Tuesday 11 November 2014

Rebuilding old Power pack, LiPo-edition (Hitachi B-2 for DRC-10K)

Old cordless drill with a bew batterty pack
Drill with the first LiPo -pack I made some months ago.
Apparently my fathers venerable old Hitachi DRC-10K cordless drill fails to die, so once again there are need for battery refreshment build. Earlier days (see earlier post ) I did use NiMh cells, but the times have moved on.

When I checked the prices of those tabbed Sub-C cells from net, I noticed that can get a decent capacity LiPo power pack at the same price. So i decided to give these new packs a go, thus I did the first conversion some months ago.

Realistically it's only 1200 mAh if even that.
During this time I've been listening feedback from my father about this idea. And since he's happy with drill performance and lighter weight, I'm now making second LiPo battery pack, and finally taking couple pictures for this blog as well.  :-)

LiPo vs NiMh
So much smaller and lighter.
One of the biggest advantage for using LiPo batteries in this kind of project is the wide availability of ready build packs. These packs are used on various radio controlled systems, like cars, planes and quad copters. On first version I  purchased brand battery and balancing charger from RC hobby shop.

For this second build, I decided to cut some corners (and cost). This LiPo battery shown on these pictures is a cheap Chinese knock-off, with obviously incorrect markings printed on. Also the size is somewhat smaller so I'd expect the real capacity being lower as well. Chinese sellers are well known on their inflated markings, but I'd expect this battery to serve well at this application.

Ready to close zip ties, balance port at left.
While comparing LiPo to NiMh, the size difference is quite noticeable.  One needs six NiMH's cells for 7.2 volts while there are only 2 cells at this LiPo pack I used as an replacement. And even thought those old cells had more power in them (over 2000mAh) while on optimal charge, the NiMh cells high self-discharging meant those batteries were rarely well charged the moment of need.

Charging with Turingy balancing charger.
In the end this conversion was not a big job. Electrically it's just desoldering and removing the old cells and connecting the new ones to existing battery pack contacts. The biggest part of work went onto B-2 -pack case modification. These rechargeable lithium batteries need to be balanced while charging, so there's need to have their balance port (white 3-pin connector shown) routed out of the pack.

At the time of this writing I have no idea how long these batteries will last, but I'd expect them to stay good at least few years. If the new charger behaves, there should be no problems for some time. 

Friday 19 September 2014

OLPA . Or a project long not done..

This hiatus has been a lot longer than I have hoped, but there's finally a new project at my desk. It's an OLPA (One Laptop/lunchbox Per Adult), named after the famous OLPC from some years back.

Ever since the OLPC project was announced I have wanted an homemade portable laptop/lunchbox type computer with OLPC like Screen, with black/white reflective reading for outdoors use. The same screen works as a regular color LCD when used indoors with the back light connected. Even thought this type of screen uses more power than E-paper,  it's  also is a lot faster to refresh so it's better suited to web surfing etc.

Currently I have tested the display, controller, Pi alongside wireless KB/Mouse and system is quite usable even though as fast as today's commercial laptops/tablets. Next part of the project will be making/purchasing suitable enclosure and installing these components to make basic system.

Picture of Reflective LCD screen
Reflective display + parts.
While deciding the enclosure there's also the aspect of making this portable computer system battery operating. I hope to get system running at least 3 to 5 hours on lithium battery, but 10 hours might be achievable without adding too much bulk or weight. Raspberry needs 5Volt feed and the screen+controller can run on almost anything between 5-24 Volts DC, so it would be quite easy to use one of those premade battery packs sold for laptop/phone users.

On future upgrades.. I allready have wifi -module that's supposed to be supported by Raspbian which I have not yet tested. Using wireless net will eat power, but on the other hand it eliminates the need for ethernet-cabling. Also alternative CPU board is under consideration. At the moment this system is based on Raspberry Pi, but Banana Pi looks like valid alternative since it offers overall better performance with less power use.

At the moment, there's not much point making any video of this. When(if?) I get this to packed and working ... perhaps.

Wednesday 19 February 2014

A new scope.. Well... At least it's new for me.

Hi, this is an update of my long time project. I finally got myself an oscilloscope. Well, technically I have had a real old one for some years but finally I have a DSO. Rigol 1052, there at the left corner atop of my old Russian cathode ray tube scope.



And my electronics desk needs some cleanup, too.


Rigol 1052 is over 5 years old model, so it's nothing new nor fancy. But it works... mostly that is. I got a good discount on this one, since the long memory (1Meg) is partially broken. As you see at the picture, the second trace has nasty noise overlaid atop of probe calibration signal.  I took some time to test the scope and noticed that there must be something really odd at the digital part of the system. Analog front-end and AD converters work just fine and the problem appears only at one of two channels. (see picture)
Direct capture of Rigol screen.

When I got the scope, I even took some time/effort to open it up and look if there could be anything obviously wrong at the board. But as far I could see everything looked like it should be. So, I just closed it back up and keep it as it is. In theory there could be a cold joint or short right under Atmel FPGA, that causes bad data to be moved to external ram. But it's highly unlikely I could do anything to fix the board here. So it's better to use the scope as it is and not break it any further.


Not much to see here.



All in all, the good thing is that this error does not appear at all if the Rigol is set using short sample memory. So, even though this scope has some (non typical, I may add) issues, it's still quite usable for  the needs I have. And the price was right for me, so I can't (nor want) to complain.