1
\$\begingroup\$

I was trying to generate a biphasic pulse for TMS application. I was referring to this website.

The waveform is given below.My circuit implementation is described as follows.

I am making a flyback converter which will convert 300 V DC to 500 V DC. This 500 V output is used to charge a capacitor. Once the capacitor is charged fully, it will be connected to an inductive load for some time using a switch.

My circuit and its output is given below. I need the biphasic pulse for only 200 us.

What did I do wrong?

enter image description here

enter image description here

enter image description here

EDIT-1

I tried to simulate the circuit provided by periblepsis.I am not getting the waveform as expected.May I know where I went wrong.

The circuit diagram and waveform is given below.

enter image description here

enter image description here

EDIT-2

The issue is solved and circuit is working as expected.

\$\endgroup\$
8
  • 2
    \$\begingroup\$ What is the graph showing? Node numbers are fairly meaningless. Label nodes and use label names in plots please. \$\endgroup\$
    – Andy aka
    Nov 4, 2023 at 19:08
  • \$\begingroup\$ Did you scroll down to see the schematic? Everything needed is there, and nothing more. Pay particular attention to the series resistors between supply and pulse capacitor; the choice of switching device (SCR); and the diodes (or other SCRs) around it. Spartan, yet comprehensive, as Sam's work always is. \$\endgroup\$ Nov 4, 2023 at 23:44
  • 1
    \$\begingroup\$ Hari, here's a simplified example that works in LTspice. Just for educational purposes. \$\endgroup\$ Nov 5, 2023 at 4:27
  • \$\begingroup\$ @periblepsis,Thank you very much.I tried to simulate the same circuit,But it is not working for me as expected.I edited the question.If you don't mind, could you please check \$\endgroup\$
    – Confused
    Nov 5, 2023 at 7:20
  • 1
    \$\begingroup\$ @Hari It's not the same circuit. A switch is not an SCR. The design requires an SCR (one way valve, so to speak, that turns off automatically when its current falls below a certain point.) So that just won't work for you. Which means you still aren't getting even the most basic ideas presented in the linked article. That fact is going to be a problem. \$\endgroup\$ Nov 5, 2023 at 7:25

2 Answers 2

1
\$\begingroup\$

Not a slightest idea how your L3 could be in accordance with what you want if L3 really has 10 henries.

You want a 200 microseconds long pulse - that's a single period of sinusoidal oscillation. The shown (and obviously the wanted) pulse is the current of a coil, it's not a voltage. To get it your resonant circuit C2L3 should have resonant frequency 1/(200us) i.e. 5 kHz. The switch should be ON exactly that 200 us (as it looks to be already). Let the simulator plot the current of L3.

But the resonant frequency in your simulated circuit is a little too low. You turn the switch OFF during the pulse and get a voltage peak I*Roff where I is the current of the coil at the turn-off moment and Roff is the OFF-state resistance of the switch. Inductor current never stops if a series switch is suddenly turned off. The inductor generates just as much voltage as needed to let the current stay ON and decay gradually.

Some hints of the practical circuit are shown in the linked source. There a triggered thyristor discharges a capacitor to a coil. The voltage of the capacitor gets inverted. That's the first half of the pulse. The thyristor turns off automatically after the current becomes zero, but it has a parallel, oppositely directed diode which allows the capacitor discharge backwards. The inductor keeps the current going and the original charging state is restored (minus resistive losses, radiation and losses caused by induced eddy currents in the material nearby the coil). That's the second half of the pulse. The situation freezes if the thyristor doesn't get a new triggering.

\$\endgroup\$
0
\$\begingroup\$

To start with, you can use just about any step-up high frequency transformer, including audio transformers, connect the low-turn-count primary to an audio amp, and drive the input of that with any desired waveform from an audio output on your PC. That's the easiest way to go from "zero to working circuit" - there's a minimum number of components you have to deal with. An off-the-shelf audio amp (10W is plenty), a transformer - speaker coupling type used with tubes will work, or a transformer from a switching mains power supply, and a connection from the PC to the amp, and a divider circuit on the transformer output to couple it to an oscilloscope.

Then replace the free-standing audio amp with a TDA7293.

Of course, you can (and should) experiment with resonant circuits and whatnot. But it really helps to start with something that is essentially guaranteed to work without trying hard. And in that case - trust experimental results, no need for SPICE.

\$\endgroup\$

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.