Ultra Low voltage input DC-DC converter / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Ultra Low voltage input DC-DC converter

1 2 next

Move Thread

LAN_403

EEYORE

Wed Aug 16 2006, 06:19AM

Registered Member #99 Joined: Thu Feb 09 2006, 06:10PM
Location: florida, usa
Posts: 637

Hello all,
I need to take a very low input voltage and turn it into about 3volts. It is for a hydrogen fuel cell...The fuel cell states an output of 0.6 volts to 0.95volts and current of 350mA. I am going to plan on the output being 0.6volts. Any tips?

Fuel Cell store, sells a dc-dc converter that will run on 0.7volts, but they charge 42$ !!!! It is just a simple 8pin DIP IC with a few capacitors, resistors, and what looks to be a transistor.

I could always go with a "joule thief", but would like something regulated if possible. Right now, I just want to power LEDs, but would later like it to be capable of more.

Thanks!
Matt

Bjørn

Wed Aug 16 2006, 07:12AM

Registered Member #27 Joined: Fri Feb 03 2006, 02:20AM
Location: Hyperborea
Posts: 2058

http://www.maxim-ic.com/Power.cfm
They have many different ones that needs about 0.7V to start but will run at lower voltages.

Dr. Slack

Wed Aug 16 2006, 07:14AM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

Two ways spring to mind, the difference being are you allowed to cheat. Cheating involves having a higher voltage available to start the operation.

If you aren't allowed to cheat, then 0.6v is not going to turn a silicon transistor on, so nothing's ever going to happen with them. Germanium transistors are still available, and they need only 0.3v, so any sort of oscillator built from them would do.

If you can cheat, then my approach would be to use 5v to drive some HCMOS logic to drive a low threshhold FET, and make a boost converter. Once it's running, the output power will provide all the 5v required and then some, so only a flea-power 5v is needed.

If a manual start is allowed, then parallel the boost FET with a push switch, and run the switcher manually for a few cycles to build up the 5v bias.

Steve Conner

Wed Aug 16 2006, 09:51AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

I remember trying to build a circuit that generated 5V for a PIC and LCD display off a single (0.7V) solar cell. I used a MAX-Something-Or-Other boost converter chip

along the lines of the ones linked to by Bjorn. I think I ended up having to use two cells in series though, otherwise it wouldn't start reliably in the conditions it was to be used under. (it was part of a display case at a renewable energy expo)

EEYORE

Wed Aug 16 2006, 08:15PM

Registered Member #99 Joined: Thu Feb 09 2006, 06:10PM
Location: florida, usa
Posts: 637

Yes,
no cheating allowed

I want this thing to run on its own..Otherwise its nothing special...

Also, i refuse to work with SMT...My fingers lost their nimbleness awhile ago...Seems the joule thief is the only way to go...But I can make a modification. I will use a germanium transistor for the joule thief rather than silicon. Should allow it to go much lower no?Im sure the 0.6volts - 0.95volts figure is when its running under ideal conditions, such as a lab supply of H2...

I wonder if there is a way to clip the spikes some of the joule thief?Such that, it remains regulated?How about a zener+capacitor or something?
Matt

Marko

Wed Aug 16 2006, 08:23PM

Registered Member #89 Joined: Thu Feb 09 2006, 02:40PM
Location: Zadar, Croatia
Posts: 3145

You can use some small, sensitive mosfets in astable instead of germanium transistors.
Way more efficient, since 300mV drop is still almost half of your operatingvoltage..

Dr. Slack

Wed Aug 16 2006, 08:36PM

Registered Member #72 Joined: Thu Feb 09 2006, 08:29AM
Location: UK St. Albans
Posts: 1659

OK, go with a geranium Joule Thief. 300mV is the bias required to turn the device on, not the VCEsat drop. In fact IIRC germanium VCEsat is better than Silicon, which is one of the reason power germaniums are still around.

No mosfet, even a sensitive type, is going to conduct with 0.6v on the gate, so they won't self start, a FET needs to be running or cheated to work. By all means use germanium to start the process then switch to a boost FET when 5v is available, taking the bias off the germanium tranny. It sounds complicated, but it's probably the most efficient way to go.

Conundrum

Wed Aug 16 2006, 09:48PM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4059

Aand a good source of power germaniums are those ooooold car radios you find in ancient cars.
Amazingly enough some people keep the old radio in their car because the quality is far better than the modern junk.

Apparently the electret microphones out of tape decks have a very low voltage JFET in them which can be handy for solar/low power applications.

Then there's the 2N3819/3820 or J309/2N3820 "lambda diode"- i've got these to work down to 1.3V before.

My current line of research is based around a "magnetic feedback" system, using a transistor, and small tapped coil wound around a tiny NdFeB magnet out of a dead CD drive.
The hysteresis effect may reduce the drive frequency and the presence of the magnet increases the effective inductance of the assembly so a relatively small coil can emulate a larger one even increasing the size of the back EMF...

This effect has been used in camera viewfinders to multiply the effective size of a small inductor in the vertical drive :)

I've also used this technique to affect CCFL inverters before, though making the magnet part of the resonant circuit would seem the next logical step.

-A

EEYORE

Thu Aug 17 2006, 01:01AM

Registered Member #99 Joined: Thu Feb 09 2006, 06:10PM
Location: florida, usa
Posts: 637

Ok, i tried out a cheaply made fuel cell lashup...Two carbon rods from batteries, some drano crystsals (40% NaOH) and a battery to make H2Bubles. ..With the joule thief, it BRIGHTLY lights an ultrabright LED(battery removed of cource

). This is with a silicon transistors still...I need to find a germanium type.

So, seems the jouke thief deffonately is the way to go!And, hurray! I made a fuel cell...More updates to come, but will be posted under the fuel cell thread.

Matt

Conundrum

Wed Aug 30 2006, 09:14PM

Registered Member #96 Joined: Thu Feb 09 2006, 05:37PM
Location: CI, Earth
Posts: 4059

Hi, I had an idea about using a JFET such as the 2N3819 or J309 with small resistors between d-g and g-s (say about 1K) to simulate a normal bipolar transistor.

Also JFETS will switch down to fractions of a volt

-A

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob