AVR-Chat

At 09:16 AM 7/30/2004, C Sizer wrote:

>Hi guys,
>
>I am designing a university project, a hand-held printer using the 
>AVR and a small thermal printer. For portability I want to use 4 x 
>2300mAh AA batteries in serial/parallel to provide me with 3V.

Shuttle or line mech?  Line is much faster, and much more demanding.
You need roughly the same horsepower as a laser printer of the same size would.

>The printer runs on 5V but consumes over 3.5A peak, so I require a DC-
>DC converter circuit (or at least a starting point) to provide 3V at 
>5A peak.
>
>Has anyone designed anything like this, or can give me some advice? 
>Maybe someone has an idea for a better battery for me to use?

You'd be far better off to put your batteries in series, and then buck regulate the 3V.
There are pulse-skipping converters that are very efficient. If the current is low enough, you can win with analog again..  Then again, why not run on 5V?

Also, do the math on burn energy, you'll find that the faster you print, the less energy you spend.  A significant amount of your print energy goes into heating up the printhead mass.  Wake up, go like hell, shut down.

Dedicate a timer to the stepper, and another to the burn.
Use the burn timer like an alarm clock, interrupting you when the first head group is done, then reload for the remainder of the burn on the second group, etc.

You can start the burn pulse with a nominal value, do the calculations (FMULT is your friend) and then change the timer setting to the calculated burn width.

The stepper will need to be accelerated, and decelerated, depending on how much rasterized data you have built up. 

The SPI port is great at blowing out data to the head at high speed. 

I did one early this year, using a Mega-8 with on-chip osc, and a stepper driver IC.
It was capable of running the mech at full speed, which ended up to be faster than the 1.7 GHz laptop could send the data out.

You'll find the details of getting this going properly, quite interesting.