Simple 12V SMPS for small powered devices

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I was thinking of a super-simple way to run low power stuff from 12V, without needing TL494, SG3525 or similar IC's. Just a simple oscillator made from common parts.

The idea was to take an Astable Multivibrator circuit similar to below but with different timing resistors and capacitors for 30khz or so.

Then take each NPN output and directly drive a MOSFET, or a Totem Pole+MOSFET from each side for push-pull. Even could be modified for Push-pull with bipolars.

Example Oscillator Type (Ignore the component values)
An externally hosted image should be here but it was not working when we last tested it.

I was thinking of taking a toroid inductor already wrapped with lots of turns I already have, add on a push-pull primary with a low turn count on it or so, and use the "secondary" to drive a fluorescent tube.

Another idea for audio use, was to make a simple split supply for preamp.

Any input on this?
you might also try royer push-pull type designs, though these need more care to the ferrite. such designs self-oscillate, and rely on transformer saturation to cause the switching. low power for this reason.

you must ensure you have a very symmetric oscialltor as the core can saturate if one switch is on consistantly longer then the other in a push pull. in which case you are relying on the source impedance to balence the volt-seconds.
Toroid Transformer

I just wound the Toroid. I used a 1.5 inch toroid that already had about 100-120 turns of green 24 gauge wire wrapped on it.

I just wrapped 7+7 turns of thick red 15 gauge wire for a primary side.

I'm wanting 170-200V of HF AC from this thing.

Any idea of how to limit current and power to 80-100W in case the lamp is too bright? I was going to use a power resistor, but I would like to limit it with the osc directly for less heat. 120V CFL uses small inductor like a ballast to limit current, but I want to avoid having to make another inductor if possible.

I want to use SMP60N06 MOSFETS with a small heatsink. I wonder if PWM IC is a better choice for CFL Lamp......
HOT Transformer

I hooked up the transformer, and 60A MOSFETS, 4700uf cap to filter 12V in, and used a TL494 driver board I already have to drive the MOSFETS.

When making it run, it was getting a little over 11.5V from a computer PSU I used to test with. The MOSFETS ran cool. It wasn't loading down the PSU with no load. Acts like it runs fine.

However, with NO load, the transformer is getting pretty warm, , not HOT, but pretty close. I'm wondering what I can do about it. Primary 7+7 turns of single solid #15 coil wire. Frequency 93khz.

I did rectify with a capacitor and measured 217V!!! after the diode, so I am getting HV out. I'm going to try with a CFL tube tomorrow, but I want to know if the transformer is supposed to get warm first.
Does it mean something is wrong if the transformer dissipates heat with no load on secondary? Nothing else gets warm but the toroid transformer. I'm still pretty new to SMPS design and theory.

*I'm just about to test the CFL with it, I'm just concerned about the transformer temp and/or saturation.
So far, so good, it runs the bulb, but the heaters stay on.....I didn't leave the bulb on long, so I didn't burn out the heaters. Also, I could adjust the brightness with the dead time adjust.

I also used a 25W 150 ohm resistor for current limit, and it got really hot :hot: So this thing has got some power.

I'll lower the frequency and see what happens.

I'm getting about 220V AC from this thing, the transistors stay cold.

The transformer is warm, load or not. Probably normal.
EWorkshop1708 said:
Does it mean something is wrong if the transformer dissipates heat with no load on secondary? Nothing else gets warm but the toroid transformer.

Nope, the volume power dissipation (in mw/cm^3) of the transformer is determined by material and flux density, not how much power you are drawing from the secondary. (that is limited by 'window area' all else being equal - but completely filling the window will smother the core and not allow it to dissipate heat well) To make the transformer dissipate less heat, add more primary turns to lower the flux density. This of course, deteriorates the ability of the core to shed heat, which is a tradeoff. Smaller cores can usually hand more flux density because there is not as much inner material to get real hot, they dissipate heat well.

A typical safe place to run a ferrite core is dB = 1500 gauss.

and for a push-pull converter, dB = [(V - 1)(0.8T/2)x10^8]/[Ae*N]

V = input dc voltage
T = switching period = 1/fs (seconds)
Ae = cross-sectional core area (cm^2)
N = primary turns

(taken from Pressman, Abraham. "Switching Power Supply Design," 2nd Edition, McGraw-Hill, 1998)

another thing to do if you have no current-mode control is to add small resistors to the drain of each driving mosfet. This will help prevent core runaway. If the core become saturated in one direction, the resistor will drop more voltage because more current is being drawn, therefore dropping the effective applied primary voltage, reducing the saturation.
Update, I will be making a few changes.

I got the 100W bulb to glow, but it glowed best at high frequency, but I had to play with cap/resistor values. However, ony the ends would glow, only one time nearly the whole bulb glowed. My Mosfets got really hot, and the transformer was even hotter.

Seemed to work, then one of my 75N06's shot a 3 inch flame :hot: out of it and burned the board! I was running this from a 14.4V NiCd battery, I didn't realize it was powerful enough to burn a 75A transistor :eek:

It's a bit hard to get it right, and drive a CFL directly. So instead I may just use this supply to get High Voltage DC instead.

After reading the UBA2021 Data sheet for CFL driver, I may instead just use this SMPS to get a high DC voltage, and use the HV DC to run the UBA2021 and MOSFETS. I have some parts from some broken 120V ballasts I can use. That way I can use the proper circuit to drive the CFL, and all I have to do is give it power. Less complicated that way. :)

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.