Simple High Voltage Converter for Geiger Tubes and other uses

[madexp]

I'd like to present a simple high voltage converter primarily designed for powering Geiger tubes, though it can be adapted for different voltages and purposes.



The circuit is based on the Resonant Royer topology, which is a self-oscillating converter known for its simplicity and efficiency. It uses two transistors operating in push-pull configuration with a feedback winding to maintain oscillation.

Circuit Description
The design uses a pair of MPSA42 transistors, though common alternatives like BC547 or 2N2222 can be used. The 2.2k Ohm resistor provides base bias for both transistors and determines the current draw. This value should be selected by monitoring both output voltage and input current - if the resistance is too high, the output voltage will drop. The optimal point is reached when you achieve the desired output voltage with minimal input current. In my implementation, the circuit draws 40mA at 5.7V input.
The resonant frequency is determined by the LC parallel circuit formed by the transformer's primary inductance and the 100nF capacitor connected between the collectors.



Transformer Design
The transformer design isn't based on complex calculations but rather on empirical rules that I've found to work reliably:
Primary Winding

Always use 10+10 turns
Wire gauge should be selected based on required current capacity
Current can be calculated from input voltage and output power requirements

Feedback Winding
Turns ratio varies with input voltage:

For 5V input: 10 turns
For 9V input: 6-8 turns
For 12V input: 4-5 turns

The feedback turns are crucial: too many turns risk damaging the transistors with excessive base drive, while too few turns result in incomplete switching, causing higher power consumption and poor conversion efficiency.

Secondary Winding
The secondary turns calculation follows this logic:

Being a sinusoidal oscillator, the input waveform has a peak-to-peak amplitude of twice the supply voltage (due to the push-pull configuration)
Calculate the turns ratio based on required output voltage




Example: For 400V output from 10V input:

Ratio needed = 400:10 = 40:1
With 10 turns on primary, secondary needs 40 × 10 = 400 turns



Note: While the primary actually has 20 turns total (10+10), calculating based on just 10 turns empirically works. This might be because the actual transformation ratio depends on the inductance ratio between secondary and primary rather than turns ratio, but I haven't investigated this aspect further.

The beauty of this design lies in its simplicity and adaptability - it can be modified for different voltage requirements while maintaining good efficiency.

[opengeiger.de]

Thanks for presenting this interesting type of HV supply circuitry! When the Z-diode DZ1 is not attached to the output what voltage would be reached at C5 in this configuration?

What would be very interesting is the averaged power consumption of the circuit including DZ1. Did you measure it?

[madexp]

Hello! Oh actually it's a terrible circuit! It's terribly good if you don't care about current but you have no time/parts to do something better. It can be built from really any kind of scrap part! You can even wind primary windings around an axial 3.3mH choke and it will act as a transformer!
It's kind-a kalashnikov of converter!

That zener acts as a crude voltage stabilizer, draining current when Vout > Vz.
Without zener, with 5.6k Ohm bias, 5.7V @ 40mA input the diode bridge outputs 386V.
Without zener, with 2.2k Ohm bias, 5.7V @ 70mA (including 9mA status led) output is 400V neat.
With zener I get 396V without current increase.

If you need to increase efficency, you can connect the zener anode to the base of a transistor 2N2222 with a 10k series resistor and connect the 2N2222 collector to the bias point of the oscillator (2.2k/base/winding point). That way the circuit shut-off when Vout > Vz.

I've obtained, with carefully tweaking of windings/bias/transistors, to make such kind of circuit working from just 2-2.4V and 20mA.

[opengeiger.de]

Thanks for this info! One last question: when you load the circuit e.g. with 1Gohm, does the circuit have enough "steam" to provide the additional current at 400V or will the voltage drop?

[madexp]

So far, I have tested the circuit with a 4.7 MΩ load (consisting of a 10 MΩ resistor and a 10 MΩ voltmeter in parallel), and the output remains stable at 396V. The circuit can be adjusted to power PMTs with a low-value base divider through some modifications to the circuit and windings.

Based on theoretical calculations, using a 400-turn secondary, 5.7V input, 40 mA current, and an EFD25 core with N87 ferrite, the output should ideally reach 228V on the secondary with a short-circuit current capacity of 1 mA. After rectification, this would result in approximately 321V DC. However, in practice, the observed voltage is higher, and the current capacity is lower than these theoretical values.

Additional Details:
-This design was developed to create a simple and safe-to-use high-voltage (HV) converter for a school project. The circuit is intentionally designed with low efficiency (around 10%) to ensure safety, preventing potential harm to students while still providing sufficient HV to power two Geiger tubes.
-The input voltage is set to 5.7V instead of 5V, as the output with 5V yields around 330V. To adjust the input voltage, I added a diode in series with the ground pin of the 7805 regulator.
-The capacitor in parallel with the primary results in an oscillation frequency of only 5 kHz. For optimal power efficiency, N87 ferrite requires a working frequency of around 50 kHz.

Improving Circuit Efficiency (and Increasing Potential Danger):
-Increase the magnetic flux in the core by raising the primary turns to at least 40+40 turns.
-Adjust the circuit to operate in the 50-100 kHz frequency range by selecting an appropriate capacitor for the primary.
-Increase the secondary winding turns by at least 25%; for example, use 2000 turns when employing a 40+40 primary configuration.
-Implement a feedback control loop to regulate the output voltage by toggling the oscillation on and off according to the load.

In previous projects, I used these adjustments to drive heavier loads, resulting in significantly higher output voltage and current—which could be dangerous if touched. If you're interested, I can provide a detailed explanation of the theory and non-empirical calculations behind this circuit and present a higher-efficiency, higher-power prototype.

[Radioquant98]

Hallo madexp,

danke für die Ergebnisse deiner Experimente.

Ich weiß jetzt nicht, ob Du meinen kleinen Experimentier-Royer kennst. Das hatte ich hier vorgestellt.
https://www.geigerzaehlerforum.de/index.php/topic,2126.0.html

Zur Zeit baue ich für mein Gammaspektrometer ein kleines Kästchen mit einem Royer für 850V, Impulsanpassung und Akkustromversorgung nebst Ladekontrolle für 2...3 parallele 18650 Li-Ion-Akkus.

Auch ich habe den Royer nicht berechnet. Ich habe mich nur an Beispielen im Internet orientiert und paar Probewicklungen gemacht.
https://www.mikrocontroller.net/articles/Royer_Converter
http://www.sophia-electronica.com/Baxandall1959JM.pdf

Noch eine Anmerkung zu den MPSA42, die Du verwendest. Der ist eigentlich sehr ungünstig, da er als Hochvolttransistor nur geringe Stromverstärkung und eine sehr hohe Uce-Sättigungsspannung hat.
Wenn dein Aufbau noch steht, dann probiere mal normale Transistoren mit höherer Stromverstärkung und spezielle Low-Sat-Transistoren. Bei Letzteren sollte der Wirkungsgrad besser werden - besonders bei kleinen Eingangsspannungen, wie einer Li-Ionzelle mit 3...4,2V.

Noch etwas: Für Zählrohre verwende ich lieber Sperrwandler, da deren Ruhestrom wesentlich geringer ist. Beim  Auslösezählrohr ist der Störabstand viel größer als bei PMT.
Beim PMT nehme ich lieber den Royer, da der Störabstand sehr gering ist und nicht so viel Aufwand mit der Entstörung ist.

Viele Grüße
Bernd

[madexp]

I read the thread, and the circuit has a very nice and clean design. Congratulations! Regarding the transistors, yes, the MPSA42 are not the most efficient choice, but I personally like them a lot because in the past, when I accidentally touched both the HV section and the driver section with my hand, I broke the transistors many times. The MPSA42 are very robust and rarely break as they are designed to handle high voltage. Of course, this is not the norm, but when experimenting, in addition to using components I have on hand, I try to take into account, as much as possible, the errors that can occur during experimentation.

To improve efficiency, I also used MOSFETs in the past and was very satisfied with them. As for the choice of the converter, for Geiger tubes, there is usually no need to pay special attention to the output ripple of the converter. I use coupling on the anode, and a few tens of pF capacitance couple the negative pulses generated by the discharge in the Geiger tube effectively, even ignoring up to 5V of bias ripple without issue. This, at least, is my practical experience and is not intended to be considered good practice.