Fusione fredda Renzo Mondaini—trascrizione

  • Arun Luthra

    It could be that the small transformer shorted out somewhere and that it is not able anymore to boost input voltage.

    In my tests, despite the small heat sink mounted on the switching transistor, that seemed to be the component that heated up the most.

    This is the main reason why I used a multimeter in series with one of the outputs to measure current—the aim wasn't measuring excess heat: according to specifications the device is only capable of providing 200 mA to the outputs continuously and I strived to keep it below this level during prolonged operation. The device has a load-limiting function, but it did not seem to be extremely effective when for example the reaction reverted to electrolysis when immersing the electrodes too much. It might also be that the very small internal multimeter resistance prevented its destruction in my case.

    From recent virtual experimentation with electronic circuit simulators, boost converters can very easily provide high peak voltage but are prone to failure. Other converter types with the outputs galvanically isolated from the inputs are more fault-tolerant (and safer) but it seems more difficult to reach high peak voltages with them—at least with a small number of cheap off-the-shelf components.

  • My boost converter came with a Ruichips RU7088R N-channel MOSFET (datasheet attached in the comment).

    It should be simple to test if it failed shorted and interesting to know if that is indeed the component which failed: https://electronics.stackexchange.com/a/36697


    Shorted gate to drain is a very common and easily tested failure mode. It'll often be a dead short or 10s of ohms. Mosfets failed in this way also tend to destroy whatever IC was driving them. When suspecting dead mosfets that's the first thing I look for.

  • In other news I just tested feeding a microwave oven transformer with a variac. The test load was four, 1 megaohm resistors (total of 4 megaohm). I had a voltmeter across one of the 1 MOhm resistors. It seemed to be fine at 400V (total 1600V) where the load current should have been 0.4 mA, and 0.16W per resistor. The 10A fuse popped in the variac as I dialed up around 500 V per resistor.

    I'm not sure what the problem was. Maybe there was some temporary short in the voltmeter (which is rated to 500V) or some issue with the variac transformer interfering with the MOT transformer. The voltmeter seems to working still. Perhaps the "absolute voltage" (rather than the voltage difference) inside the voltmeter which could be much higher than 500V was to blame.

    Edit: I think the voltmeter screen went blank or at the moment of failure so perhaps it is to blame.

    Also, there is a full bridge rectifier between the MOT and resistors.

    I am probably content to stay below 1000V with this setup...

  • A CRT or oil heater igniter flyback transformer is much safer than a MOT

  • I don't consider myself competent enough with electricity to be playing with wall voltage directly, let alone microwave oven transformers. Still, I find this "quest" for obtaining useful (i.e. more than nano/micro amps) high voltages in the order of 1–2 kV from typical DC voltages in a more accessible way instructive for my own purposes.

    While experimenting with an electronic simulator I came up with a sort of hybrid between a boost converter and a flyback converter. I'm sure I'm reinventing the wheel and very likely something along these lines already exists or has been conceived.

    In short, it uses the inductance of a small 1:1 transformer at resonance with a capacitor to boost a small voltage to high voltage like a boost converter, but also providing inherent short-circuit protection between its primary and secondary windings like a flyback converter. However, since apparently the components on the primary side too "see" the high voltages (as in a naïve boost converter) it's likely not safe to use for both the operator and the components themselves unless suitably rated for them.

    Is there a name for this?

    Link to Simulation/CircuitJS

    The switch can be used at 250 kHz, but works better at 125 kHz.

  • (new comment because the one above was "liked")

    This other one needs a high-frequency H-bridge (both a disadvantage and an advantage—respectively for greater overall complexity, but ready-made H-bridges exist) to produce the DC square wave, but appears to have the advantage that only the capacitor "sees" the high voltage produced.

    It also seems to be much more sensitive to the resonant frequency at least with the parameters used.

    Link to simulation on CircuitJS

    And a possible application here

  • With a 4 MOhm load, the MOT provides the expected ~20x voltage multiplication.

    When I reduced to 10 kOhm, there is a buzzing noise and it becomes a useless 0.90x multiplier. I guess the transformers in the variac and MOT are fighting each other?

    Next I will try wall voltage attached directly to the MOT once I have a suitable, safe arrangement for that.

    Disclaimer: MOTs are deadly, use proper high voltage safety.

  • Arun Luthra

    It sounds like there might be some form of built-in current limiter in the MOT, but I do not know well how such transformers are generally built. Besides, personally I think the way to go for these experiments is with brief intense pulses, and using directly one such transformers to power the electrodes will likely result in nice arcs, lots of heat and used current.

    See anyway for example on Quora: "What is the purpose of the magnetic shunt that is in most microwave power transformers? Can the shunt be safely removed when winding a single replacement secondary?"

  • The MOT saturates very easily, so it needs a ballast or it will probably burn. Often additional MOTs with shorted secondaries installed in series are used for that purpose.

    Edit: A few years ago while scrounging around for MOTs, two appliance repairpersons from different shops told me the same thing:

    "If you are messing around with MOTs to get sparks, you need two more things on hand at all times: Someone to give you CPR and someone to call the ambulance while CPR is going on."

  • I'm not sure what failed. It is not providing voltage, and it is drawing the max current from the DC power supply, where I set the current limit to 4.6A. Seems like something is shorted out. I think I was driving it too hard trying to maintain a continuous plasma.

    Today after I tried a few tests at maximum voltage (800V across both electrodes) I found that the DC boost device was emitting strange noises and not working as intended when about a couple hours later I wanted to make another test. That was the kind of noise it usually emits when current is being limited due to excess draw. It turns out that one of the 400V capacitors failed shorted: it's reading 0 ohms across both leads.

  • After replacing the faulty capacitor of my DC boost converter with an equivalent one I could finally use it again at 800V.

    (the brown 400V capacitor on the right)

    I tried repeating the Mondaini-style plasma discharges in a potassium electrolyte solution of about 90 g water and 9g K2CO3+a few large flakes of KOH using various cathode materials. The main difference this time was adding a 15 kOhm ballast ceramic resistor to the cathode wire. This made the reaction acquire a kind of glow or arc quality that wasn't apparent with the previous setup. Splashing does not occur here. Actual discharge voltage while the arcs (?) appear is quite likely much lower than 800V, but I have not measured this specifically.

    This video is composed of a few selected ones stitched together one after another. The first one had an AM radio running close to the camera. EMI was at times elevated enough that I could hear it on other equipment, which is also the noise that can be heard sometimes in the exact moment when arcs (?) are generated.

    Water or the electrolyte does not seem to be a requirement for these discharges. They also occur across dri(er) portions of the counter electrode above the water surface (not shown in this video). The 15 kOhm ballast resistor gets hot quickly so I couldn't run the discharges continuously.

  • Interesting patterns, again using the 15 kOhm ballast resistor (unfortunately running it for too long caused the wire attached to it to desolder; so no more tests until possibly tomorrow), but this time on a semi-dry surface. The movable electrode is the cathode, the bottom plate the anode.

    Inter-electrode voltage during the discharges was about 300–350V (x2 = 600–700V). I think this qualifies as a glow discharge rather than an arc.

    Raw notes I took:

    • I tried again using a negative screwdriver and a positive steel bracket.
    • The inter-electrode voltage is about 390V * 2 (I can only measure between one electrode and ground or the multimeter will go off scale above 600V).
    • That the discharge occurs still at a high voltage seems to point out that it is indeed a glow discharge.
    • The electrodes are dry and no significant EMI seems to be occurring (that I can hear on speakers nearby).
    • Discharges smell acidic (nitric acid).

    As for the round spots, possible similarities to this phenomenon ...?

    From https://safireproject.com/ewEx…SAFIRE-Project-Report.pdf

  • Or possibly ozone?

    It's a pungent and disagreeable smell, I did not think of ozone. Nitric acid can be produced by electric discharges in a damp atmosphere.



    [...] In 1806, Humphry Davy reported the results of extensive distilled water electrolysis experiments concluding that nitric acid was produced at the anode from dissolved atmospheric nitrogen gas. He used a high voltage battery and non-reactive electrodes and vessels such as gold electrode cones that doubled as vessels bridged by damp asbestos.[29]

    The industrial production of nitric acid from atmospheric air began in 1905 with the Birkeland–Eyde process, also known as the arc process.[30] This process is based upon the oxidation of atmospheric nitrogen by atmospheric oxygen to nitric oxide with a very high temperature electric arc. Yields of up to approximately 4–5% nitric oxide were obtained at 3000°C, and less at lower temperatures.[30][31] [...]

    I uploaded a longer video of the tests:

    The first section here shows the patterns of the above GIFs.

    These are similar tests I did just before those.

    A high-res photo of the atmospheric glow-or-arc (could actually be something inbetween; I'm not really sure).

  • This probably deserves a different thread as it's starting to get a tad bit too out of scope for this one, but the apparently regular surface spots/dots I mentioned earlier seem to appear mainly when the discharge looks more like a glow discharge rather than an arc. This has likely to do with the characteristics of the DC power supply, which does not have a stable output. Perhaps by using a higher resistance (and heavier duty) ballast resistor, constraining the discharge more on the glow region, it will be easier to see them.

    Below I'm using the flattened portion of the tip of a graphite cathode on a 1 cm steel ball in electrical contact with the anode at about 800V (open circuit voltage). Photo shutter speed was 1/250s at 50ISO.