I have designed a board for CC1101 (868 MHz) and MAX2430 125mW amp. After testing everything I will make another board that will have a final amp stage with MAX2602 for 1W of power.
The law said that 1W on this band is legal when using freq. hopping with a hop rate of 50 hops-per-second.
I am about to order these bords from BatchPCB and I would like you RF people to let me know if I’ve done something wrong with this board design and I am thinking about the RF. [The DRC bot said that design has passed the required BachPCB check]
Regards :roll:
UPDATE:
I am giving away few of my PCBs so if someone would like them, please contact me on PM.
I was think of making a board almost exactly like this soon (but with a pic added).
The only obvious problem I see is that it doesn’t look like your using RF traces with a characteristic impedance of 50 ohms. This is only important for electrically long traces so it probably wouldn’t be a huge problem on your board since it’s so small.
Look up Microstrip Traces and grounded coplanar waveguides for info on making 50 ohm traces. Calculators are available online but the microstrip trace widths needed on two layer board are very large (~112 mils).
I have used 54 mil traces with wider ground traces on either side and underneath, connected on each side every 100 mils by a via. With a 10 mil gap between the middle trace and the ground traces this works out to be 50 ohms.
In short it mostly looks good but do some reading on transmission line theory.
Thanks for the info. I looked up those transmission lines a while ago, and downloaded HP’s software for those things: Agilent Technologies AppCAD.
In every documentation of parts I used on my board: CC1101, MAX amplifier, and HCS RF switch, they say that the input and output impedance is matched at 50 ohms. As you mentioned, when using very short RF lines as in my case, impedance shouldn’t be a very big problem. These lines could be much shorter, but I will go with this design.
In datasheet of MAX2602 1W amp that I will use in future as final stage, they say that I really should use transmission lines of 50ohms because the transistor’s input and output is not matched at 50ohms. This is a problem because the board will be very large and that is just wrong :?
I guess I will have to figure out how to make this 50ohms impedance with RLC components at 868 MHz. I think that AppCAD can calculate those values, but we will see
I have built the device. It works, sort of. I am having some problems with firmware now. I am using Atmega16 for testing purposes and strange things happen. For example, in CC1101’s documentation there are some registers that should allways and only have their default value. For example: Chip version should be 0x04, but I get something else. I allways get the same response so the return-values are not random.
Something good IS happening but in the wrong way
I tried to find some sample C-code for Atmegas but I only found library and samples for TI’s 430 DSP micro controller.
Is there anyone with some Atmega’s C that could help me, I am going crazy here
with regards to the 1 w amp you know that the antenna will increase the power as well. unless you are going to use a really shit antenna you are going to be exceeding the 1 W pretty easy.
every 3dBi you are doubling the power so with your 125mW amp you can only have a max of 15dBi antenna before it is illegal, sure there will be losses but you dont want to go above 1w.
If you are doing this on the cheap as a home project. have a look at cantennas here is a website http://www.turnpoint.net/wireless/cantennahowto.html you will get around 10dbi maybe more if your lucky and it is pretty directional as well.
Other wise look into just getting a good antenna rather than increasing the power with a 1w amp.
Muris:
The law said that 1W on this band is legal when using freq. hopping with a hop rate of 50 hops-per-second.
I just want this 1W of power to see the range I will get…
And, I am doing this only for fun
Thanks for the antenna info. I’ve built a few cantennas already and they are cool.
In the future it might be possible that I will make a few base stations with omni antennas, and a few mobile units packed into a Nokia 1100 package. This will be used for an SMS-like application - an open source&hardware for free chatting-fun.
The CC1101 modems I sampled from TI.com last year, I think.
I just can’t get them to work properly. Whenever I write something to CC1101’s registers and read back, I never get the expected result. For example when I programm the channel number I never read-back the value I just programmed in. I assembled 2 units, the first one is the one you see in the picture (please note: tantalum capacitor is turned the other way around, it has been fixes before powerup) and it just didn’t work.
Then I assembled the second with only CC1101, Crystal, Voltage stabilizer and other passive components that CC1101 require to work, and I get the same result.
After powering up the CC1101 modems I issue the SRES command (reset). After that, when I read the initial register values I don’t get the values that are in the technical documentation. For example:
I read the “0x31 (0xF1): VERSION – Chip ID” and the PDF say that the value should be: 0x04, guess what - I get different chip version.
I read the “0x2B: AGCTEST – AGC Test” and the PDF say that the value should be: 0x3F, I get 0x3E.
I read the “0x2A: PTEST – Production Test”, the PDF say that the value should be: 0x7F, I get 0x7E.
And yes, everytime I read some register I allways the same result (for that register) so I don’t get any random results. It all kind of works, but not the way it should :oops:
This is making me crazy. Can someone help? Do you have CC1101 or CC1100 modems that really work so you could test my firmware??
chipcon modules - no. For 1/2 mile and less 250Kbps I just spend my $20 on XBee modules for 802.15.4 on 2.45GHz. Plug and play, to include wireless UART emulation.
On 900MHz, I have used Digi’s modem modules. More like $75ea, FHSS, 1W. Lots of range.
Have also done FHSS on 900MHz with 2MHz QPSK and got 5-8 miles’ range, using a 9dBi yagi on one end and a 5dBi omni on the other end (a boat).
A guy mentioned that you could contact him for how to set registers if you have trouble with it. Since the CC1101 and CC2500 seem register and pin compatible, this could be of help.
