TPS61023: Vout leakage

Hi Chris,

Thanks for choosing TI boost converter products.

Is the target output votlage 3.8V?

What's the 2A short peaks lasting time? If it is over 1mS, for TPS61023, it could be seen as a steady state load current, not a instant load current because the TPS61023 load transient respond time is usually less than 1ms. In such case, the needed input current would be 3.8V*2A/1.8/0.9=4.69A, higher than TPS61023 switch valley current limit spec. 

And what's the reason of adding the the device LM66100?

Hi Zack,

thanks for the quick reply.

Per Standard, peak current for GPRS is for 577us.

And, wouldn't the big caps buffer the transient load?

The LM66100 is for muxing. If GPIO disables the TPS61023, LM66100 will enable and deliver the current directly. I suggested this to avoid the Vout leakage current of TPS61023 (which was my other question). If in the described scenario where VIN>=VOUT the TPS61023 can simply be disabled with drawing less than 1uA from Vout, I don't need the LM66100. But the datasheet of TPS61023 is not clear on this

Hi Chris,

When TPS61023 is disabled, the leakage current into Vin and SW pin is 0.1uA typically. Leakage current into FB pin is 6nA and leakage current into Vout pin is 1uA typically. You can find these information in datasheet.

So you hope that when Vin < 3.8V, the TPS61023 is enabled to boost the voltage up to 3.8V, when Vin > 3.8V, TPS61023 is disabled and LM66100 will enable and Vin=Vout, right? Actually TPS61023 or TPS61022 has the feature of pass through mode when Vin > Vout. You don't need the LM66100.

The figure 19 load transient waveform in TPS61023 datasheet shows this device load transient respond time is less than 100us. So 577us is pretty long for TPS61023 and it could be seen as a steady state load current. The calculation result in last post shows it would trigger TPS61023 valley current limit spec. A big caps (hundreds of uF) is needed to buffer the transient load. Another solution is using TPS61022, which has higher current capability.

Thank you Zack.

Given the fact that the GPRS modem will be in sleep (< 9uA) most of the time, the Iq of TPS61022/61023 is relatively high. That‘s another reason for the idea to bypass it with a LM66100 for light loads