Re: TI-H: TI85-irlink

[Prev][Next][Index][Thread]

Re: TI-H: TI85-irlink

pea@dalnet.se wrote:
>>Well, there have been at least two announcements of new IR link designs,
>>but I don't think they're worth much..
> 
> Don't underestimate our stuff before testing it. I can say to to you my
> friend that if you want a I2C protocoll
> just write to me and tell me how you want it and I made it (I now how it
> work!). This product that I have made is using the TI85-protocol. And do you
> really want the expander SF protocol? Let me remind you: Any protocol that
> waits for the other part and isn't sending duplex can be done with the ir.

By any protocol I really meant any protocol, because we might have a whole
new protocol sometime in the future. The only way to have such extensive
support is to make the link completely emulate the link cable without
interpreting the protocol somehow. So you need to send the current states
of the link port at regular intervals to the other side. This way, when one
calc pulls one pin down, it is pulled down by the IR link device on the
other side as well, regardless of the protocol in use. Especially in I2C it 
is also important that more than one IR link device can receive the signal 
and act accordingly. This way we can have an IR network of calculators.

Besides, I2C is a quite complicated protocol. I think it's easiest to make
an I2C IR link with this type of generic protocol support, not I2C-specific
support, because the I2C routines themselves would need much programming
for the PIC (or whatever microcontroller) and take up much of the ROM code.

> About the distance I have tested up to 15 min daylight ( No problems! ).
> That is with a 100 mA ir-led. I have seen a siemens 300 mA ir-led and that
> one I guess reach with out problems 30m. 

Light spreads in two dimensions, so double power doesn't mean double range
(okay, you had triple power and double range, but that's still too
optimistic). I had a sort of IR link, and when I tested it with two IR LEDs
the range was only a few percent better than with one LED.

> You sofware nerds are so limited. I
> have looked at components that can reach a the speed of 12000 bits/s wich is
> a TI85-transmit with new batteries. This speedlimit you talk about is only
> depending on the component SFH-506 x khz. With a special transistor I can do
> any frequency I like. In future I not going to use a ir-modul. But with
> SFH-506-30 I can soon reach at least 1000 bit/s. 

I know the raw speeds can be quite high. But you need to encapsulate the
link port protocol in another, and that soaks up much of the bit rate. At
least this is the case if you do it in the non-protocol-specific way, as I
explained before in this message.

> And don't underestimate the
> power of a PICxxxxx. It can do almost anything. About the size of it all I
> can say: With a 3V small model of Pic you can make it really small with two
> 1,5 V small batteries as supply! 

It can do almost anything its programmer can do. :)

> I have a friend that is developing a pure hardware irlink without a
> microprocessor. Magnus Svedin is his name. He has helped me a lot. His
> hompageadress: http://www.ts.umu.se/~el97msn. 

I tried to, but then I gave up. Maybe I should try again, I've since bought
some components that should better suit the design.

-Ozone

> P.S With you I mean you all! And I sleep 10 h every night else I can't work
> so good with my stuff : )!

Sleeping is one of the lowest priority tasks that I have...I'm very busy
with schoolwork and an organization I'm working for...
I just slept five hours after school. =)

--
*** Osma Suominen *** ozone@clinet.fi *** http://www.clinet.fi/~ozone ***
References: