[SOLVED] Low Quiscent Current DC-DC Converter Suggestion

[crackerboy]

Low Quiscent Current, High Efficiency when Low Loaded DC-DC Converter Suggestion?

I am using a MCU and a GSM module in my project. MCU consumes 250mA (max), GSM module consumes 400mA(average) and while transmitting RF signals consume 2A (peak).
And also MCU consumes 1mA, GSM module consumes 1,5mA while IDLE mode.

I have searched for ultra low quiescent current DC-DC converters, find some possible ICs that is useful for me.

I want to use a 800mA DC-DC converter IC for GSM module. But i want to support it via capacitor to achieve 2A current supply.

Question: Can i use a High Value capacitor to sink 2A instantaneously from 800mA DC-DC converter ? Which capacity is suitable?




Actually i'm using LM2576 that has 10-14mA Quiescent current, it is not suitable for my project.

Basicly i need:
1- Efficiency>60% while 1-5mA Current Consumption,:!:
2- Low Quiescent DC-DC converter IC to supply GSM module (Peak 2A),:roll:
3- 4.2 and 3.3V outputs for GSM Module and MCU.:sad:


Here is my solutions but not yet decided to use:

1-**broken link removed** -> 1.5A 13.2V, 0.15mA Quiscent, -> Very Low Quiscent Current but relatively complex circuit.

2-**broken link removed** -> 2.5A,600mA 13.2V, 0.6mA+0.07mA Quiscent, -> Very Low Quiscent Current but relatively complex circuit.

2- **broken link removed** MAX5087B -> 45V 400mA Qis-0.07-5.3mA -> 400mA is not enough for peak 2A current requirement
3- TPS5420 -> 2A(3Apeak ), 3-4.4mA Quiscent ->

 

[dick_freebird]

You don't say the duration of the 2A peak, or how much
voltage sag you will tolerate.

I think you are unlikely to hit your no-load efficiency goal
and your max current goal simultaneously. Big FETs (2A)
need big drive current pulses. Maybe something with pulse
skipping, but you have to tolerate more and more variable
ripple amplitude and frequency variation. Radios usually
don't like that kind of thing.

 

[crackerboy]

I found LT1529 is the best choice for me! Now i will order a sample and test it.

 

[FvM]

You don't say the duration of the 2A peak, or how much
voltage sag you will tolerate.

According to post #1 it's a GSM Tx current wavefrom with 4.6 ms period and 1/8 (or probably 1/4) duty cycle. You can easily calculate, that at least several 1000 uF are needed to keep the voltage drop within reasonable limits.

 

[crackerboy]

Please look at the picture,

1-)For a GSM Tx current wavefrom with 4.6 ms period, is LT1529 DC-DC converter's transient response enough?

It responses about within 30µs~150µs to Δ2A load difference with a Cin=3.3µF and Cout=47µF capacitors!

Any idea?

2-) And finally i have to ask: I decided to use will use Tantal or Ceramic capacitors with low ESR. I know ceramic caps lowest ESR values under 100uF. DO they draw a current like a resistor while working DC-DC converter because of the ESR?
I want to know which one is good choice for low power design, i decided to choose LT1529 for low Quiescent current and Low Ground current, high efficieny while low current loads.

How many milliWatts a 47uF capacitor consumes while supplying a 1mA or 2A to a loan on a DC-DC converter's output? For example a Tantalum or a Ceramic cap?

 

[FvM]

As a first remark, supplying a 3.6 or 4V GSM module from a 12V battery with a linear regulator sounds like a waste of energy. Buts the most simple way, of course.

I would probably use higher value ceramic output capacitors than 22 or 47 uF to reduce transient voltage steps. As said in the datasheet, LT1529 is stable with higher values and low ESR capacitors which isn't the case for many linear regulators.

I don't understand, why you worry about mW possibly dissipated in ESR when loosing W in the linear regulator? ESR losses with ceramic caps are really low and will be a problem, if at all, only with frequency high ampere AC currents.

Ceramic caps are superior for bypassing in my opinion, although you should think about piezoelectric behaviour, that can give an audible sound with GSM pulse current.