Maker Pro
Maker Pro

Self-oscillating circuit for 5V power supply

Miguel Lopez

Jan 25, 2012
255
Joined
Jan 25, 2012
Messages
255
Hello everybody.

I'm trying to make the circuit shown bellow using the components of a CFL (HV transistors, ferrite core, HV diodes, HV capacitors, etc) in order to obtain a 5V source to supply a TTL-made timer which demands up to 500mA

I need some help with the cicuit configuration (for the self oscilating circuit) and how many turns (approx.) should I coil in the ferrite core. And how would be the marks in the transformer winding.

I want to rectify the 110 V AC line supply and using one of the capacitor of the CFL, obtain 150 V CD (non-regulated), and then the self-oscilating circuit with one independent (and isolated) secondary to obtain 5V using a single diode and a small capacitor. I've read that 40 kHz oscilation would be suitable for this purpose.

Thanks in advance.

s20113302203978.gif


An explanation on how this circuit works could be found here:

http://www.gzhambo.com/NewsView.asp?ID=109&SortID=22

Thanks in advance.

Miguel

PS: If there is a thread about this I apologise. Please redirect me to that thread.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
The major problem with this circuit is that the output is unregulated. That is an issue for TTL because it is very intolerant of supply voltages outside of about 4.75V to 5.25V.

At this relatively low power, you may be best scavenging a transformer and making a linear power supply.

If you can get stuff from China, there are a plethora of 110V USB "chargers" available for a couple of $US. This is one.

Perhaps you can salvage an old cellphone charger? Not sure what options you have.

If you're going to build a mains operated SMPS, you really need feedback (via an optocoupler) from the output to ensure it is regulated.
 

Miguel Lopez

Jan 25, 2012
255
Joined
Jan 25, 2012
Messages
255
Steve said:
If you can get stuff from China, there are a plethora of 110V USB "chargers" available for a couple of $US. This is one.
Sorry, I can not.

Steve said:
you may be best scavenging a transformer and making a linear power supply.
I did it once. The transformer was not suitable (too small) and it got hot a lot. Despite that, it worked well, more or less. But there isn't another transformer available at suitable size. Too big or too small. I'm trying to find a suitable solution, technically and economically.

Steve said:
Perhaps you can salvage an old cellphone charger? Not sure what options you have.
Not many options available. I don't even have a cell phone. I pretend to use the CFL because they are available at affordable prices.


If I can make it self-oscillate, then I will consider the option to add a regulator, such as 7805, in the output.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Sorry, I can not.

Maybe I should try to send you one?

I pretend to use the CFL because they are available at affordable prices.

If I can make it self-oscillate, then I will consider the option to add a regulator, such as 7805, in the output.

I'm not sure "pretend" is the word you were looking for...

In any case, the normal solution is to have an optocoupler sense the output voltage and control the inverter (essentially stopping it when the voltage is too high)

An alternative method that may work is to have another secondary and sample the voltage on that one. This could be used on the primary side without removing the isolation of your output on the main secondary winding.
 

Miguel Lopez

Jan 25, 2012
255
Joined
Jan 25, 2012
Messages
255
Steve said:
An alternative method that may work is to have another secondary and sample the voltage on that one. This could be used on the primary side without removing the isolation of your output on the main secondary winding.
Any diagram?
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Actually, I was mis-reading that diagram.

It looks like it actually does use 2 secondaries as I suggested above.

OK, you'll probably have to re-wind the transformer. The next issue is ensuring the transistors you have are specced for high enough voltage. In that circuit you could see something like 350V across the CE if your mains is 220VAC, something like 175 is more like it if you have 110VAC mains.

The problem with switchmode power supplies (especially discrete ones) is that you can generate lots of smoke before you fix all the bugs in your design.

It might be worth looking at the circuit used in the CFLs and see if rewinding a secondary and leaving the rest pretty much the same is an option.

This site may provide you with some information.

I'm not sure you could rewind those transformers for a lower voltage (it would require more primary turns and fewer secondary ones. I think you'd probably end up with the core saturated :(
 

Miguel Lopez

Jan 25, 2012
255
Joined
Jan 25, 2012
Messages
255
Thank you Steve.

I had already visited that site. In fact I have all of that diagrams in my PC. I also reversed engineered one of the CFL we have in Cuba. My first idea was to rewind the transformer in the CFL but they work with the load connected. I don't know how to replace the load (fluorescent tube)

I have the components of the CFL, MJE13003 high voltage transistors, 350V capacitors (we use 110V power supply, high voltage diodes), and a huge need to make this circuit work. I'm asking for help before get myself involve in some kind of "home fire with plenty of smoke" but I don't mind. It won't be the first time :D.

Steve said:
I'm not sure you could rewind those transformers for a lower voltage (it would require more primary turns and fewer secondary ones. I think you'd probably end up with the core saturated
Maybe with a couple of cores put together?

Well, I know cell phone charger use this principle. In fact that diagram is from a cell phone charger. My motto is: "There's nothing you can do that can't be done". I'll keep trying.
 
Last edited:
Top