Buyer Beware! - Your power adapter may be cheating you!

If you have a small format (65W) AC power adapter, your notebook may be underpowered and thus not giving you 100% of its performance.

A simple test may reveal that on some of the Lenovo notebooks, especially the ones equipped with a high end processor (tested on X61 tablet 7764-CTO Intel Core 2 Duo processor L7700 1.8 GHz), the small format (65W) AC power adapter is inadequate. This may happen primarily on custom configured notebooks, but it would not hurt to test yours as well.
In contrast, a 90W AC power adapter has the capability to bring your notebook to its full potential, sustaining the maximum Processor Frequency speed (100%).

Lenovo does not offer options and refuses to deal with the issue, as this looks like a gross and costly engineering mistake.

The test:
Remove the notebook battery and power your notebook only through the provided small format (65W) AC power adapter. Otherwise, as someone else pointed out, the battery will act as a big capacitor supplying extra current and the notebook will perform normally as expected.

How to get the performance parameters on your notebook:
A) If you have Vista, simply run the Resource Monitor tool (the button is available on the Performance tab in Task Manager). You may see that the Maximum CPU Frequency of your machine is reading 66% (not 100% as it should be). This is regardless of any OS and BIOS settings you may have (that is even if you have set your machine for maximum performance). You may also follow the directions specified below, as they are general and not just valid for Windows XP.

B) If you have Windows XP (or would like to run a more conclusive test), you can use any of these utility programs available for free (just Google and download it):
1) Intel Processor Identification Utility (On the frequency test tab - Reported Core Speed will not get above 1.2 MHz)
2) CPU-Z (On the CPU tab - Reported Core Speed will not get above 1.2 MHz)
3) RMClock (On the CPU info page - Core Clock and Throttle will not get above 1.2 MHz)
4) SuperPi (This tool will not reveal directly the Core Frequency of your processor. However, being a very processor intensive program, you can actually time your machine when running with a fully charged battery attached and then without it - you may notice a difference in performance ~33%. For example try to calculate pi using a large number of decimals, like 1 Million).

 

Also, there is a Vista desktop gadget available on the Microsoft site that continuously monitors CPU speed and usage, per core. That is an interesting finding about the 65 w adapter. When I ordered my t61p, after I had set up the CPU, video card, and ram, the 65 w was not available as an option. Only the 90 w was available for either the main supply, or the travel charger. Lenovo must be aware of the power usage after the options are specified.

 

The 65W adapter is designed for non-p models. The 90W is for the p models. Everything was as it should have been .

The power supply is meant for conditions which fall under the TDP of the processor and the other components. I believe Intel calculates the TDP as 85% of processor usage. Everything is working as intended, Lenovo just doesn't expect the normal user to tax the processor 100% (Though I run MATLAB simulations all day long on mine, and have never had any issues).

It's not an engineering issue, but an assumption about the average user's processor usage. I would be shocked if this affected any users.

 

I have a 14" T61 with the 140M and I have a 90W power adapter.

 

I just checked Lenovo's site. It's 65W adapter for a T61 with integrated graphics and 90W for a T61 with discrete graphics.

 

Huh, that's odd. According to Lenovo's website, the 65W power supplies are:

"Compatible with all ThinkPad X60 Series notebooks and with select ThinkPad R60, T60 and Z60 Series and Lenovo 3000 notebooks ( with integrated graphics only). Also compatible with the ThinkPad Essential Port Replicator."

 

Makes sense then. I guess the 65W is only for the integrated graphics.

 

Same config here, I also have the 90W adapter. Another thing I've found, if you have the T61 plugged in and power manager is in the "balanced" mode, it Resource Monitor's CPU Maximum Frequency will read under 100%. When I changed the mode to "High performance" my max freq. went to 100%.

 

Where did you find this max frequency information?

 

Notebook draws power from fully charged battery when its connected to AC? I doubt it.

 

As per the first post:

How to get the performance parameters on your notebook:
A) If you have Vista, simply run the Resource Monitor tool (the button is available on the Performance tab in Task Manager). You may see that the Maximum CPU Frequency of your machine is reading 66% (not 100% as it should be). This is regardless of any OS and BIOS settings you may have (that is even if you have set your machine for maximum performance).

 

Guess that's why Window's WEI cpu score falls with battery out.

No matter how much fiddling around I did with bios, settings an' so forth, that cpu perf score always fell - machine felt less responsive too.

Of course, "solution" is just leave the battery in.

 

Erm, mine is only 50% maximum :O is something wrong? But as I observe the chart (real-time) there are occassions that it cross 50%...

So what's going on? I'm using 65w adapter, which came with the lappy.

 

Here is my finding on my new X61 with Penryn T8400 2.4Ghz:

Battery alone (on Max performance): 100%
Battery and 65W adapter (on Max performance and charging): 100%
65W adapter only (on Max performance): 49%!
-Althought, I have seen the CPU spike to 86% when opening programs.

So what does this all mean?

 

Ok..

14.1" ThinkPad T61 w/ T7300 (2.0GHz), Quadro 140M, 2GB RAM, and 90W adapter.

Core Speeds recorded with CPU-Z:

AC & battery (charging) on Max Performance scheme: ~2000MHz
AC & battery (not charging) on Max Performance scheme: ~2000Mhz
Battery alone on Max Battery Life scheme (lowest CPU speed): ~1600mHZ
AC alone on Max Performance scheme: ~2000Mhz

Conclusion: I get a maximum core speed of 2.0GHz on AC or AC+battery and a minimum core speed of 1.6GHz on battery. Basically, I lose 20% of the max CPU core speed when on battery; 1.6/2.0 = 80%. Not bad.

 

I downloaded CPU-Z and here are my stats for my 2.4Ghz Penryn (I'm not impressed with the 65W AC only):

All set to Max Performance:
AC Only: 1200 MHz
AC + Battery (charging): 2400 MHz
Battery only: 2400 MHz

Set to Max Battery Life:
Battery only: 1600 MHz

So basically, my laptop performance with AC only is slower than my Max Battery Life setting. That's just wrong.

Is there any way to fix this? I don't like to leave the battery in the laptop when I have it on AC power because I don't want it to drain the battery.

Also, how can I reduce the CPU speed even more than Max Battery Life?

 

"Of course, "solution" is just leave the battery in."

I've been dubious since the beginning that this thread is representative.

Who removes their battery when attached to a charger?

I have a maxed-out t61p with a nine cell battery. At home I use a 90 watt adapter and at work I use a 65 watt adapter. At work, the 65 watt adapter is quite adequate to run the system and slowly charge the battery over the course of a day. I leave my battery in the system at all times and I achieve 2.6 ghz with both cores.

A processor draws relatively current when it is saturated with compute demands. It's quite possible that the 65 watts charger sans battery is not capable of supplying those current spikes. When in the system, the battery looks like a huge capacitor across the charger supplies extra current for high compute situations.

I don't understand how it's meaning to remove the battery and then say the 65 watt power supply is robbing cpu power.

 

If you leave the battery in while connected to AC, the power will go from the AC to the battery, to the laptop. So the laptop will continually draw power from the battery. This will greatly shorten the life of the battery.

 

Greatly shorten? I think that's a little overkill. I've always kept the battery in since day one and after over half a year, I've only lost about 3-5% of the maximum capacity of a 7-cell. IMO, that's not a great loss at all and is quite understandable. I'm on my 112th battery cycle by the way.

I'm on the same boat with Renee; I don't ever use my notebook without the battery in.

 

Interesting. Maybe they have improved the battery/charging performance. Because with my old Toshiba laptop, I left the battery in all the time and after 1.5 years, the battery couldn't hold a charge for more than 2 mins.

With my next laptop, an Asus, I always took the battery out when on AC power, and now, after 2 years, the battery is still holding 80% of its charge.

What do you set the battery charge setting on?

 

Like my charge thresholds?

Well, I go to school and I'm a commuter. Before going to bed, I set it on full charge and charge it 100%. Then I wake up, drive to school, use my laptop, and drive home. When I arrive home, my battery life is at about ~10%. Then I repeat the process. Apparently, I've repeated it 112 times.

On the weekends and days off, I use my laptop with the battery charge lingering around 40% by using the charge thresholds.

 

I just noticed under Vista 64 the Lenovo Power Manager is not showing battery cycle count. My system is about eight months old and I never take the battery out and the battery is still in good condition. I work concertedly to not cycle it very much.

 

I'm honestly not sure why this issue gets brought up with the same old info every time. Really the only reason to take the battery out while on AC is if you are either using your laptop on a HOT surface (as heat kills li-ion cells much faster than normal) or if you have a MBP and game a lot as the location of the battery and the case material of the MBP are not ideal for keeping the cells cool.

The reason that some are seeing the CPU locked below 100% on AC with the battery removed has nothing to do with whether you have a 65W or 90W adapter. For some reason, Intel and probably AMD as well have put this logic into the chipset themselves, most likely as a precaution for those situations where people might be using a third-party, underspecced AC adapter. In those cases, yes, the battery would provide the extra wattage needed.

In the case of the 65W or 90W adapter "problem", the reason it downclocks is solely due to the chipset trigger mentioned above. If you want to check and see for yourself, charge your battery to 100% and then run something that pushes the system hard for an hour or two. Your charge percentage should not change (give or take a percent or two for normal idle charge loss over time).

This isn't just with Lenovo's notebooks. Any somewhat current notebook will exhibit the same behavior with regards to downclocking on AC without the battery.

 

Why don't you run a poll and find out? Who are you to tell us what to do and what not to do?

If you lack understanding for other people's way of life, that does not mean you get to say what the norm is.

 

If you are not sure, it means you don't see the entire picture and you do not understand. Who are you to proclaim there is only one reason to take the battery out while on AC?

I, as many others, for various reasons, want to use the notebook without having the battery attached.
How can you tell me I have to use the battery, when my aim is to have the smallest format (no bulging/protruding battery), a lighter equipment in my hands and no compromise. No one can dictate my preferences.

This is not a matter of seeing the processor locked down. It's a matter of wanting to get full power on your notebook and not being able to do so, no matter what, if you used the 65 W adapter alone.

The information you are trying to pass on is false, and here is how wrong you are:

1) Booted with the battery (no AC power adapter). OS started.
2) Plugged in the 65W AC power adapter and then removed the battery.
3) Ran Super Pi with 1 Million decimals once: 44 sec.
4) Ran Super Pi with 1 Million decimals second time: 43 sec.
5) Ran Super Pi with 1 Million decimals third time: 43 sec.
6) Attached the battery, then switched to the 90W power adapter and then removed the battery.
7) Ran Super Pi with 1 Million decimals once: 28 sec.
8) Ran Super Pi with 1 Million decimals second time: 28 sec.
9) Ran Super Pi with 1 Million decimals third time: 28 sec.
10) Attached the battery, then switched back to the 65W power adapter and then removed the battery.
11) Ran Super Pi with 1 Million decimals one (last) time: 43 sec - it never fails.

Note: No reboot was done between the sets.

What do these tests tell you? There is a clear difference between the two AC power adapters. If you choose not to see it, there is nothing much I can do about it.

Even if the cause of the downclocking is the built-in logic of the chipset, the fact remains that you cannot get 100% performance on the 65W adapter without the use of the battery.

The "normal idle charge loss over time" plus the fact that the battery is already connected to a power source, will trigger short charging cycles every time the threshold limit is reached. You end up with unwanted lost cycles if you never remove the battery from the notebook.

Because I made sure not to leave the battery attached to the notebook unnecessarily, I am still able to use it after more than 5 years, with plenty of cycles to go through.

I have not heard of the "any somewhat current" notebook brand yet. Do you have a name and a model type so people can test it? Since when has this become the norm? Where is this information made officially/publicly available?

 

You're awfully angry, for no apparent reason.

The computer's BIOS is dynamically downclocking the chip probably based on the input wattage, perhaps using the battery to condition the power input.

If you're really needing every spare cycle you'll probably want to avoid losing your work if the computer loses power, and thus would want the battery in anyways. If you're not pushing the processor, who cares? It's not running any slower if there's no load.

This is all academic for the vast, vast majority of users.

 

Please keep remarks such as this to yourself. You are in no position to judge me or anyone else.

My notebook was bought in North America. How often do you lose power on the AC adapter?
Maybe we should have a second battery connected, just in case the first one fails... after a power loss.
Again, a matter of feeling and preference...

You continue to say things that cannot be verified. Is there any material published about how people are supposed to power their notebooks? Is there a warning if you do not comply with the supposed "norm"?

As in many other cases, when I bought my notebook, I did not have an option to chose the type of AC adapter.
Even today, if you want to buy a X61 tablet, you can only chose the type of your second (optional) AC adapter.

 

even if you are correct ibm1, there is no need to go ****ing angry.
btw, SPI isn't that great of a CPU benchmark, its more a memory one.

 

"Who are you to tell us what to do and what not to do?"

Can you locate where I have told anyone what to do?

 

I have been noticing the same 'problem' on my R61, and have been wondering why the core speed of my T7100 never goes beyond 1.18 GHz. I run my notebook without the battery attached almost all the time as well. I will post more after running more tests. Thanks for the post.

 

Here are my results. All tests were ran with the bios profile set to 'maximum performance'.

With only 65 W AC adapter
Maximum core speed: 1.18 GHz
Super Pi 1M: 46 s

With 65 W AC adapter + battery
Maximum core speed: 1.8 GHz
Super Pi 1M: 30 s

Wow. I wonder if Lenovo has this clock throttling 'feature' published anywhere. 'Coz I'm feeling kinda shortchanged.

 

That's correct, but "some" are seeing? Some? Well, sir, I don't know how close you looked but include me in. The speed lock "feature" is well known and has been around for a while. In Linux you can get around it by hacking the DSDT table (difficult but feasible) or by resorting to some quick and dirty kludge (boot, hibernate, unplug battery, resume). In Windows, I have no idea.

That's an interesting conjecture. It is based on... what exactly, your faith in a commercial enterprise pursuing a benevolent agenda too profound to be expressed in words? And what chipset are you talking about? There is nothing in the Speedstep or ACPI specifications that calls for capping the CPU when the battery is unplugged. It makes no sense whatsoever. Have you seen this "feature" advertised anywhere? By ANYONE? Of course not, because the "feature" is a blatant rip-off and something to be ashamed of. It doesn't come from Intel but from the OEMs. It's in the BIOS. No, you can't turn it off, and good luck waiting for Lenovo or others to fix it. Why should they? It motivates the user to buy an expensive battery pack every -what- 1-2 years, right?

You know, UltraC, it may well be that Lemonovo (and Apple, and Dell) is doing you a favour by crippling your CPU "for some reason". But sometimes the simplest explanation is the correct one. Your battery will eventually deplete and die no matter what. Have you considered the possibility that Lenovo may have something to gain by forcing you to purchase MORE batteries rather than LESS? Assuming you want to run your laptop at the speed you thought you paid for, that is.

 

Try and say that to my face again after reading the following:

This is what actually happened...

When I called Lenovo Technical Support in January 2008, reporting this problem, the agent I talked to, said the behavior was odd/wrong and they did not get any reports of this kind. He also said there may be something wrong with the motherboard and they may need to replace it (!).

Knowing there is nothing wrong with the motherboard, I told him I would like to try using a more powerful (90W) AC adapter. On the 25th of January I placed an order online for such an adapter. It "only" took 10 weeks and a formal complaint to be delivered.

After I have completed my tests with the more powerful adapter, I have called again Lenovo Technical Support (end of April 2008). This time a different agent acknowledged the problem , created a case number and told me I will receive a replacement adapter within 2 days.
When I asked him what type of adapter he was going to send me, he said he has no way of knowing, and that I have to wait and see.

One week later I called again to ask about the status of my case and got to talk to a different agent. He said there is no such case number under my name. I asked to talk to the supervisor of Technical Support, who told me this is not a warranty case, since the adapter is not defective, and I have to sort it out with Sales.

I called Sales right away and they said it's not within 30 days of purchase and I have to call Service. When I called Service , the person I talked to, told me this is clearly a case of Warranty and referred me to Technical Support. The new person I got to talk at Technical Support, checked with the supervisor and hung up on me saying there is nothing they can do about it.

When I get the chance I will tell you all about my experience with Lenovo Customer Complaints department.

 

@Ibm1

If I understand correctly, you suspect that Lenovo is shipping laptops with an underpowered 65W adapter, leaving it up to the battery to fill the gap. I doubt that's the problem, unless you are running a fully populated maxed-out system with all ports connected to power-hungry devices.

Can you try the following:

1. Disconnect all peripherals (USB devices, printer, etc.).
2. Remove the battery.
3. Plug in the 90W adapter.
4. Boot your machine WITHOUT the battery.

Do you get the full CPU speed? Does the frequency scale correctly?

Needless to say, you should check by running a benchmark routine (i.e. don't rely on CPU meters, load indicators, or numbers returned by calls to the hardware).

 

He and other people here did it and that is the problem really. CpuZ AND benchmarks show 100% CPU when on 90w brick only.

And something's wrong with the 65w combo, no doubt. The only question is, are you killing your battery by not removing it?
Someone suggested that this may be a bios protecting from low power 3rd party power supply and that this 'feature' is disabled by plugging your battery in (in this scenario system doesn't really use the battery like ibm1 suspects). I guess that the only way to test this is to use the 65w brick, plug the battery in, let your machine work at maximum capacity (long benchmark for example) for a relative long period of time and see if battery is fully charged all the time during this test. If thinkpads rally use battery to help 65w adapter to power them, this would be manifested by battery discharging.

My r61 is on its way and after reading this thread, i'm thinking of buying my self a 90w one (because of this and less carrying around).

 

I called Lenovo Canada yesterday about my concerns over this problem. They are now sending me a 90W adapter no charge since I am still under warranty. The CSR was very nice about it.

 

Just FYI, I'm not sure why there's so much sudden finger pointing in this thread. No one is trying to tell anyone what to do with their notebook, that's always up to each user. All some in this thread have done is discuss the topic that the OP posted and tried to bring some new info to light.

That aside, the clock lockdown on AC with battery removed *is* a known behavior of practically every modern notebook out there, it's NOT simply Lenovo notebooks that do this. Check around on NBR or any other hardware forum that deals with notebooks if you doubt this, it's been talked about and checked before.

 

Then how do you explain that the laptop doesn't exhibit the clockdown behaviour with the 90W adapter? Anyways, I'll post my findings after I get my 90W adapter from Lenovo.

 

That would be something worth finding out if possible. Speculating here, but I would guess there's something set differently in the BIOS as to what behavior it follows depending on what the adapter is reporting as max wattage to the notebook.

 

That's not what the guy did. He booted and THEN removed the battery. It may or may not make a difference, either way I'd be very appreciative if he put a final word to it by booting without battery and see what happens.

Make that yes doubt for me.

There is no question that the clamping of the CPU is real, what is unclear is what triggers it and why. I strongly suspect a whitelist routine in the BIOS similar to the one for the mini-PCI card. It is a form of vendor lock-in and it has nothing to do with power requirements. The BIOS looks for the battery; if it doesn't find it, it puts you in the slow lane until you recognize the value of paying premium dollars for an original battery pack. Lenovo wouldn't be the first to play this kind of games behind the user's back. It's the industry dirty little secret and is widespread.

I'd be happy to be proved wrong.

 

Just to chime in here: My T61 (T9300, integrated graphics, 65w adapter) will have its CPU clock ramped up to maximum - 2.5GHz - when running Prime95. The ONLY discrepency is with CPU-Z: It says the clock speed is 2.4GHz instead of 2.5, but RMClock reports the CPU ramping up to 2.5GHz under load.

I'm no electronic/electrical engineer, but I'm quite sure that having to coordinate drawing power from the battery AND AC adapter at the same time requires some pretty fancy circuitry. It should be much cheaper to throw in the 90 watt adapter in the first place.

@ibm1: I also feel you have been acting hostile to some of the members here. We are not attacking you; we're all in this together trying to figure this out. If this problem is really true, you might want to get angry at Lenovo instead of us.

 

Why not go into the BIOS and set the power settings for BATTERY/AC to MAXIMUM instead of AUTOMATIC?

 

That's what I did for AC. It's the SpeedStep options in BIOS. Set both AC and Battery to Maximum Power.

 

tallshorty, which number did you call, Technical Support, Service or Sales?

disco_inferno, my initial tests were all without using the battery (booting without the battery). This is how I discovered the problem in the first case. Never have I tested the notebook with peripherals attached.

 

Have you used the battery... you did not specify this detail in your post.

My intention was not to offend anyone.
I have all the reasons in the world to be angry at Lenovo, but in this forum I felt I got the same treatment like on their official website. It made me feel like I talked to the same people, who were not listening and just interfering with nothing much to add, but things like: I doubt, maybe, who takes out their battery ... and so on...). All this after I presented facts. I did not ask... "what do you think about this, could it be possible?".

And then, there were people who got angry because (they said) I was angry (?!). I have reasons to be angry, how do they justify their anger? They are angry because someone else is angry?! This would constitute a priority case for a psychiatrist.

So, by a raise of hands... how many Lenovo employees are here?
For you and everyone else, I intend to post an even more incredible story about Lenovo Customer Complaints department, and then you will see more reasons to be angry.

Something has to change. The fact they've already sent to someone a 90W adapter as a replacement, due to the mentioned problem, is a step forward. At the same time, it means I have a discrimination case on my hands, since they refused to provide the same service for my notebook which is still under warranty. This goes beyond belief.

 

Just an UPDATE:

I called Lenovo Canada regarding this issue and they pleasantly agreed to send me a 90W adapter at no charge I got the adapter today.

It is HUGE compared to the mini 65W adapter. Also, it came with an adapter for charging via the cigarette outlet in cars and on airplanes and a pouch to carry everything.

I ran CPUZ and here are my findings with Maximum Performance setting:

1) With charging battery and 90W adapter: 2400 MHz
2) With only 90W adapter: 2400 MHz
3) With only battery: 2400 MHz

With my 65W adapter:
1) With charging battery and 65W adapter: 2400 MHz
2) With only 65W adapter: 1200 MHz
3) With only battery: 2400 MHz

CONCLUSION
So this pretty much proves that using the 65W adapter without battery reduces the CPU performance for some unknown reason. Maybe it does not provide sufficient power, maybe it's all a conspiracy, who knows.

BUT

Either way I'm happy Because of this issue, I now have 2 adapters. The 90W adapter I'll leave at home at all times connected to my X6 ultrabsase dock and I'll keep the 65W adapter with me on the go.

 

Sorry I didn't reply sooner; haven't been keeping track of this thread.

Just tested WITHOUT the battery: CPU-Z reports 12x multiplier and 2.4GHz speed; RMClock reports 2.5GHz. According to my kill-a-watt meter, the laptop is drawing about 50-54 watts of power at this time (screen at max brightness). I'm not sure how much power the integrated x3100 graphics requires @ max power, but it appears to me that the adapter is suitable for the laptop's hardware.

 

Glad the adapter did the trick for you. I will have to try eventually, although I'm a bit hesitant to sink more money into the Lenovo pit at the moment.

For the technically minded, here's a good overview of the problem:
http://bugzilla.kernel.org/show_bug.cgi?id=7060

The bloody chunks:

Without disputing your findings (as I said, I don't have a 90W adapter), the above explanation seems quite convincing to me. It is consistent with my amateurish attempts to troubleshoot the problem on my Thinkpad. Incidentally, my TP also has 3 p-states. Do all thinkpads have 3 states?

Other than that, the only people who know for sure are the Lenovo developers. As far as I know, they aren't talking. Or even acknowledging that the problem exists.