# Why do single transistor oscillators work to power a transformer, if a transformer needs AC?

I have been making some high voltage projects lately, such as a spark gap Tesla coil, slayer exciter, and a flyback transformer HV supply.

The flyback transformer uses a single transistor circuit, just like the slayer exciter circuit.

I know that solid state Tesla coils use a pair of transistors in order to create an AC signal to power the primary.

How can only 1 transistor create an AC signal to power the primary of a slayer exciter or flyback transformer?

• Could you please provide references from where you got these images if they are pulled from the web? Mar 16, 2021 at 18:31
• AC and DC can be very vague terms that mean different things to different people. More clear would be to think about signals that only consist of a 0Hz component, those that consist of only components greater than 0Hz, and those that consist of >0Hz components and a moderate 0Hz component, and those that consist of >0Hz component and a very large 0Hz component. Mar 16, 2021 at 21:09
• The presence of a DC current through a winding doesn't change the operation of a transformer in principle. The DC component just doesn't get "transformed" to another winding. However, in practice, such a DC current magnetizes the core of the transformer, so the presence of such a current has to be factored into the construction, in order to avoid saturating the core. The usual countermeasure is a gap in the core. You'd need to read about transformer construction for the gory details. Suffice to say that the transformer transforms only AC, but can tolerate some DC while doing it.
– sh-
Mar 16, 2021 at 22:01
• @sh- I like to think of DC as being gum in your stomach as your body transforms food to poop. Mar 17, 2021 at 18:01

Transistors can "generate AC", in fact with the right filters and switching you can get transistors to generate sinewaves (this is how most DC/AC inverters work).

In today's modern power supplies transistors are switched on and off rapidly to create changing current that can be passed through transformers. Many switched mode power supplies (SMPS) do this to convert AC/DC or vice versa.

This is typically done with some kind of feedback or control loop. The control loop either excites the voltage (like in the top diagram) or controls the voltage to a certain level (like when we want a constant 5V out of a 5V power supply). The transistor acts as an amplifier to provide positive feedback. In SMPS an analog or digital control loop is used to provide negative feedback.

• so a signal created by a single transistor that oscillates from 0V to 5V is AC? If so, how come solid state tesla coils use 2 transistors to form a signal that oscillates from -V to +V, such as the one in the image? Mar 16, 2021 at 16:25
• Anything that is not "steady" or 0Hz is considered AC in most contexts. AC is also associated with AC mains power and 60Hz. The circuit in the image uses two transistors in bi-stable mode to continually switch the current from positive to negative Mar 16, 2021 at 18:32
• @VoltageSpike: I think you're missing the point of the question. The original poster already knows (as shown in his 3rd schematic image) that "two transistors" can "switch the current from positive to negative"; his question appears to be more about the "single transistor" circuits in his first 2 schematic images. Mar 16, 2021 at 20:04
• It doesn't change the thrust of the answer, but surely most inverters are the cheaper "modified sine wave" variety, indeed probably the rather nasty hi-zero-low-zero waveform rather than one with more steps. Mar 17, 2021 at 13:33
• @davidcary You can generate a sine wave (not a very good one, or low impedance) with a single transistor. The answer was written with the first two schematics in mind, and less so of the third one Mar 17, 2021 at 15:39

Transformers don't require AC to operate. They require a changing current. This can be a changing DC current just as easily as an AC current.

• what is "changing current" other than "alternating current", if the integral over the current needs to be zero, lest you plan to build a saturated electromagnet instead of a transformer? Mar 16, 2021 at 16:04
• I agree, if it's not 0Hz (DC), then it's probably AC Mar 16, 2021 at 16:05
• @Marcus, "Is switched DC = AC" has been asked before. My thoughts are that if the current direction does not alternate then it's DC. Of course a varying DC can be considered as AC with a DC offset. Mar 16, 2021 at 16:14
• @Joren, technically no. The current direction is always in the same direction, varying in amplitude. I know that for analysis it can be considered as AC on DC. Would you consider a current that never changes direction as alternating? Mar 16, 2021 at 16:30
• @VoltageSpike Do you say that the output of a bridge rectifier is AC? How do you rectify your AC definition with the definition of a rectifier? Mar 17, 2021 at 4:48

how can only 1 transistor create an AC signal to power the primary

There are two things happening here.

1. If you have an "AC" signal on top of a constant "DC" offset:

The constant DC part will go straight through the transformer coil and create a constant magnetic field. This does not induce any voltage in the secondary. All it does is saturate the core a bit and increase losses.

However the AC part will still cause a varying magnetic field, which will induce a voltage on the transformer secondary.

In other words, transformers only work on AC but can ignore some amount of DC offset current on the primary. The transformer core still has a maximum field limit above which it will saturate, and this depends on instantaneous current, thus having a DC current on top of your AC current reduces the maximum AC current the transformer can handle before its core saturates.

1. It's a flyback transformer

When the transistor turns on, current rises in the primary. Since this is a transformer, this causes a corresponding rise in voltage on the secondary, but in a flyback transformer this part of the secondary waveform is not used. The voltage rises, but no current goes into the load, usually because there is a rectifying diode in the output.

When the transistor turns off, the primary is disconnected. But the magnetic energy in the core has not disappeared, which means current has to find somewhere to flow. If this was just an inductor, voltage would rise until something breaks and lets the current through, for example the transistor would avalanche. However this is not an inductor, but a transformer instead, it has two coils, so voltage rises very quickly on both primary and secondary until something breaks. On a flyback transformer, this is used to obtain high voltage on the secondary. The turns ratio of the transformer is chosen so magnetic energy stored in the core discharges through the secondary coil into the load before the transistor avalanches. Once the energy is dumped in the load, magnetic flux is back to zero and the cycle can restart.

So, in a flyback transformer, having DC offset on your AC is an intentional part of the design, it does not create problems like it would on a "normal" transformer. The goal is to ramp up primary current to store energy in the core, then dump it in the load via the secondary.

As long as the transistor stays on, the current and magnetic flux will keep increasing. This can't go on forever of course, first copper losses are in i^2 so they increase dramatically at high current, second at some value of flux the core will saturate. So, input voltage, ON-time, primary winding turns and copper gauge, and of course the core itself, should all be chosen together for best efficiency and lowest cost. Since a core with a higher saturation flux will be bigger and more expensive, peak primary current also influences cost.

I know that solid state tesla coils use a pair of transistors in order to create an AC signal to power the primary, so how can only 1 transistor create an AC signal to power the primary of a slayer exciter or flyback transformer?

If you want to more efficiently drive a tesla coil, then you would use a push-pull arrangement of transistors and "hard drive" the transistors as switches. This minimizes transistor losses and makes a more power efficient solution. However, you have to watch out for "killer" back emfs but, that's another story.

If you used a single transistor (as in your first circuit) it is likely that the transistor will operate in class A and be handling signals that are more akin to sinewaves. Because the transistor operates in class A, there will always be a standing collector current in the transformer primary and, the voltage on the primary will have an average value of zero because it will rise above the power rail as much as it falls below the power rail.

An example I have to hand is a Colpitts oscillator that uses an inductor in the collector but, this could equally be a transformer in the collector: -

Pictures from this Q and A.

As you can see, Vout (blue trace) rises equally above and below the 5 volt rail hence, the transistor remains in class A and, the average voltage across the inductor is largely zero. So, even though the supply is DC, the actual input voltage across the "primary" is virtually pure AC and, of course, with any inductor or transformer it has to be to avoid saturation.

These circuits use positive AC feedback to oscillate by pulsing DC into AC to generate the AC output. The 3rd coil on the primary side is a reliable method. The other methods rely on the flyback of proper polarity of magnetic and parasitic coupling to create positive feedback with gain >1.

DC positive feedback merely makes it a latch. but AC positive feedback with delay and proper polarity makes it an oscillator.

In order to make a great Tesla coil, Tesla did not use any transistors. because they weren't invented yet. (unlike @Andyaka 's)

So how did he maximize dV?

By minimizing dt into a small inductance and reasonably primary high arc voltage to create a low capacitance 2nd stage transformer winding with low turns ratio <100 but a primary step transformer with the highest step up ratio possible so the inductance won't affect the result using the purest 99.9% best insulation, Mica to couple the trigger to the 2ndary transformer and flyback the biggest > 1 MV arcs.

It is hard to simulate with Falstad's limited 2kB trace memory as aliasing occurs when the rise time is faster than 1/2 the sampling rate yet try to fit the repeated arc interval in one trace. But it looks something like this. Slow down the simulation just before it reaches 10.0 kV the trigger threshold of the gas tube I chose to trigger the arc. This will produce > 120 kV. Anything is possible in theory, but practise has physical limits of interwinding capacitance and L/R ratio of a 1kW high inductance primary transformer (huge).

## result

V = L dI/dt as dt goes to zero you get the biggest arc voltage but with low L for low open circuit stage 2 secondary capacitance at some resonant frequency. .= LdI/V using an arc gap switch with the smallest dt arc rise time in xxx picoseconds.

HV arcs required a rapid negative resistance to collapse in voltage in picoseconds to nanoseconds to make a UHV flyback coil to generate an ultra-high voltage from the follow-on change in current.

The arc density determines how low the ESR becomes and how quickly the arc extinguishes to determine how fast dI/dt is the result. So by using a gas arc gap, you get far faster rise times than trying to force the current turn-off with transistor capacitance.

It is ultimately the lower capacitance and Q of the RLC resonance of the gas arc extinguishing that results in the hi V. But along with high Q is an envelope rise time so the peak voltage is not always on the 1st cycle but after the risetime of envelope as the primary discharge cap decays with a low negative ESR and the Tesla Coil DCR that affects the result.

Shown with 1ns sampling resolution you can see the rise of the output envelope with a reasonable L/R ratio of 32 uH/0.1 Ohm for Litz wire in a large air core and using a donut termination to reduce e-field gradient.

Now 175kV

Don't try this at home.

Your 3 versions have different operating principles. Also the transformer is used differently. The 1st and 2nd circuit use the transformer both to rise the voltage and to make feedback in the oscillator. Both circuit use a pulse oscillator principle where the transistor at first conducts, but gets suddenly closed to unconductive (=no Ic) state. The conduction starts again quite soon. How soon - that depends on many things, even parasitic capacitances affect. Well predictable (lower) operating frequencies can be got by inserting an explicit RC-time constant circuit instead relying on parasitic capacitances.

But the essential idea in getting high voltages in circuits 1 and 2 is to charge magnetic energy to the coil with growing DC current through the primary. Oscillator feedback is only a side function. When the primary current suddenly stops an inductive kickback pulse is generated. It's a circular electric field around the core in the transformer and so strong that at least one of the windings pushes current to somewhere. The voltage simply jumps as high as needed to allow gradual magnetic field decaying. Stepwise magnetic field change is as impossible as stopping a moving mass immediately from full speed to zero. It doesn't happen, every braking allows more or less motion because infinite forces do not exist. Read this old case for reference: https://electronics.stackexchange.com/questions/282053/how-does-the-inductor-really-induce-voltage?r=SearchResults&s=1|63.8725

The transistor stays hopefully unconductive at least as long as it's is needed to dissipate all of the magnetic energy stored in the transformer. Otherwise the current gradually cumulates and there's a short circuit.

version 2 differs from version 1 in one important area. In ver. 2 the transformer is loosely coupled. The secondary can well operate as resonant circuit (C=the stray capacitance) and oscillate some time after the inductive kickback. That's the same idea as in Tesla coil.

The 3rd circuit is different. Its output transformer is also loosely coupled and rings substantially after every input pulse. The mosfets feed alternating polarity pulses (=AC) to the primary. The primary isn't used to generate inductive kickbacks the transformer rises voltage more like a normal transformer. But the ringing due the resonance is essential to make the same high frequency content as ver 2.

A single transistor forward drive of a transformer with DC pulses. DC pulses work as well if there's an auxiliary flyback or "DC-reset" circuit which returns the magnetic energy from the transformer back to the power supply and thus prevents DC current cumulating and developing a short circuit.

Most several hundred watt power supplies I have seen in PCs use this principle which feeds the primary with DC pulses. The output in the secondary is in use at the same time as the transistor in the primary circuit conducts. The next Wikipedia image presents the principle:

The auxiliary winding returns the inductive kickback pulse to the DC input supply through D3.

For more examples search for "forward DC to DC converters". In flyback converters the output gets a new pulse when the primary circuit switch stops conducting. In forward converters the output gets a new pulse when the primary circuit switch is ON. Turning the switch off generates a pulse which is returned to the input DC voltage source.

how can only 1 transistor create an AC signal to power the primary of a slayer exciter or flyback transformer?

Its an AC signal + a DC bias. A transformer can have some DC bias without affecting its output, provided the core doesn't saturate.