Driving a CRT monitor flyback transformer with squarewave

[Wiebe Cazemier]

Hi,

I'm working on designing a flyback transformer driver and I have a few
problems. It's been a while since I've been in sci.electronics.* and I've
come to ask for the expertise here again

My driver is an Atmega 168 mictrocontroler, driving a IRF840 (500V 8A)
MOSFET through two stages of BC547/BC557 transistors. It drives the flyback
in pure flyback mode (so no resonant circuits with capacitors and such).

My problem is dealing with the back-EMF. Right now, I have 150R 15W
resistance with a UF4007 75 ns diode. With this, not only does it create an
enormous amount of current in the back-EMF ballast when the secondary
winding is open, it also creates high voltage spikes, often destroying my
500V 8A MOSFET. And, I feel that it is severly limiting the output of the
secondary winding. My squarewave is from 0 to 50V 90% dutycycle, yet
breakdown distance of the transformer's output is about 1.5 cm. When the
flyback was still in the CRT, it was about 2 to 2.5 cm, and that is with a
sawtooth drive signal, putting less energy in the core than a squarewave.

In an old style CRT TV, the flyback driver circuit does not just generate
high voltage by collapsing the magnetic field, but it creates a resonance
circuit with the retrace timing capacitor (also called safety capacitor).
This prevents having large back-EMF problems.

My first question is: how does a multiscan CRT monitor (any not-very-old VGA
and better) drive its flyback transformer? These things scan at 30-150 kHz,
so it's kind of difficult to make a resonant circuit with a capacitor,
because that would only work at one frequency.

My second question is dependent on the answer to the first, I guess, namely:
how do I handle the back-EMF efficiently? Can I do it like a CRT monitor
does? I'm going to try 'transorbs' (I've ordered some). These are basically
very fast bipolar zener diodes, designed to suppress transients. Perhaps if
I use a 400V version to shunt the back-EMF (meaning limit it to 400 volts),
there will still be enough energy to give my decent secondary winding
output. I don't know whether I still need resistance in the back-EMF path,
though.

Any insight is welcome

 

[Wiebe Cazemier]

reduce the duty cycle to 50% max.

Most people building these things, mostly based on 555 timers, use
dutycycles of 90%. With 50%, it's impossible to get any kind of spark out of
it.

 

[Wiebe Cazemier]

The flyback isn't used to drive the yoke in multiple 'H' frequency
monitors.

But the flyback is driven by the horizontal output stage, so the flyback
itself operates at different frequencies. So, what driving method is used
and how is back-EMF handled?

 

[petrus bitbyter]

Hi,

I'm working on designing a flyback transformer driver and I have a few
problems. It's been a while since I've been in sci.electronics.* and I've
come to ask for the expertise here again

My driver is an Atmega 168 mictrocontroler, driving a IRF840 (500V 8A)
MOSFET through two stages of BC547/BC557 transistors. It drives the
flyback
in pure flyback mode (so no resonant circuits with capacitors and such).

My problem is dealing with the back-EMF. Right now, I have 150R 15W
resistance with a UF4007 75 ns diode. With this, not only does it create
an
enormous amount of current in the back-EMF ballast when the secondary
winding is open, it also creates high voltage spikes, often destroying my
500V 8A MOSFET. And, I feel that it is severly limiting the output of the
secondary winding. My squarewave is from 0 to 50V 90% dutycycle, yet
breakdown distance of the transformer's output is about 1.5 cm. When the
flyback was still in the CRT, it was about 2 to 2.5 cm, and that is with a
sawtooth drive signal, putting less energy in the core than a squarewave.

In an old style CRT TV, the flyback driver circuit does not just generate
high voltage by collapsing the magnetic field, but it creates a resonance
circuit with the retrace timing capacitor (also called safety capacitor).
This prevents having large back-EMF problems.

My first question is: how does a multiscan CRT monitor (any not-very-old
VGA
and better) drive its flyback transformer? These things scan at 30-150
kHz,
so it's kind of difficult to make a resonant circuit with a capacitor,
because that would only work at one frequency.

My second question is dependent on the answer to the first, I guess,
namely:
how do I handle the back-EMF efficiently? Can I do it like a CRT monitor
does? I'm going to try 'transorbs' (I've ordered some). These are
basically
very fast bipolar zener diodes, designed to suppress transients. Perhaps
if
I use a 400V version to shunt the back-EMF (meaning limit it to 400
volts),
there will still be enough energy to give my decent secondary winding
output. I don't know whether I still need resistance in the back-EMF path,
though.

Any insight is welcome

In the old CRT TVs/monitors the LOPT is not driven by a sawtouth, it
produces the sawtouth current for the horizontol yoke. This is mainly done
by tuning the (complex) load so the current in the yoke raises lineair
during the ramp. At the end of the ramp, the primary supply is switched off
by a short pulse to retrace the line. The collapsing field produces a
back-emf which is used to produce the high voltage, roughly 10kV to 25kV.
Though the current provided is small, there's a real load. Often some other,
less high voltages are also produced by the LOPT causing a real load.
Together they prevent the back-emf from rising indefinitely. Nevertheless,
when facing a failing CRT, the line output transistor is a primary suspect.

To protect your switching transistor, you can make an artificial load that
only comes in when the collector voltage becomes too high. See below using a
fixed font.

--->|-----+----+-----+
fast | | |
high | | -
voltage | .-. / zener diode
diode --- | | - p.e. 150V
--- | | A
| '-' -
| | |
| | |
+----++----+
| | |
| | |
| | -
| .-. /
--- | | -
--- | | A
| '-' -
| | |
| | |
+----++----+
| | |
| | |
|
repeat as often
as required.

created by Andy´s ASCII-Circuit v1.24.140803 Beta www.tech-chat.de

petrus bitbyter

 

[Sylvia Else]

Hi,

I'm working on designing a flyback transformer driver and I have a few
problems. It's been a while since I've been in sci.electronics.* and I've
come to ask for the expertise here again

My driver is an Atmega 168 mictrocontroler, driving a IRF840 (500V 8A)
MOSFET through two stages of BC547/BC557 transistors. It drives the flyback
in pure flyback mode (so no resonant circuits with capacitors and such).

My problem is dealing with the back-EMF. Right now, I have 150R 15W
resistance with a UF4007 75 ns diode. With this, not only does it create an
enormous amount of current in the back-EMF ballast when the secondary
winding is open, it also creates high voltage spikes, often destroying my
500V 8A MOSFET. And, I feel that it is severly limiting the output of the
secondary winding. My squarewave is from 0 to 50V 90% dutycycle, yet
breakdown distance of the transformer's output is about 1.5 cm. When the
flyback was still in the CRT, it was about 2 to 2.5 cm, and that is with a
sawtooth drive signal, putting less energy in the core than a squarewave.

In an old style CRT TV, the flyback driver circuit does not just generate
high voltage by collapsing the magnetic field, but it creates a resonance
circuit with the retrace timing capacitor (also called safety capacitor).
This prevents having large back-EMF problems.

My first question is: how does a multiscan CRT monitor (any not-very-old VGA
and better) drive its flyback transformer? These things scan at 30-150 kHz,
so it's kind of difficult to make a resonant circuit with a capacitor,
because that would only work at one frequency.

My second question is dependent on the answer to the first, I guess, namely:
how do I handle the back-EMF efficiently? Can I do it like a CRT monitor
does? I'm going to try 'transorbs' (I've ordered some). These are basically
very fast bipolar zener diodes, designed to suppress transients. Perhaps if
I use a 400V version to shunt the back-EMF (meaning limit it to 400 volts),
there will still be enough energy to give my decent secondary winding
output. I don't know whether I still need resistance in the back-EMF path,
though.

Any insight is welcome

It's not clear to me what you're trying to achieve, but if by "decent
output" you mean a high voltage, I think you have a fundamental problem.
In an ideal transformer, all the flux links both windings, and the ratio
of the emfs in the two windings has, at all times, to equal the turn
ratio. So you cannot get a high voltage on the secondary without having
that same voltage divided by the turns ratio on the primary. Real
transformers, being less than ideal, will simply make the situation worse.

With the ballast resistor in place, and the secondary open, the voltage
is going to rise initially to a level equal to the current that was
passing through the primary immediately before switching multiplied by
the ballast resistance. In your case, if you were actually driving 8A,
then the voltage would rise to 1200V, and it would be no surprise if
that killed your transistor.

Putting in a zener diode to protect the transistor just has the effect
of throwing away some of the energy that was stored in the magnetic
field in the transformer, which might as well not have been stored there
in the first place (i.e. by using a lower current).

Sylvia.

 

[Jamie]

Michael A. Terrell wrote:




But the flyback is driven by the horizontal output stage, so the flyback
itself operates at different frequencies. So, what driving method is used
and how is back-EMF handled?

Saw tooth wave with a damping component? (Diode)?

Jamie

 

[Grant]

They are separate circuits in modern monitors, so the flyback
operates at a fixed frequency. Other than that, it's a traditional
TV/monitor HV supply.

Ya think? Not in the dozen or so monitors I dismantled in the last six months,
fair bit of MOSFET power switched capacitors around that HO section to tune
it to selected rate. Some monitors you can hear relays click when they go high
speed. But the ones I had were about a decade old.

I saved some of the triplers, maybe build a Jacob's ladder or so.

Grant.

 

[Wiebe Cazemier]

It's not clear to me what you're trying to achieve, but if by "decent
output" you mean a high voltage, I think you have a fundamental problem.
In an ideal transformer, all the flux links both windings, and the ratio
of the emfs in the two windings has, at all times, to equal the turn
ratio. So you cannot get a high voltage on the secondary without having
that same voltage divided by the turns ratio on the primary. Real
transformers, being less than ideal, will simply make the situation worse.

With the ballast resistor in place, and the secondary open, the voltage
is going to rise initially to a level equal to the current that was
passing through the primary immediately before switching multiplied by
the ballast resistance. In your case, if you were actually driving 8A,
then the voltage would rise to 1200V, and it would be no surprise if
that killed your transistor.

Putting in a zener diode to protect the transistor just has the effect
of throwing away some of the energy that was stored in the magnetic
field in the transformer, which might as well not have been stored there
in the first place (i.e. by using a lower current).

Sylvia.

I'm trying to achieve about 20 kV at as much current I can get out of the
flyback. In the device that I'm going to power, there will be a spark gap,
but not to protect, but to give me sudden discharge of high voltage
capacitors (long story).

Anyway, at 20V, it draws about 5 amps at 90% duty cycle. So with 150R
ballast, that gives me 750V.

I hadn't really though about what you're saying about throwing away energy;
I've been increasing the voltage on the primary winding to get higher output
voltage, but I might as well increase the back-EMF ballast to reduce the
amount of energy I throw away.

More testing is indicated

 

[Wiebe Cazemier]

Ya think? Not in the dozen or so monitors I dismantled in the last six

months, fair bit of MOSFET power switched capacitors around that HO
section to tune
it to selected rate. Some monitors you can hear relays click when they go
high
speed. But the ones I had were about a decade old.

I saved some of the triplers, maybe build a Jacob's ladder or so.

Grant.

I'm looking at one of those CRT's right now, that clicks when you switch
some resolutions.

Anyway, could you elaborate on these MOSFET power switched capacitors? Do
you mean that modern CRT's do have circuitry to achieve a resonance?

 

[Sylvia Else]

I'm trying to achieve about 20 kV at as much current I can get out of the
flyback. In the device that I'm going to power, there will be a spark gap,
but not to protect, but to give me sudden discharge of high voltage
capacitors (long story).

Anyway, at 20V, it draws about 5 amps at 90% duty cycle. So with 150R
ballast, that gives me 750V.

I hadn't really though about what you're saying about throwing away energy;
I've been increasing the voltage on the primary winding to get higher output
voltage, but I might as well increase the back-EMF ballast to reduce the
amount of energy I throw away.

If I'm visualising your circuit correctly, that will push up the voltage
that the transistor sees, and 750V is already way above your
transistor's rating.

Perhaps you should look at voltage multipliers.

Sylvia.

 

[Wiebe Cazemier]

If I'm visualising your circuit correctly, that will push up the voltage
that the transistor sees, and 750V is already way above your
transistor's rating.

Perhaps you should look at voltage multipliers.

Sylvia.

For more information about my schematic, see my reply in another thread. It
was a reply to whit3rd's message, who replied without a reference, so the
message didn't appear in this thread.

 

[Sylvia Else]

For more information about my schematic, see my reply in another thread. It
was a reply to whit3rd's message, who replied without a reference, so the
message didn't appear in this thread.

http://www.electronics-lab.com/projects/misc/001/index.html

OK. That's pretty much what I thought.

With that circuit, it seems to me that the maximum output voltage is the
transistor's voltage rating multiplied by the turns ratio of the
transformer.

To get any higher, you need a way of preventing the transistor from
seeing higher voltages from the primary *without sinking current from
the primary*, which is something of an ask.

Sylvia.

 

[Jamie]

http://www.electronics-lab.com/projects/misc/001/index.html

OK. That's pretty much what I thought.

With that circuit, it seems to me that the maximum output voltage is the
transistor's voltage rating multiplied by the turns ratio of the
transformer.

To get any higher, you need a way of preventing the transistor from
seeing higher voltages from the primary *without sinking current from
the primary*, which is something of an ask.

Sylvia.

This maybe true if the input to the coil was a sine wave or some sort of
slow ramp down after the coil is magnetized. Using the square wave the
signal is being turned off immediately and thus the field is allowed to
collapse abruptly. This as you know, is going to generate more voltage
than what you see in the turn ratio.

We have plasma generators at work used to treat the surface of wire
for printing and the coil in those are being driven via a saw tooth,
where output is regulated by monitoring current in the primary side of
the coil. We set current limit to what we need for correct output.

So when viewing this on the scope it may look like a trapezoid of
sorts at times. But the important part is when the driver drops to zero
is when we get our high voltage we need to generate the plasma.

Jamie.




Jamie

 

[Sylvia Else]

This maybe true if the input to the coil was a sine wave or some sort of
slow ramp down after the coil is magnetized. Using the square wave the
signal is being turned off immediately and thus the field is allowed to
collapse abruptly. This as you know, is going to generate more voltage
than what you see in the turn ratio.

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.

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.

We have plasma generators at work used to treat the surface of wire for
printing and the coil in those are being driven via a saw tooth, where
output is regulated by monitoring current in the primary side of the
coil. We set current limit to what we need for correct output.

So when viewing this on the scope it may look like a trapezoid of sorts
at times. But the important part is when the driver drops to zero is
when we get our high voltage we need to generate the plasma.

Jamie.

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.

Sylvia.

 

[Grant]

I'm looking at one of those CRT's right now, that clicks when you switch
some resolutions.

Anyway, could you elaborate on these MOSFET power switched capacitors? Do
you mean that modern CRT's do have circuitry to achieve a resonance?

I've only looked at the PCB, not the circuit, so while I see some MOSFET
switched caps, I don't know what they're connected to, but one of the MOSFETs
is mounted on a heatsink. No photos and nothing here at the moment.

Got an old monitor in the car, it's 1995 though, may check it later.

Grant.

 

[Warren]

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.

Warren

 

[Spehro Pefhany]

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.

Warren

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.

 

[Warren]

Spehro Pefhany expounded in

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.

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

Warren.

 

[Spehro Pefhany]

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.

Warren.

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.

 

[Warren]

Spehro Pefhany expounded in

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.

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