do some testing later this week to see how my system reacts.

I notice your setup is the WLED with a lower resolution and SSD, which may give you some clearance to actually run smoother without throttling with the 90W.
Just speculation, but some of the throttling might be based on voltage dip and not current draw, and a 90W supply is likely to dip voltage at the higher current draw, where the 130W would not, but the xps BIOS may still limit the current to ~90W.

 

would be nice to be accurate. no wonder dell ignores like this. there are some good testers here, and a bunch of half-assed testers who don't choose correct words or sentences. maybe because i have a WLED only 1600x900 and SSD may mean i have diff power req than some others. if youre going to pretend to be a troubleshooter, at least make your findings and advice clear. this is coming form someone who worked for MSN/Microsoft doing Tier 2 tech support for years.

"does not help. "

should read: does not help some users....

maybe people with my specs don't need throttlestop, maybe we do.i haven't noticed much of drop in multipliers since getting the 130W with or without Throttlestop.

but the main point is that there doesn't seem to be one fix, and if people keep making definitive statements, this whole thread is nothing more than a bunch of users rambling about issues.

on another note if we hope to get dell to give a thread credit, and maybe get some media attention, there needs to be a thread with nothing but benchmarks, and results, keeping the average joe consumer's opinion in this thread.

i did update the bios, but i had received my adapter before that, which got rid of the crackling, and the fps lag.

about the bios limiting the draw to 90W, i dont know if thats correct. i thought i read that someone had a kill-a-watt and a 130W, and they saw the unit draw more than 90W under full load.

about my config, ya i thought that too, i actually made the rest of this post before i read your whole post to quote you. ^^

 

I got my 1645 today [had 130W adapter from before] I'm reinstalling windows to get rid of all the bloatware and then I'll try out a few games with throttlestop and see

 

max no offense, but have you done proper testing to see if your not still throttling while gaming? because you yourself are making statements like you have.

what i mean is, the horrible slideshow effect may not be there. Perhaps modulation is not kicking in due to better stability. But your multipliers may not be in the 15x-21x range they should be while gaming. Like you in tf2 things where barely playable at lower resolutions without huge problems, though it was being throttled to 8-10x multiplier, severely limiting the gpu load. With throttle stop giving the full 15x-21x multipliers, gaming is now possible at 1080p resolutions. Before throttlestop at 1080p i was 25fps-30fps, now it averages around 55fps or higher. HUGE difference in smoothness and awesomeness.

its proven these laptops need the 100-110w to run at its best. And with kill-a-watt meters its been shown the laptop is limited to 90w regardless of adapter, unless using throttlestop.

And yes your system is probably the least effected, given the low resultion and wled screen.

 

do dell make a 150w power supply because i have the rgbled screen and i heard the 130w isnt enough for it

 

i have done tests with throttlestop an dprime95 furmark, running prime95 until all cores are maxed out, its giving me the 12.5-13 which is my base speed of 1.73...which is what it should run at under full load. iahevnt tested with less cores, i have time ot do that today.

im getting 35-40 fps in crysis at 1600x900

the issue i have with the statement that the bios is limiting it to 90W is that when you unplug the laptop, it appears to run at full speed, same with 130W but when on 90W the problems pop up.

so if it cant pull more than 90W, why are there no problems on 130 or battery?

also, this is from a friend who runs an Authorized HP service center.....he asked me how is anyone sure that the mobo cant pull 90W? the BIOS detects my adapter as 130W, and it's not unreasonable that since the BIOS detects it, it could probably use it.

 

Just a piece of info for what is worth. Not sure you are familiar with my story, but I initally ordered a 1645 and later changed my mind and ordered a Envy (due to power issue). Dell refused to cancel my order so I received both machines.

Today I decided to call Dell to return their machine (still within 21 days). Had to go thru tech support in order to avoid the re-stocking fee.

I explained the whole thing (throttle issue) to the tech person. He put me on hole twice for about 10 mins each. He then came back and offered to replace the motherboard. He said the cirtcuitry in the new motherboard would fix this issue. I asked whether he was sure what the issue was and whether Dell was aware of it. Or whether this was just an attempt to "fix" my machine. I told him this is a known problem in the community and not my machine. Pointed him to threads here.

He again said Dell knew about it and that he was sure the new motherboard circuitry would fix the problem. Knowing better I declined and said I wanted to return the system.

I thought I'd share here since I haven't heard this one before. I am not sure whether they have found out something or whether it was just a shot in the dark by this guy. I certainly wouldn't believe anything from Dell until it is verified by folks here.

 

you need to have realtemp log the multipliers while you are running your games.

no one know for sure/without a doubt the mobo can handle the 100-110watts the system needs. Bus so far dell has just said they havent tested, and reasoning if the battery can supply it just fine, it should be okay. Other people have friends that apparently designed laptops saying its fine as well. but like i said we do not know 100% for sure. so far has been working great though.

unclewebb program disables dells throttling that they do. with a kill-a-watt adapter we can see that the system is limited to 90w even with a 130w adapter. and the way it behaves looks very much like an algorithm in the bios. when it hits 90 it will jump down to 70-ish watts and slowly un-throttle the system til it hits that mark again. with throttlestop program, under gaming the system gets up to 100 watts and peaks at 110watts for me.

 

The BIOS will have different modulation tables if you’re running off battery or an AC adaptor. Which shows there is no problem when running off battery.

Buy a Kill-A-Watt and learn how your system runs for yourself.

With a 130W adaptor and stressing this laptop out it has pulled up to 147W. While gaming on most games, it pulls around 110W at this point the video card is maxed out and you’re not going to really pull anymore watts unless the game is CPU heavier and will max out more cores. When I play games it’s always the video card maxing out before the CPU anyway.

The BIOS does not have a modulation table for 130W or greater ac adaptor and so its set to always modulate the system to 90W NO MATTER what ac adaptor you plug into the laptop, until a BIOS is released to support modulation tables for 130w, 150w, etc adaptors. This system will continue to modulate to 90w max draw on ad adaptors.

Throttlestop removed that modulation being done in the BIOS to allow greater than 90w.

Your system is designed by dell to throttle at 90w regardless if you don’t see it or feel it.

 

then i guess im off to buy a kilowatt. if my multi stays high, and i don't take more than 90W, then im thinking my specs play quite a heavy role in this issue.

another option, is that yes, perhaps it is throttling, and i just dont notice it. although, i have no explanation as to why the FPS dips stop when im on battery and 130W but they don't stop when on 90W.

doesn't that suggest that the battery, and the 130W are helping. that doesn't really mesh with your bios modulation theory and the lack of 130W table.

you do seem to be quite concise and well versed with the issue though, whats your take on what someone earlier mentioned, and it being a voltage dip and not watt issue? i don't know enough about electricity to have an opinion on that matter.

yes, that why i think there should be another thread with just results. the first time i mentioned this thread to Dell, the tech guy was excitied to hear about this, and mentioned he would forward it. the latest time i mentioned it, they laughed at me, and pretty much said it's a joke.

that may have been just diff reps opinions, but it may also be, that they look at all this babbling like its a joke.

 

if i were to guess, the 90w gets extremely hot when ran at 90w, suggesting it is at its limit. The computer also seems to take a split second to throttle, so you can very well see power spikes. As with the kill-a-watt meter i could see it hitting 130w+ for a split second before throttling down, that was on the 90w adapter i had. So stability in the power could be an issue with the spikes on a loaded 90w adapter.

or it could also be that it does slightly help, and your computer was just on the very edge. As the bios may have slightly different tables for the 130w, since it does detect it as a 130w adapter. It still limits to the 90w though, but perhaps the 90w profile was changed to be more aggressive to limit the adapter heat etc later on?

I will mention not all games throttle the computer. Torchlight, which isnt really a full game like modern warefare etc, its not very intensive at all, and it barely peaks over 90w's at 1080p. its very dependent on config and game.

let us know what the multiplier logs tell you, your case is interesting, and ill make a note of it in the new thread. which i hope yall go to eventually.

 

It just amazes me that Dell don't inform their staff of the current status of the problem.

How hard can it be to send around emails to each manager and all staff including sales.

Because of this we get all the lies with each Dell agent telling us whatever comes into their mind.


Dell can't be that stupid to figure out whats going on.
It must be in Dells plan to keep it like this and frustrate the customer hoping they will give up.
Maybe there is no reliable fix without the lifespan of the laptop being reduced.
If a bios update and larger power suppy will be the official solution why are they so slow to acknowledge or implement it.

I hope that the throttlestop/larger power supply will not damage the laptop but how can we know for sure.

It does not bode well for anybody looking for an official solution.

I still keep hearing people commenting about the 21 day return policy.
The 21 days refers to a laptop which is not defective.
The 1645 is defective under the law and you are entitled to a refund even after 21 days.

This laptop should be withdrawn form sale until its fixed, but that would be the right thing to do and Dell dont seem to like to do the right things.

As I said before, send back those laptops and maybe Dell will stop treating us like fools.

This is really turning out to be a hopeless situation.

 

For DC sources wattage = Voltage * current. AC calculation is slightly more complicated because the voltage is not constant.

The laptop uses a constant voltage supply of 19.5V. This provides power to the various laptop subsystems. The current is going to vary up to the maximum the power supply can send. 130W can supply up to 6.67 Amps at 19.5V (assuming 100% efficiency) vs 4.6 Amps for 90W.

Typically power is calculated by reading the voltage and current. If you have constant voltage source and you see a dip in that voltage, you know that the system is having power issues as you are drawing more power than the supply can provide to maintain that voltage. Dipping slightly is usually allowed by spec, but the system will likely react to prevent dipping lower.

 

Does anybody know how the issue with the 1645 was not spotted by laptop reviewing websites?

 

romills - The same formula is true for AC power. You just use RMS voltage which is what folks are talking about when they say 120vac or 220vac. RMS is peak to peak times .707 if my memory serves me. Using one of these plug in watt meters that go inline with the AC is measuring the AC power draw and not the DC. The power adapters are rating the wattage on the DC side. We would have to know the efficiency of the power supply to know how to convert to the actual DC power. My guess would be no more that 75 or 80% efficiency based on my ancient experience and the assumption that things have gotten more efficient over the last 20 years since I was involved with electrical engineering.

 

i had wondered about that briefly as well. to get correct DC draw measurement to the laptop, wouldn't the device have to be between the brick and the laptop, not the brick and wall?

 

In a way yes, but so far its so far 110watts is far above the range for accuracy. The adapters are very efficient though, their not like desktops, according to the markings the 90w has at least a 85% or better Average efficiency.

The 130w adapter at 90w isnt very taxing for it and will be very effecient, which is why i see it throttling to 90-91 watts on the meter. But with the 90w adapter at 90watts, i see it stabilize to throttling to 93-94w, most likely due to the inefficiency at max load for the 90w adapter.

the difference in wattage, will be in excess heat. That extra power used in the adapter will go to creating heat. the 90w does get hot, but it heats up slowly ive noticed. given the reduced surface area etc, it cant really get rid of heat all that fast letting it get really hot. free convection is a very slow form of heat transfer, and really the only side giving off a lot of heat will be the top of the adapter, as the bottom is conducting through the ground.

http://forum.notebookreview.com/showthread.php?t=446193 max please post your results with multipliers etc in this post here, and i again hope we move on over there. As some people suggested earlier

 

With this laptop being so "new" and back ordered for so long ... I am not sure if there are any real reviews of this guy out in the wild yet.

 

Regarding to this person, who just talked to Dell and it seems that Dell has acknowledged that the new motherboard need to be changed to solve the issue. However,

Regarding to this person, whose friend runs an Authorized HP service center, saying confidently that Motherboard is not the problem here but AC adapter.

I'm not sure which one I should believe. Since I dont believe Dell and also HP customer service much

 

you seem to have misread. he asked me how can anybody be sure. he didn't make a definitive statement.

im trying to get him to create an acct here, but he's playing the "too busy" card.

 

First let me say thank you for all of the information provided.

My 1645 arrived 10 days ago and I am enjoying the system.

I have been playing games such as Arkham Asylum and Crysis but find that after 30 minutes or so, the FPS will dip and the screen will freeze for a short period of time. I am guessing that this is due to the throttling issue?

Additionally, sometimes Firefox will freeze (especially when I am typing). Another strange occurrence is that the sound on a video I am playing will stop working and the mouse cursor will become huge (annoying). I have to restart in order to resolve this. Is this also due to the throttling issue?

Has anyone had any success in calling Dell and demanding the 130W power unit for free? I do not wish to pay for an issue that is Dell's fault.

Many thanks for your help. These questions have probably already been answered but there are too many posts to read!

 

that's how i got my 130W, but they sent me the huge one. also, i didn't send my 90W back, and i wasn't charged anything.

 

Hello guys, i've been watching this thread for a couple of weeks now, and i received my studio xps 1645 yesterday. So far i have not noticed anything wrong with my system using the 90w adapter dell sent me, but i was hoping to be able to find a sure test to see if everything is really ok. I've downloaded unclewebbs throttlestop, real temp and prime 95. I already had the 130w powerbrick from my old xps m1710, so i used this one for testing.

I did 2 tests:

130w adapter:

http://img63.imageshack.us/i/130wt.jpg/

90w adapter:

http://img694.imageshack.us/i/90w.jpg/

Edit:

Did one on battery also.



Can anybody explain what these numbers mean? Do i have better performance with the 130w adapter?

Thanks a million guys!

 

If I don't have a i7 and instead a normal C2D will I encounter these issues?

 

I also have that problem of the laptop freezing for a moment when I'm typing sometimes. It's incredibly annoying. What gives?

 

@Quietcat and theoak...

A review of this system at notebookcheck.net did catch these problems, although exact explanations weren't found (which is fine, notebookcheck does very in depth reviews as it is and I wouldn't expect more).

But it was identified in the course of review that the 90w power adapter got very hot, and it also lost marks in the performance tests to the extent that it couldn't be recommended for gaming - only because there were unexplained drops in frame rates the likes of which they'd not seen before....now we have only since discovered over time that these two issues stem from the same problem!

 

You only need three tools to reproduce the issue:

* Run i7 Turbo (included with realtemp). This will show you the current multiplier on your 4 cores.

* Run FurMark and IntelBurnTest at the SAME time. On FurMark, uncheck the "full screen" option, and choose benchmark.

On AC Power, you'll see the multiplier dropping to 9 or below (not sure what processor you have).

On battery, you'll see the multiplier staying at 13 (or 12, not sure what processor you have).