• Please review our updated Terms and Rules here

You get what you pay for

NeXT

Veteran Member
Joined
Oct 22, 2008
Messages
8,095
Location
Kamloops, BC, Canada
Right, about that TransNote.

When I got it the battery sucked. Being the original battery it was in no mood to see life beyond five minutes. Unlike almost all the other Thinkpads made after 2000 however the battery for the TransNote never saw enough demand for chinese batteries to surface on ebay so at best all you can find are other dead batteries or NOS batteries on Amazon which I wouldn't hold hope for being all that good. My first great idea was to send the battery in for a rebuild and leave the dirty work for them. Half a year and two returns later it was apparent that Radio Shack could not do the job and I couldn't find anyone else in town who would do it either. I was left with no choice but to rebuild it myself.

CGS_0397.jpg


This sounds really easy given that there's pages all over the internet on recelling your pack but the TransNote had to be different from the others like a retarded kid eating gravel on the playground. Unlike the others which use the super common cylindrical "18650" cells I use the flat rectangular "103450" cells, not to be confused with their lithium polymer cousins. These are apparently not only harder to source but more than twice the price ($25) compared to other first-brand 18650 cells. I was forced to look around and see if I could buy other chinese battery packs that might hold them because then at least I had some sort of assurance the cells would have a high amp rating. After failing to find laptop packs holding the right cells (sure there are photos but no model numbers) I looked elsewhere and found the NP-120 camera battery. It was basically a single 103450 cell with a charge controller. They could be found for amazingly cheap on ebay so I took a gamble and bought four for $4 each.
The first batch gave me a few useful bits of info. For starters once gutted you could see the batteries were brandless wonders. I had no hopes of a real life boost at this point because off-brand chinese batteries generally suck, especially when they are not rated for computer use. The second thing was that unlike 18650's, it's illegal for battery stores to spot weld 103450's in my part of the world. Apparently their chemistry is different and more unstable so unless the tabs are added at the factory nobody is going to help you at the risk of the cell exploding. This also limits me to buying cells that are pre-tabbed so brand new first brand cells off ebay or wherever were out of the question and I wasn't about to purchase a spool of ribbon and a spot welder for one job. In my first batch of cells I had already ripped the old tabs off before knowing this. Oops... $18 later I had another four on the bench.

CGS_0926.jpg


>>IMAGE<<

Again in typical chinese fashion the second batch of batteries were different. This time the tabs for both polarities were on the same side. The wiring inside the pack would have to be rearranged to fit this change.

THIS...
Transnote_old.jpg


BECAME THIS...
Transnote_new.jpg


The tabs would be used to attach everything together. Electrical tape was used to separate the cells to keep them from shorting. A strip of cardboard was used to also prevent shorts. Once the pack was reassembled it dropped into the old plastic enclosure.

CGS_0928.jpg

CGS_0929.jpg

CGS_0930.jpg

CGS_0931.jpg


From here I trimmed back the leads and carefully reinstalled the charge controller. Once everything was soldered in (and a notch was cut in the tape so the temperature probe contacted the physical cells) we were ready for a charge.
CGS_0932.jpg


>>IMAGE<<

When dropped in sans AC adapter the computer POSTed so the rebuild was a success so far but the battery was in need of a charge. I plugged in the AC adapter and left it to do its work, occasionally checking to make sure the cells were not overheating.
CGS_0934.jpg


The final results were amazingly lackluster. With the defective cells the rebuild service had given me before later blowing one of my safety fuses and trying to give me the run around the warehouse I at least had 45 minutes of life. With these mysterious chinese cells the charge rate never exceeded 690ma and after an hour had bottomed out at 10ma. I might be putting too much blame on the cells there. It could also be the charge controller was completely confused with the new cells and it had not yet seen a full cycle.

>>IMAGE<<

Another 45 minutes later I finally decided it was ready and give it a discharge.

>>IMAGE<<

30 minutes tops. The life is even worse than before. Sure we can conclude out of this that new cells will work but unless you can get cells from a more reputable brand like Samsung or Sanyo or Panasonic AND you know someone who can spot weld the tabs on for you I doubt you'll ever get optimal life out of the battery again.
 
I'm getting a fair number of mixed signals from the Hackaday article. For starters I submitted this a week ago and in the past few days I made a few discoveries that originally I did not mention or find. For the record:

1 - I don't know why they said that 103450 cells were rare. I only said that they were very uncommon to find both new and with tabs already spot welded on.

2 - The original pack was dismantled and my wiring diagrams were based from that. Polarity was also properly noted. There's no miswiring or reversed cells.

3 - The cells are not rated for the job. The key thing that the charge controller is expecting is one specific discharge rate.
BM-S23-Fig1-300x287.jpg

In comparison to the original Sanyo cells, these new cells level off one to two volts lower. The controller it seems cannot modify the pre-programmed and expected curve so it prematurely gives battery level warnings which brings me to the next note...

4 - My initial tests did not have hibernation turned off. It was still set to hibernate at 7%. The new cells would level off at 6% and hang there for another 35-45 minutes. In all I do not have 30 minutes of life but a hour and a half. Using properly rated substitutes for the Sanyo cells will probably give me both an accurate reading and the two and a half hours I should see.

5 - The charge controller is properly adjusting itself without a hard reset or EEPROM modifications, even with the incorrect cells. Originally my battery capacity was 45 minutes and about 5Wh. With the new cells and several complete power cycles this increased to 45 minutes and 8.5Wh which when you calculate the average consumption is almost spot on with the hour and a half I was seeing with all power management switched off.

6 - Some noted that my extremely low charge rate was because the controller had slowly notched it down over time and a controller reset was required to fix this. This is false.
survival.png

As a test I removed my new cells and reinstalled the old pack. The charge rate jumped back up to 1.5Ah and remained steady. On discharge I also watched the controller change my capacity back down to 5.85Wh so it's averaging what it's seeing and trying to guess what the capacity of the pack is now. Again I'll have to put blame on the new cells for the terrible charge rate. Something about them seems to make the controller cautious and as such refuses to let it charge any faster.

7 - Yes the pink cells pictured are from Samsung. Before I sent the pack off for rebuild it was filled with Sanyo cells. When it came back and had a moment to photograph the innards it had Samsung cells. When it came back from rebuilding the second time it still had the same Samsung cells but the idiot doing the rebuild must of shorted the pack (which is why I got a call from the shop saying the pack could not be repaired) because the thermal safety fuse was blown and had to be replaced. I'm just glad I got my $85 back. If they had blown the controller I would of murdered someone....

My initial point still stands. You get what you paid for. I decided to cheap out on lithium cells and opened a can of worms. Have I leaned my lesson? Yes.
 
Back
Top