Driving a CRT monitor flyback transformer with squarewave

[Spehro Pefhany]

It is probably not designed to run continuously but it never
seemed too strained when I used it. If your oil burner has
trouble it can be active for fairly long periods of time
(several repeated startup cycles). Mine never got warm or hot.
But then it was the kind of thing that was never casually left
on anyway.

Warren

The one I found at the dump as a teenager weighed about 10lb and was
potted. It didn't get particularly warm, IIRC.

 

[[email protected]]

Spehro Pefhany expounded in


An oil burning ignition transformer does that very well. I
don't know how that compares to a neon transformer though.

Then I was in high school I used an oil ignition transformer to drive a Tesla
coil. It was 10KV @10mA. 100W should be enough to get the air moving in a
Jacobs ladder. ;-)

 

[Grant]

What I've argued is that the emfs in the primary and secondary are
always in proportion to the turn ratio, on the grounds that the change
in flux linkage is in that proportion.

You're completely missing the point of flyback! It's not about turns ratio
as more about charging energy into the inductor and letting the flyback
generate the high voltage. Others have pointed out that the OP is on the
wrong track driving the thing open loop like that. The web circuit reference
is bad in so many aspects, from burning out pots or the 555 on user adjustment,
to having no safe control of the power in the inductor.

Attempting to stop the current in the primary abruptly will cause higher
voltages in both primary and secondary than a slower change would, but
it doesn't alter the relationship. So if you want a high voltage in the
secondary, you have to handle a correspondingly high voltage in the primary.

Yeah, sorta. Close but not quite...

The OP's problem is not an inability to create high voltages, it's an
inability to cope with the high voltage that appears in the primary,
which destroys his transistor, as I would expect it to.

OP does not even have proper snubber network, if he follows that 555
circuit.

If you're not doing the power supply method of applying a voltage until
current rises to a turn off point, then wait for field collapse, you would
need to do a tuned oscillator topology where you add power at just the
right point in the cycle. Sort of saying the same thing in two different
ways, I know. But power circuitry like this is not rocket science.

Says me, has some bits in a box, not yet produced a nice sparking voltage
source -- though I can draw little sparks from an ioniser I built couple
weeks ago to clean the air in here.

Grant.

 

[Grant]

Wiebe Cazemier expounded in


The best Jacobs ladder I ever had was using an ignition
transformer from an old furnace oil burner.

I took that thing to high school once and demonstrated it in an
electrical class (motor/generator stuff). I recall holding some
magnets near the arc, to influence the arc. Then I must have got
some charge from being a bit too close because suddenly my pant
legs balooned out from the static charge collected. That
classroom had a nicely varnished wood floor.

 

[Grant]

Just get an old neon transformer.

Yes, if I needed one for a stage prop, or quickly, but sourcing a s/h
neon xformer in a country town? I dunno where to look...

Grant.

 

[Grant]

Mains frequency neon transformers have a magnetic shunt that gives
them a relatively high output impedance (to act as a ballast for the
negative resistance tube). You'll get a much more intense arc out of
an old ignition transformer, but you might toast the xfmr.

Over here we don't have oil furnaces in common usage, and oil heating
virtually disappeared after the '70s oil price rise, in favour of
natural gas heating. Central heating likely gas or electric reverse
cycle airco.

Neon xformer commonly recommended for Jacobs ladder, otherwise it's
an auto ignition coil or old TV tripler with some 12V high current
power driver circuit.

I've got some high voltage gear recovered from laser printers, but the
current used in them there is very small. No good for drawing big
sparks.

Grant.

 

[Grant]

I think for a good Jacobs ladder you need to put lots of power into
the arc to cause enough heating so that the arc rises on the
electrodes.

Or put the ladder inside an acrylic tube to shield it from the breeze?

Benefit too of limiting ozone generation, or emission, I imagine.

Grant.

 

[Sylvia Else]

You're completely missing the point of flyback! It's not about turns ratio
as more about charging energy into the inductor and letting the flyback
generate the high voltage.

I don't believe I've missed the point at all. The voltage appears
because the current cannot change instantenously. But the ratio of
voltage on the primary and secondary remains.

Others have pointed out that the OP is on the

wrong track driving the thing open loop like that. The web circuit reference
is bad in so many aspects, from burning out pots or the 555 on user adjustment,
to having no safe control of the power in the inductor.

Yeah, sorta. Close but not quite...

In what respect is it wrong? In particular, is there any conceivable
circuit that allows the secondary to present a high voltage while not
having the primary present that high voltage divided by the turns ratio?

Sylvia.

 

[Grant]

On 5/04/2011 9:32 AM, Grant wrote: ....

I don't believe I've missed the point at all. The voltage appears
because the current cannot change instantenously. But the ratio of
voltage on the primary and secondary remains.

Yes, but...

Others have pointed out that the OP is on the

In what respect is it wrong? In particular, is there any conceivable
circuit that allows the secondary to present a high voltage while not
having the primary present that high voltage divided by the turns ratio?

The other keyword you overlook: tripler.

CRT HO drives that high voltage tripler, what's inside of that? A little more
than the crappy so-called efficient 555 circuit from a web-site that holds one
of the worst collections of circuits I've seen on the Internet. Not a place of
learning.

So the simple steps to driving a CRT are to drive primary to some current limit,
You may need to add your own low voltage primary around the exposed ferrite for
12V, I've seen 12V circuits, so it can be done.

So, turn current old until some current limit before saturation, turn off
current, collapse of magnetic field generates very high voltage, repeat at
some frequency that suits how you see the circuit, watching on a CRO is
good, or simply tuning for minimum input for big output.

The open loop 555 or whatever drive is not do good, as you really do want
to switch off input power before saturation of the ferrite, otherwise you're
simply going to burn something.

You can see saturation as unreasonable current draw, or with a CRO a breakpoint
from gentle linear ramp to steep almost vertical ramp of input current.

Maximum efficiency is charging the ferrite so some value below saturation.

Thus driving the primary open loop from an oscillator is bad. Better is using
a latch that you set at a regular rate, and current limit resets latch, Q drives
the output transistor, and voltage feedback tells you how long before setting
the latch again for variable frequency, ort also clears the latch for fixed
frequency. This is standard power switching topology stuff.

Other way is at a fixed frequency into a tuned circuit which is how the CRT
does it, I think. There it was a controlled oscillator with a tight frequency
spec with a BTW we can also make some EHT here too, due to the large amount
of energy circulating to produce horizontal scan driver to the yoke coils.

This pulse train pumps up the energy through the tripler and makes your high
voltage of 26kV max for low xrays from the CRT in original app.

Efficiency comes from driving a tuned circuit so you need a proper snubber
and/or tuning to work with the primary side issues. Of course adding you own
low voltage primary sidesteps the noticed high voltage seen on existing driver.

But before you argue about the thing cannot work, it did work as a CRT EHT
source until OP pulled the parts from PCB.

I've written the above from memory, so the might be minor errors. But I'm trying
to explain the general picture, not give a precise working guideline -- for that
I'd have to drag out gear I got here and have a play. I have other things on
my todo list right now.

As far as voltage multipliers go, I used them very recently to make a 40V IC
generate over 100V to drive 50 LEDs in series at 20mA, 36V chopped DC stepped
up from 12V though a tripler with over-voltage feedback and LED current regulation.

This stuff is not rocket science. But EHT adds some corona and visible sparks
that very much add to the excitement. I hope the above ramble has some useful
points to help OP.

Grant.

 

[Sylvia Else]

Yes, but...

The other keyword you overlook: tripler.

I've already suggested to the OP that he look at voltage multipliers,
so I cannot be overlooking triplers.

CRT HO drives that high voltage tripler, what's inside of that? A little more
than the crappy so-called efficient 555 circuit from a web-site that holds one
of the worst collections of circuits I've seen on the Internet. Not a place of
learning.

So the simple steps to driving a CRT are to drive primary to some current limit,
You may need to add your own low voltage primary around the exposed ferrite for
12V, I've seen 12V circuits, so it can be done.

The issue is not the voltage driving the circuit, but the voltage
tolerance of the switching transistor.

Specifically, the OP is seeing that when the transistor turns off, the
primary voltage rises to a level where the transistor is destroyed.

The OP is trying to get a particularly high voltage *on the secondary of
the transformer* without the voltage on the primary destroying his
transistor. I'm suggesting that this goal is misconceived (aside from
getting a transistor that tolerates a higher voltage).

Sylvia.

 

[Wiebe Cazemier]

I've already suggested to the OP that he look at voltage multipliers,
so I cannot be overlooking triplers.


The issue is not the voltage driving the circuit, but the voltage
tolerance of the switching transistor.

Specifically, the OP is seeing that when the transistor turns off, the
primary voltage rises to a level where the transistor is destroyed.

The OP is trying to get a particularly high voltage *on the secondary of
the transformer* without the voltage on the primary destroying his
transistor. I'm suggesting that this goal is misconceived (aside from
getting a transistor that tolerates a higher voltage).

Well, if the voltage is equal to what it was in the CRT monitor, I'm happy.
I'm not trying to do anything beyond what the flyback was meant to do.

I think I have that kind of voltage now, but not very efficiently, since my
MOSFET driving the thing does get very hot. Perhaps it's over-saturation. I
will give the scope reading Grant suggested a go, see if I can detect the
saturation point.

One thing I have noticed, is that when I have an arc, the arc gets a bigger
orange glow when I reduce the duty cycle. At 95% for instance, the arc is a
purple line. At 80, it's almost a flame. It also feels hotter (the air above
it)

 

[Jamie]

Sylvia Else wrote:




Well, if the voltage is equal to what it was in the CRT monitor, I'm happy.
I'm not trying to do anything beyond what the flyback was meant to do.

I think I have that kind of voltage now, but not very efficiently, since my
MOSFET driving the thing does get very hot. Perhaps it's over-saturation. I
will give the scope reading Grant suggested a go, see if I can detect the
saturation point.

One thing I have noticed, is that when I have an arc, the arc gets a bigger
orange glow when I reduce the duty cycle. At 95% for instance, the arc is a
purple line. At 80, it's almost a flame. It also feels hotter (the air above
it)

You need to saturate the Mosfet to reduce heat. This also goes to say
that you need to select a mosfet with the lowest turn on R you can find.

You really should be current limiting the primary side. There are a
couple of ways I can think of to do this. One way not to do it is
biasing the transistor, this will generate heat in the Mosfet but can
be done if you so desire.

Years ago I had to make a 20kv dielectric tester to generate 1 ms
transient pulse across mica capacitors. I was able to achieve this
using an automobile coil. I used 2 555's. one was a current regulated
switching supply and the other was for gating that end point supply
to the primary side of the coil. Mosfets were used on the forefront here
since it made perfect sense. I elected to use a low voltage design since
it was easier to find Fets that had much lower Ron and a car coil was a
perfect candidate for the job.



Jamie

 

[Grant]

You need to saturate the Mosfet to reduce heat. This also goes to say
that you need to select a mosfet with the lowest turn on R you can find.

Same time you don't want to saturate the transformer, unless you're making
a self-oscillating LV switcher that uses that magnetic saturation point to
turn off the gate drive. Then you could regulate the output like a CCFL
backlight inverter circuit, there's lots of those on the 'net too. Though
they regulate to constant current, you can see the topology. Some here would
name it, I forgot the name.

You really should be current limiting the primary side. There are a
couple of ways I can think of to do this. One way not to do it is
biasing the transistor, this will generate heat in the Mosfet but can
be done if you so desire.

Years ago I had to make a 20kv dielectric tester to generate 1 ms
transient pulse across mica capacitors. I was able to achieve this
using an automobile coil. I used 2 555's. one was a current regulated
switching supply and the other was for gating that end point supply
to the primary side of the coil. Mosfets were used on the forefront here
since it made perfect sense. I elected to use a low voltage design since
it was easier to find Fets that had much lower Ron and a car coil was a
perfect candidate for the job.

Yup. TV EHT tripler has the xformer, tripler built in, but I've not
looked at the driving circuit for years. There's space to wind one's
own 12V primary around them. I collected a couple more monitors to
dismantle, see if I find a more recent one.

Grant.

 

[Sylvia Else]

Well, if the voltage is equal to what it was in the CRT monitor, I'm happy.
I'm not trying to do anything beyond what the flyback was meant to do.

The (bipolar in that case) transistor I removed from my own colour TV
recently had a rating in excess of 1000V.

I think I have that kind of voltage now, but not very efficiently, since my
MOSFET driving the thing does get very hot. Perhaps it's over-saturation. I
will give the scope reading Grant suggested a go, see if I can detect the
saturation point.

One thing I have noticed, is that when I have an arc, the arc gets a bigger
orange glow when I reduce the duty cycle. At 95% for instance, the arc is a
purple line. At 80, it's almost a flame. It also feels hotter (the air above
it)

Well, you'll only get a high voltage in the secondary while the
transistor is off (and then only while there is energy remaining in the
field).

At 95% I wouldn't be surprised if the transistor turns on while there is
still energy in the field, which would also reduce the amount of energy
that can be added, and increase the likelihood of driving the
transformer into saturation.

Sylvia.

 

[Wiebe Cazemier]

The (bipolar in that case) transistor I removed from my own colour TV
recently had a rating in excess of 1000V.


Well, you'll only get a high voltage in the secondary while the
transistor is off (and then only while there is energy remaining in the
field).

At 95% I wouldn't be surprised if the transistor turns on while there is
still energy in the field, which would also reduce the amount of energy
that can be added, and increase the likelihood of driving the
transformer into saturation.

Sylvia.

I've determined the inductance value of the primary coil of my flyback. I
charged it up to 20V through 150R (with a squarewave). It took 30uS for the
voltage across the coil to reach about zero. With T=L/R and assuming it
takes 5 time constants to charge it, inductance is 900 uH.

By calculating the back-EMF ballast resistor with the same formula and the
15 uS it took for it to discharge, it gave me 300R, which is about right.

Now on to find the proper value for a capacitor to make a tuned circuit. And
read up on tuned circuits

You need to saturate the Mosfet to reduce heat. This also goes to say
that you need to select a mosfet with the lowest turn on R you can find.

The IRF840 has Ron of 0.85 Ohm. At 5 A, it would dissipate 21W. I have a
feeling mine does more.

I drive it with a two stage of bc547 and bc557. Those have a maximal current
of 100 mA (200 peak). Perhaps that's not enough to saturate the MOSFET. But
I don't understand this transconductance thing of MOSFETs... BJT's are
easier...

 

[Grant]

I've determined the inductance value of the primary coil of my flyback. I
charged it up to 20V through 150R (with a squarewave). It took 30uS for the
voltage across the coil to reach about zero. With T=L/R and assuming it
takes 5 time constants to charge it, inductance is 900 uH.

By calculating the back-EMF ballast resistor with the same formula and the
15 uS it took for it to discharge, it gave me 300R, which is about right.

Now on to find the proper value for a capacitor to make a tuned circuit. And
read up on tuned circuits


The IRF840 has Ron of 0.85 Ohm. At 5 A, it would dissipate 21W. I have a
feeling mine does more.

That's fairly ordinary. But than you don't get high voltage and low Ron so
easy. Just had a look at IRF840, that's too small for this work, IMHO. IR
are very optimistic on their claimed specs, I'd be looking at much higher V
and I ratings for a MOSFET doing 5A high voltage switching.

I drive it with a two stage of bc547 and bc557. Those have a maximal current
of 100 mA (200 peak). Perhaps that's not enough to saturate the MOSFET. But
I don't understand this transconductance thing of MOSFETs... BJT's are
easier...

If you're using the common pnp + npn emitter follower for gate drive, allow
for amp or more drive current, try a pair of 1A transistors?

Grant.

 

[Wiebe Cazemier]

That's fairly ordinary. But than you don't get high voltage and low Ron
so
easy. Just had a look at IRF840, that's too small for this work, IMHO.
IR are very optimistic on their claimed specs, I'd be looking at much
higher V and I ratings for a MOSFET doing 5A high voltage switching.

The problem is, that the high current ones (like 40A) have lower voltage
rating, like 200V. And there is a problem with the IR product selector. I
can't find any >300V MOSFETs with it, even though they exist (like the
IRF840).

If you're using the common pnp + npn emitter follower for gate drive,
allow for amp or more drive current, try a pair of 1A transistors?

I'll see about adding some extra drive transistors.

 

[Jamie]

Why not cascade a bjt transitor with your fet? Maybe one of those tv
deflection types might work.

Then simply use a IGBT Transisor.
Jamie

 

[Sylvia Else]

Then simply use a IGBT Transisor.
Jamie

If the OP tries cascading, then he needs to put resistors across the
drain and source of each transistor, to balance the voltage across them
when they're turned off.

Also need to watch the drain-gate voltages.

Sylvia.

 

[Grant]

If the OP tries cascading, then he needs to put resistors across the
drain and source of each transistor, to balance the voltage across them
when they're turned off.

Also need to watch the drain-gate voltages.

Look up the switching transistor from a monitor circuit, 1500V peak, 25A
peak (800V open base, 10A continuous, BU2520). What OP is using is nowhere
near that rating. Mind you, with a peak base turn off current at max 6A,
peak base drive 9A, continuous base current 6A, these puppies wont be the
easiest to drive at speed either But they'll do it for years.

MOSFETs not necessarily the best fit.

Grant.