# Voltage Multiplier Question

Discussion in 'Electronic Design' started by [email protected], Aug 25, 2008.

1. ### Guest

My Usenet reader is playing up so if this is a double post I
apologize:

I am experimenting with a wind turbine home built with a brushless,
permanent magnet,
three phase Fisher and Paykel washing machine motor.

To do some tests, I have used a variable resistive load and, by
manually
varying the load on the wind turbine, I have found that a 6km/h wind
can
generate 10W, a 12km/h wind can generate 80W, a 23km/h wind can
generate
640W and a 46km/h wind can generate over 5KW.

The problem is when I use batteries instead of a variable load.
At lower wind speeds, the produced voltage of the motor
is not higher than the batteries, so the turbine just spins and spins
but
does not make any useable power. The turbine only starts charging the
batteries at 20km/h winds, which is a shame as a lot of power just
flies by
without being collected. This is a battery impedance mis-match
problem.

So I think it would be very worth while designing a circuit that
doubles
triples or quadruples the voltage at lower wind speeds, so that I can
get
every last bit out of the wind.

I have looked at voltage multipliers, and since the F&P motor is three
phase, I wonder what your opinion is of this circuit right at the
bottom of
the page:

I was thinking of using a microprocessor to control mosfets to select
the
optimum output voltage for the RPM of the wind turbine. Each mosfet
would be
connected to each subsequent voltage doubler point.

I am not too familiar with voltage multipliers, hence if anyone has
any
opinions on their use in this purpose, I would be very interested to
know.
Voltage ripple is not much of an issue, I am more interested in what
the
efficiency of such a circuit would be or any drawbacks that I have not
forseen.

2. ### MooseFETGuest

The power available from wind goes as the cube of the wind speed.
This makes it a lot harder to get useful charging in a light wind.

[... snip about voltage multipliers ...]

A voltage multiplier will boost the voltage but I think there are
better ways to go. A simple flyback booster will boost by a variable
amount. In the wind generator case, things are fairly easy for the
flyback design. The MOSFET doesn't have to handle a large amount of
power.

At low wind speeds, the input voltage to the booster is low. At low
wind speeds, the torque is also low so the current you want to take is
low as well. This means that the MOSFET only needs to pass a lowish
current when it is on.

At higher wind speeds, the voltage is high. The torque and hence the
current is also high. In this case, the MOSFETs on time will be short
making the power in the MOSFET low.

At very high wind speeds, the boosting isn't needed at all so the
MOSFET never turns on.

If you wind your own inductor, you can go a little light on the core.
It doesn't matter if the core saturates in the high wind case because
the MOSFET isn't going to turn on.

Chances are a very simple controller will do what you want. I suggest
a home brew one that does this:

If the drain on the MOSFET is at a lower voltage than the battery,
trigger a one shot.

The one shoot turns on the MOSFET for a fixed time in most cases.

A safety circuit consisting of an NPN and a resistor in the source of
the MOSFET forces this first one shot to reset if the current goes too
high. This circuit only acts in a fault case.

A second oneshot enforces a short delay after the first oneshot.
During this delay, the first oneshot is prevented from firing.

This second oneshot prevents the first from firing again during the
time it takes to gat the MOSFET actually turned off. It also helps in
the very light wind case where the logic given above would turn the
MOSFET on for too great of a duty cycle.

3. ### leggGuest

As you have forseen, a multiplier stage cannot be easily deactivated,
largely because the control switch has to deal with AC current.
Controlling mosfets in the diode positions will not be effective
because they only control in one direction.

You already have rectifiers in the circuit. By locating suitable
inductors on each phase, before these rectifiers, a single PWM'd boost
switch located after the rectifiers can be used to transform the
apparent impedance of the battery load. This is easily disabled at
higher input voltages, simly by turning the switch off completely.

Examples in this survey by Jang and Jovanovic:

http://www.deltartp.com/dpel/dpelconferencepapers/Apec98-comp3p.pdf

RL

4. ### JoergGuest

Go with Ken's suggestion. This can later be expanded into something like
maximum power point tracking where the conversion ratio is adjuested on
the fly when the wind changes. This makes sure the blade and motor
always see a somewhat optimized load.

5. ### neon

1,325
0
Oct 21, 2006
DO YOU REALY BELIEVE that you can change a battery like a resistor load?
and on a real system it is not unsual to get comercial power to start spinning the wheels to begin with. and doubleing voltage will get nothing but a less efficient system. you may change energy from a to b but you cannot create b from a from nothing.

Last edited: Aug 25, 2008
6. ### MooseFETGuest

There is a way around this problem.

If the MOSFET is in series with a capacitor and is switched off. the
substrate diode will conduct for a couple of cycles and charge up the
capacitor. At that point no further current will flow. This would
allow you to switch out a tripler without too much trouble. Turning
the MOSFET back on will be trouble, however because the stored charge
must be dealt with.

Another way to attack the problem would be to gear up the speed so
that the voltage is always higher than desired. You could then use an
SCR based circuit to reduce the voltage back down. The required
inductor will be physically large but since the windmill is big and
heavy, nobody will care.

7. ### leggGuest

The real problem is the presence of the capacitor in series, and the
fact that rectification is performed in the circuit. You can't short
it out - only an open circuit deactivates it - meaning a whole new
rectification path is needed in it's absence.

The multiplier has advantages that may be mis-applied in this
situation.

It might be easier to process the energy and to control the effective
load impedance, if the control function is positioned after the
rectifier. Easier to do at home, anyways......

RL

8. ### MooseFETGuest

If the rectifier is a tripler, just opening up the capacitor leaves
you with a simple rectifier.

If you make a bridge rectifier, you can make it a doubler by
connecting one side of the source through a capacitor to ground.

The MOSFET can be turned on with a photovoltaic isolator. This will
turn the MOSFET off and on very slowly.
I'm not suggesting that a multiplier is a good idea. I am just

I don't know:

Go to the junk store and buy a Variac. Replace the knob with a
pulley. Put a smaller pulley on a small electric motor. Buy some
weights. The electric motor turns the Variac down and the weights
pull it up. When the voltage to the motor makes just the right torque
to balance the weights the Variac setting is just right.

9. ### neon

1,325
0
Oct 21, 2006
The point is was and still is. POWER if he cannot generate the power to begin with doubling the voltage sticking mos fet is not going to do it. A battery don't give a damn about voltage[ as long as is higherthen itself] it wants current if the current isn't there it will never work.

10. ### leggGuest

There are many things you can do with a big ball of string, but please
keep in mind that the OP is working with a 3ph, 'variable' frequency,
'variable' voltage source. Variable, in this case, being as
predictable as a mule.

RL