It is an ATMEGA8L-8PI pdip:
onebitmusic.com/media/images/TristanPerich_OneBitMusic_DeviceOnWhiteCorrected_PhotoByTristanPerich.jpg http://tinyurl.com/9xjvb
The mp3's are quite interesting. I was expecting a single note two
channel system, but he's actually implemented a little bit of
polyphony. It doesn't appear that the pins are PWM, so he's doing it
in software, which is quite reasonable given that it's running at
8MHz.
On Thursday 15 September 2005 11:09 am, John Nall scribbled:
> Bob Blick wrote:
> >http://onebitmusic.com/
> >
> >Looking at where power is connected, I think this is a PIC.
>
> Well, sure, that is an easy one -- but WHICH pic?? :-)
>
> John
>http://onebitmusic.com/
>
>Looking at where power is connected, I think this is a PIC.
>
>Anyway, he fits 11 songs on it.
>
>Cheers,
>
>Bob
>
>
>
>
Nice case, too.
> It is an ATMEGA8L-8PI pdip:
> onebitmusic.com/media/images/TristanPerich_OneBitMusic_DeviceOnWhiteCorrected_PhotoByTristanPerich.jpg
> http://tinyurl.com/9xjvb
>
> The mp3's are quite interesting. I was expecting a single note two
> channel system, but he's actually implemented a little bit of
> polyphony. It doesn't appear that the pins are PWM, so he's doing it
> in software, which is quite reasonable given that it's running at
> 8MHz.
>
> -Adam
>
> On 9/15/05, Bob Blick <.....bblickKILLspam@spam@sonic.net> wrote:
>> http://onebitmusic.com/
>>
>> Looking at where power is connected, I think this is a PIC.
>>
>> Anyway, he fits 11 songs on it.
>>
>> Cheers,
>>
>> Bob
> Ahh, that's a much better picture!
>
> > It is an ATMEGA8L-8PI pdip:
Lots of processing power, and pretty good code space and ram.
I did a high speed thermal line printer on this chip once.
The only other significant componentry was a dual bridge to drive the
stepper motor, and a big mosfet to turn the printhead on and off.
It ran 832 pixels per line, and IIRC 2000+ lines/sec. I had to measure
the voltage and temperature, and calculate burn times while they are
happening, though a moderately complicated algorithm. Seven terms I
think, their FMULT instruction made that much nicer than it would have
been otherwise.
Another complication was that the PC driving it couldn't keep up, so I
had to smoothly throttle the print speed depending on the buffer state.
I could only hit the full rated speed of the printmech when running from
internally generated data, and generating the data took about as long as
printing it.
It's pretty cool watching the printed paper come out so fast that you
really think "It's not really printing that fast, it was pre-printed,
right?"
Assuming this was a 208dpi print head, then yes, 10" per second does
seem rather fast!
I knew that complex printheads required thermal managment for fast
printing, but hadn't considered the mechanical aspect - you can't go
10"/sec and then stop instantly (well, at 1/208th of an inch, anyway)
I can see a market for fast printers at best buy, though. They seem
to print out about 4 feet of receipt per customer (credit card,
purchase, contract and/or return policy, multiple rebate forms,
etc...)
I always enjoyed the cutter head testing. Advance one line, cut,
advance one line, cut, etc. Pull the strip out and it develops a
nice twist where the center was only partially cut.
I imagine for the fast printing you'd end up with either a seperate
mechanism to handle the cuting (and paper that backs up during the
cut) or you'd have to watch out for the cut command in the data stream
and slow things down as you got closer.
> > Ahh, that's a much better picture!
> >
> > > It is an ATMEGA8L-8PI pdip:
>
> Lots of processing power, and pretty good code space and ram.
>
> I did a high speed thermal line printer on this chip once.
> The only other significant componentry was a dual bridge to drive the
> stepper motor, and a big mosfet to turn the printhead on and off.
>
> It ran 832 pixels per line, and IIRC 2000+ lines/sec. I had to measure
> the voltage and temperature, and calculate burn times while they are
> happening, though a moderately complicated algorithm. Seven terms I
> think, their FMULT instruction made that much nicer than it would have
> been otherwise.
>
> Another complication was that the PC driving it couldn't keep up, so I
> had to smoothly throttle the print speed depending on the buffer state.
> I could only hit the full rated speed of the printmech when running from
> internally generated data, and generating the data took about as long as
> printing it.
>
> It's pretty cool watching the printed paper come out so fast that you
> really think "It's not really printing that fast, it was pre-printed,
> right?"
>
>
>
>
>
> Assuming this was a 208dpi print head, then yes, 10" per second does
> seem rather fast!
Yup.
> I knew that complex printheads required thermal managment for fast
> printing, but hadn't considered the mechanical aspect - you can't go
> 10"/sec and then stop instantly (well, at 1/208th of an inch, anyway)
No, that doesn't work well. Even with all that head friction, inertia
still wins.
The Seiko docs for the printhead are just plain wrong, their math
doesn't evaluate for one thing. I went through the pain to put it all
into mathcad, and mathcad didn't like it much. It seems they don't want
to really tell you what's going on, because they also make printers, and
they want to keep you in the dark a bit.
One thing I did that really helped, was to measure the printhead voltage
right at the printhead, 100uS into the burn cycle. Then I applied the
remainder as a second timed burn. This gets fun when you have 13 head
segments, and might need to fire as many as four full burns (note that
13/4 is not an integer!) and as few as one. You also have to keep the
motor in sync with all this, so it doesn't step until all the burns are
done, AND it's own timer has expired. If the motor timer expires before
the burn, then you can recalculate where you need to be in the motor
acceleration table, etc.
You can imagine what happens when you breakpoint in the middle of a burn
with the printhead drawing 9A @ 7V. Keep lots of spare printheads
handy!
> I can see a market for fast printers at best buy, though. They seem
> to print out about 4 feet of receipt per customer (credit card,
> purchase, contract and/or return policy, multiple rebate forms,
> etc...)
This was a small portable printer.
Same print mech as this one, but it was designed to be an embedded
system component. http://www.mobilecommand.net/
> I always enjoyed the cutter head testing. Advance one line, cut,
> advance one line, cut, etc. Pull the strip out and it develops a
> nice twist where the center was only partially cut.
I've not played with cutters yet, they always end up too expensive for
my clients/employers. I've thought up some other ways to do the cut,
but haven't had the opportunity to try them.
> I've thought up some other ways to do the cut,
> but haven't had the opportunity to try them.
>
Using the printhead? I imagine testing that would go through a few
printheads as well :)
> > I've thought up some other ways to do the cut, but haven't had the
> > opportunity to try them.
> >
> Using the printhead? I imagine testing that would go through
> a few printheads as well :)
That actually works, but the results are not pretty.
>> > I've thought up some other ways to do the cut, but haven't had the
>> > opportunity to try them.
>> >
>> Using the printhead? I imagine testing that would go through
>> a few printheads as well :)
>
> That actually works, but the results are not pretty.
>I've not played with cutters yet, they always end up too expensive
>for my clients/employers. I've thought up some other ways to do
>the cut, but haven't had the opportunity to try them.
Like leave the printheads turned on and burn through the paper??? ;))))
Oh, no, not another fire alarm - we'll have to slow those checkout girls
down, they have set off the smoke alarm again ...
sonic.net> wrote:
