RE: programming an rf synthesizer?

On 03-Feb-2003 David Young wrote:
> Seems like there are a few of you on the list who are programming wireless
> drivers, so I ask: has anybody had to program an RF synthesizer yet,
> especially the Si4126?

Programming a synth isn't that hard. As you say below you need to bit
bang a serial interface. Getting the numbers to programme it is harder!
I assume that you have no other documentation on this radio?
 
> Seems the ADM8211 does not provide a register for selecting the channel,
> but it does have registers for you to use to bang bits at the RF
> synthesizer over a serial bus.  Makes sense.  These ADM8211-based radios
> contain the Si4126 RF synthesizer. Also, I see a 44MHz crystal. Am I
> way off-base, or does the Si4126 divide the 44MHz crystal to produce
> carriers in the 2.4GHz band?

This is really nasty of them. I can't find the data sheet for a Si4126, but
there is an Si4136 (www.silabs.com). The really nasty thing is that this
contains three oscillators - only one of which will be used to change
channel. As there are two candidates on the Si4136 I cannot really guestimate
which one is to be used. Do you have docs on the ADM8211 (who makes it?)?
I would be able to figure it out from that.

Just download the Si4136 data sheet and it covers the Si4126 too. Ahh, Si4126
only has one RF synth on it. Okay. We need to programme RF2. Looking at the
datasheet there is a lot that needs setting up that cannot be guessed at.
do you have an initial programming set for any of the registers? To change
channel we only need to programme register 4, but to set the IC up we
need to programme registers 1, 2, 5, 7, 8. I can guess at some bits, but
I need a schematic (circuit diagram with component values)
of the radio itself to guess at some other values. This might not be too
easy to get...

As to how the PLL works - it will take the 44MHz, and divide it to 500kHz
(probably) using the R dividers. This 500kHz signal is used as the input
to a block of circuits that contain a 2.3GHz oscillator and some feedback
circuits. The circuits can be thought of as an amplifier that amplifes
frequency (it is actually a phase locked loop), and it's "gain" is
set by programming the N divider registers - to get an output of 2.342GHz
N would be set to 4684. You may ask why it isn't 2.4GHz, but something a
bit lower - it's all to do with how you design a superhetrodyne radio
(try google).

Duncan

-- 
________________________________________________________________________
Duncan Barclay  | 
[email protected]   | 
[email protected]| 
--
*bsd wireless list, a bawug thing <http://www.bawug.org/>
[un]subscribe: http://lists.bawug.org/mailman/listinfo/bsd-wireless/