Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Capacitor Charger Questions

  1. Jan 13, 2010 #1
    With all courtesy to Mr William Nechefski for this circuitry.

    I have gone ahead to build this circuit. Everything is the same except the transformer.
    I had a old transformer with was used to step down from my 50hz mains.

    Primary 12-012
    Secondary 0-230V
    Power 12VA

    My transformer has a average ratio of 1:10 turns,
    The self made transformer is 50/2000 turns, means a ratio of 1:40 turns
    I played around with the frequency using the trimmer and got around 550-600V pulsed output. 1:40 supposedly gets 2000V output, so my 1:10 gets 550-600V. Pretty straightforward.

    Things turned out pretty well for a good 30 secs and the output dropped to 80V.

    A transformer designed for 50hz would work for this type of application?
    It seems the output voltage dropped due to the saturation of the transformer.
    If i switched it on for a longer time, i think the tx would eventually burn out due to the high frequency oscillations.

    Even when i reduced it around 2.5khz from 20khz, it was eventually the same.

    HOW to overcome the saturation problem then? If that is the problem.


    Attached Files:

  2. jcsd
  3. Jan 13, 2010 #2


    User Avatar
    Science Advisor
    Gold Member
    2017 Award

    Saturation would occur each cycle - not after 30s.
    I suggest that a spike in the waveform may have caused arcing in the secondary winding and you now have shorted turns. Or, as the frequency is too high for the iron core, the input reactance was so low that the driver circuit has burned out due to too much current.
  4. Jan 16, 2010 #3
    Hey, u r right! I have burnt 2 MOSFETS up to now. Smoking hot.

    My dc power supply is limited to 1A output but the MOSFETS still smokes.

    Right now, i changed to resistors to get a different frequency.
    The SAME transformer can now safely bring the output to a stable pulsed 600Vdc after rectification.

    Frequency is at 10kHz, really beats me man. When i reduce or increase the frequency even just a bit. The output drops to the same 80V.

    As what the author mentions, the output is truly dependant on the secondary windings or should we say transformer. But he should have added the frequency plays a big part.
    Most step up transformers can't take the high frequency anyway.

    2 manufacturers have rejected to make a 1:40 ratio transformer for me, citing their iron can't take the high frequency. Guess i gota make a new one by myself then.
  5. Jan 16, 2010 #4


    User Avatar
    Science Advisor
    Gold Member
    2017 Award

    I think the reason it works at just one frequency could be a resonance. A lowish value of L can resonate with a capacitor in parallel to increase the input impedance. This capacitance could be in the construction of the txformer. If you put another C across the txformer primary you may find it will lower the frequency of resonance. Perhaps 0.1 or 0.01 uF or less? I can't estimate the self L of the primary so you'd have to experiment. I think that tuning transformers is quite a common technique at highish frequencies.
  6. Jan 16, 2010 #5


    User Avatar
    Science Advisor
    Gold Member
    2017 Award

    The orig author was presumably using a ferrite core and the frequency dependence might not be so marked.
    Ferrite transformers are available or just cores. Not too pricy!
  7. Jan 16, 2010 #6
    Making 1 is not pricey . . . ferrite cores are readily available at the local Farnell or RS Component distributors. The insulated copper wire is the killer, local suppliers sell at 40kg minimum. I need 2 sizes, gona blast a hole in my pocket to get them.

    I will just get it from any tx manufacturer locally. IF they wana make one for me.

    I have not thought of meddling with the primary at all. GOOD reminder!
    I'll try it and let you know ya.
  8. Jan 16, 2010 #7


    User Avatar
    Science Advisor

    Radio Shack have this :
    * 40 feet of 22-gauge wire
    * 75 feet of 26-gauge wire
    * 200 feet of 30-gauge wire
    as a fixed deal for US$6.59

    I wonder about those diodes, too. 1N4007 are medium speed recovery diodes.
    Maybe you should be looking for some faster diodes.
    The UF4007 is a 1000 V PIV 1 amp diode with a fast recovery time.
  9. Jan 17, 2010 #8
    No no . . . Radio shack is unavailable in my country.
    For me, RS or Farnell would be great choices locally.

    I also had qualms about 1N4007. But they are working fine for now.
    I have not got 2000V yet, so problems may surface when i do get that output voltage.

    Anyway, a faster diode helps in what way? Able to rectify at same frequency of the output?
    I wonder if using slower diodes actually makes the pulsed dc output lag with the output of the tx....
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook