Two questions about dc-dc converter.

[sami_r]

I'm not so familiar with dc-dc converters. Here are my two questions, maybe stupid ;)

1. How to get a steady 3.3v from 3 alkaline batteries whose output goes from 4.5v to 3v? Are buck-boost converters designed do this job?

2. How to estimate the input current of a dc-dc converter given output voltage and current? I know it must be different from an LDO. Is it reasonable by estimate the total power consumption, such like: P=Vo*Io/efficiency, then Ii=P/Vi.

 

[berkeman]

I'm not so familiar with dc-dc converters. Here are my two questions, maybe stupid ;)

1. How to get a steady 3.3v from 3 alkaline batteries whose output goes from 4.5v to 3v? Are buck-boost converters designed do this job?

2. How to estimate the input current of a dc-dc converter given output voltage and current? I know it must be different from an LDO. Is it reasonable by estimate the total power consumption, such like: P=Vo*Io/efficiency, then Ii=P/Vi.

Welcome to the PF -- both are very good questions.

-1- Yes, use a buck-boost or "Sepic" converter. You can find info on them at websites like National Instruments, TI, ON Semiconductor, Maxim, etc.

-2- You are correct. A DC-DC is basically a constant-power converter, and your equations are correct. The only modification would be to remember that the efficiency generally varies with the power level (see the DC-DC converter IC datasheets for typical curves of efficiency versus Iout or in terms of some other variable).

 

[sami_r]

Yes I read some datasheets, but didn't notice such varing efficiency curves. Now it's much clearer to me. Thank you!

 

[berkeman]

Glad to help. Here are a few more thoughts in case they help you or others:

** Boost/Buck topologies are very common in cell phones or other battery-powered electronics where the battery voltage can vary near the operating voltage of the electronics. The good news about this common application is that the ICs and components to do this function are very cost optimized (cheap for us!)

** Boost/Buck topologies are also used for Power Factor Correction -- I'll leave this as an exercise for the interested reader

** The efficiency of a DC-DC converter consists of pretty much two components, which have varying importance depending on what you are trying to do. The first is the efficiency of the switching components themselves, which would push you to minimize the voltage drop across the switching elements like the transistors and diodes (and lead you toward synchronous rectification). The second has to do with the quiescent power consumption of the control circuit that drives the switching elements.

When you deal with true micropower DC-DC battery-powered circuits, you think about all of these things...

 

[sami_r]

Well, I have more questions now =)

. Do we only consider the two components of the efficiency only when we are designing such converters? Most of the time, we just pick dc-dc converters sealed in chips because of the cost/size/... Do we still have work to do with them in such applications?

. In some micropower battery-powered devices, the variance of battery voltage which is fed into chips directly could lead to circuit mulfunction, especially when interfacing different chips (MCU, RAM, or sth else). I just want to know if the devices are so cost-constraint that no DC-DC converters are allowed, is it possible to make the circuit robust (to some degree) against such power variance?

. Could you recommend some articles on DC-DC converters? For Power Factor Correction, I just found one by On-semi. Hope it helps.

 

[yungman]

Well, I have more questions now =)

. Do we only consider the two components of the efficiency only when we are designing such converters? Most of the time, we just pick dc-dc converters sealed in chips because of the cost/size/... Do we still have work to do with them in such applications?

. In some micropower battery-powered devices, the variance of battery voltage which is fed into chips directly could lead to circuit mulfunction, especially when interfacing different chips (MCU, RAM, or sth else). I just want to know if the devices are so cost-constraint that no DC-DC converters are allowed, is it possible to make the circuit robust (to some degree) against such power variance?

. Could you recommend some articles on DC-DC converters? For Power Factor Correction, I just found one by On-semi. Hope it helps.

There are a lot of DC to DC converters manufacturers around...Power One, Pico etc. Too many to remember. On Semi are IC manufacturer, you still have to put external components around it...At least in my days which is few years back! I would tend to buy complete DC to DC converters rather to get an IC and add inductor and close loop feedback.

I see quite a few cases IC type circuit occilate...Occationally! That is the deadest case because you might get into a wild goose chase. I was involved in a noise problem of a gun sight camera. People had spent months chasing the problem. The project was dump on my lap! It turn out two of the Linear Tech converter IC occilates! They are closed loop feedback circuit. Cap at output is usually the guilty party that cause occilation because caps introduce a pole and putting bigger cap will not fix the problem, just kick it to lower freq. You really have to go full stability analysis to fix this. It is not as simple as matching poles and zeros...You don't know where are they! You have PWM, big caps, inductors etc. in the circuit. Application circuits are not all reliable! So let the manufacturer take care of it.

I have not have problem with input voltage variation as long as you guarante the lowest input voltage is higher than the converter's spec at max load. They start to act funny under full load when the input voltage is close to the edge of the spec. Particular if the input voltage can dip below the min. spec when a surge happen. Once this happen, the converter can go into a surging mode that continue on.

One other thing you have to be very careful on choosing a converter. Get one that is spec for minimum current draw in your circuit. ALL DC to DC converters tend to be unstable when you draw below the min. current spec. We have more trouble those days of converter occilating because of that. So don't just think the bigger the better. Our curcuits draws more current in certain situation and idle for long time. We had so much problem with this.

We resort to buy the converter and actually test it out at lowest current we use. Buy converter that have multiply manufacturers. A lot of converters have same footprints so if one don't work out, you can easily replace with another manufacturer. Some do work better, not occilating as easy.

For high reliability circuits, I tend not to get multiple output converters. Because something wrong on part of your circuit and draw more power will affect the other voltage. It is also harder to find second source if you have multiple outputs.