Starting off of..

I am trying to build one of these using an air-core coil. I got it oscillating but have to cut power off after 4 seconds as it starts to heat up (can't hold it with my fingers.) The transistor is not on a heat sink.

I have searched everywhere (probably not everywhere, but extensively) and could not find answers to how to get it working without so much heat dissipating and a using MOSFET.

I started this project only to understand how it works using (cheap) oscilloscope and obviously safer voltages than what this could produce. Secondary as calculated is producing a voltage of 150V. The ocilloscope has a limit of 400V.

As per the schematic above, I am using a power supply which is rated 16V 625mA. When I checked current going to the circuit (using just an ordinary DMM) it is actually 2.38A. I don't know where this is coming from.

When the circuit oscillates I am getting 25V on gate-source if not more. Reading other comments on other threads, for full turn on of this MOSFET I need to have at around 10V on the gate even if thr MOSFET turns on at 5V and is capable of 20A.

When I connect a resistor between the power supply and the circuit to lower current and voltage, the resistor goes black in 2 mins and the circuit doesn't perform as expected.

What are other options to lower the voltage without dissipating extra voltage as heat? Say 13V?

The secondary circuit is open and no oscillation (or at least not intended oscillation) is happening on secondary. Is this required to get it working with lower power consumption?

Back EMF from L1: There are suggestions to use a Zener diode to suppress this back EMF. A suggestion to utilize a back EMF snubber is mentioned here.

Figure 16. Regenerative snubber:

Regenerative Snubber

What would be reasonable values for this using a 12-16V power supply? Is Lx a simple choke?

Second picture I found here.

Blocking Oscillator

This option appears better as feedback is provided separate from oscillation. It doesn't protect the base from back EMF from L1, but a snubber could be used as stated above. Is this a simpler solution?

I don't have capacitance on source. Is that required? What value would be reasonable value/type?

I have tried to build a simple Joule thief circuit as per MeTube videos and folks are getting amperage at 10mA using 1V battery. I've built one and could not get it lower than 35mA. Different topic, but I guess this is similar on a larger scale. Could it be lacking oscillation?

In article above "Wireless Energy Transfer" they claim 0.8A maximum drawn from the power supply and mine sucks 2.4A.

Parts used in this circuit were at hand. The other power supply I could use is a car battery I have at hand, but I can't imagine what power could be drawn from that using just 6 turns on primary L1. It could melt the speaker wire I used.

The IRFZ44N again was available at hand with great values of fast switching, high amperage capability and 50V should be sufficient if used correctly. Hoping if I could go over limits of normal transistor like 2N3904 or 2N2222 I should be fine, however it seems more complicated.

What I am expecting is power consumption below 1 ampere The lower, the better, since I'm not trying to transfer this power.

Since this circuit has been discussed from 2013, I hope someone would be able to provide a solution/idea how to improve this using MOSFET.


Idea, but unsure if it is a viable setup:

enter image description here

According to this video here , I need Snubber on primary L1.

Now with this setup I don't see voltage going above 8V on the gate and if it does, that should be regulated by Zener diode which connects Gate to ground.
With Snubber explained in that video, I should be fine with 4.7nF capacitor.
LED should light up if it is oscillating.

--- Final Edit ---

After a month digging through internet and trying to find answer to my own question, since very low contribution to this thread, I found out, that my primary was too close to secondary. It's short answer. Long answer won't be posted here.

Circuit redesigned from scratch.

Thanks to everyone who contributed to this thread and posted views and possible solutions.

127 views and 27 days since post created.

  • \$\begingroup\$ The circuit you have taken as “ready to use” is an idea, a conceptual representation. It’s on the implementer to add the bits that make it work, like gate biasing and removal of DC from the primary winding. A lot of oscillator circuits are conceptual like this: they are used to explain an idea and assume working knowledge of electronics. They are more “notes to those skilled” than “circuits to copy to a breadboard”. And sometimes people who post them just lie about the circuit ever working. They assume they did something wrong, not that the circuit has no chance of working as is. \$\endgroup\$ Commented Apr 8, 2022 at 17:31
  • \$\begingroup\$ You’d best find a YouTube video of someone running the circuit and explaining it, from a reputable source since there is all too many “electronics tutorials” videos on the web that are completely fake – just copied information and circuits that aren’t shown in a form that works, or just barely work, etc. \$\endgroup\$ Commented Apr 8, 2022 at 17:35
  • \$\begingroup\$ @Kubahasn'tforgottenMonica Thanks for your input. If you know about such YouTube video, it would be much appreciated rather than skimming through 1000s of them only to find out, that doesn't answer my question. I have posted here my problem with hope that someone who is building these (or similar), have knowledge to point out how it works and how can I solve that heating problem. \$\endgroup\$
    – Old Navy
    Commented Apr 8, 2022 at 17:43
  • \$\begingroup\$ besides, most of YouTube videos are showing just results and not how they encounter problems during their build of this and that circuit. \$\endgroup\$
    – Old Navy
    Commented Apr 8, 2022 at 17:46
  • \$\begingroup\$ Excellent example here: [link] (youtu.be/AehL3BodduE). Person showing that his transistor burned, and solution was to put five of them in parallel??? Not an answer I'm after. \$\endgroup\$
    – Old Navy
    Commented Apr 8, 2022 at 18:05

1 Answer 1


Here is a version with a MOSFET.
Two cases for the "protection" of VGS voltage. Note the current drawing from the voltage source.

Made with microcap v12.

enter image description here

enter image description here

  • \$\begingroup\$ I only assume that current is limited by that resistor on the Source of MOSFET and Zener Diodes are for trimming amplitude to keep it within range. Don't see much difference when one Zener D is shorted. \$\endgroup\$
    – Old Navy
    Commented Apr 8, 2022 at 16:13
  • \$\begingroup\$ The resistor is for "limiting" current a little. The two Zener are needed for limiting the voltage across g-s to 12 V (simple caution). But one can see that this limits also the average supply current by a factor of 2 (for C=10 nF, 97 mA vs 45 mA). The output voltage is also higher (134 V vs 156V). \$\endgroup\$
    – Antonio51
    Commented Apr 8, 2022 at 20:09
  • \$\begingroup\$ I think, that overheating is caused by back EMF rising above PSU voltage. Would you be able to simulate my last post? On secondary (L2) can connect capacitor and resistor like you have done in your post. \$\endgroup\$
    – Old Navy
    Commented Apr 8, 2022 at 21:32
  • \$\begingroup\$ Ok. Will try it. \$\endgroup\$
    – Antonio51
    Commented Apr 9, 2022 at 5:59
  • \$\begingroup\$ @Antonio51, I found very interesting your simulation results. (1) Which software you have used for this simulation and results? (2) I understood you used new L3 to provide feedback. But that is significantly different than original layout using just L1 & L2 windings; is this later unfeasible or problematic with MOSFET? \$\endgroup\$
    – EJE
    Commented Apr 10, 2022 at 12:09

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.