What's a "charge pump"?

Discussion in 'Electronic Basics' started by Steve Evans, Nov 9, 2004.

1. Steve EvansGuest

Helo,

whats a charge pump, what does it do and how does it work? Is it true
it can be used to step-up DC volatages?

tnx,

Steve

2. John LarkinGuest

Very generally, the idea of charging capacitors and carrying the
charges here and there. You can, for example, charge a number of
capacitors in parallel from, say, a 9 volt battery, then disconnect
them and rearrange in series and get N*9 volts, at least for a while.
Simple diode-capacitor circuits (voltage doublers or multipliers) can
be considered charge pumps, too. A Marx generator is an extreme charge
pump.

PLLs sometimes use "charge pump phase detectors" which shoot short
current blips, hunks of charge, into a capacitor to pump its voltage
up or down.

I guess it's any circuit that depends on quantized charge to define
its operation. This, like a lot of other circuit terms, is fairly
vague.

John

3. Steve EvansGuest

Oh yeah Ive heard of diode voltage multipliers someplace.
So say I've only a pretty crappy low voltage available, like 5V say,
and I want to momentarily whack a capactior up to say 15V, could I do
it by pumping a short but powerful 'packet' (or *packets* ) of charge
from the 5V source into it? Would the capactior store the excess
charge in the form of excess voltage accross its plates if you get
what i mean?

tnx

4. John FieldsGuest

---
It won't, because no matter how "powerful" the packets are, they've
only got 5V from the supply across them, so you're stuck with 5V.

BUT...

You _could_ charge up a bunch of capacitors in parallel, and then hook
them up in series, and discharge them into the cap you want to whack
up to 15V, like this:

CHARGE
+5V---+-----+-----+-----+-----+-----+
| | | | | |
[C1] [C2] [C3] [C4] [C5] [C6]
| | | | | |
GND---+-----+-----+-----+-----+-----+

DISCHARGE

+--[C1]--[C2]--[C3]--[C4]--[C5]--[C6]--+
| |
| |
+---------------->[C7]<----------------+

Since all of the caps were charged up to 5V, when you hook them up in
series all the voltages add up, (the same as if you were to hook
batteries in series) so at the instant you connect all the caps in
series to C7 there'll be a total of 30 volts across C7. It won't stay
at 30V for long though, since it's like filling an an empty pitcher
with a full one.

5. Robert MonsenGuest

Another way to do this is with an inductive spike. Inductors like to
keep current flowing through them. Thus, if you get current flowing, and
then cut it off suddenly, the voltage at the point you cut it off will
spike due to the sudden extra electrons (or lack thereof).

This is how DC-DC boost converters work. They have an oscillator that
turns on and off current through an inductor, and tap the point where
the inductor and transistor meet (the point with the voltage spikes)
with a diode pointing away from the junction. They then charge up a
capacitor with the 'packets' of charge that the inductor keeps sending
in, regardless of the voltage. A feedback mechanism turns off the
oscillator when the voltage reaches the point the circuit is designed to
get to.

The amount of power that can be taken from a circuit like this is
defined by the frequency and duty cycle of the oscillator, the
inductance of the inductor, and the voltage across it when charging.

You can buy these dc-dc converters off the shelf for various voltages,
or build them pretty easily. White light LEDs require about 4V to light,
but they can be lit from a single 1.5V AA battery using a simple circuit
like this.

--
Regards,
Robert Monsen

"Your Highness, I have no need of this hypothesis."
- Pierre Laplace (1749-1827), to Napoleon,
on why his works on celestial mechanics make no mention of God.

6. peterkenGuest

yes, it IS possible to get 12V from 5V, i've done it in a circuit needing
+12V/-12V for a few mA supply and only having 5V available
you'll need an oscillator circuit giving a 0-to-5V wave (say a 555 circuit
or something similar), and one voltage doubler for every output
Carefull, using a TTL output is NOT enough due to the diode-drops and TTL
usually going only swinging from 0.8 to 3V

(view in notepad using fixed font)

|\|
+--||------------+-------| |--+--o +12V
| | |/| |
| --- |
| / \ |
| |\| --- |
+--||--+---| |---+ |
| | |/| --- ---
| | --- ---
| --- | |
| / \ | |
| --- === ===
| |
| |
| o
| +5V
|
|
| |/|
+--||------------+-------| |--+--o -12V
| | |\| |
| --- |
| \ / |
| |/| --- |
+--||--+---| |---+ |
| | |\| --- ---
o | --- ---
(osc) --- | |
o \ / | |
| --- === ===
=== |
===

Oh yeah Ive heard of diode voltage multipliers someplace.
So say I've only a pretty crappy low voltage available, like 5V say,
and I want to momentarily whack a capactior up to say 15V, could I do
it by pumping a short but powerful 'packet' (or *packets* ) of charge
from the 5V source into it? Would the capactior store the excess
charge in the form of excess voltage accross its plates if you get
what i mean?

tnx

8. Steve EvansGuest

Tnx, Terry, thats not quite what I was getting at.
lets say I wanted to contrive a VCO where controle voltages of 4.3,
10.25, and 18.35V were needed at certain intervals; would a charge
pujmp do the trick? is it accurate enough from a 5V source to bump up
that much precisely?