Suggestions for books/articles on building a DC to DC converter

[Corn-Picker]

My grandfather used to be a volunteer fireman, and while rummaging around his workshop I discovered a fire truck horn The problem is, my car is a 12V system, and the fire truck horn is a 24V horn. What I want to build is a simple 12VDC to 24VDC converter. This is for a horn that shouldn't take more than 500ma, and will be used very intermittently, so I'm not concerned about efficiency.

I can buy a converter for $39.99, but I'd rather build my own. Any suggestions on books or articels that could get me started would be a great help. Thanks a ton guys.

[Netlist]

Here is something. Hope you can work yourself trough the math’s.

/Hugo

[Corn-Picker]

Thanks for the link, it got me thinking. I've found a .pdf article that seems a little more well written, or easier to understand anyway. It's located at

http://www1.jaycar.com.au/images_uploaded/dcdcconv.pdf

At the bottom of this message is a diagram of what I need to build, taken from the above website.

I need 2x voltage multiplication, which implies a switching duty cycle of 50%. My question now is, how do I control the duty cycle, what type of high speed switch do I use? The diagram has a black box that simply says "switching control circuit." Is this something I can easily build or should I simply buy an IC that does the same thing? I would rather build it myself if it's feasible.

I'm imagining the frequency of my switch will determine the size of my inductor and capacitor. I can envision the system and do the math, but I don't know what any of the parts are called, as I am a chemical engineer by training. We need a universal language among engineers, if wires were called electron pipes or check valves were called water diodes we could understand each other easier

[Netlist]

I haven't got a clue what's in the black box.
Meanwhile, I moved this thread to power supply design where the attention might be higher.

/Hugo

[Bakmeel]

Well, Corn-Picker, you are on the right track for starters, 'so keep up the good work. Take a look at http://www.ipes.ethz.ch/ipes/dcdc/e_Boost.html for a nice visual way to understand Boost Converters.

To answer some of your questions:

- The high speed switch is generally a Mosfet, or IGBT for higher power levels. A single mosfet would suffice in your application.

- The "switching control circuit" you refer to basically does the following: It regulates the duty cycle as a function of the measured output voltage, compared to a reference voltage you preset in it (like 24V)

It is feasible to build it with single components, but it may be quite a hassle. Buying a single control IC would make things easyer and faster.

Do ask your questions and I'll try to get you along.

Bouke

[Corn-Picker]

Thanks guys. I re-read some sections of my my old Scherz "Practical Electronics for Inventors" book, and I think I'm closer to a solution. For the black box called "switching control circuit," I'm going to substitute a TLC555 timing IC. This IC will be setup to generate a 50% duty cycle square wave, which will control the MOSFET. The 555 timer IC has a 15 Volt limit, so I'll have to make sure it doesn't see the boosted voltage of 24VDC.

One question: Assume I make my resistors R1 and R2 such that I get a frequency around 500 kHz. Would it be a problem for a MOSFET to keep up with a frequency that high? I want to make my frequncy as high as possible, so that I can make my inductor and capacitor from the first circuit diagram I posted as small as possible.

I'll test the square wave generator on my ebay special oscilloscope, and if everything works well I'll post impressions on how loud the fire truck horn is

[N-Channel]

CP-

The 555 is a great chip, BUT- if you can get your hands on UC3842 or a UC3843 IC, that would be MUCH better. These chips, made by ONsemi and others were made exactly for this application. These are THE preferred chips for flyback and non-isolated boost topologies. They can be salvaged from almost any flyback-type AT or ATX power supply rated under 200W. They are current-mode chips, and are good for at least 500kHz. The difference between the two is the operating voltage range. The 3842' turn-on and turn-off threshholds are 16V and 10V, respectively, while the '3843's are 8.4V and 7.6V, respectively. Since your car's 12V bus will never see 16V, the '3843 is the logical choice for your application. Try looking at it at this link:

http://www.onsemi.com/pub/Collateral/UC3842B.PDF

One question though? Why 500kHz? I realize you said you want to keep the reactive components (coil & cap) as small as possible, but you will run into switching losses that might be on the high side. Also, if these two components are TOO small, they will not be able to handle the power levels being pushed through them. If you try, say, 250- or even 100 kHz, your components will still be pretty small. For 24V @ 500mA (let's say 1A for some headroom) and 75% efficiency, that's still only about 30W. A powdered-iron toroid (like #26 mix, colored yellow and white) with an outside diameter of about 0.800 in. should do. You can easily salvage one of these from the same AT or ATX flyback box you get the '3843 from.

I hope this helps. You've come to the right place- there is LOTS of help here, and as Bakmeel said, DO ask your questions here, there's plenty of help here.

Goos Luck and '73,

Steve, N8XO

[N-Channel]

Corn Picker,

I just thought of another way you can do this, and with much ease...

Try National Semiconductor's LM2577 SimpleSwitcher IC. This is a PWM Boost IC with the power switch integrated into the chip. It comes in a 5-pin TO-220, and is rated at 3A. So, at 12V input (3A being the max you can switch), you should be good for about 1.5A at 24V output. Go to this link for more:

http://cache.national.com/ds/LM/LM1577.pdf

In their most basic forms, the step-up SimpleSwitchers require only an input cap, an output cap, a coil, a hi-speed switching diode, and an RC compensation combo. For the adjustable version, add 2 resistors to the FB pin, and you're good to go.

Samples are available from Nat Semi, in either the LM2577T or the LM2577S surface-mount version. I have made many DC-DC "uppies", as I call them, for friends in my Amateur radio club, so they can use them to power their 19-22V laptops from 12V with no problem.

This, as well as all of the SimpleSwitcher chips, are very easy-to-use and forgiving chips to work with. The datasheet also has a design procedure that works well, and requires minimal tweaking of initial designs.

Hope this helps.

Steve

[mflorin]

I think that a cheaper solution is to use MC34063 from Motorola (ONsemi). All you need to do is to download the data sheet and the application notes. You will see that this chip is very easy to use in several converter topologies including the "boost" one.

If you don't have to much experience with current mode control chips like UC3842 and others, the MC34063 is the best solution.
Of course you can use N-Channel's solution, but I think MC34063 is cheaper than LM2577.

Regards,
Florin

[Kenshin]

To a new learner, simple is cheap. Use a expensive device to success is cheaper than exploding many cheap ones.

Note: the feedback is necessary to turn off the swiching pulse when output voltage reaches desired level. If feedback is disconnected, high voltage may appear at the output, it may cause either circuit damage or human accident.

Originally posted by mflorin

If you don't have to much experience with current mode control chips like UC3842 and others, the MC34063 is the best solution.
Of course you can use N-Channel's solution, but I think MC34063 is cheaper than LM2577.

Regards,
Florin