S-XPS 1645 AC Power Throttle Issue Investigation

I will consider "false" my hope just when they'll release the bios and see that they haven't fixed properly the whole issue. Not a second before

 

My first post, like so many of you i've been lurking this thread for a while.

Misleading advertisement Dell Europe The Netherlands:

'Funny' thing: the i7 cpu isn't mentioned al all. Should we state that the 1645 core duo is superiour over the i7?

(translated: great deals for our gamecomputers)
'Funny' thing 2: When you click the banner you are being forwarded to dell xps 8000 ...

 

@lorthirk
I admire your optimism, i really do

Nothing would make me happier than if i'm wrong on this, and that dell really deliver a miracle BIOS fix that manages to tweak enough power from our 90w adaptors to unveil an unleashed system. But given the documentation from testers with powermeters, sometimes showing power consumption at 110w + how is that even possible?

 

Wouldn't a better analogy be GM voiding your warranty for installing a third-party engine management chip?

 

I'm not asking for such a miracle, just for the freedom to use a new 210W adapter with the official Dell's permission.

 

Hi,

I saw a chatlog of an agent who was affirming that he knew there was a problem with the 1645 and that they were investigating this issue. I also saw a chatlog of an agent who accepted to send a 130w adapter. Can you please repost those chatlogs because I can't find them anymore and here in Montreal they honestly seem to be unaware of the problem and the ungoing investigation from Dell engineer.

Thanks.

 

The problem with fraud class actions is that reliance (a key element of a fraud claim) turns out to be an individual issue and almost always makes class treatment inappropriate, except for narrow exceptions such as fraud on the market in the securities context. The SCOTUS has narrowed the types of actions suitable for class treatment considerably since the late 90s.

I am not saying Dell is without blame here; indeed, quite the contrary. Dell should suffer some ramifications for putting out a crippled product. Just saying that if I were on the defense side in a Dell "fraud" class action, I would feel pretty good about my chances of busting the class.

 

Does anyone have an XPS 1645 with a Core 2 CPU that they can test with Furmark + Prime95? I'd be very surprised if these didn't throttle as well when pushed hard. You can also use ThrottleStop to monitor one of these.

You're right Midnight_Voice. No matter what analogy you want to use, it's still a sad situation. It's hard to do much "hard-core gaming" when the frame rates are crawling in the low single digits because your i7 has been throttled back to a fraction of its intended capabilities.

 

This is from http://www.dell.com/us/en/home/notebooks/laptop-studio-xps-16/pd.aspx?refid=laptop-studio-xps-16&s=dhs&cs=19, today 01-07-2010, 23:51 CET, please note the word they highlighted:



Of course, they are referring in general to Studio XPS 16, C2D versions included, but ... they aren't stating any exception for the (presumedly more powerful) 1645.

 

I think we are agreed that you can safely use a larger power adaptor with a 1645, with Throttlestop not installed, or just monitoring, as the laptop will only draw what it needs, and this will never go over 90w anyway. Dell, though, are making a fine distinction between 'safely use' and 'will guarantee'.

However, with Throttlestop turned on, the laptop can draw more than 90w, and one of three things will happen:-
(i) if you are using a 90w adaptor, this will likely burn out. Or maybe a protection circuit will kick in; who knows?
(ii) if you are using a 130w or bigger adaptor, the laptop can draw more than 90w, many components will be stressed beyond Dell's design intentions, and some will be stressed beyond the design intentions of the individual component suppliers.
(iii) if you are using a 130w or bigger adaptor, the laptop can draw more than 90w, many components will be stressed beyond Dell's design intentions, but none will actually be stressed beyond the design intentions of the individual component suppliers.

Even if (iii) is the case, Dell will, as in my engine management chip analogy, likely decline to guarantee the laptop any longer as the solution is, at best, Not Invented Here.

Now let's look at what the testing of A04 might tell Dell. Maybe they have tried for (iii), and found it's unattainable for some reason, and they always get (ii). Or maybe they know that already....

So we have the choices:-
(a) They unthrottle the laptop completely, and make it able to still run on only 90w
(b) They futz with the throttling so it's rarer, but can still happen, and hold it to 90w still.
(c) They have the BIOS detect 90w or 130w, and unthrottle completely when they see 130w.

They will only do (c) if they can do (iii) above. Otherwise they will do (a) - or would if it wasn't likely contrary to the laws of physics - or (b).

They might even do (b) even if they could do (c), gambling that the hit on returned 1645s, and to their reputation, would still be cheaper than issuing 130w adaptors.

If Dell do actually do (c) - and the omens don't look too good for that at the moment - what they do may be like Throttlestop, better than Throttlestop (were that possible) or not as good as Throttlestop.

But because it would be their solution, validated in their way, to their standards, Dell would guarantee it.

Me, I still haven't put away the packaging this 1645 came in, and I won't while the jury is still out on it.

 

I really pray for (iii). I would be satisfied that way. But I think anyone would be. However, with my XPS delivering to me tomorrow, i won't throw away any packages too.

By the way, here's the XPS16 pdf i talked about before.

http://rapidshare.com/files/331923354/xps16.pdf.html

 

So took my XPS into our computer R&D department and they found the throttling issue. They suggested I put a number of questions to DELL. I began with asking if they knew and acknoledge the issue, which they said they did. Our technical guys also believe a BIOS fix may be possible without an upgraded powersupply, although this is unlikely and the easiest solution will be to get a larger one.

They also stated a number of questions I put to DELL:
1. My Technical teams response:Using higher W could be problematic for the battery (when charging) and other components.
Question to DELL:Do other dell models ship with the same battery and a higher then 90W PS?
Dell reponse: "I dont know"
Note: Does any here have an answer to this?

2. My Technical teams response: The 90W throttling threshold may be in place for more then just the power supply. Typically then limit the power supply supply based on the heat as well.
Question to DELL: Do you anticipate any issue with using a higher W powersupply?
Dell Response: Yes, we believe you may experience over heating issues. We would only allow a higher W PS if you agree to use a cooling matt. You should not experience any other issues, however this has not been tested.

3. My Technical teams response: It is clear that when on battery power, more then 90W can be drawn at any given time. However when drawing from battery a number of BIOS configurations change.
Question to DELL: When on battery power it is clear the system requires more then 90W of power. If the battery is drawing more then 90W how do you expect a BIOS fix to address the issue?
Dells response: I dont know, I will have one of our infrastructure engineers call your technology team on monday.

I also enquired about warranty and using programs such a as throttlestop. They refused to answer on the grounds that the new BIOS would fix the issue, and to wait until then...

They also mentioned that it was not a gaming machine. (I pointed out that I was not using it for gaming and that the system does not preform to its epcification).

I also said if the issue could not be 100% rectified I expected:
1. A new PS
2. A BIOS fix with the new PS (which does not just mask the issue)
3. That a cooling mat twould not be required or would be supplied for free
4. Compensation

If not I expected:
1. A full refund
2. An apology
3. Compensation

I will reevaluate once i get the new BIOS. But if it doesnt work I will be suggesting to our technology team we re-evlauate our contracts with dell and take my issue to consumer affairs.

Thanks,

M

 

The more I read about this issue, the more depressed I am. I should be happy as a jumping bunny since tomorrow I'll meet my XPS for the first time, instead I find myself wondering if I can even keep it. Help.

 

if your battery is fully drained and your system is idle, it has around 40W to pull on a idle system, or 81W to pull on a computer turned off. running a 90W adaptor.

i have never had the system pull over 70W while charging the battery, which tells me they limit how fast the battery can be charged - which means more w to the system would not cause any issue to the battery.

 

I wouldn't be surprised if the January BIOS will do nothing at all to address the issue. They were telling us the Dec 15th update would fix the issue, and it had nothing to do with it.

 

I am not suggesting this is a fraud case. I am suggesting that the item does not perform as advertised because the cost department and the performance department at Dell did not work this out. They now know it's wrong, and thus far have not fixed it.

 

Just got off the phone with "the xps experience" technical support and they "promised" me that the bios would be released by the end of the 2nd week or beginning of the 3rd week of January.

The representative also assured me that the update would address adapter issues as well as power consumption issues.

Amusingly he recommended I buy a 3rd party adapter that is more than 90W....

Dell starting to change their stance on the adapter maybe? we can only hope.

 

I'm pretty tempted to return this thing. The fact that dell is telling people that the thing isn't designed for games is pretty crappy. I don't see how the new bios will fix this without a new adapter. The thing throttles since there is not enough power for all the components. It's been show repeatedly. In order for me to play a game with the included 90W adapter I have to reduce brightness to minimum and turn off the keyboard backlight. And even then the multiplier sits at 7. Unless they can magically reduce the power draw of the 4670 and i7 720 while keeping the same performance, it's not going to happen with the 90W.

I just spec'd the envy 15 with similar config, but includes ATI 5830 and it's pretty close in price. And they ship with a 120W adapter.

I'm getting close to my 21 days (delivered on 12/14), so not sure what I should do.

 

EDIT:

All right...after some thought (and a very helpful message!), I've realized maybe I shouldn't have posted that direct quote from the email.

The gist: there is more to this issue than we think. People are working hard.

It's ambiguous and you won't like it, but there is more to this issue than meets the eye. My laptop is on it's way to Dell. They're working. No one's forgotten us yet.

 

whether its fud or not who knows, but they sure arnt saying much. I really hope these "engineers" have help this time, as their first time and letting this problem through is pretty bad. It would be like heres your corvette but with square wheels getting through testing.

does give some hope, maybe. At least he could give was an actuall eta? or did he say and im just confused by all the other eta's thats been givin out?

I have no problem being wrong, but what weve seen, they have a mountain of explination to climb when the day arrives.

 

I'll see if I can finagle a date out of him after they receive the laptop and run the tests. As of now, anything a representative says is all we've got. But there is work being done on a BIOS.

~Ibrahim~

 

So the fix about buying a 130/150W won't work fully? (because it kills the battery)

I understand that the specific components are not designer for the larger charge, but I was just clearing up some ends...

 

When my battery is charging while my system is turned on but idle, my adapter puls a very steady 89W from the wall. I usually work until the battery is at 20% and have it recharge from there. I have not yet measured it with the system turned off.

 

It seems that 130/150W + throttlestop burns battery when switvhing from power source to battery and "nerar 130w" of actual power consumption. But I think that TS has been update to avoid this problem disabling itself when on battery (can someone confirm this?). I hope that the next bios will prevent such a risk.

 

I ordered my laptop on 3rd Jan and its shipped today?!?! I heard stories of people waiting a month for their laptop to arrive and mine looks like it will be within 2 weeks!

Unfortunately I was actually hoping it took a while to arrive so these issues were/wern't resolved by the time it dropped on my door step! I wish I had researched the machine and identified this issue before I bought....I'm pretty sure I wouldn't have gone through with the purchase...

I don't know what to do really, bit p'ed off by the whole situation....

 

I'm in a similar position as you i ordered mine on the 2nd of January and it got shipped on the 7th, i was hoping for it to take some time before i get it as well. I do wonder how they have built it in just 5 days? i guess i'm getting someone else's returned laptop

I've decided to return it anway when it comes.

 

Your probably right!!

I figure it will arrrive the start of next week. Looks like people having been getting refunds when requested on the issue so I'm gonna stick it out until the end of the month to see what dell come up with Re: the bios update.

As has been previously mentioned, what dell could reasonably do is:

1) produce a new bios which detects the rating of the adapter and throttles accordingly.
2) agree to provide a 130W power adapter free of charge to all those customers who request one.

This way they can get away with continuing to ship with the (presumably cheaper) 90W adapter and yet put a stop to all the returns they must be dealing with by providing a free of charge solution. Whether it will happen or not is another question entirely though....

 

On another note, has anyone tested the length of time the battery will last on a full charge while under a high intesity load but without TS running?

The battery is a 85Wh unit so using that length of time we could calculate the power being drawn by the Mobo from the battery and put to bed Dells myth that the Mobo can't handle a 130W supply....

I'm guessing the battery will last a fair bit less than an hour and therefore a power draw of greater than 85W.....

 

But if you are not switching between battery mode/PS mode, everything would be ok?

Seems like the battery can't handle huge switches in power drawing. It can handle if you are going from 90W then higher up while running on battery all time. But if it experience a huge blast in pwr drawing it will die. Don't know if it's the truth but may be an idea.

 

Running Crysis Warhead and nothing else I'm around 50 minutes. That is with High Performance, Powerplay set to Maximize performance, keyboard backlight on max, wireless and bluetooth on. Crysis was 1280x800, with gamer (high) quality. I played for about 10 minutes and when finished I was around 40 minutes left (I'm not being 100% accurate). CPU temp was 71 and GPU was 77. Multiplier was floating between 12 and 13 (I have the 720), even though only about 2 cores were busy.

I switched to balance just to see what would happen. Everything is set to defaults even GPU (300MHz/300MHZ). Game FPS dropped (don't have hard numbers, just perception), CPU multipliers were hitting 15 and battery life increased to 1hr 12min. CPU temp dropped to 60 and GPU dropped to 58.

So the battery is 84WHr, if 50 minutes is accurate then I'm pulling around 100W from battery.

 

I was under the impression that someone (atlstang) actually experienced a dead battery while using the standard 90W adapter and NOT using ThrottleStop. If that's the only reported case of the 1645 battery going bust here, then I'm not sure I understand the panic about batteries.

unclewebb stated explicitly that running ThrottleStop in tandem with a full-on stress test (e.g. Prime95 + Furmark) and then abruptly switching from the adapter to battery could cause a system crash... but that under normal conditions, including gaming, that shouldn't occur. Stressing the hell out of the system is not representative of real world conditions.

I believe unclewebb did add in some functionality to disable ThrottleStop when switching from the adapter to battery, but I imagine there may be some latency in detecting the switch, so it's probably not foolproof if running the scenario I described above.

My only concern with using the 130W adapter thus far is that I've noticed occasionally there is a long delay (as many as 5 seconds) between me pushing the power button and the Dell logo actually appearing on screen. This doesn't seem to be the case when booting off the battery and it doesn't happen all the time... so for all I know, it might actually occur with the 90W as well or be related to the setup I've got going at the time. There are so many variables including things like whether I'm using the PowerShare port, etc. As such, no one should jump to conclusions and run around saying the 130W is dangerous. My only thought regarding that is me wondering whether or not the delay is caused by the BIOS attempting to detect the adapter... which is does correctly as 130W.

At any rate, the sky isn't falling just yet. I'm definitely in the camp of people frustrated with Dell and was I not past my 21 days, I'd seriously consider just returning the 1645 and getting the Sony F11 (I had been looking at the FW590 at the time of my purchase)... but it's not like I'm suffering here. I'll have as much faith in Dell as I can muster that they will properly address this. Admittedly I will be waiting with bated breath for someone brave enough to flash the upcoming A04 BIOS. Good luck to those early adopters

 

Thanks for that. Can I ask what CPU and screen you have?

 

Here is what a found on the dell community
Our Engineers have been looking into what is causing the “7x BIOS BEEP” and a temporary fix that they have been able to replicate is to reseat the CMOS battery. Here is a link to the Service Manual that will explain how to do this: http://support.dell.com/support/edocs/systems/studio1747/en/sm/coinbatt.htm#wp1182375 The issue that is causing this problem to occur is the BIOS and it is not hardware related. Please keep in mind that replacing parts for this issue will not be a permanent fix. Dell’s Engineers are currently working on a BIOS revision that will address this issue.

We will be monitoring this thread and will provide updates as they become available.

 

This has nothing to do with throttling does it?

 

Apparently the new BIOS update is coming out on monday just read it on a Dell blog.
The dell rep also stated:

1. That using 130W power supply will NOT void your warranty and that they have shipped mupltie 130W power supplies already.
2. That a 130W power supply was found not to be resolving the issue and that subsequently a BIOS fix is being developed. (due on Monday)
3. He also stated he was happy for me to have th 130W powersupply if the BIOS upgrade did not resolve the issue, but only on the condition that I use a cooling matt , as the 130W PS would cause 'overheating'.

here is the link: http://en.community.dell.com/forums/p/19306277/19596373.aspx?PageIndex=8

 

i know they know about the throttling essue as well and they are working on both of the pblms.
Brad just posted twice , at least they are reacting to our posts.

 

I've had another think about this. Is anyone with an RGB screen and 820qm able to run a really high load test (super pi and future) to work out worst case scenario for battery drain?

 

Does any one know if they will provide the 130W and the cooling mat for free?

 

Sorry about that, the 720qm and rgbled. Also have the 120GB ssd.

 

I will try the new BIOS, with the adapter in my Watt-meter. I have a file for A03, so I suppose I can go back if I want to. I also have a spare 90W adapter, so if it breaks, I can keep working while Dell sends me a new one.

 

Just been given a call back from the guy I spoke to the other day, apparently from Dell Corporate office. He said he went back and researched what I was telling him the other day. He told me about the 130w workaround but he doesn't recommend it.

He then said it is a top level priority and they are working on it just now. He expects it to be available in 3 to 4 weeks!

Everyone seems to be getting different dates for a fix!

Told him I expect to be able to return my laptop if they dont fix it with the update, he said he's documented my calls and if i'm not happy I can return it but he assures me they will fix the problem!

 

The Auto Turn Off option was introduced to ThrottleStop a while ago which turns off ThrottleStop immediately if it senses you've gone from AC to battery power.

Even with this feature, I still don't recommend running Furmark + Prime95 and then pulling the plug. It may take a while after ThrottleStop stops before the bios is able to take over power management again. It's not worth burning through some batteries to test this feature further but during normal use or while gaming, it might help keep your computer running safely without killing the battery.

It's possible that I could build in an additional option so at the exact moment you pull the plug, clock modulation would be set to 50% which should be enough to keep your battery healthy. ThrottleStop would then hand over control back to the bios but there's no guarantee that the bios would restore your clock modulation setting to 100%. If the bios didn't introduce the modulation then it might take a while before it would restore it to 100%.

Thoughts?

 

Just received my delievry confirmation call.....mine is coming on tuesday morning!!

That's tuesday 12th so since Dell will 'definitely' be releasing a 100% fix on monday 11th I'm sorted!!!

 

With 130w AC adapter and TS, I resolve any throttling issues?
Or it is necessary a 150w power adapter?

Alienware 210w it works?

Thanks

 

For most games / applications, 130 watts + TS will be enough. To recharge your battery quicker maybe 150 watt would be better if you do a lot of gaming.

 

Thanks, I choose for 150w

 

Alienware's 210 is too much? Are the chance to damage something bigger than 130/150? (I think no, but I'd like to hear from you that for sure has better knowledge than me )

 

Hi unclewebb, please pay attention to this, it could be interesting:

Yesterday, after billion of tests, i started to get good framerates with PA-3E, while in the past three weeks i got just low ones with that power supply (and almost good on battery).

Don't ask me why, but yesterday i got 30 fps on Crysis (1920x1080, medium), 26 on Resident Evil 5 (1920x1080, high), while before i had half this rate.

Obviously OS profile was set on max performance on both (battery and AC).

I really can't explain this behaviour, but it's gone like this, global stuttering (for a month and a half!) and then good results (just yesterday!), the same good results i had for a month and a half on battery use.

I didn't change driver, bios or settings at all.

This video explains the situation before this strange boost of performance, i play a bit on AC (unplayable), then i unplug it and bioshock starts to reach 100 and more fps:

http://www.mediafire.com/?ntwmnwyjtm4

Take a look at the gpu (5th field from above), low load on AC but it loads fully on battery (675 MHz on both).
We don't see cpu clock clock mod (let's assume there was a hard throttling on that 8-9-10x multiplier), but i think it's more interesting the gpu behavour.


Well, this morning i received the PA-4E (130W) and downloaded last version of TS to give it a try, but it seems i have the same results i had yesterday with PA-3E, even setting 13x and clock mod to 100%.

It means good results, but not better than yesterday.

Could you (or anyone else) post a cpu or cpu/gpu benchmark (ie 3dmark vantage cpu score, or cinebench, or sisoftware sandra, or Resident Evil 5 internal gameplay benchmark, or whatever) in order to give me an idea of your performances with PA-4E and TS enabled?

For example my cpu score on 3dmark vantage is about 10000, while it should be about 26-28000.

I'm afraid my cpu still throttles...

Thanks in advance.

 

What's the model number of the ac adapter on the 150w?

So far I know that

pa-3e = 90 watt

pa-13/pa-4e = 130 watt

pa-7e = 210 watt

 

I think TS should carry a warning that it can void your warranty and cause potential damage/thermal problems. It overrides the softwares ability to regulate the processor clock and temperature to a degree. It's not worth using to me because of that and the fact that it doesn't seem to fix the problem.