This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

LMK04828: Locking problem with external VCO

Part Number: LMK04828

Hello,

We are using the LMK04828 with an external VCO and loop filter (only PLL2 is in use).

The external VCO is used in the frequency range of 1.2GHz-2.4GHz.

OSCin changes between 75-150MHz.

In some devices, we see that the LMK04828 can’t lock on frequencies near 2400MHz (the problem starts around 2380MHz up to 2400MHz which is the highest frequency used). OSCin is 148.75-150MHz (for VCO 2380-2400MHz). PLL2_DLD indicates PLL2 is locked.

For 2400Mhz, Vtune entering the VCO is at its highest value (rail of the LF opamp), and the actual frequency is 2470MHz.

The problem seems to be constant on the devices it appears on.

I’m attaching the schematics and the register values used.

Can you advise of a possible reason for this to occur?

Thank you,

Yulia

LMK_registers.txt
======================================================================
LMK4828 [A] Registers 
======================================================================

System-Functions Registers: 
--------------------------------------------------
R00 (RESET, SPI_3WIRE_DIS):  0x00
R02 (POWERDOWN):             0x00
R03 (ID_DEVICE_TYPE):        0x06
R04 (ID_PROD[15:8]):         0xD0
R05 (ID_PROD[7:0]):          0x5B
R06 (ID_MASKREV):            0x20
R0C (ID_VNDR[15:8]):         0x51
R0D (ID_VNDR[7:0]):          0x04

Clocks Pair [0-1] 
--------------------------------------------------
R100:  0x28
R101:  0x55
R103:  0x01
R104:  0x00
R105:  0x00
R106:  0x71
R107:  0x01

Clocks Pair [2-3]  
--------------------------------------------------
R108:  0x60
R109:  0x55
R10B:  0x01
R10C:  0x20
R10D:  0x00
R10E:  0x70
R10F:  0x55

Clocks Pair [4-5] 
--------------------------------------------------
R110:  0x08
R111:  0x55
R113:  0x00
R114:  0x00
R115:  0x00
R116:  0x79
R117:  0x01

Clocks Pair [6-7] 
--------------------------------------------------
R118:  0x08
R119:  0x55
R11B:  0x00
R11C:  0x00
R11D:  0x00
R11E:  0x79
R11F:  0x01

Clocks Pair [7-8] 
--------------------------------------------------
R120:  0x08
R121:  0x55
R123:  0x00
R124:  0x00
R125:  0x00
R126:  0x79
R127:  0x01

Clocks Pair [9-10] 
--------------------------------------------------
R128:  0x20
R129:  0x55
R12B:  0x01
R12C:  0x20
R12D:  0x00
R12E:  0x70
R12F:  0x55

Clocks Pair [11-12]  
--------------------------------------------------
R130:  0x24
R131:  0x55
R133:  0x01
R134:  0x00
R135:  0x00
R136:  0x71
R137:  0x05

SYSREF, SYNC, and Device Config.  
--------------------------------------------------
R138:  0x40
R139:  0x03
R13A:  0x03
R13B:  0xC0
R13C:  0x00
R13D:  0x08
R13E:  0x03
R13F:  0x00
R140:  0xA0
R141:  0x00
R142:  0x00
R143:  0x11
R144:  0xF3
R145:  0x7F
R146:  0x18
R147:  0x12
R148:  0x06
R149:  0x0B
R150:  0x42
R15B:  0xF4
R15F:  0x0B
R160:  0x00
R161:  0x01
R162:  0x48
R163:  0x00
R164:  0x00
R165:  0x10
R166:  0x04
R167:  0x00
R168:  0x08
R169:  0x59
R16A:  0x20
R16B:  0x00
R16E:  0x13
R172:  0x02
R17C:  0x15
R17D:  0x33
R182:  0x00
R183:  0x04

  • Yulia,

    A few things:

    1.Realize that if you lock the VCO, but then change the OSCin frequency, the VCO will only track it a little bit.  The VCO is divided into many smaller bands and this will not track the reference until you reprogram the VCO.

    2.   I think that 2.7 V sounds like the maximum voltage.  This part has internal LDOs and maybe this is the maximum core voltage for this device. 

    3.  If you are saying that 2.7V is the highest frequency of the op-amp, yet the VCO is only putting out 2470 MHz, does this not imply that the VCO might be limiting you.

    4.  When I have done active designs for loop filters, the op-amp input rails and output rails can often cause this sort of problems.  Is it possible that you are getting too close to the rails of this op-amp?

    5. For this configuration, if the op-amp is not unity gain stable, sometimes you can also have flakey results.

    In general, the op-amp in the active filter is somewhat of an art.  So I would not be surprised if this issue is op-amp related.

    Regards,
    Dean

  • Dear Dean,

    Thank you for your reply.

    Adding information to the above issue:

    1. When the circuit has fan-convection on it, locking is achieved (weaker convection doesn't help).
    When there is no fan, locking is not happening, the LF opamp output is in saturation at its supply voltage (VTUNE) and the VCO follows VTUNE.
    When bringing the fan back, without any LMK programming, locking is achieved within seconds.
    Increasing opamp VCC only increased saturation voltage and VCO frequency.
    LMK, VCO and LF  opamp case temperatures were measured (with no fan). Temperatures do not exceed 50deg.

    2. We measured LMK CP output and VCO VTUNE input:
    On the unlocking circuit, it seems the locking process starts and immediately stops. Thus VTUNE is as described above.
    See enclosed scope shots:



    3. Temperature is a trigger. We are not sure it is the cause.

    What is the root cause for this unlocking issue?

    Thank you

    Gil

  • Make sure it is not a soldering issue with heating PCB flatness changed. How many boards had been found with such issue?

  • Hi Shawn,

    Three out of five.

    We have few shields on the board and eight GND layers (out of 16). Heat convection through the GND is significant.

    Simple air flow (which lowered the GND temperature) resolved this in few seconds.

    We don't think it's a soldering issue.

    With the same conditions, frequencies lower than 2300MHz, do not raise this issue.

    Thanks,

    Gil

  • Then continue to study Op Amp output and VCO Vtune requirements, as Dean mentions.