0
\$\begingroup\$

I want to transfer power via induction.

I have done this successfully multiple times with multiple designs. For various reasons I am steering clear from more traditional methods, for one, so that I could potentially have a lot more current flow through my induction coil (which is just a single hollow copper loop) reliably.

Using a crystal oscillator and a volt regulator a have gotten a Flip Flip to perfectly oscillate Q and notQ to switch on the gates of my fast switching Logic Level MOSFETs. I have confirmed the Flip Flop outputs with an oscilloscope. I smoothed the power supply (10V) with a 2200 uF electrolytic Cap and added a parallel tuning cap for the LC tank. My circuit is working and transfering power wirelessly.

However, I am not opperating at the efficiency I expected. Previous models of mine that only use one MOSFET and one output of the flip flop are even performing much better. This design however is "off" half of the time, when the volts over the gate are zero. The MOSFET has to block a lot of current, gets really hot and prevents current flow at the level I would like.

My logic tells me that my design belowe should be on almost 100% of the time. First allowing current to flow top to bottom on my schematic, then vice versa. The flip flop first puts on the two top MOSFETs with Q, and then the bottom two with notQ.

I have built this design and it does work, but not as well as I hoped. I believe there must be some noise or feedback in the system preventing it from smoothly oscillating. Any help how I could potentially fix this problem would be of great help. Many thanks. enter image description here

\$\endgroup\$
  • \$\begingroup\$ hm, is the same power supply driving the flip flop? \$\endgroup\$ – Marcus Müller Jan 15 '17 at 10:16
  • \$\begingroup\$ and, is the flip flop definitely able to quickly charge and discharge the gate capacities? \$\endgroup\$ – Marcus Müller Jan 15 '17 at 10:17
  • 5
    \$\begingroup\$ Yo have 2 low-side switches and 2 high side switches, all N type MOSFETs. But there's no evidence that you're taking the proper measures to drive the high side switches properly. Without the proper gate voltage ( i.e. 10V above the source) when ON) you'll waste power and heat them. Think what those source and gate voltages should be ... then read all about high side drivers. \$\endgroup\$ – Brian Drummond Jan 15 '17 at 10:29
  • \$\begingroup\$ Marcus, yes, same power supply is driving the flip flop, but indirectly via a voltage regulator and then 5v crystal oscillator. \$\endgroup\$ – Marcel Hattingh Jan 15 '17 at 11:50
  • 2
    \$\begingroup\$ Ignore that "logic level", that's marketing mungo jumbo. How are you producing a voltage above the rail voltage for your top side Nfets to give them positive Vgs? \$\endgroup\$ – winny Jan 15 '17 at 12:39
2
\$\begingroup\$

Two problems. If you took away the inductor you are charging a capacitor and then discharging it. That is just 100% wasted energy given off in heat. The second problem with your H bridge is that the high side N channel MOSFETs are source followers and will never switch on to a low on resistance: -

enter image description here

Shoot through might also be a big problem. Gate charge and discharge times due to the gate-source capacitance will also mean inefficient switching and this will burn energy. A lot of MOSFET drivers are capable of dumping about an amp into the gate to turn it on sub 100 ns; a regular flip=flop probably can't produce more than 20 mA.

Efficient H bridge design is never accomplished with a simple circuit. Try looking up MOSFET ZVS circuits for an increased efficiency: -

enter image description here

There are only two MOSFETs and the power feed to the coil (via the inductor at the top of the circuit) allows the coil and capacitor to freely resonate and hence efficiency is much better. If you don't like the split centre tapped coil you can use two inductor feeds from the positive supply to each MOSFET drain.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Thanks Andy for an great answer. The large cap over the power supply is just to smooth the DC, as I am picking up some small noise. Do you think I could be fine without? \$\endgroup\$ – Marcel Hattingh Jan 15 '17 at 12:00
  • \$\begingroup\$ I'm talking about the capacitor across the coil. \$\endgroup\$ – Andy aka Jan 15 '17 at 12:02
  • \$\begingroup\$ Also, thanks for the MOSFET driver idea. Will definitely give that a test. Do you think more amps (and maybe volts) could help drive my IRL540 MOSFETs? \$\endgroup\$ – Marcel Hattingh Jan 15 '17 at 12:02
  • \$\begingroup\$ Ahh, I see, I'm doing that to tune my LC circuit, do I need it? \$\endgroup\$ – Marcel Hattingh Jan 15 '17 at 12:04
  • \$\begingroup\$ You can't effectively drive a parallel tuned circuit with a H bridge. With coil and capacitor in series you can but watch out for extremely high currents due to the Q of coil. You can also drive a parallel LC via an extra series C from a push pull output but not a H bridge as is your driver. \$\endgroup\$ – Andy aka Jan 15 '17 at 12:16
0
\$\begingroup\$

Inductance in the drain circuit is key to getting Reasonable efficiency at 500KHz .The circuit that Andy provided will do ZVs .You could adapt The circuit to your full bridge proposal .The full bridge will make more power so it could be worth the effort .Your Flip Flop I guess is driven from a sig gen at say 1MHz ? Remember that tuning is critical for low switching losses .This tuning precision is augmented by high tank Q which is good for low loss inductive power transfer .Self Oscillating schemes when designed properly are fundamentally self tuning .If you use a gen then in production it should be locked to the LC tank so you will need some sort of phase detector to form a PLL .Now you can see that the self Osc scheme is much simpler.

| improve this answer | |
\$\endgroup\$
  • \$\begingroup\$ Autistic :) you guessed correctly. 1 meg oscillator. I was hoping that if I chose the appropriate capacitor, based on the inductance of the inductor, I could make the LC circuit resonant frequency 500KHz. It sounds like you're saying that the LC circuit resonant frequency is dependant only of the indoctor itself and the cap is only used to tune it to that frequency. Am I understanding you correctly? \$\endgroup\$ – Marcel Hattingh Jan 15 '17 at 12:16
  • \$\begingroup\$ Even then, could I use a 555 to potentially create the exact frequency required? \$\endgroup\$ – Marcel Hattingh Jan 15 '17 at 12:18
  • \$\begingroup\$ The choke inductor is usually chosen to have high impedence a twice the resonant frequency .In yoiur case it must have high impedence at 1 MHz .The transformer inductance is resonant with the capacitance. \$\endgroup\$ – Autistic Jan 16 '17 at 10:47

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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