LTC3265

Hello

We are using LTC3265 DHC Package for driving power rails of ADI's AD8139. 

- The AD8139 is configured for I-V differential converter with supply rails of (+/-5.6V Approx)

- LTC3265 is configured as follows (ref typical application circuit pg 16/20  of datasheet): 

VIN_P = 5;  RT= 0; 

R1 = 36.5k, R2 = 10k 

VLDO+/- = +/- 5.6 V (approx)

Additionally, we have provided 20pF cap across LDO and ADJ for both LDOs. 

The VLDO+/- measurements are as expected.  

However, the input supply current (80 mA) of LTC3265 is approximately twice the current demanded by AD8139 i.e 42 mA, as no input single is applied.  We verified this value by powering  AD8139 directly from Power supply (21 mA each supply rail) and then from LTC3265. Is this behavior true for LTC3265 ?.  Plz help us in this regard. Thanking in anticipation. 



moved and added tag
[edited by: elizabeth.uva at 2:13 PM (GMT -4) on 27 Apr 2021]
Parents
  • 0
    •  Analog Employees 
    on Apr 29, 2021 1:03 PM

    Hello,

    The boost charge pump is a doubler. It basically charges the charge pump up to VIN and stacks it on top of VIN to provide ~ 2VIN depending on the mode. The input current will be ~2x of the output current of the boost. In this circuit the boost load is the positive LDO plus the inverting charge pump which supplies the negative LDO for the total of 42mA as mentioned. Much of the output power in this case is dropped in the LDOs creating less efficiency from input to LDO outputs. 

  • Hello,

    Thanks for replying.

    Keeping in view the settings mentioned earlier, how can I optimized the efficiency of this circuit? Are there any  recommend settings to  improve LTC3265's efficiency? 

  • 0
    •  Analog Employees 
    on May 4, 2021 6:24 PM in reply to SAli810

    Hello,

    About the only thing that can be done to improve the efficiency is to reduce the switching frequency to maybe 200kHz. This will cause the LTC3265 to switch less and reduce the Iq current. This will also cause more VIN to VOUT loss and less power loss in the LDOs. This will also cause a higher VOUT ripple on the charge pumps. The over efficiency be slightly less and the temp rise in the IC will be less. Make sure there is plenty of LDO headroom margin to make sure sure the LDOs can regulate at ±5.6V no problem. 

  • Hello,

    Thank you very much for reply and suggestions for improving efficiency of LTC3265.

    I have still a little confusion about the Iq current as mentioned in the datasheet. On pg1 of datasheet it reads "During Burst Mode operation, the boost charge pump regulates its output (VOUT+) to 0.94 • 2 • VIN_P while the inverting charge pump regulates its output (VOUT–) to –0.94 • VIN_N. In Burst Mode operation the LTC3265 draws only 135µA of quiescent current with both LDOs on". It seems quite low and more demanding for power management on PCBs. 

    My question is that iif I configure the ltc for Burst Mode now, will I be able to reduce the input current as compared to the constant freq mode? kindly help me in this regard.

  • 0
    •  Analog Employees 
    on May 10, 2021 1:50 PM in reply to SAli810

    Hello,

    Where burst mode really helps out is at light loads. At light loads the charge pump will boost for small periods until VOUT reaches a threshold and then the oscillator goes to sleep until VOUT reaches a lower threshold saving Iq current and reduced switching which also saves current at the cost of a higher ripple on the charge pump outputs. As the current increases and the voltage loss from VIN to VOUT approaches 6% (VOUT = 94% of VIN) the charge pump will start to run continuously and the benefits of burst mode will be lost. 

    With 42mA on VOUT+ and 21mA on VOUT- the voltage loss between VIN to VOUT will be greater than 6% of VOUT and both charge pumps should be running continuously anyway. I do not believe this will be a benefit unless the loads are reduced at times. 

Reply
  • 0
    •  Analog Employees 
    on May 10, 2021 1:50 PM in reply to SAli810

    Hello,

    Where burst mode really helps out is at light loads. At light loads the charge pump will boost for small periods until VOUT reaches a threshold and then the oscillator goes to sleep until VOUT reaches a lower threshold saving Iq current and reduced switching which also saves current at the cost of a higher ripple on the charge pump outputs. As the current increases and the voltage loss from VIN to VOUT approaches 6% (VOUT = 94% of VIN) the charge pump will start to run continuously and the benefits of burst mode will be lost. 

    With 42mA on VOUT+ and 21mA on VOUT- the voltage loss between VIN to VOUT will be greater than 6% of VOUT and both charge pumps should be running continuously anyway. I do not believe this will be a benefit unless the loads are reduced at times. 

Children