Boost converter help - burned wires.

Thread Starter

late_bloomer

Joined Feb 22, 2017
21
Hello everyone..

I wanted to build a boost converter, wanted to see how it works, what are the possibilities. I did not wanted to use an IC, so i build from parts that my friend had, used his equipment, and so on.
Edit: VDD is -9V.

U1 created frequency, around 80kHz, u2 provided PWM. Set to 80 KHZ, 80% duty cycle (80% low, 20 % high). This part worked great, verified on osciloscope. Massive heatsing on Q1 to prevent overheating. Q1 was IRF3415 not 3710 as in schematics.
The problem happened from battery, through inductor, through mosfet. All wires got burned after few seconds. Insulation melted, wires melted off. I was able to measure the voltage before disaster to be around 180V without load applied.

boost converter.jpg

What did i do wrong? No load was applied, so there should have been low current draw.
 
Last edited:

Thread Starter

late_bloomer

Joined Feb 22, 2017
21
Most of these parts were just laying around in component bin, nothing is properly tested.
The inductor in question was one of those parts that was removed from some circuit board. Looked like one of those with much more windings, labeled 8-20 uH. (I do not know who it was measured)
 

crutschow

Joined Mar 14, 2008
25,117
Most of these parts were just laying around in component bin, nothing is properly tested.
The inductor in question was one of those parts that was removed from some circuit board. Looked like one of those with much more windings, labeled 8-20 uH. (I do not know who it was measured)
Then you can't be surprised if the circuit doesn't work as expected.
If that inductor saturates, the current is limited only by the circuit wire and MOSFET on-resistance.
 

Thread Starter

late_bloomer

Joined Feb 22, 2017
21
Resistance of inductor is 0.2 ohms, so putting 12V through it was probably too much.
Crutschow, can you provide me a link or page, on where I can find calculations on what role the inductor, mosfet and frequency does to the circuit? I found few, but it involves calculus, and I am blank on it. Other sites did not give me enough information, just half of calculations that i needed.
 

ian field

Joined Oct 27, 2012
6,539
Most of these parts were just laying around in component bin, nothing is properly tested.
The inductor in question was one of those parts that was removed from some circuit board. Looked like one of those with much more windings, labeled 8-20 uH. (I do not know who it was measured)
Op-amps are about the lowest on my list for an SMPSU PWM generator.

The 555 is a viable choice and there's even a few workable designs using CMOS gates.

My first choice would be the MC34063 - there are online tables, and even an interactive calculator for working out the values.
 

Thread Starter

late_bloomer

Joined Feb 22, 2017
21
This project is not intended to be used for anything except a demonstration, a tool to learn. Everything was build on breadboard, and is already torn down. It was just an attempt to see if i can do it.. Apparently I did not foresee the problems that i experienced.

Thanks everyone for your help.
 

ian field

Joined Oct 27, 2012
6,539
This project is not intended to be used for anything except a demonstration, a tool to learn. Everything was build on breadboard, and is already torn down. It was just an attempt to see if i can do it.. Apparently I did not foresee the problems that i experienced.

Thanks everyone for your help.
It appears to have successfully demonstrated that you went about it the wrong way................
 

Thread Starter

late_bloomer

Joined Feb 22, 2017
21
Yes, i jumped into project without realizing what could go wrong. It worked great on simulation, so I assumed it would work in practice.

I have no problem admitting my mistakes, once I see them and are explained to me.
 

ian field

Joined Oct 27, 2012
6,539
Yes, i jumped into project without realizing what could go wrong. It worked great on simulation, so I assumed it would work in practice.

I have no problem admitting my mistakes, once I see them and are explained to me.
The breadboard can be part of the problem on SMPSU projects.

My method of prototyping is basically "dead bug" with a few refinements. A good start is a sturdy supply rail/frame with 16 or 18SWG tinned copper wire, the ends are closed by adding supply decoupling capacitors.

Most of the interconnects utilise the component leads, you may need to add a few more - 20 or 22SWG is good for that - sometimes when you're trying to wriggle a leaded component into a small gap, an SMD part will do just nicely.

Another method that will work uses an un etched board as a ground plane and small segments of copper clad are glued on for connection points - some places sell the segments ready cut, but you can make the blanks youself with a heavy duty hole punch.
 

Thread Starter

late_bloomer

Joined Feb 22, 2017
21
I didn't calculate that when mosfet is on, it would pull so much current through inductor and mosfet it self. Wires melted, inductor was superhot. Mosfet was fine since it had a massive heatsink.

Back to the drawing board.
 

MrAl

Joined Jun 17, 2014
7,665
Hello everyone..

I wanted to build a boost converter, wanted to see how it works, what are the possibilities. I did not wanted to use an IC, so i build from parts that my friend had, used his equipment, and so on.
Edit: VDD is -9V.

U1 created frequency, around 80kHz, u2 provided PWM. Set to 80 KHZ, 80% duty cycle (80% low, 20 % high). This part worked great, verified on osciloscope. Massive heatsing on Q1 to prevent overheating. Q1 was IRF3415 not 3710 as in schematics.
The problem happened from battery, through inductor, through mosfet. All wires got burned after few seconds. Insulation melted, wires melted off. I was able to measure the voltage before disaster to be around 180V without load applied.

View attachment 122826

What did i do wrong? No load was applied, so there should have been low current draw.
Hi,

First you should never run a boost converter without some load. The voltage goes up too high and blows something out.

Second you have to turn the Vcc up slow when running for the first time, so you can see if something goes wrong like the current goes too high. You can stop before anything burns out.

Third, 20uH is kinda low for 80kHz. 50uH would be much better at that frequency. The peak current could go over 6 amps so the inductor has to be able to handle that current. If it cant, the current can go up to 20 amps or more.

There are several things that can go wrong, but the first is you need some load.
 

ian field

Joined Oct 27, 2012
6,539
I didn't calculate that when mosfet is on, it would pull so much current through inductor and mosfet it self. Wires melted, inductor was superhot. Mosfet was fine since it had a massive heatsink.

Back to the drawing board.
The inductor could be saturating - see if forcing a lower on time takes it from too hot to cold.

In this situation - boost converters usually damage the PSU and buck converters damage the load.
 
Top