View Full Version : Questions regarding DIY boost converter

09-11-2011, 05:37 AM
I have always wanted to make my own boost converter(Just to know how to to), so i bought a few 220 uH inductors from mouser and plugged them into a breadboard setup... I have a 555 timer wired in astable mode. I have a 1.5k pot hooked up to adjust the output waveform(which forms a voltage divider with another 1.5k resistor and charges a 220 uF cap). The boost converter consists of a TIP3055, a diode and a couple of capacitors. When the 555 is outputing a lower frequency, i can measure up to 112 volts on the output. On the highest frequency, i can get about 13 volts or so. The transistor gets really hot, but the circuit worked well, until i discovered something. The output voltage was decreasing and when i applied a load it dropped right off and stopped working, until i took the load off. I read something online about the circuit needing some kind of feedback system to adjust the output frequency to keep the voltage constant? Or maybe i need a true PWM and not just a square wave that has a fixed duty cycle? I am also going to conclude that i need a bigger value capacitor on the output to supply voltage while the inductor is charging?...

Also, my understanding of a Henry(Unit of inductance) is a voltage of one volt being produced, if the current is changing at 1 amp per second? How does that relate to boost converter design? I just can't wrap my head around it.... Thanks for any info!!! I would submit a schematic, but i can't draw at all:) I hope i am not shooting in the dark with this one?

09-11-2011, 09:40 AM

09-11-2011, 07:57 PM
First off, welcome to the forum...I have already briefly read that tutorial, but i will look at it once again. The circuit that they show in the schematic looks a little more complicated than mine. It looks like this one has some sort of feedback mechanism though... I might just hook up a uC pin to the boost converter to simplify the oscillator part. That way i could directly control the PWM duty cycle and frequency in software. Thanks for the thought:)

09-11-2011, 09:29 PM
A 2n3055 is not suitable for a switching type power supply. WAY to slow. Use transistors and diodes designed for the purpose. A cheap power fet would be best for your experiments. A switching power transistor from a discarded pc power supply would work. Depending on where you put the diode, one of the schottky rectifier diodes might work. There might be a fast high voltage diode on the primary side that would work.

09-12-2011, 04:40 AM
@RickB... The TIP3055 says that it has a 2 MHZ switching speed(On the package, it said "high speed switching"). Is that not fast enough? Do you have any specific transistors in mind?

09-12-2011, 06:10 AM
The TIP3055 must surely be a modern day equivalent of the old 2N3055 NPN power transistor. I mean "old" as in I was using them back in the early 1970s. So I suspect "high speed" in that context is relatively slow by modern standards.
One major consideration in PSU design is efficiency. Especially in switchers where they can do rather well. As you said your TIP3055 was getting hot.
Why is that? Well clearly if it is fully OFF no current flows and no power is consumed. Also if it were fully ON max current can flow but the voltage across it is low, again little power is consumed. Problem comes when it is transitioning from ON to OFF and OFF to on. If that switching is slow it will spend a significant amount of time with both a large voltage across it and a large current flowing through it. That is to say consuming power and getting hot. Power W = IV.
Another problem is how well ON is it when it is on. In that state you want the voltage across it to be as small as possible such that the IV is low and power is not wasted.
So, that transistor has two requirements here:
1) A high switching speed.
2) A minimum ON voltage.

Many modern MOSFETs can do this. Look for the Ron parameter in the data sheets. As the flyback article linked to above says:

For the transistor I have used a BUZ41A. This transistor is rated at a maximum Vds=500V and an on resistance of 1.5ohm at 4.5A. Equivalent or better types like the IRF730 will also perform well. The diode should be a fast switching type like the BYW95C or better. An old (computer) power supply will yield you most of these components.

Whilst we are at it, it's good to not have much resistance in the inductor so as to minimize power loss. Again from the article:

The inductor I picked from a catalogue and is 100&microH with a few tenths of an ohm series resistance capable of handling several Amps of current.

Now, if you ever get this working nicely with the Prop replacing the 555 as the control element I would be very interested in the results.

David B
09-12-2011, 04:01 PM
It would be great to have a low-current 5 volt supply available from a prop. I use LCD displays a lot, and they only need about a milliamp at 5 volts.

I wonder if you could make a simple voltage detector with just a resistive divider feeding a prop pin? A 5 to 1 divider might be able to detect about 8 volts at the input based on detecting the prop input pin switching at half the 3.3 volt CMOS supply. Then the 8 volts could feed a conventional 5 volt regulator to power low current 5 volt parts.

Mike Green
09-12-2011, 04:14 PM
You could feed a Prop I/O pin to a simple voltage doubler or tripler (http://www.kronjaeger.com/hv/hv/src/mul/) using a couple of diodes and capacitors. You could either use a linear regulator or the Prop could do PWM based on feedback via your voltage divider and another I/O pin.

09-12-2011, 04:21 PM
@Heater... Thanks for the help. So, i need a faster transistor, specifically one that needs a small trigger(Base) voltage? And you are saying it gets hot in its transitional phase between On and Off? Could you develop a little more on that thought? I think i get what you mean, but i just want to be sure. Oh, and by minimum ON voltage, do you mean the transistor's minimum trigger voltage? My inductor can only handle a little over an Amp(1.2 i think), but i thought it would be ok, seeing as my power supply can only output 500mA's anyways. Also, i wonder if the output capacitor has anything to do with it? I should probably get a bigger one to supply power while the inductor is charging? I will definitely wire this up to a Propeller and do some experiments:)

09-12-2011, 04:22 PM
The TIP3055 datasheet suggests that to get fast turn-off you need to drive the base terminal several volts negative---I suspect its taking several us to turn off without this hard turn-off drive. This would limit its efficient range to a few kHz. A device with switching times of 200ns is more what you want for faster PWM. An oscilloscope would be invaluable to diagnose exactly what is going on...

BTW bipolar transistors are inherently slow to turn off from saturation due to charges stored in the junction. MOSFETs are universally used for fast switching these days as they don't have this 'minority carrier' issue.

09-12-2011, 04:24 PM
Interesting. I have an old 2 by 16 LCD that needs a minus 5v or so supply to get any contrast out of it. Looks like a cap and diode or two on a Prop pin could do it.

09-13-2011, 01:43 AM
Also, i wonder if the output capacitor has anything to do with it? I should probably get a bigger one to supply power while the inductor is charging? I will definitely wire this up to a Propeller and do some experiments:)

I think the output capacitor has a great deal to do with the circuit being able to store the charge until the next swing is approaching.

09-13-2011, 09:33 AM
I remember reading the datasheet for the 2N3055, way back in the late 80s, and it was stated a transition frequency of 150 kHz. Maybe the TIP version is faster... Depending on the current that you want to switch, a smaller transistor will be more suited (faster, higher beta...). The 3055 has a very low current gain (you will need D'arlington configuration). Get yourself a MOSFET instead. 50 kHz for such a large inductor is everything you need, more frequency is not adecuate (find why!) (you can of course simulate the circuit using (LT)Spice from the linear.com site).

09-13-2011, 11:41 AM

Well, imagine that the switching transistor is like a resistor connecting your
inductor to ground. That resistance is controlled by the current into the base
of a reguralr transistor or the voltage applied to the gate of a MOSFET.

Ideally when the thing is ON, the resistance from collector to emitter (or
drain to source for a MOSFET) is zero. Thinking in terms of Ohms law V = I * R.
So if R is zero so is V. Then the power dissipated is W = I * V. Which is also
zero as V is zero. All is good.

Alternatively when the thing is OFF ideally the resistance it presents is
infinite and we have from Ohms law I = V / R. So I is zero because dividing by
infinity gives zero. Then the power dissipated is W = I * V again which is
again zero. All is good.

So we see that if the transistor can switch instantly the power dissipated in
it is zero.

But what happens if it switches slowly? Then it spends a lot of time with some
intermediate voltage across it's terminals and significant current flowing.
Resulting in significant power dissipation and heat.

Of course no transistor is perfect. So for example in the ON state it's
resistance, collector to emitter or drain to source is not zero. So in the ON
state it will be dissipating power and generating heat. Best to find transistors
with low ON resistance specs, Ron.

Actually when I said "minimum ON voltage" I meant the voltage, drain to source,
when it is switched ON. Which as we see above we would like to have as low as

09-14-2011, 03:37 AM
Oh, wow... Sorry for not replying back. I have had a VERY busy couple of days(Whole week actually)... I have read every comment and i am plotting my next move:) Thanks to everybody.