1
\$\begingroup\$

I am working on a wireless charging project that needs a three-phase inverter. I built the inverter (with flexible wiring using connectors) myself and tested it as a single-phase full bridge converter at first. The inverter worked fine for resistive load (although there were so many noises in the waveform), but when the coil was connected, the voltage across the coil was totally distorted (not a square wave and I don't know what is that). I would like to ask what could be the possible reason(s) for this?

enter image description here Inverter test for coil (left) and for resistive load (right)

enter image description here This is the circuit of the inverter test for coil. For resistive test, the circuit was just similar to this one but with R1 being removed and L1 being replaced by resistor.

-Coil inductance = 3.9uH
-Switching frequency = 128kHz

At the moment of switching, there was no secondary coil. The coil connected to the inverter just acted like an inductor. Thank you!

\$\endgroup\$
7
  • \$\begingroup\$ You should post your complete schematics. Please also show how you have dealt with the induced voltages in the coils the moment you are switching. \$\endgroup\$
    – cx05
    Commented Jun 27, 2016 at 16:31
  • \$\begingroup\$ Its called "ringing". You normally will need a capacitor or capacitors resistor combination across that coil to dampen the ringing. Efficient wireless charging systems' primary normally employ auto tuning resonant drive, thus the RC or RLC elements must be matched. The voltage across the elements will be sine wave although they are driven by square wave driver. \$\endgroup\$
    – soosai steven
    Commented Jun 27, 2016 at 16:38
  • \$\begingroup\$ Some explanation of what each photo is may help too, and measurements of the coil inductance and DC resistance, and switching frequency. \$\endgroup\$
    – user16324
    Commented Jun 27, 2016 at 16:44
  • \$\begingroup\$ @cx05 I am sorry for missing the information, it have been/will be attached. Thank you. \$\endgroup\$
    – K. Man
    Commented Jun 27, 2016 at 17:14
  • \$\begingroup\$ @Brian Drummond I am sorry for missing the information, it have been/will be attached. Thank you. \$\endgroup\$
    – K. Man
    Commented Jun 27, 2016 at 17:14

2 Answers 2

Reset to default
0
\$\begingroup\$

The RD15HVF1 MOSFET is totally unsuited to the application and may indeed have been damaged by overheating. That device is specifically designed for VHF/UHF High power amplifiers applications.

It is not a good "switcher" and has an on-resistance of about 0.5 ohms on a good day. Driving it with Arduino outputs is not recommended either because of the gate capacitance takes quite large peaks of current every time you "switch" - this may damage the Arduino.

But it's even worse than this - the top three MOSFETs are also driven from Arduino outputs but are configured as source followers and this means the maximum positive voltage that can be driven from the source is pretty poor - maybe a couple of volts maximum. You NEED a proper driver circuit to avoid damaging your arduino and / or the MOSFETs (on load).

I'll also echo what was said about needing a capacitor from top drains to bottom sources to catch back-emfs when MOSFETs switch off.

\$\endgroup\$
6
  • \$\begingroup\$ Thank you very much! I think these are some of the problems of my design. In fact, I found that the typical input power of the MOSFETs was in the thousand watt range, but I kept using the MOSFETs since they sometimes work. Then, I forgot about this problem and it finally turned out that the performance was really dissatisfactory. I would consider searching for a better MOSFET model, and of course improving the design of my circuit with your opinions. Thank you so much! \$\endgroup\$
    – K. Man
    Commented Jul 4, 2016 at 1:56
  • \$\begingroup\$ Basic fundamental problem is that you need top side drivers for the three MOSFETs M1, M3 and M5. Gate needs to be several volts higher than source to fully turn it on therefore, if gate is only (say) 5V then source cannot be greater than a couple of volts. \$\endgroup\$
    – Andy aka
    Commented Jul 4, 2016 at 9:25
  • \$\begingroup\$ Thank you very much! I just reviewed on some information about MOSFETs and discovered this, which I have neglected before. Thank you so much for your advice. \$\endgroup\$
    – K. Man
    Commented Jul 4, 2016 at 16:08
  • \$\begingroup\$ The convention for "thanks" is to upvote the answer or for the OP (you) to accept the answer formally. \$\endgroup\$
    – Andy aka
    Commented Jul 4, 2016 at 18:37
  • \$\begingroup\$ In fact, I have done it before but it seemed that my reputation was not enough. I am really sorry that I may need to do it later. \$\endgroup\$
    – K. Man
    Commented Jul 7, 2016 at 7:25
0
\$\begingroup\$

To run this circuit properly you will need a high side driver for M1, M3 and M5. The gate voltages must be higher than the source, so they must be higher than the V1 rail in order to turn on properly. Many manufacturers make chips you can use for this purpose.

An app note to explain: AN-6076

\$\endgroup\$
3
  • \$\begingroup\$ Thank you very much for your answer! But I don't understand why the gate voltage needs to be higher than the source. I just found that in my circuit, when one of the MOSFETs in a leg is turned off before the other one is turned on, the middle connection of the two MOSFETs seem to be at unknown potential. Is this the reason? Or is it because of the design of the MOSFETs? Thank you. \$\endgroup\$
    – K. Man
    Commented Jul 4, 2016 at 1:46
  • \$\begingroup\$ Anyway, thank you so much for your opinions and information! I would find and include a suitable driver to improve the performance. \$\endgroup\$
    – K. Man
    Commented Jul 4, 2016 at 2:00
  • \$\begingroup\$ I just reviewed on some information about MOSFETs and figured out this point. I have forgot about this before. I am sorry for not double checking it before asking and thank you again for your help! \$\endgroup\$
    – K. Man
    Commented Jul 4, 2016 at 16:11

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.