Post Go back to editing

(ADV7390)EXT_LF resister

Hi !

At the datasheet , EXT_LF's resister is 170ohms.

Is there any specification about 170Ohm?

1)

Can you tell me which type of 170Ohms should we use?

170 ± 10%(153~187Ohms) is OK?

2)

Is 10% is OK , can we use 180Ohms ± 1%(178.2~181.8Ohms)?

3)

I read the discussion.

=> https://ez.analog.com/message/156258#156258

But I couldn't understand what is "PLL stability"?

Would you explain it more concretely?

Best regards.

Kawa

  • Hi GuenterL.

    Thank you for your answer.

    I think maybe some of our customer are using 170Ohms so we have to check about this.

    So, I understood that 169Ohms 1% is the spec.

    It well help us that if you can modified the datasheet and other documents.

    I understood what will happen .

    But what will happen if they already not using 169Ohms 1% ?

    I mean we think one of our customer are not using 160Ohms 1%

    ( I mean maybe 169Ohms 10% or 160Ohms + 10 Ohms)

    and already mas-production.

    So it is very difficult to change the resister which is already mas-production.

    Can you say if there is no problem with there system not using 169Ohms 1%,

    they can use that?

    I think it is difficult to say....

    I will say to our customer to change the resister next project.

    Best regards.

    Kawa

  • Hi,

    We have asked an expert on this part to take a look and reply.

    Best Regards,

    Jeyasudha.M

  • 1) Use 169 Ohm 1%.  There are no 170 Ohm 5% or 10% resistors.  In my opinion the reference schematic should be modified to 169 Ohm 1%.  I've used 169 without issues on other designs

    PLL stability is basically how much jitter the PLL output has on it when referenced to a jittery source.  Normal goals of PLLs are to output a frequency locked to the source frequency and/or reduced jitter compared to the source jitter.

  • Hi GuenterL

    Thank you for your reply.

    I want to confirm it for the last time.

    1)

    So did the designer simulated with 170Ohms 5%?

    I mean which value will the device Permit?

    2)

    So,

    169Ohms 1%   ; 167.31~170.69Ohms => OK

    169Ohms 5%   ; 160.55~177.45Ohms => NG

    169Ohms 10%  ; 152.1 ~185.9Ohms   => NG

    170Ohms 1%   ; 168.3 ~171.7Ohms    => OK

    170Ohms 5%   ; 161.5 ~178.5Ohms    => OK

    170Ohms 10%  ; 153  ~187Ohms        => NG

    180Ohms 5%   ;  171  ~189Ohms        => NG

    * 170Ohms = 160+10 Ohms E24

    Is this correct?

    3)

    If they already use 170Ohms 5% or 10%  and they are already mas-production ,

    and there is no market failure can we say that

    they can use that value under their responsibility?

    Also we will say to our customer to change the value at next model.

    Best regards.
    Kawa

  • The original chip designers simulated the PLL with a 170 Ohm resistor but they did not know that 170 Ohm is not a standard resistor (an oversight).  The PLL has more then enough tolerance to handle 169 1% resistor.  Unfortunately this 170 number got pulled into the data sheet. 

    The simulations were done with 170 but it should be OK with 169 1%.  If the opportunity ever arises I would have this one customer switch to 169 if possible.

    We'll put the change to 169 on the data sheet in the queue but it is unlikely to happen soon.

    BTW 169 1% resistor falls within the 170 Ohm 5% range.

  • 1) I don't know exactly what tolerance the designers used for their simulations.  But what you need to keep in mind is the capacitor tolerances effects on the filter will be greater then just the resistor tolerance.  My guess he may have user 170 Ohm 5% for simulations.

    2) All the 1% and 5% combinations would be OK except for the 180 Ohm.  I would stay away from the 10% combinations

    3) If they are using 5 or 10% 170 Ohm resistor combinations for production and are not having any problems then all is good.  The capacitor tolerances are going to swamp out resistor tolerance effects.  Capacitor tolerances are already 5-10% at best.

    In my opinion the designer simulating the circuit simply forgot about what is available in the E96 and E24 resistor sets.  He obviously would not want the customer to use two E24 resistors in series to achieve the 170 value.  And over all the capacitor tolerances will swamp out the resistor tolerance effects.  The fact that he used 170 for simulation just rippled through

    In conclusion: if the circuit works currently don't change any thing.  In the future for re-spins or new designs I would use a 169 1% resistor.  In reality all resistors on the board can be 1% resistors, the cost differential between 5% and 1% are minimal. If you look at the reference schematic bom, the resistor in question is labeled 170 Ohms but they actually installed a 169 Ohm 1% resistor.

    Another issue to worry about is the temperature coefficients of the capacitors.  Depending on the system operating temperature range and the coefficient the cap value and move out of working range.  X7R caps have the worst temperature coefficient.  For wide temperature ranging system I would use NPO caps (yes, they are large but are much more stable).  Again if they have not seen any problems, don't worry about it.  Just keep this fact for future reference.

    Hope this helps

  • Hi GunterL

    Thank you for your kindly answer.

    That will help us very much!!

    1)

    I understood what you said.

    Yes we can say that the CAP has more bigger value error.

    2)

    169Ohms 1%  ; 167.31~170.69Ohms => OK

    169Ohms 5%  ; 160.55~177.45Ohms => OK

    170Ohms 1%  ; 168.3 ~171.7Ohms    => OK

    170Ohms 5%  ; 161.5 ~178.5Ohms    => OK

    * 170Ohms = 160+10 Ohms E24

    I got it!

     

    3)

    I got it.

     

    I have another question.

    I think this is the last question...

    ===========================================================================

    "He obviously would not want the customer to use two E24 resistors in series to achieve the 170 value."

    Did he say this because the PCB layout is critical?

    ===========================================================================

     

    As you said. I asked the value of the resister at here because I saw the ADI's EVAL-Board BOM list .

    I would say to all of our customer to see the EVAL-Board  and use the same value.

     

    Best regards.

    Kawa

  • Hi GuenterL

    Thank you for your reply.

    I understood.

    Our customer are using 160+10Ohms because as you said, there is no 170Ohms,

    So I think they will use 160+10Ohms at recently model.

    Best regards.

    Kawa

  • Using 2 resistors versus 1 is not that critical.  What is more important is Pvdd must be very clean.  And why use 2 resistors when 1 will do the job.

    Yes, always follows the schematic parts list.

  • Hi GunterL!

    Thank you for your support.

    > 3) If they are using 5 or 10% 170 Ohm resistor combinations for production and are not having any problems then all is good.

      > The capacitor tolerances are going to swamp out resistor tolerance effects.  Capacitor tolerances are already 5-10% at best.

    As you said 170Ohms 10% is OK , I think they can use 180Ohms 1%.

    Because 170Ohms 10% is 153~187Ohms

    And 180Ohms 1% is 178.2~181.8Ohms

    Is my understanding is correct?

    I want to ask this because our customer are using 170Ohms 10% at 2014 model.

    And they think they can use 180Ohms 1% at 2016 model.

    Also 169Ohms is not registered at our customer and also they want to reduce the number of parts .

    It is very difficult to explain that they should use 169Ohms 5% because there is no spec at the datasheet

    and also they are already using 170Ohms 10% ,

    so if we can say 180Ohms 1% will work good , it will help us very much.

    Best regards

    Kawa