Maker Pro
Maker Pro

SMPS boost regulator using a schmitt trigger oscillator

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
Cool Steve :) This is quite interesting.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
Here are some quick resuts for a 8.95V input, and approx 22V output.

attachment.php


As you can see, the voltage falls as the current exceeds about 10 mA. This is due to the max 50% duty cycle, the frequency, and the high value of the inductor. There are a lack of data points between 10 and 20 mA, I expect you would see a nice curve here.

At 2.2mA output current, the output ripple is around 7 mV. The duty cycle of the oscilator is between 12% and 20%

At 64mA (and at 14.2V, the ripple increases to 27mV)

It is intersting that this regulator is effectively current limited and would apear to be fairly tolerant to overload.

Edit: R2 was removed from the circuit. Not removing it from the circuit diagram was an oversight.
 

Attachments

  • eff22.PNG
    eff22.PNG
    12.6 KB · Views: 441
Last edited:

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
After ironing out a few bugs in my original graph, you can see a similar behaviour:

attachment.php


I found that the only zener diodes I had in non surface mount packages were 11V and 3.3V. The 3.3 V ones have such a sloppy knee that each one only added 1V to my output voltage -- the forward voltage from was almost indistinguishable from the reverse voltage drop!
 

Attachments

  • eff11a.png
    eff11a.png
    13.5 KB · Views: 363

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
The 3.3 V ones have such a sloppy knee that each one only added 1V to my output voltage -- the forward voltage from was almost indistinguishable from the reverse voltage drop!

I guess that's because the transistor in the feedback loop will respond to a pretty low base current. I was thinking that you might want to add a base-emitter resistor there, to make it more immune to the zener's leakage current. Do zeners become more leaky with age? I guess they'll be more leaky at higher temperatures, at least.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
The 11V zener, I calculate as having about 10.5V (possibly slightly more) across it at the minimal currents used here (0.1uA calculated)

I was actually contemplating using a TL431 for a short while, however those tend not to work too well under 500uA or so.
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
After making a litte adapter for a 15V zener, I have the following results:

attachment.php


This is the design voltage and 30mA represents the maximum load.

I'm probably going to use the other gates in this package to create an "on" timer which keeps the circuit enabled for only a short time
 

Attachments

  • eff15.PNG
    eff15.PNG
    13.5 KB · Views: 359

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
And just to try out digikey's schematic drawer...

attachment.php
 

duke37

Jan 9, 2011
5,364
Joined
Jan 9, 2011
Messages
5,364
A very neat drawing, I will perhaps look at digikey when I have time.

What does R2 do or IC1b for that matter?

I am wanting to produce about 400V, I will start a thread, asking for help.
 

KrisBlueNZ

Sadly passed away in 2015
Nov 28, 2011
8,393
Joined
Nov 28, 2011
Messages
8,393
R2 seems to be left over from the design in post #4 when feedback was applied differently. I don't see any need for it now.

IC1B is needed because when Q2 conducts and disables the oscillator, IC1A's output will go high, but you want the MOSFET to turn OFF.
 

jpanhalt

Nov 12, 2013
426
Joined
Nov 12, 2013
Messages
426
I am interested in seeing duke37's thread. Could he be thinking of using it for a CDI?

John
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
Good one Steve, I use that sort of oscillator all the time in non critical timing applications. I did something similar a few years back when I made use of the high voltage produced by the coil to charge up a capacitor for an electro magnet.

Adam
 

(*steve*)

¡sǝpodᴉʇuɐ ǝɥʇ ɹɐǝɥd
Moderator
Jan 21, 2010
25,510
Joined
Jan 21, 2010
Messages
25,510
A very neat drawing, I will perhaps look at digikey when I have time.

It can be neat and convenient, but it's not particularly quick to use.

I am wanting to produce about 400V, I will start a thread, asking for help.

For low current, using a transformer instead of an inductor will allow you to get higher peak voltages. As I mentioned earlier in this thread, this design topped out at 90V, probably due to the diode or the mosfet (it can't have been good for the capacitor either).

What does R2 do or IC1b for that matter?

R2 seems to be left over from the design in post #4 when feedback was applied differently. I don't see any need for it now.

Kris has it exactly right on both counts. I think I removed it at some point, but it wouldn't have made any difference.

[/quote]IC1B is needed because when Q2 conducts and disables the oscillator, IC1A's output will go high, but you want the MOSFET to turn OFF.[/QUOTE]

Yeah, all I wanted was the mosfet off when the output voltage was too high. There are probably several ways of doing it, but this was the simplest and the inverter was already just sitting there.

I could have used another transistor and a couple of resistors to pull the oscillator's gate input high, but that's more components. This was initially designed to be as simple as possible and use the smallest number of components.

Good one Steve, I use that sort of oscillator all the time in non critical timing applications. I did something similar a few years back when I made use of the high voltage produced by the coil to charge up a capacitor for an electro magnet.

My original idea came from a suggestion to use the output pin of a microcontroller to generate a higher voltage. And yes, I like to use schmitt triggers for various non-critical timing, pulse shaping, pulse extending, missing pulse detection, etc. applications.
 
Top