Hi,

this topic is about the PCB for a wireless master in the OpenHR20
project.
A thread about the OpenHR20 firmware can be found at
http://embdev.net/topic/118781

A german article about OpenHR20 can be found at
http://www.mikrocontroller.net/articles/Heizungsst...
(information might be a bit outdated).

Some schematic of a PCB can be found in the SVN repository at
http://openhr20.svn.sourceforge.net/viewvc/openhr2...

The 10-pin connector is the same as the one on HR20, so there is JTAG
and a RX/TX. For connection to a router as WRT or FritzBox, a
serial-to-USB converter might be necessary. But I'm not sure whether a
FTDI driver would work on a FritzBox.


Feel free to contribute if you are interested in designing or if you
already have a prototype and want to share your findings.

Regards
Marco

Ok, seems that the project Freetz [1], a custom firmware for FritzBoxes,
supports the FTDI serial converters [2].

Would it be better to integrate a FTDI converter on the master board, or
connect two boards using the 10-pin connector?

And is it ok to supply the master board from USB?



[1] http://trac.freetz.org/wiki/freetz (german)
[2] http://trac.freetz.org/wiki/libs/ftdi (german)

my first try (I obviously not completed routing yet ;)

ok, next version :)

If it is connected directly to a serial interface, the FTDI USB
converter part can be cut.
All parts except the coil are mounted on the top, and even the coil
could be mounted on top if it is isolated a bit.

I've no idea what kind of antenna should be used on the RFM12 modules,
so feel free to contribute, also if you find some errors :)

br
Marco

Hi Stan,

Yes your board looks good. I have add ,at my master, a second led which
show the transmission of the sync packet. This looks usefull for
diagnostic purposes. Maybe it could usefull to connect some additional
portpins to pad for expansions.

bye

Ronny

Why not use a mini USB ? like
http://www.reichelt.de/?;ACTION=3;LA=444;GROUP=C12...

Ronny

Hi Ronny,

great idea! Are these four enough? Is it ok to use the INT-inputs? They
are located at a free corner ;)

br
Marco

Hi,

here are the changes for version 0.4:

jdobry wrote:
> - current SW need some communication buffers and use ATmega32 not
> ATmega16. Pinout is same, difference is volume of RAM (and Flash)
Atmega16 changed to ATmega32.

> - RTS# signal for RESET# can be problem if we want to use simple
> terminal for debugging. Can you add jumper into this signal?
done

> - Antenna can be lambda/4 wire, for 868 it is 86mm (L[mm] =
> 75000/frequency [MHz])
> - Antenna can be path on PCB, due to PCB effect it must be 10-20% shorter
> than wire. Wire is usually better.
ok, I keep the the via to solder a wire.

> FSK pin is not used and in this case it need pull-up resistor. My
> original design use pull-up inside MCU, it save soldering of one
> component. Just a trick, nothing more.
external pull-up resistor added and connection to ATmega removed.

> Please remove R2+C5 components. ATmega contain better startup reset
> circuits inside and it complicate external reset signal connection
> (capacity load).
well, if they are not populated there's just 1 mm of extra wire on the
PCB.

Test point number four is replaced by an extra LED as suggested by
Ronny.

Any more suggestions?

BR
Marco

here's some kind of preview :)

Of course L1 (at the bottom) is not a resistor, but it is the most
similar package.


Another idea:
would it be better to rotate the RFM12 by 180° so that the antenna is
located at the left edge of the PCB? Or is it ok right now because a
vertical antenna can be used?
What do you think?

Thanks Jiri :)

I'll add a point for soldering a GND connection and tomorrow I'll place
my order.

Hello People,

I will modify my code to work with your LED port and send a patch in the
evening.

Ronny

Marco G:

You can compare rendered image with real here:
http://openhr20.svn.sourceforge.net/viewvc/openhr2...
Only one difference except missing FTDI is used inductor. I was use
10uH.

Could you send me source files for update documentation?
Today I plan to replace old prototype board with your PCB and in this
case source code will not fit into schematics in SVN. Thanks, Jiri

OPPS!

I found problem. This design can be used without FTDI with 3.3V power.
But with FTDI and USB power RFM12B will be overpowered.
Recomended power range for RFM12B is from 2.2V to 3.8V
5V from USB will not damage it (maximum is 6V) but it is outside range.

Jiri

Yes you're right, I intended to use a RFM12 (without 'B') on master
because of the 5 V USB supply.

AFAIK there's no difference between RFM12 and RFM12B, but I'm not sure.

here is differences between RFM12 and RFM12B:
http://news.jeelabs.org/2009/05/06/rfm12-vs-rfm12b...

Its easy to fix: remove L1 and use the 3.3V Ftdi Internal
Voltageregulator it drive up to 50mA - is that enough for the rfm12???

@Ralf H.:

50mA is enought

- RFM12B have 26mA maximum
- ATMEGA32 (5V/10MhZ) need 15mA maximum
- and remaining 19mA for LEDs

Problem is that you must crop VCC layer on back side, because FTDI must
be powered from 5V, remainnig part from 3V3

PS: I was also use RFM12B and I miss information that pull up on FSK is
not needed. It is not problem on this PCB, but I was can save soldering
of one wire and resistor inside valves.

Jiri Dobry wrote:
> Could you send me source files for update documentation?
> Today I plan to replace old prototype board with your PCB and in this
> case source code will not fit into schematics in SVN. Thanks, Jiri

I'll send them to you this evening :)

Ralf H. wrote:
> Its easy to fix: remove L1 and use the 3.3V Ftdi Internal
> Voltageregulator

It's not that easy:
VCCIO must be connected to 3V3OUT so that the UART pins are driven with
3.3 volt. Otherwise TX would still have 5 volt high level which would
cause problems with ATmega.

> VCCIO must be connected to 3V3OUT so that the UART pins are driven with
> 3.3 volt. Otherwise TX would still have 5 volt high level which would
> cause problems with ATmega.

I thing that it is easy, see to the pictures.

- 1 wire to crop.
- 2 new wires
- do not use inductor

Connection on bottom side for modification.

sorry, I haven't got any left.

But I would fix the bug mentioned above and order some new ones. The bad
thing is that they need at least 2 weeks to arrive...

Hello,

i have one board more then i need. but not the simular design. its a
little bit smaller. and some components are in 0402. I get these boards
next week.

Ralf

I've added my schematics to the repository, including a fixed version
0.6