Flyback transformers are found in
monitors, TVs or anything with a CRT, and are sometimes known as Line
OutPut Transformers, or
just LOPT. They are used for generating high voltage for the CRT,
which is needed to create an electric field, which in turn accelerates
electrons towards the screen, which finally excite phosphors and create
the image you see. Flybacks are
designed to work best anywhere between 15 to 150 kHz, so some
experimentation is required to find the intended operating frequency.
TV flybacks are generally designed for
upper audio frequencies, which is the cause of the high pitched noise
heard from a muted TV. (If you're over 40 you will need to confirm this
with your kids.) The optimal operating frequency can have many
harmonics, which will work as well as the
frequency to some extent. Since flyback transformers use a ferrite core
vastly different operating conditions than an iron cored mains
transformer. In fact flyback transformers aren't really conventional
transformers at all, but coupled inductors which means they should be
driven differently. Fyback transformers are generally either driven in
"flyback mode", or some push-pull topology. The first two drivers on
this page drive the flyback in flyback mode, while the last two use
push-pull topologies. To obtain a high frequency variable duty cycle
drive signal we can use the 555 timer. This simple driver circuit is
quite efficient if tuned correctly, and in some cases quite powerful.
It is currently set to run between 17-50 kHz, which should be a large enough range to sweep through any
harmonics a flyback may have.
This is a pretty standard
design. All parts except the timer and mosfet are non-critical. Input
should be 12-16 volts, the current draw can reach a few amps. For the
mosfet I used an IRFP450, though any mosfet with a breakdown voltage
above 200V and "avalanche rated" will work. Make sure you use a mosfet
and not a bipolar junction transistor, the symbol in the datasheet
should resemble the one in the schematic. For a different
frequency range you can use a 555 timer calculator or just experiment
to find new capacitor and resistor values. As mentioned above this
driver drives the flyback in flyback mode. What that means is that the
mosfet is turned on by the timer, and current starts to flow through
the primary winding. After some time the timer will turn off the mosfet
again and the current will be forced to stop. However, this is not
possible since the primary has significant inductance. The current then
causes the voltage at the mosfet drain to increase in an attempt at
allowing current to flow. The voltage will rise up to the breakdown
voltage of the mosfet, where it stops (since the mosfet is avalanche
rated this does no harm, and only produces heat in the mosfet). The
voltage at the mosfet drain will potentially be equal to the breakdown
voltage of the mosfet, meaning the primary voltage will be hundred of
volts now. Due to the large turns ratio of the flyback the few hundred
volts at the primary become several thousand volts on the secondary.
Since some energy is avalanched in the mosfet, adequate heatsinking of
the IRFP450 is required.
Winding your own Primary
I recommend winding your
for several reasons. For one thing you don’t have to worry about
finding the built-in
and you can adjust the primary turns according to the drive voltage or
desired output voltage. Also you don't need to worry about destroying
the internal primary during the experimentation phase.
The primary must be wound directly onto the exposed ferrite core. The
number of turns varies, and is determined by operating voltage, on-time
and core cross-sectional area. For general use, 3 to 10 turns should be
this driver. Fewer turns mean higher voltage, but increased mosfet
power dissipation. Start with 10 turns and remove them until
the MOSFET gets too warm or the spark too big.
Standard Monitor Flyback Transformer with new primary.
For those of you who have
never seen a
flyback transformer before, it may be a bit tricky to know where the
primary, ground and other pins are. The ground pin can be found by
the pin the HV arcs to the most. Simply take the HV lead and bring it
near the pins on the bottom. The internal primary can be found by
measuring resistance. It should be around 1 ohm. Some flybacks may
have several winding which will appear to be primaries, in this case
the real one can only be found by measuring inductance. A typical
primary inductance is often 300µH.
Sparks at 5v. This was before I knew anything about electricity, so 5V was all I considered safe for the driver.
Of course something can go wrong, so if you're unlucky check these points.
If you hear a
high pitched whine but you don’t have any HV,
you're close. This is caused by a false primary or wrong
phasing. Try switching the primary leads, since the flyback is
rectified by a single diode, which makes polarity is important. If the
is weak, the primary polarity may need reversing. Check to see if you
have found the correct ground pin.
Silence. This is what I hate the most. Basically just check that
everything is wired correctly, if it is, check that all the parts are
functioning. You can check for a frequency by replacing the flyback
with a speaker. If you get a high pitched whine it's alive.
If you're getting a drive signal out of the circuit, you are
probably not connected to the correct primary. Wind your own if you
fail to find the internal primary.
that the 555 is still functioning. Use a 555 timer tester circuit.
555 Driver MKII
About 18 months later I decided to
try this again, only this time I knew what I was doing. I whipped up
this new 555 driver, which works quite well. The max voltage is 50V
with an IRFP450, due to the primary energy becoming to much for the mosfet to avalanche
without dying. If you want to power it from an even higher voltage, use
a snubber or stronger mosfet instead. The additional circuitry simply
isolates the 555 from the "power" supply, so it can be increased beyond
the 16V rating of the NE555. Keep in mind standard 7812 regulators
should only run from 30V maximum, so for 50v you need to cut out the
7812 and run the logic section from a dedicated supply.
Make your own PCB!
I've designed a PCB for this driver which can be edited with ExpressPCB. A pdf of the
copper traces and components layout is also included for those without ExpressPCB.
I recommend using a 5.45mm block connector for the IRFP450, in case it
needs changing. As with the basic driver, heatsinking of the IRFP450 is
required. I've used an old processor heatsink with a small 12V fan. If
powering the fan the 7812 regulator will need a heatsink as well.
24V input, 3 cm+ sparks
24V in juicy spark
70kV from 50 volt supply (driver was modified)
Tired of measly 2 cm sparks? Want more
power? Try the Mazzilli ZVS flyback driver! This driver is capable of
pumping upto a kW of power, so impressive arcs can be made. Normal
operation from 12V only results in about 100W, but that's still several
times what the 555 drivers will process. The circuit is has very
few parts, is simple, and very elegant. If it weren't more deadly than
the 555 drivers I would have made it my recommended newb driver. The
arcs produced by this driver
are very hot, the copper ground wire goes white
anything brought close to the arc in incinerated. The current in these
arcs can kill, so be carefull. The primary winding is
center-tapped, so two windings must be wound in the same direction. I
usually intertwine two wires, and then simply wind one winding with the
pair. The start of one wire and end of the other are connected and
serve as the center-tap. The primary should be
between 3 + 3 and 10 + 10 turns, depending on voltage. The
amount of turns depends on supply voltage, and resonant frequency of
the circuit. The highest voltage I've heard of people running
Mazzilli driver at is 100V. As you can imagine, the arcs were insane!
The resonant frequency is determined by the total primary inductance
parallel capacitor. It is a simple parallel resonant circuit, so the
resonant frequency is easy to find with the parrallel resonant LC
Circuit designed by Vladimiro Mazzilli.
The circuit works by one mosfet turning on due to differences in the gate
resistors or internal structure of the mosfet. Once on, the opposite
mosfet will be held off by the fast diodes. The voltage across one
primary half will rise up an fall again in a half-sine wave. Once at
zero the mosfet that was on will be forced off, and the mosfet which
was held off will be allowed on. The cycle repeats in opposite this
time, before returning to where it started. The large inductor serves
as a "current capacitor", providing constant current to the driver.
Thanks to the resonant action of the circuit, it benefits from ZVS, or
Zero Voltage Switching. This means that the mosfets switch on with no
voltage across them, so while they transistion from off to on they
won't dissiapte power. (P = I * V)
Arc at 25V supply voltage provided by my
MOT PSU. Susposed to be 50V but it dropped to 25V when pulling an arc,
I should have done a better job of rewinding the MOT!
Off-Line Flyback Driver (The wall's the limit)
Running power through multiple
transformers just to power another transformer seems a bit absurd, so
I though it was time for a direct mains powered flyback driver. This
driver rectifies mains and produces a 320V DC source. The circuitry
from the TL494 to the GDT creates alternating pulses for the
IRFP450s at any desired frequency and duty cycle. The IRFP450
half-bridge feeds a square wave at +/- 160V to the flyback
primary. 160V at considerable current, stepped up to some kV and mA
produces some impressive arcs. The driver itself is my "multipurpose inverter".
Contrary to the 555 drivers this driver provides actual AC to the
primary, and due to DC flybacks being half-wave rectified you'll likely
have problems with saturation regardless. I've run DC flybacks offline
several times, but they have all failed eventually, and they cause the
driver to heat up. Unrectified AC flybacks are much more suitable for
this driver, and cause very little heating of the mosfets. See my
article on making HV
transformers for details.
The frequency and number of turns are a matter of
tweaking and design, but for safe testing 30 turns and 100kHz have
worked fine for a wide selection of flybacks. The main issue here is
saturation of the transformer, which prevents the core from becoming
further magnetized. Practically speaking, it
means the transformer inductance will drop suddenly when saturation
occurs, causing incredible current
draw from the inverter and most often destroying it. (Or with my
multi-inverter, simply tripping the OC protection.) I've put a
calculator in the multipurpose inverter spreadsheet for determining the
minimum number of
turns required to achieve a specific flux density. Use 0.25T as a
starting point if you don't know the saturation flux value of the core.
Since you won't be running the flyback transformer in flyback mode anymore you can
remove the air gap to decrease the idle current drawn. Remove the metal bracket, and pull the core halves out.
In-between the core halves is a thin plastic spacer which creates the air gap.
Entire driver assembled in case. This was before it was upgraded into the Multipurpose Inverter.
Awesome arcing action.
Some interesting things to do with HV
arcs are coloring them with salts, or making magnet vortexes. Common
table salt will give red/orange arcs, and boric acid will give green
arcs. Check the Internet for flame tests to see which salts will give
which colors. The
effect is hard to capture since the camera saturates and can't see
the green color. Arcs drawn with boric acid are much brighter than
usual, and are uncomfortable to look at even with sun glasses on. Arcs
drawn with salt (or iron, I'm not sure which is causing the arcs to
change appearance) on the other hand, are actually pleasing to look at
as they are dimmer than usual. Of course, coloring plasma with salts in
more impressive with Tesla coils, since they aren't as bright. Another
thing to do with DC arcs is to make a plasma vortex. Arcs consist of
plasma which is easily affected by magnetic fields, hence the awesome
vortex. See the 4HV thread for more info and videos.
This blurry picture is all I have to suggests that the arcs are green.
A magnet plasma vortex can be made with the round ferrite magnets found in microwave ovens.
I do not take responsibility for any injury, death, hurt ego, or other
forms of personal damage which may result from recreating these
experiments. Projects are merely presented as a source of inspiration,
and should only be conducted by responsible individuals, or under the
supervision of responsible individuals. It is your own life, so proceed
at your own risk! All projects are for noncommercial use only.