Auxiliary Power Supply for Welder

[GeorgesWelding]

I needed two 15V isolated auxiliary power supplies for the SMPS Welder I'm currently trying to design/build. This is the simplest design I could come up with. This power supply will fit on a 1 7/8" x 2 3/4" board. One of these will be powering a TL494, a few LM324 Op-Amps, and a couple UHS Opto-couplers(5mA each) and the second one of these supplies will be powering two IR2110 gate driver ICs each driving 2 IGBTs.

It's using a TL431 instead of a zener diode so that the output is slightly adjustable. The simulation of the opto-coupler and the TL431 works fine, but will this all work properly in the real power supply?

Here is the schematic: View attachment 15V Adjustable Supply (Schematic).pdf

 

[BradtheRad]

Do you need bipolar +15 -15V supplies?
You might get by with a single 30V coming from your converter, then split it by using a resistor divider.

 

[GeorgesWelding]

No no, I need two separate +15V supplies that are not only isolated from each other but isolated from the mains as well and they needed to be regulated. These supplies are super simple and I harvested the ICs, transformer cores, and the diodes from PC PSUs that I had laying around. I was just curious as to whether using the TL431 in the way I did would work in practice the same way it works in simulation. I assume it does, but sometimes Multisim does things that don't work in the real world. The datasheet for the TNY266 mentions using the TL431 instead of a zener but provides no example.

 

[BradtheRad]

The TL431 will work the way you have it hooked up. I just tried it with hardware and a variable supply.

I find it's a good idea to install a safety resistor inline with the led. In your schematic I would start by making R4 470 ohms.

You will have to adjust values, in order to find a range where led brightness changes sufficiently to affect the photosensor inside the optocoupler.

 

[GeorgesWelding]

You're right, in simulation the current draw through the diode was over the 80mA max rating in the datasheet when using the 100R resistor. I stepped it up to 470R and that dropped the diode voltage down to 1.227V@21.8mA which was still plenty to trigger the photosensor in Multisim. If it isn't enough in practice, I also have 390R & 330R resistors I can try which will give more brightness while still being within the operating limits of the LED. Thank you for catching that. That's why I post these things here before I go building them. Still learning.

-Brad George