QwikRadio

tm

10

July 1999

MICRF002

MICRF002

Micrel

1. Selecting REFOSC Frequency ft (FIXED Mode)

As with any superheterodyne receiver, the difference

between the (internal) Local Oscillator (LO) frequency flo

and the incoming Transmit frequency ftx must ideally equal

the IF Center frequency.

Equation (1) may be used to

compute the appropriate flo for a given ftx:

flo = ftx

± 1.064 * (ftx / 390)

(1)

where ftx and flo are in MHz. Note that two values of flo

exist for any given ftx, distinguished as “high-side mixing”

and “low-side mixing”, and there is generally no preference

of one over the other.

After choosing one of the two acceptable values of flo, use

equation (2) to compute the REFOSC frequency ft:

ft = flo / 64.5.

(2)

Here ft is in MHz. Connect a crystal of frequency ft to the

REFOSC pin of the MICRF002. 4 decimal-place accuracy

on the frequency is generally adequate. The following table

identifies ft for some common Transmit frequencies when

the MICRF002 is operated in FIXED mode.

Transmit Freq. ftx (MHz)

REFOSC Freq. ft (MHz)

315

418

433.92

4.8970

6.4983

6.7458

2.

Selecting REFOSC Frequency ft (SWP Mode)

Selection of REFOSC frequency ft in SWP mode is much

simpler than in FIXED mode, due to the LO sweeping

process. Further, accuracy requirements of the frequency

reference component are significantly relaxed.

In SWP mode, ft is given by equation (3):

ft = ftx / 64.25.

(3)

Connect a ceramic resonator of frequency ft to the REFOSC

pin of the MICRF002. 2-decimal place accuracy is generally

adequate. (A crystal may also be used if desired, but may

be necessary to reduce the Rx frequency ambiguity if the Tx

frequency ambiguity is excessive.

See Application Note

TBD for further details.)

3. Selecting Capacitor CTH

First step in the process is selection of a Data Slicing Level

timeconstant.

This selection is strongly dependent on

system issues, like system decode response time and data

code structure (e.g., existence of data preamble, etc.). This

issue is too broad to discuss here, and the interested reader

should consult the Application Note 22.

Source impedance of the CTH pin is given by equation (4),

where ft is in MHz:

Rsc = 118k

Ω * (4.90 / ft).

(4)

Assuming that a Slicing Level Timeconstant TC has been

established,

capacitor

CTH

may

be

computed

using

equation (5):

CTH = TC / Rsc.

(5)

4. Selecting CAGC Capacitor in Continuous Mode

Selection of CAGC is dictated by minimizing the ripple on

the AGC control voltage, by using a sufficiently large

capacitor. It is Micrel’s experience that CAGC should be in

the vicinity of 0.47

µF to 4.7µF. Large capacitor values

should be carefully considered, as this determines the time

required for the AGC control voltage to settle from a

completely discharged condition. AGC settling time from a

completely discharged (0-volt) state is given approximately

by equation (6):

∆T = (1.333 * CAGC) – 0.44

(6)

where CAGC is in microfarads, and

∆T is in seconds.

5. Selecting CAGC Capacitor in Duty-Cycle Mode

Generally, droop of the AGC control voltage during

shutdown should be replenished as quickly as possible after

the IC is “turned-on”.

Recall from the section “AGC

Function and the CAGC Capacitor” that for about 10msec

after the IC is turned-on, the AGC push-pull currents are

increased to 45X their normal values.

So consideration

should be given to selecting a value for CAGC and a

shutdown

time

period

such

that

the

droop

can

be

replenished within this 10msec period.

Polarity of the droop is unknown, meaning the AGC voltage

could droop up or down.

Worst-case from a recovery

standpoint is downward droop, since the AGC pullup current

is 1/10

droop is replenished according to the well-known equation

(7):

I / CAGC =

∆V / ∆T

(7)

where I = AGC Pullup current for initial 10msec (67.5

µA),

CAGC is the AGC capacitor value,

∆T = Droop recovery time

(<10msec), and

∆V is the droop voltage.

For example, if user desires

∆T = 10msec, and chooses a

4.7

µF CAGC, then the allowable droop is about 144mV.

Using the same equation with 200nA worst case pin leakage

and assuming 1uA of capacitor leakage in the same

direction, the maximum allowable

∆T (Shutdown time) is

about 0.56 seconds, for droop recovery in 10msec.