Attempt at a simple boost converter (3V--> 5V)

[Mr RB]

...
I need to give credit to Mr RB, who posted practically the same idea in this reply.

Thanks for giving me SOME credit!

And seriously now congratulations! You have done a lot of work and got a result to be proud of. It was only a couple of threads back you were working with Colin's converter and you've come a long way.

You still might want to experiment with a much bigger inductor and RC values etc changed to suit. If you can get the frequency down to 20kHz or so the efficiency should be quite nice.

 

[carbonzit]

Thanks for giving me SOME credit!

You deserve more than a little.

After a night's sleep and playing with the simulation a little more, I realized that not only do I have a very believable 3V--> 9V converter, but it really is a "poor man's Minty Boost" (3V--> 5V converter usable for USB devices), simply by changing the regulator voltage. I believe it can give that device a run for its money. By reducing the load resistor (to 20Ω), I was able to get almost 250mA out of it at just a hair under 5 volts.

It's time to order parts and build me a couple of these! Many thanks again to all who helped.

 

[carbonzit]

Try this:
Q2=2N4401
L1=200uH, Resr=0.1 ohms
C2=10uf
RL=33
R1=50
C1=300pf

and note the output current level.

I did, and unfortunately it was pretty near zero. In other words, the circuit simply did not work with these values.

First of all, the transistor. I tried many of the ones I have in LTspice's parts box; the only one I could get to work reliably was the 2N2222.

However, having said that, I did get this version of the circuit to work:

**broken link removed**

I should say a few things about this. Not only does it work, it works amazingly well!. In fact, what you have here, my friend, is a real Minty Boost contender. It actually works better than my converter here. Note load resistor value; it can deliver the full 1/2 amp into that load at as close to 5 volts as never mind, so it really qualifies as a very usable USB adapter powered by 2 AA cells. Not bad!

Of course, it needs regulation. How would you suggest regulating this? Can one use the same simple scheme I used (zener diode and transistor)? Where would you connect the regulator transistor?

I think I'll get parts for this as well and play with it in Real Life.

Simulation posted below (do you have LTspice?).

 

[MrAl]

Hi again,

Try editing your .trans dot comand to this:
".tran 0 .02 1e-7 UIC"

without the quotes.
These circuits depend on a more natural start up.

We'll work on regulation next if you like.
Try this first:
2nd 2N2222 transistor (Q3) collector to base of Q2, Q3 emitter to ground. 10k resistor on Q3 base to ground, 2k resistor from Q3 base to anode of 1N750 zener diode, cathode of zener to output.
Regulation should be around 5v or so with the following other circuit values:
C2=50uf,
RL=33 ohms,
R1=50
C1=300pf
L1=200uH, 0.1 ohms ESR

We probably have to go to a higher current Schottky eventually, maybe one of the 3 amp packages.

 

[carbonzit]

Well, dang, that's an improvement:

**broken link removed**

Now I'm getting about 6V @ 600mA into the same load resistor. Kewl!

LTspice question: it looks as if all that command does is 1) delay the time to start saving data and 2) skip the initial operating point solution. So how would this allow that other transistor to work? (I suspect it's mostly due to the UIC parameter?)

I haven't measured the current flowing through Q2, so I wonder if we're coming close to its 600mA limit. I'll check. Also, should there be another diode in the same position on Q1?

Uh, oh: I just checked it (in the sim), and Ic for Q2 is about 2 amps! not good. I think we need another part here ...

One note is that this higher-output version also produces more ripple. The 2nd filter cap reduces it, but there's still about 49mV of ripple (a little less than 1%). How bad is this? (Depends on the load, I know: I'm thinking of a typical USB load here.) I'll play around with filter caps some more.

And yes, I'd like to start working on regulation. (Hey, I just thought of something: could we incorporate a "poor man's capacitance multiplier" into the regulator to reduce the ripple? Just an idea ...)

 

[MrAl]

Hi again,


Gee i'm sorry, the .tran command is
.tran 0 0.02 0 1e-7 UIC

note the extra zero between the 0.02 and the 1e-7.

The diode you have there goes on Q1 not on Q2. Q1 base emitter reverse voltage goes too high without it when the output is higher than about 8v with a 3v battery.

Didnt i suggest a RL of 33 ohms to start with? Maybe i forgot that.

To start on regulation, use the 33 ohms resistor for RL first, then connect the transistor, resistors, zener, as i suggested in the previous post. You've be happy with this i think

The "poor mans cap multiplier" is a linear regulator of sorts, so it will smooth out the voltage but it will eat up the efficiency. Interesting idea though.

 

[carbonzit]

Gee i'm sorry, the .tran command is
.tran 0 0.02 0 1e-7 UIC

note the extra zero between the 0.02 and the 1e-7.

That's better. You still didn't explain just how this affects the simulation.

The diode you have there goes on Q1 not on Q2. Q1 base emitter reverse voltage goes too high without it when the output is higher than about 8v with a 3v battery.

Actually, it's needed in both places. I noted some really nasty negative spikes on Q2's base, which are gone now.

To start on regulation, use the 33 ohms resistor for RL first, then connect the transistor, resistors, zener, as i suggested in the previous post. You've be happy with this i think

Yes I am. Yes I am.

**broken link removed**

Couple things: 50µF wasn't enough to get rid of that excessive output ripple. 100µF still leaves about 28mV of ripple (about 0.3% at 9 volts output), which is tolerable.

I also had to add that 1K resistor between Q3's collector and Q2's base. Without it, the output "weaved" really badly; apparently hysteresis or some other odd behavior. It's very steady now.

The other odd thing is the regulation: I ended up using a 6.8V zener plus a 1N4148, which still gives me a little over 9 volts output. This didn't work quite the same as my other boost converter. The regulation works fine; it's just a matter of sizing the parts to adjust the output. And I really don't see why you wanted a resistor between the zener and Q3's base: why would that be needed?

You never addressed a major problem with this circuit, which is that Q2's Ic is a full 2 amps! That's gonna kill that poor transistor. How do we solve that?

Except for that, it seems close to ready for prime time ...

 

[MrAl]

Hi again Carbon,

Well, i seem to have explained building the regulation part of the circuit a little different than that, but no problem, instead i'll post the circuit so you can try it yourself. The collector current looks ok and no spikes on the base of Q2.

 

[markalic]

great thread, thisi also a big help for me

 

[MrAl]

Hi,

Why you want to build a boost converter too?