ThinkPad X220 power connector repair

My ThinkPad X220 at some point started to complain that there was no power when I plugged in the DC power adaptor. After a while, the power would come and go (signalled by an ominous sound signal) as I moved around with the laptop in my lap, or even when I didn’t move at all. After a while, using the computer became an exercise in finding a «good» position for me and the laptop, and then not to move, at all.

My expectation was that this problem was due to the traditional broken solder joint with a power connector sitting directly on a fixed PCB. When I opened up the X220 however (see guide here, do clean out the fan while you are in there), I found that the X220 power connector was beautifully designed with the connector itself being clamped down and then connected to the motherboard via flexible wiring. I therefore concluded that the connector itself must somehow be bad, even though I was unable to confirm this visually or by measurement.

I ordered a new connector, and when it arrived I did the whole process again, fitting the new connector. That solved the problem… for about two hours. To my amazement, the exact same problem reappeared almost immediately after my «fix».

This called for another, more open-minded inspection of the problem. That is, I opened up the laptop once more, and tried to plug in the power connector while I could see what was happening. This revealed the actual problem:

I first clamp down the power connector, then try to insert the plug.

The connector simply isn’t fixed down properly, and moves away when you try to insert the plug. In other words, the connector itself is not the problem, rather there’s a mechanical problem with holding it in place.

It appears that the root problem is a tiny tab of metal on the laptop housing that is broken off. This tab would be so thin that I’m not even sure it was ever there, but this is the explanation that makes the most sense to me. In that case, even just a moderately excessive push on this connector would break off this tab, causing this issue. You can verify this problem without opening up the laptop simply by poking at the connector from the outside with a pen etc. to see if the connector disappears into the casing.

Having identified the problem, repairing it is in principle trivial. However in practice I found it not so easy. There’s no obvious way to weld or glue a new «tab» or back-stop back on. I ended up just stuffing the area with some material from earplugs, which leaves the connector flexible while still supporting a proper connection. This has worked well for me for a few weeks, although I suspect there are better solutions.

My solution to fix the loose power connector on the X220.

ASUS tf201 tablet battery revival

We have a (by now) old ASUS Transformer Prime tf201 tablet that has been collecting dust for many months, because it wouldn’t charge. I figured I’d make an effort to revive it, and found it was impossible to turn it on, even when plugged in. Only occasionally I’d see a red battery symbol indicating low or bad battery. Also, the little LED in the power button would do triple blinks rather than stay solid red that would indicate it’s charging. No amount or combination of button-presses would convince the tablet to even attempt to boot up.

Some googling indicated that the triple blink of the LED means that the battery level is so low that the tablet considered it broken and would not charge it.

I assumed I would buy a new battery to replace the broken one, but thought I’d open up the tablet and have a look first. There are youtube videos showing how to open up the tablet, and picking out the battery was fairly easy.

I figured there would be no harm in making an attempt at reviving the battery. That is, there is of course potential harm in messing about with lithium batteries, so do proceed with caution. The battery unit consisted of two battery cells and a logic board. After pulling back some pieces of tape, two flat «wires» to each battery cell were exposed. A multimeter revealed a voltage of about 2.7 volts for each cell, which is certainly on the low side, but also not beyond rescue. The multimeter also identified the positive and negative terminals.

For the kids’ RC cars I have bought an «Imax B6» smart charger. It has modes for charging LiPo and NiCd batteries, among other technologies. I connected this charger to one of the battery cells, using this type of test probe connector:

Note I did this outside, wearing safety glasses and spider-alert-sense on «high».

Firing up (pun intended) the Imax B6 in LiPo mode, unfortunately it refused to try to charge the cell due to «too low voltage». Sigh. Well, a new quizzing of google revealed a trick to use the somewhat dumber NiCd mode to charge the LiPo cell up to a minimal voltage, before proceeding with a LiPo-mode charge. So I set the charger to NiCd and 0.5 amps, and let it run for an hour or two until a voltage of about 3.3 volts was reported, while making sure that the temperature of the cell was not going high, nothing was smoking or turning into a hitherto unknown lithium-based lifeform. Then, switching the Imax charger back to LiPo mode, I’d charge the cell for a few more hours up to some 3.8 volts, and repeat the whole process for the other cell.

After reassembling the tablet, lo and behold the thing booted up happily, reporting about 70% charged. To my considerable surprise, the battery has also been quite good ever since, holding its charge for a very reasonable amount of time. I’m tempted to say it’s good as new. I’m assuming this procedure might work for similar tablets and maybe also telephones.

Finally, when I booted up the tablet the first time, the root of the problem also revealed itself. Trying to charge the tablet by normal means, i.e. with the original charger (the tf201 requires a special 15 volt USB charger…) and cable proved to be very difficult. Turned out it was the USB cable itself that was half-broken, only making a good connection if held in a very carefully selected position. I’ve now bought a new cable, and everything works quite well. (To give you an idea, a good hour of youtube took about 10% battery.)

Again, to perform this procedure you should be very careful, and know your batteries, amps, volts, watts and wattnots.

Toggle your linux monitor by DPMS

I have at long last configured our MythTV appliance to use an remote control via a homebrew serial IR receiver and LIRC. One problem was how to make the «power» button turn the monitor on and off. I finally figured out that the following litte script did the trick, at least on my ubuntu machines it will toggle the monitor on and off by DPMS.

#!/bin/sh
if (xset q dpms | grep -q "Monitor is On"); then
        xset dpms force off
else
        xset dpms force on
        sleep 0.5
        xset s reset
fi

Veetle on 64-bit Linux Ubuntu Natty Narwhal

It seems that the way to get Veetle running on current 64-bit Linux Ubuntu (Natty Narwhal) is to run the following «sort-of-beta» self-extracting file, as announced on the Veetle forums: http://veetle.com/download.php/veetle-0.9.17plus-linux-install.sh

Lenovo Thinkpad T400s Ubuntu Linux CPU frequency scaling

I have a Lenovo Thinkpad T400s running Ubuntu Linux 10.04 lucid. While running the mythtv frontend there was exessive video stutter, apparently because of performance issues. The CPU load was very high. At the same time, I noticed that the CPU frequencies were being scaled down–the 2.4 GHz rated CPUs were every second or so dipping down to 800 MHz, and at times running more at reduced than full speed. This made no sense, the CPU was struggling while being held back by the power-saving subsystem.

I tried to fiddle with the various CPU scaling options, «Conservative», «OnDemand», «Performance» and so on. None of them seemed to «take», and the actual CPU frequencies remained erratic. Even chosing the absolute «2.40 GHz» option didn’t have any apparent effect.

The problem appears to be that the BIOS CPU scaling is fighting against the Linux kernel counterpart, making the aggregate effect erratic. I went into the BIOS setup (hit F1 at bootup) and in the «Power» settings set to «max performance» (or something to that effect) for both battery and AC modes. After this, CPU frequency scaling seems to work well. Video playback is now nice and smooth, also in «OnDemand» mode.