STV0056 데이터 시트보기 (PDF) - STMicroelectronics

부품명

상세내역

제조사

STV0056

STV0042/STV0056 APPLICATION NOTE

STMicroelectronics 

STV0042/STV0056 APPLICATION NOTE

3 - APPLICATION NOTES (continued)

c - How to Use the Results

of the Auto-calibration Routine

Let us keep in mind that the final objective is to get

a PLL gain or a lock range as close as possible to

its optimum value.

Coming back to relation (1) :

ωL = KOTyp. ⋅ Ktable ⋅ KD = KOact ⋅ Kcorrected ⋅ K

KOTyp. : is the VCO slope when there is no dispersion.

Ktable : is the required coefficient (Reg 05 b0-5),

when KO has no dispersion.

KOact : the actual measured value

(with the auto-calibration).

Kcorrected : the coefficient that will be set after the

auto-calibration.

How to select the optimum setting in the FM

deviation table (Reg 05 B0-5)

In the FM deviation table of the STV42/56 circuits,

each step produces a 11% increase :

KN+1/KN = 1.11

Consequenctlythe maximum allowed dispersion α

of the slope KO before a correction is necessary, is

the dispersion which respects the following relation :

(1+ α) = (1 - α) ⋅ 1.11 ⇒ α = 5.26%

Repeting the same idea, it is possible to build a

table of the required correction versus the disper-

sion of the VCO slope. Proposed table :

In this table R is the ratio : KOact / KOTyp.

RMin. ≤ R ≤ RMax.

Required correction

------

0.692

-4

0.692

0.768

-3

0.768

0.853

-2

0.853

0.947

-1

0.947

1.0526

No correction

1.0526

1.168

+1

1.168

1.297

+2

1.297

1.44

+3

1.44

------

+4

+ 1 means : for a given deviation, + 1 is added to

the typical value which is loaded in Reg 05 (exam-

ple for ± 50kHz, typical value is 36h and the cor-

rected value is 37h).

Remark : In general the correction will not exceed ± 2.

In practice

In practice the microcontroller which are used in the

satellite receivers seldom have advancd arithmetic

processing functions (such as division). So, we

propose to compare the actual result (Fe - Fo) to its

expected typical value δth.

δth = ∆V ⋅ KOTyp.

∆V : 1V for Reg 05:23h and 2V for Reg 05:03h

KOTyp. : 460kHz/V

To select the corection to be implemented, a table

similar to the above described one can be used.

The selection range of this new table is :

δth ⋅ RMin. ≤ Fe - FO ≤ δth ⋅ RMax.

d - Additionnal Advices

about the Auto-calibration Function

d.1 - When to implement the auto-calibration

function ?

It is suggested to implement the auto-calibration

function when the end-user turns-off the satellite

receiver (remote control command or front panel

command).

This position of the auto-calibration function offers

two advantages :

- The receiver is auto-calibrated when the part is

already warmed-up. In this way potential drift

related to temperature are taken into account.

- The auto-calibration is done when the end-user

no longer wishes to use the receiver, conse-

quently the duration of the auto-calibration rou-

tine (about 1s) is no longer an inconvenience.

d.2 - To minimize the errors coming from

voltage offsets

There are two settings of Reg 05, 03 and 23 which

can be used for the auto-calibration, but in preac-

tice we suggest to use Reg 05 : 03 because it

produces a higher voltage increase in the VCO

drive. In such a case the offset voltages have

relatively less influence.

d.3 - Compensation of the drifts appearing on

CPUMP (see Figure 15)

To be efficient, the auto-calibration routine must be

as precise as possible.

For this reason we would suggest to compensate

some errors which may happen during the auto-

calibration routine.

More precisely, it has been noticed that small leak-

age currents produce a drift of the CPUMP voltage

which also drives the VCO. Consequently,because

of the required delay between the 3rd and 4th steps

of the auto-calibration routine, a small error ε may

be induced in the drive voltage of the VCO.

Suggestion :

To repeat twice the 4th step (frequency watchdog).

The first time about 200ms after the 3rd step. The

second time after a well known duration after the

first time.

Assuming that the drift is linear, it is possible to

correct the first reading.

Additionnaly to minimize the leakage currents

which would induce a drift of the CPUMPvoltage, it

is suggested not to connect the PLL circuit to the

incoming RF signal, but to keep the phase detector

input to bias (ex: if left channel is calibrated

Reg 06 B2 = L and Reg 06 B4 = L).

17/37

Share Link: