DC 20kV flyback converter schematic in multisim

[HelloCthulhu]

I found this schematic while watching the following video on building a DC 20kv power supply:



I wanted to test the circuit in multisim, but I'm having trouble finishing it. None of the transformers available look like the one used in this schematic. Below are the transformers available on multisim, but none of them worked above millivolts. I included one of my last attempts. Any help is greatly appreciated!

 

[Baluncore]

The circuit is for an NPN BJT that conducts base current.
You are modelling it with a 2N7000 MOSFET that does not conduct base current.

 

[alan123hk]

but none of them worked above millivolts

What does "none of them work above millivolts" mean?

Rated voltage of the simulation model ?
Oscillation signal amplitude observed in simulation result ?
1 millivolt = 0.001V and only the order of millivolt ?

 

[HelloCthulhu]

The circuit is for an NPN BJT that conducts base current.
You are modelling it with a 2N7000 MOSFET that does not conduct base current.

Thank you for quick the response! That definitely got the circuit running, but the max voltage and current are much lower than I thought they'd be (1.26kv, 1.65mA). Can't figure out what else I'm doing wrong.

 

[Baluncore]

How have you set up the transformer parameters?
What turn counts or ratios have you set up for the transformer windings?
What primary inductance?

 

[HelloCthulhu]

I couldn't find any information on the turn ratio of a computer flyback transformer, so I used the turn ratio of an ignition coil: primary at 135 and the secondary at 11000.

 

[Baluncore]

Notice there is no capacitor to set the frequency. The frequency of operation is therefore determined by the inductance of the primary, the supply voltage, and any stray capacitance.
What inductance have you specified for the transformer primary winding ?

 

[HelloCthulhu]

What inductance have you specified for the transformer primary winding ?

I'm very new to transistors and transformers, so I'm not sure how to calculate the inductance for this circuit. I'm working on it now.

 

[Baluncore]

I'm very new to transistors and transformers, so I'm not sure how to calculate the inductance for this circuit. I'm working on it now.

With Multisym I believe you specify turn counts and the primary inductance. With SPICE we specify the inductance of each winding so it can calculate the turns ratio.

You used; Primary n=135. Secondary, n=11,000. Feedback, n=5.
I don't believe they are correct for an 18 V to 20 kV transformer.
The primary has too many turns.

I would have started with about 1 or 2 volts per turn, here let's say 2 volts / turn.
Primary V=18 V; n=9; n² = 81;
Secondary; 20 kV; n=10,000; n² = 100e6;
Feedback, V=5 V; n=3; n² = 9;
I computed n² also since inductance for SPICE is proportional to the turns count squared.
Use these n values for the turn counts on your transformer windings.

I guess the magnetising current in the primary will be about 20mA.
18Vdc ≈ 6.5 Vrms.
XL = 6.5 Vrms / 20 mA = 325 ohms, reactive.
XL = L · 2 · π · freq; ∴ L = XL / ( 2·π·f )
Guess at a frequency, maybe about 5 kHz.
An estimate of the primary inductance; L = 325 ohms / ( 2 · π · 5 kHz ) = 0.010 = 10 mH.

With those values, my LTspice model produces 25 kV pulses at 4.85 kHz.

 

[HelloCthulhu]

Thank you so much for your help on this. Other than the issue with the number of turns, I figured out that the schematic connections for the transistor as I understood them (which is probably wrong since the transformers don't look the same) was incorrect. After connecting the bottom of the top transformer (which I've always thought of as the primary) to the base of the transistor and the bottom of the secondary to the collector I finally reached 26kV in simulation. I greatly appreciate all of your help with this!

 

[Tom.G]

The below is Information I dug out of US Patent #3,665,288 dated May 23, 1972. As such, it applies to a vacuum tube circuit, rather than to a transistor circuit. I would expect the same physics applies to a transistor circuit, but would require a higher primary-secondary turns ratio.

A flyback TV transformer takes advantage of the rapid collapse of the magnetic field when the Horizontal output tube is abruptly turned off at the end of a sweep.
The retrace time is only around 10-11uS, or 200kHZ. The transformer is often tuned to resonant at the 3rd or 5th harmonic. This ultimately yields a voltage step-up of around 100, even with a pri-to-sec turns ratio of 2 or 3 for a vacuum tube circuit.

So it seems you would want your circuit to be a blocking oscillator, that is a circuit that hard switches the primary Off after the primary current reaches maximum.

Here is a table from the patent showing some transformer flyback transformer info (remember, it's for a tube circuit). 'Tertiary' in the table is the high-voltage winding.
--

--

Here is a link to the Patent Office filing. I find the patents easier to read if you click on the "Image" button on the page. This gives you a .PDF version, better formatting and easier to navigate.

http://patft.uspto.gov/netacgi/nph-...50&s1=3665288.PN.&OS=PN/3665288&RS=PN/3665288

Cheers,
Tom

 

[Baluncore]

After connecting the bottom of the top transformer (which I've always thought of as the primary) to the base of the transistor and the bottom of the secondary to the collector I finally reached 26kV in simulation.

You still have some way to go. You have swapped the base and collector connections.

The primary of the transformer must be connected between the positive supply rail and the collector of the transistor. There should not be a resistor in that collector circuit.

The two resistors make a base bias circuit. The midpoint must be connected through the feedback winding to the base of the transistor.

Change your turn counts to those I calculated earlier.

 

[Tom.G]

--
https://components101.com/sites/default/files/inline-images/BJT-Symbol.png
--
Hmm... that was supposed to be an image of a transistor symbol with the terminals labelled, try clicking on it.

 

[HelloCthulhu]

You still have some way to go. You have swapped the base and collector connections.

The primary of the transformer must be connected between the positive supply rail and the collector of the transistor. There should not be a resistor in that collector circuit.

The two resistors make a base bias circuit. The midpoint must be connected through the feedback winding to the base of the transistor.

Change your turn counts to those I calculated earlier.

The turns were adjusted during the 26kV sim, but I've made the necessary changes to the transistor connections. Hope they're correct. New simulation just under 20kV.

 

[Baluncore]

That now looks good.

 

[shjacks45]

Note the gain of power high voltage bipolar transistors is very low, and impedance will be low. Even with 100V CEO transistor it'd be add a snubber circuit to protect the collector. (I blew out a 600 V Triac in one inverter). For a more mundane 12VDC to 110VAC inverter always push-pull, and add a small emitter resistor to improve stability. It will get hot with a few amps. Rem heat sink. 26000 V x .1 mA output would be 5+ watts in the transistor.