DC Transformer Via PWM

[gravypod]

Hey everyone. This will be my first post on the EEVBlog. I'm a computer science student at NJIT and I've been trying to branch out into something a little different, a bit of EE.

I've got a question about the simplest way to create a large amount of different DC voltage rails. I'm working on a project that needs 3-4 rails. They will probably be in 5V, 3.3V, and a few others. I was thinking of building a transformer that will have more then one tap on it. I was all excited to get started until I realized that for a transformer to work I need a changing magnetic field!

I was wondering if I would be able to use some kind of PWM signal to create a variable magnetic field on the supply side of the transformer. This will probably be needed anyway due to the extremely variable supply power I'll be using. It should be over what I'll need in the transformer but it will also be changing variably.

I was thinking about using a buck converter for this sort of project, probably an lm25011. If the transformer idea doesn't pan out, then my fallback idea would be to just use more lm25011 chips.

Is there a better way to do what I'm doing?

Thanks for everyone's time!

[MosherIV]

Hi

Welcome to the forum.

I've got a question about the simplest way to create a large amount of different DC voltage rails.  I'm working on a project that needs 3-4 rails.

The transformer idea may work but is far too complicated/too much work.

Depending on what the application is, ie how much noise it can handle on the supply rails.
For very clean voltage, you need to start with a clean voltage rail and then you can drop the rails
with a number of voltage regulators.

If the application is not fussy about voltage rail noise then
Use a number of buck/boost converters such as the lm25001.
There are also some multi-rail buck converters but I do not have any part numbers at hand.

[kfitch42]

While a switching regulator chip like the lm25011 is interesting, I suspect you might be better off starting with something a little simpler to implement. The typical application for the 25011 ( http://www.ti.com/general/docs/datasheetdiagram.tsp?genericPartNumber=LM25011&diagramId=23302 ) has a diode, inductor, several caps, and a handful of resistors. On the other hand a linear regulator like the 7805 generally would require an input and output cap and thats it (Ok, maybe a heatsink too).

If you want the efficiency of a switching regulator with the simplicity of a linear regulator, I would look for an existing dc-dc converter brick. You have to ask yourself what the purpose of your project is. Are you trying to get some chip working (the one that requires all those rails) or are you interested in the minutia of generating a number of solid power rails? EE is a big field, don't try to do everything on your first go.

[Mechatrommer]

They will probably be in 5V, 3.3V, and a few others. I was thinking of building a transformer that will have more then one tap on it.

use one transformer and many regulators... AMS1117-5V, AMS1117-3.3V, AMS1117-ADJ and few others. building many taps transformer must have a good reason, i myself have not accomplish that.

[rstofer]

Unless we know what "a few others" voltages are, it's hard to come up with a regulator.
Further, unless we have some idea how much current will be required at each voltage, we absolutely can't come up with a recommendation.  Milliamps?  Amps?

For most chip type projects, like FPGAs, there are voltage regulators that produce all the usual suspects from a 5V input - 3.3V, 1.8V, 1.2V etc.
More specific to FPGAs, go to digilentinc.com, pick an FPGA board and then go look at the schematic to find out how they did it.
These boards can often be powered from USB so you can be certain that the currents are low.

http://www.analog.com/en/products/power-management/switching-power-converters/multi-output-regulators.html#multi-output-regulators
http://www.ti.com/lit/an/snva558/snva558.pdf

In the end, a bench type power supply, like the Rigol 832, is a pretty useful way to go.  It has one fixed 5V output and (2) 0-30V outputs.  This makes it ideal for +-15V op amp circuits with a little 5V logic (or step it down to 3.3 with a linear regulator).

[gravypod]

Unless we know what "a few others" voltages are, it's hard to come up with a regulator.
Further, unless we have some idea how much current will be required at each voltage, we absolutely can't come up with a recommendation.  Milliamps?  Amps?

For most chip type projects, like FPGAs, there are voltage regulators that produce all the usual suspects from a 5V input - 3.3V, 1.8V, 1.2V etc.
More specific to FPGAs, go to digilentinc.com, pick an FPGA board and then go look at the schematic to find out how they did it.
These boards can often be powered from USB so you can be certain that the currents are low.

http://www.analog.com/en/products/power-management/switching-power-converters/multi-output-regulators.html#multi-output-regulators
http://www.ti.com/lit/an/snva558/snva558.pdf

In the end, a bench type power supply, like the Rigol 832, is a pretty useful way to go.  It has one fixed 5V output and (2) 0-30V outputs.  This makes it ideal for +-15V op amp circuits with a little 5V logic (or step it down to 3.3 with a linear regulator).

I'm probably looking at the range of 500 mA to maybe 750 mA on each rail. Probably not too much current draw.

[gravypod]

Unless we know what "a few others" voltages are, it's hard to come up with a regulator.
Further, unless we have some idea how much current will be required at each voltage, we absolutely can't come up with a recommendation.  Milliamps?  Amps?

For most chip type projects, like FPGAs, there are voltage regulators that produce all the usual suspects from a 5V input - 3.3V, 1.8V, 1.2V etc.
More specific to FPGAs, go to digilentinc.com, pick an FPGA board and then go look at the schematic to find out how they did it.
These boards can often be powered from USB so you can be certain that the currents are low.

http://www.analog.com/en/products/power-management/switching-power-converters/multi-output-regulators.html#multi-output-regulators
http://www.ti.com/lit/an/snva558/snva558.pdf

In the end, a bench type power supply, like the Rigol 832, is a pretty useful way to go.  It has one fixed 5V output and (2) 0-30V outputs.  This makes it ideal for +-15V op amp circuits with a little 5V logic (or step it down to 3.3 with a linear regulator).

Thanks, the ADP5050 is perfect for what I'm doing! Didn't know this sort of thing existed!