Undervolting - The difference is SHOCKING - over 2000 points in 3Dmark2005

For those that are not absorbing the calculations, taking the official specs of the battery inside the M1530 and doing the numbers, here is the breakdown:

(Especially for all of you that have wasted thread space in arguing over this)

The M1530 11.1v 56Wh 4400mAH 4.4 amp battery delivers 48.84 watts.
The M1530 running full bore needs about 88 watts.

The battery in the M1530 is never going to be able to deliver the performance of the AC adapter. Power plans are irrelevant.

The 8600M GT needs 20 watts (1.02 amps @ 19.5 vdc ) from the power adaptor. It would need 1.8 amps from the 11.1 volt, 4.4 amp battery. That only leaves 2.6 amps for the entire rest of the system. The CPU all on it's
own can consume 2 amps. The rest of the system cannot run on 600 milliamps. These are laws of physics and electricity.

I am praying that we can drop this and move on.

Edit: for the sake of anyone reading this stuff, you would also have to account for what you can refer to as a 'fudge factor'. When you fudge it, you can be safe with a number like 90% in this case (I don't know the actual factor percentage, and have no idea of how to obtain it). It allows for you to account for the losses incurred with your circuit. The electronic commponents, their conductors, and their voltage conversion circuits take up a certain amount of power. This is a loss for the system. That will be our fudge factor. So at 90%, the M1530 11.1 v 4.4 amp battery can really only give about 44 watts to the system. There is no such thing as a perfect (100%), no loss power delivery system.
The actual efficiency of the power circuit in the M1530 could possibly be better than this. I don't know for sure. I can only promise that there is an amount of loss during power conversion.

For comparison, desktop PSU's are rated this way: good quality = >90% efficiency, bad quality = <78% efficiency. Power losses are incurred when a notebook system has to convert the 11.1 volts provided by the battery to the various voltages needed by the system.

 

7oby, you're killing me. Go look up the specs on the battery in the M1530.
They are : 11.1vdc 56Wh 4400mAH 4.4 amps. 4.4 amps is what that battery is rated for. It would melt trying to provide 75 amps.

Have you ever watched the UFC? Do you know what a submission is? Well I am in a Rear Naked Choke right now, I am choking, and I am tapping out.

Please, for the love of God, can you understand the physical limits of the battery?

Edit:

 

The battery isnt a bottleneck on my system, I get the same 3dmark05 score whether I use ac or battery.

 

I give up. I am sorry that I ever started this thread. You people should be thanking me for breaking down the capabilites of the M1530 battery so that we can dispell the apparent misunderstanding. I can't believe how horribly wrong this thread has gone.

Edit: 7oby, this especially for you.
The M1530 needs 90 watts at full bore.
An 11.1 volt battery delivering 75 amps would have to disipate 832.5 watts of heat.
Your machine, using 90 watts, creates a temperature of possibly 80c as measured by the ACPI thermistor.

Can you imagine the temperature that would be produced by disipating 832.5 watts?

Do you really believe this can happen?

 

Do I mis-understand - or are you calling me a liar?

My battery appears to be giving me enough juice to match my ac performance - you state this should not be possible.

 

I did. Can't find anything regarding maximum current:
http://accessories.us.dell.com/sna/...x?c=us&l=en&s=dhs&cs=19&sku=312-0663#Overview

And even if it would, I couldn't believe it until I tested it, since it's not a limit of the lithium ion technology itself. A current restriciton mechanism inside the battery wouldn't help, the laptop would simply crash if it wouldn't get the current it requires.

All Dell can do is provide some safety mechanisms that run the laptop in DC mode in a more safe way (e.g. by reducing GPU clock).

I mentioned already: DC vs. AC mode is a complicated matter.

You started the original posting with: Undervolting vs. Non-Undervolting 3dmark05 scores. I suggest we get back to this issue here. And for everything else a different thread could be started. Or change the subject of this thread.

 

mc1, I swore I was done with this thread, but this is just for you.

No. I do not state that this should not be possible. The battery manufacturer does. I have merely taken their official numbers, put them into formulas that all of us electronics students had to understand, and I have given the answers. I did not create the laws of physics.

 

I was talking about lithium ion technology itself when mentioning that number.

This is pointless since the M1530 doesn't consume 832,5 Watt. Thus the battery is never discharged by that amount.

 

The only other case he describe where the 3dmark is higher when undervolted is bunk, the guy overclocked AS WELL AS undervolting.

Does slowdown think his results are typical and everyone else is abnormal? Is he trying to flex his last semester EE/Physics knowledge? It won't matter how many formulas he throws up until we start to get people reporting similar findings on their 3dmark scores.

 

I humbly appologize on the power calculations, as I may have been misled. I saw another page with the specs for this battery and it says 6600 Wh. That would mean that it can produce 6.6 amps and not 4.4 amps. If this is true, then the battery can produce 73.26 watts. However, even if this is true, 73.26 watts is still going to fall short of the power requirements to run full speed.

Edit: I will add however that it looks like the specs for the bigger battery for the M1530.

Edit: I looke again, and that is the capacity for the larger battery. My numbers are good.

 

Man, I gotta stay away from this place. Where the h*ll are the other guys in the forum? We laughed and drank beer and helped each other solve heatsink problems. Those were the good old days.

 

LMAO. I'm DL'ing 3dmark05 right now and i'll post scores with my cpu at normal and undervolted with no overclock.

EDIT: Btw, to stop my undervolt I can just kill the RMclock program running in my taskbar, right?

 

paperkut! Hey! One of the guys I was just talking about.

I'm pretty sure that should do it. What I usually do is disable to profile in RMclock, or I just exit RMclock. I have tried both ways and it is the same effect.

Edit: paperkut, you give me hope for this dreadful thread.
Edit: I'm waitin for those numbers paperkut.

 

We are not yelling at you or anything, just trying to correct your misunderstandings. Your perceptions on what a battery can handle are completely wrong. If you have a voltmeter you can test this yourself. You would actually need two. Connect up a battery to a low resistance load, could be a resistor with a really low value and measure the voltage drop and current draw. I have li-ions with a 2Ah rating that can supply 4A or even higher current draws at the same voltage of the li-ion battery just for a short period of time.

If what you say is true please explain to me the helicopter question. My radio controlled helicopter draws about 15W and the battery has only 5Wh. I have measured with multiple different multimeters. It has enough charge to only fly for about 10 minutes but can definitely fly.

Check out this link: http://www.rcgroups.com/forums/showthread.php?t=61226
18650 cells are what is used in laptop batteries and most are stating that a 1800mah battery can sustain a 4A output.

Here is some more data showing that 18650 cells can sustain a 2C discharge meaning twice their mah rating.
http://www.batteryspace.com/index.asp?PageAction=VIEWPROD&ProdID=945
http://www.batteryspace.com/index.asp?PageAction=VIEWPROD&ProdID=4869
Those are the cells used in laptop batteries.

That thread shows 1.6Ah batteries under a load of 5A, 10A and 15A all done by computers with graphs displaying runtime. Explain to me how a battery with 1.6Ah can put out 15A with your understanding, what you are saying is that a 1.6Ah can only put out 1.6A.
http://www.candlepowerforums.com/vb/showthread.php?t=205814&highlight=18650+test

1600mAh means that is can put out 1600mA for a period of ONE HOUR.
OR 3200mA for 30 minutes etc. This is also why power supplies are rated in A and not Ah.

 

slowdown: Hate to be the bearer of bad news but I'm seeing a negligible difference with or without an undervolt (no gfx OC).

Undervolt: 8698
Normal: 8717

The response from this thread is overwhelmingly that an undervolt does not change scores that drastically, and that's what I'm seeing. There's gotta be some other factor that's influencing your scores, it's a bit of mystery to me :/

 

Think I read somewhere that on some processors that there are fractional multipliers that round down on RMClock and effectively underclock the processor? Maybe?

 

Thanks paperkut for testing. And I think that I have found part of the answer.
I decided to run the same tests under Vista. I got the same exact results as you guys. Very little difference.

no undervolt : 8333 CPU max temp = 79c GPU max temp = 80c
undervolt : 8430 CPU max temp = 71c GPU max temp = 77c

So as it turns out, for some reason XP Pro is giving me excellent performance while undervolted, and dismal performance when not undervolted.

johnny, I have already taken the time to take the official specs of the 11.1 v 56Wh 4400mAH 4.4 amp battery, do the math, and explain what that battery can and can't do. I'm not going over that anymore. And I don't know much about the battery in your hellicopter, so I will avoid that altogether.

Please, let us continue to move this thread in a positive direction.

Edit: the only way that the figure I give of 48.84 watts is wrong, is if the specs that I found were wrong. If you find any specs that are contrary please post them. Then we can create another thread that talks about batteries. I just don't want this thread dragged down anymore with talk about batteries.

 

I am glad to know that I got the same results as the rest of you in Vista. That tells me that I have a perfectly normal M1530 and normal PSU.

Now to figure out what in the world is going on in XP. I am going to try some different drivers, but my initial thought is - why would XP give different performance between undervolted and not undervolted due to a driver?

 

For those that have followed this thread.

At the begining of this thread, I trusted the results that I was getting while running XP Pro. It had not occured to me that the huge difference I was seeing between normal and undervolted could in any way be influenced by the OS.

After running the tests under Vista, as the others had been doing, I received the same results that they did - very little difference in 3Dmark scores.

So my original assertions appear to be wrong at this point. For what ever reason this is occuring in Windows XP, I do not have an answer. If I discover an explanation I will post it. But at this point I would also say that my judgement appears to only be half wrong, since this behavior of such a large difference in 3Dmark scores, is definitely occuring in Windows XP. It would be nice if someone else with an M1530, running XP could run the tests and post the results. Some consistent behavior could be established.

One thought that has occured to me is that maybe Vista has a more effective way of power distribution and it is at least possible that Vista can get the same effect of RMclock in a dynamic way. Vista could be able to detect that the CPU does not need full available power, and that the GPU does need it's full power for it's workload. This is nothing but a theory of mine, so I am not saying that this is true, only that I think it's possible.

In any event, I would take XP with RMclock over Vista any day. But I do want to know why XP has different behavior. In the end it really doesn't matter. The result is that XP is awesome when you undervolt.

This thread went horribly off track and began disussing batteries and power. I appologize for my part in derailing this thread.

 

I just wanted to mention my own personal experience with what Slowdown has mentioned here:

I had a PSU that was rated 500w, and I had an 8800GTS 320. The average scores that I was getting in XP with 3dMark06 was around 7200 give or take a few.

Anyway, the PSU died a while ago, so I replaced it with a 750w psu.

All things being equal (Drivers, Software, OS, etc) the *ONLY* change made was the PSU. Now I'm scoring around 8600 ish.

I do know quite a bit about the server/software/hardware side of machines, but I only know enough about the videocard hardware to know that drivers make a different.

Being that the ONLY thing that changed was my PSU, it would seem totally logical that the score increased because there was enough power to drive the chipset to doing what it needed to do.

I know with overclocking, to get some higher speeds you often need more voltage to keep it stable, so it would make logical sense to me that you would need more power to sustain higher performance.

Just my take on the experience anyway that I thought I would add.

**Edit**
I just wanted to add that I am also using XP on the machine that experienced this behavior
**Edit**