DC step down circuit

[South Easter]

I run 24V, but sometime would like to run 12V appliances, so I designed this circuit.  I am a bit of a newbie to electronics, so comments and suggestions would be most welcome.

[scottsAI]

South Easter,

Concept works except:

Q2 is upside down. Pin 2 & 3 should be swapped.

Additionally the circuit will be unstable. To much gain after the opamp. Due to the Mosfets.

By adding a cap from output of opamp to neg input should fix it.

If you look at the data sheet for LM7812, shows ways to boost the output current.

Might want to check it out first.

Have fun,

Scott.

[South Easter]

Well spotted Scott. I spent a while scratching my head before I figured that out! (I wonder why the design progs give you P fets 'upside down'?)

Can you tell me what sort of efficiency this would give (when drawing say 0.5Amp at 12V).

I added a 1uf cap as you suggested, and that made the voltage swing (under Spice simulation) from about 10 to 15V.  Reducing it to 0.1uf helped, but still made a more spiky output volatge.  What value do you suggest?

For interest, here is the PCB layout I used - red lines are connecting wires on the top of the board, green lines are the existing strips under the board.

[tecker]

You have zero short circuit protection in this circuit and a short will probably take out the opamp too . The Fairchild data sheet gives a switching circuit and a reverse lead protect circuit both of which are well suited for for both a lm 78 series and lm 317. And of coarse there's the 12/24 battery connections that give you what you need if you don't need peak regulation or switchin power . here's

 a link   http://www.otherpower.com/images/scimages/1431/lm317_2.pdf

http://www.otherpower.com/images/scimages/1431/lm7812.pdf

   

[scottsAI]

South Easter,

My first look said it should work, it was simple enough so simulated it.

It did not work?? I also scratched my head, then realized yes FET up side down.

Thought you might like to know!!

Efficiency is 50%, this is a linear series controller, dissipates the power you do not use.

If that is your only load, the LM7812 can supply that directly.

Remember to use heat sinks on all the power device! 6W will fry a T0-220 without.

Switching supply can do this much more efficiently, complexity is not much more.

The stabilizing cap should be something small, like 1nf, or less. This is for frequency stability control, want to slow down the system due to extra gain. I did not run a frequency plot. I leave that for you.

Have fun,

Scott.

[South Easter]

Thanks for comments.

Must admit I dont understand about the linear bit and dissapating unused power. Its seems like the PFet is switching full on, then full off, at high frequency, giving the average stable 12V output.  But like I said, I am an electronics newbie!

The 7812 would waste too much power - I want to run CFLs etc off 24V, so may draw 0.5 - 2A.

[Flux]

I haven't checked the details, the basic circuit would work but it is not rally practical. Switching converters should incorporate an inductor and free wheel diode to act as an energy converter.

Your circuit may be ok to feed a motor or an inductive load such as a relay if you add a diode but just chopping with a mosfet is only good enough for resistive loads ( heaters, lamps and charge controllers).

If this is for low power there are plenty of buck converter circuits available. If it is high power you are into rather a specialist field and you will easily find the basics but you may not find enough details to stop the smoke.

Flux

[Kevin L]

If you are looking for efficiency I would look at the LM2975, as it is hard to beat efficiency wise.  Page 5 of the data sheet provides a schematic to apply a variable voltage regulator.  The data sheet linked above is a great source for this unit and it should do what your looking for.

Kevin L

[claude]

KevinL is right although his link (and IC name) is wrong.

LM2576 can handle up to 1.5A and 40V IN but they also have LM2576HVT-ADJ that can handle up to 3A and 60V IN. A great IC, it only needs a few external components and you have yourself a switching power supply. Here's mine, the pot adjusts the voltage from 0 to (inputV-2V):

Efficiency is around 75%. Check the IC datasheet for this.

[Flux]

The more I look at this the worse it seems.

I assumed you meant it to switch. It does seem to be based on a linear regulator with far to high a gain and no hope of stabilising without reducing the op amp gain and adding a few capacitors. It then becomes nothing more than a 50% efficient linear regulator that you could produce in many better ways.

As it stands I think it will switch but not in a defined way. The ripple current into the output capacitor will be horrible and the gate drive to the mosfets will be totally unsatisfactory at any finite switching frequency. Even switching I doubt that it will do much better than 50% if the power mosfet survives a finite time.

If you want to do it linear and crude then a 2N3055 used as an emitter follower from the 12v regulator would be far simpler, but for any sensible application then a decent switcher is the way to go with inductor etc.

Flux

[Goahead]

Hi everyone

I beleive that LM5115 can be also used for a good switching regulator.

http://www.national.com/pf//LM/LM5115.html (Input till 75V, output till 20A 13.5V)

Inductor to be used in the output can be one found inside a PC power suply. (old or new power supllies)

Since it can provide 13.5Volt then it can be connected directly from windnill to baterry charger.(i beleive that windmill will not produce more then 75Volts..thats a gess..still building mine)

Can someone with electronic experience post here a schematic for such solution using this LM5115??

I only need the DC-DC converter since we can find chargers shunt controlers over the internet.

Ill design the PCB with cad software and give you all the gerber files if need.

Goahead

[miata2k]

I think it is great to design your own controls and circuits.  I'd love to do the same, but sometimes I just don't have enough hours in the day, so I run to the nearest electronics surplus store and see what they have.

You may want to look for high efficiency DC/DC converters on line and see what you could find that could be more efficient than this circuit.  You can find them on ebay or places like BG Micro or Electronics goldmine, or MPJA.  They will usually work with a minimum of external parts.

I also found a 24 to 12 volt converter at BG Micro for $18.08 and it will handle up to 5 amps, and they have a larger one that will do up to 20 amps.  

Here is the link to the manufactures product page:

http://www.vellemanusa.com/us/enu/product/view/?id=522479#

You will have to search around BG micro to find the device.(PSDC05, PSDC20)

If you find a high efficiency DC-DC converter, you may also want to find a 24vdc to 5 volts device.  Just connect the 5 volts to a Female usb cable and you can charge your USB devices.  I bought a set of cables to charge things like my PDA, Cell Phone, Bluetooth headset, etc.  No need to go up to 120vac then back to 5 volts to charge a cell phone.

-Chuck

[South Easter]

Good idea GoAhead about raiding a PSU for an inductor.  Any idea what current they can take?

I look forward to someone posting the schematic you ask for!

--

[electronrancher]

My suggestion is to start with a non-synchronous buck like the national parts mentioned by Kevin and Claude.  That LM5115 is overkill, and it will be a pain for the newcomer to design and compensate.

So go get a LM2576 or their newer LM2676 (for about 5% better efficiency).  If you buy the 2676-12, you don't need a resistor divider to set VOUT, you just tie VOUT to FB.  Very simple.  Probably, you should get the To-220 package because it will be easier to work with than LLP.

http://www.national.com/pf/LM/LM2676.html

If you download the datasheet, you can find all the component values you would need.  I suggest 47uH inductor, 3.5A-4.5A Isat rating.  And 2SK34 schottky diode for D1.  But they also have web-bench so you can try different things and see if the circuit works.

For example, they suggest input cap = 3x 4.7uF and output cap = 2x 47uF when using the inductor and diode I suggested.  And with that - you're pretty much done.

[electronrancher]

PS - I forgot to mention, but I can send you an inductor and diode if you can't find them locally.  I make DC-DC converters for a living, so the component companies are happy to send me their products for free so I will mention them in my datasheet.  ;-)

[mungewell]

Do you really need to convert 24V to 12V?

If you are able to run pairs of DC CF then you might be able to wire 2 in series (ie. '+'--[CF1]--[CF2]--'-'), each one will get 1/2 of the available voltage.

Simon.