High voltage power inverter IDLE current question

[szoftveres]

Hi All,

I'm trying to build a HV power supply to drive a scope CRT. I am using an arrangement very similar to this one:



R1, R2 = 330 ohm
R3, R4 = 68 ohm
Q1, Q2 = BD139
Transformer: Würth Elektronik 750871010 https://www.we-online.de/katalog/datasheet/750871010.pdf

I need the symmetrical arrangement (as opposed to the single transistor joule-thief -style) because I have to drive a multiplier from the secondary to bring the voltage up to around 600V.

My problem is that even thought the circuit works, it's extremely inefficient and draws 0.5A at 5V (2.5W) even when the secondary is unloaded. The transistors are getting really hot. The 2.5W power consumption slightly increases as I'm loading down the circuit. With approximately 0.6W load (3mA at 200V) the circuit is still drawing approximately 3W, so the efficiency is around 20%.

I've got a hunch that since the transformer is unloaded, the flyback effect causes 60V+ voltage peaks on the primary side (collector), thus breaking down the C-B junction in the transistors and thereby keeping them open for a longer period.

I didn't try, but I'm assuming that with appropriate load the efficiency would eventually increase beyond 50%. However for my use case, I'd like this circuit to draw as little current as possible when IDLE-ing.

How can I achieve that?

Thanks!

[David Hess]

I need the symmetrical arrangement (as opposed to the single transistor joule-thief -style) because I have to drive a multiplier from the secondary to bring the voltage up to around 600V.

Nothing prevents the single transistor blocking oscillator style from driving a multiplier, and this is how it was most commonly done in the past.  A high voltage capacitor is used in series with the output of the inverter to achieve this.

I didn't try, but I'm assuming that with appropriate load the efficiency would eventually increase beyond 50%. However for my use case, I'd like this circuit to draw as little current as possible when IDLE-ing.

How can I achieve that?

I would first try adjusting the resistor dividers to provide less base drive to the transistors, or add a resistor in series with the emitters.  Oscilloscopes use a negative feedback loop to control the base drive so that the circuit barely oscillates but this requires a separate winding.

Another alternative is to drive the primary side with a current instead of directly connecting it to the input supply.

[szoftveres]

Nothing prevents the single transistor blocking oscillator style from driving a multiplier, and this is how it was most commonly done in the past.  A high voltage capacitor is used in series with the output of the inverter to achieve this.

Could you show me a working example?

[David Hess]

There is nothing special about it.  The common voltage multiplier circuits are AC coupled at their input through a capacitor so any source can drive them and any DC bias is removed.  Even a single inductor or transformer flyback circuit can directly drive a voltage multiplier.  (1) There is no requirement to have a floating transformer winding.

https://en.wikipedia.org/wiki/Voltage_multiplier

The only reason that a transformer is commonly used is to take advantage of the step-up from the winding ratio to start at a higher voltage, and to separate the high voltage winding from the primary.

(1) https://www.arrow.com/en/research-and-events/articles/lt8365-high-voltage-micropower-boost