Low voltage step up design?

[BobG]

I have a D cell boostcap... 350 farads, 2.7V. Takes several minutes to
charge up on a 3A 2.5V bench supply. The volts tics up a 1/10v every
second or so. I want to make a dc-dc conv that will chop the cap
voltage and step it up to about 14v. I'd like it to work from 2.5v down
to .5v or as low as we can get it. Would a flyback topology work? Or
maybe the cap voltage supplies an h-bridge and a transformer primary is
driven from the h-bridge? There are commercial wide input range dc to
dc converters, but the lowest input range seems to be 4.5-18v. Anyone
have any hints, tips, ideas, problems to look out for? The secondary
could be a regular old bridge rectifier and a pwm regulator. The trick
seems to be sniffing the cap right down to .5v or so before shutting
down.

 

[Tim Williams]

Old germanium might be the only thing that'll let you run that low.. I mean
if you keep a higher voltage supply (1-5V) at all times, you could even use
FETs and get better efficiency still, but silicon BJTs need 0.7V or more to
turn on and more than 2V for MOSFETs. If it shuts down and you don't have
anything to start it below the operating voltage, you'll have to charge it
again before it can be used.

Ge gets Vf ~= 0.3V, as compared to Si ~= 0.7V. It's usually slower, so
you'll need an iron cored transformer for the inverter. Such circuits have
been used for single solar cells.

How much current do you need, anyway?

Tim

 

[John_H]

I have a D cell boostcap... 350 farads, 2.7V. Takes several minutes to
charge up on a 3A 2.5V bench supply. The volts tics up a 1/10v every
second or so. I want to make a dc-dc conv that will chop the cap
voltage and step it up to about 14v. I'd like it to work from 2.5v down
to .5v or as low as we can get it. Would a flyback topology work? Or
maybe the cap voltage supplies an h-bridge and a transformer primary is
driven from the h-bridge? There are commercial wide input range dc to
dc converters, but the lowest input range seems to be 4.5-18v. Anyone
have any hints, tips, ideas, problems to look out for? The secondary
could be a regular old bridge rectifier and a pwm regulator. The trick
seems to be sniffing the cap right down to .5v or so before shutting
down.

Search the converter sites for "single cell." There are a bunch. When
you're cut off at 0.8V, keep in mind that it's only 10% of the energy
you're leaving in the 2.5V cap, not 30%. You may be challenged to find
the current you want if it's excessive. If you're fine with low
current, you may need to boost with the single cell converter first and
boost to 14V second. You could also use the single cell converter to
bootstrap a better boost converter. Options abound.

 

[Winfield Hill]

John_H wrote...

Search the converter sites for "single cell." There are a bunch.
When you're cut off at 0.8V, keep in mind that it's only 10% of the
energy you're leaving in the 2.5V cap, not 30%. You may be challenged
to find the current you want if it's excessive. If you're fine with
low current, you may need to boost with the single cell converter first
and boost to 14V second. You could also use the single cell converter
to bootstrap a better boost converter. Options abound.

With care a two-step converter isn't necessary. Bob is starting
with a charged cap, so he can power the converter IC from the cap
to start, and from the converter's output as it starts running.

I think a flyback _transformer_ is the best configuration for Bob.
E.g, a 1:6 step-up ratio would allow for a sensible switching duty
cycle, especially at the low end of the input-voltage range. If
he wishes, Bob can add a second winding to run the switcher IC at
its sweet-spot voltage after it's operating.

 

[John Larkin]

John_H wrote...

With care a two-step converter isn't necessary. Bob is starting
with a charged cap, so he can power the converter IC from the cap
to start, and from the converter's output as it starts running.

I think a flyback _transformer_ is the best configuration for Bob.
E.g, a 1:6 step-up ratio would allow for a sensible switching duty
cycle, especially at the low end of the input-voltage range. If
he wishes, Bob can add a second winding to run the switcher IC at
its sweet-spot voltage after it's operating.

Hi, Win,

There are lots of small, cheap, stocked 2-winding surface-mount
toroidal transformers around. This can sometimes be handy...


Vin----=========--*--=========-----ak----+-----Vout
| |
| |
d c
---g |
s |
| gnd
gnd



It's a sort of autotransformer bootstrapped onto the input voltage.
You don't get a lot of choice on turns ratio, I guess.

John

 

[Joerg]

Hello Tim,

Old germanium might be the only thing that'll let you run that low.. I mean
if you keep a higher voltage supply (1-5V) at all times, you could even use
FETs and get better efficiency still, but silicon BJTs need 0.7V or more to
turn on and more than 2V for MOSFETs. If it shuts down and you don't have
anything to start it below the operating voltage, you'll have to charge it
again before it can be used.

Ge gets Vf ~= 0.3V, as compared to Si ~= 0.7V. It's usually slower, so
you'll need an iron cored transformer for the inverter. Such circuits have
been used for single solar cells.

Ge is expensive and not well suited for any serious production. You are
basically dealing with a limited stock, almost collector's items.

Another option might be to add a nifty start-up circuit around a JFET or
other depletion mode device.

 

[Tim Williams]

Ge is expensive and not well suited for any serious production. You are
basically dealing with a limited stock, almost collector's items.

I would've suggested tubes, but they typically don't go that low. <g>

That's true, he didn't mention if this was production or what...

Tim