I am attempting to drive high-current (~1000 amps) at high frequency (10~30 kHz) through a coil. I am not an electrical engineer but a student working in a small lab where we have to build everything we need ourselves (which is pretty fun), so I may not understand some basic things that might be obvious to the EE community. Prior to several months ago, I had only worked with voltages and resistors (in year one of undergrad).

My attempt: attach a capacitor in series with the coil so that it would have a resonant frequency and then drive that load with an H-bridge, supplied by a large capacitor bank. For testing, I am currently using two 400amp and two 600amp IGBTs that we already had for the switches, but will eventually upgrade to 1200amp IGBTs if I get everything working properly. These IGBTs are controlled by gate drivers that are optically isolated from their inputs coming from an FPGA. I have configured the FPGA to create 4 square waves, 2 of which are in-phase with each other and that are out of phase with the other two. This will be a pulsed system, and a trigger signal to the FPGA controls when the system fires.

The H-bridge is basically as below. In reality, there are currently two inductors, the main coil and a second to adjust (with the capacitor) the resonant frequency. The signals from the gate drivers, which are floating, are applied across each IGBTs gate and emitter. I have not included my control circuits to avoid unnecessary clutter; they work exactly as I intended.

enter image description here

Using Rogowski coils I measure the current in the two branches of the H-bridge, through the coil, and I also check the signal being sent to one gate for reference. If I charge the capacitor bank to a low voltage the current waveform looks perfect. It creates a nice sinusoidal waveform just like I want but the current is low .

However, when I increase the voltage of the capacitor bank source and drive more current (> 5amps) the current waveform becomes worse and worse.enter image description here

What is the reason I am not getting a sinusoidal current through my coil at max currents greater than a few amps? I need an amplitude of 1000amps, not 2! Am I misunderstanding something fundamental or doing something silly?

I realize this is a long post. Thank you for your time in reading and considering it!

  • 2
    \$\begingroup\$ Are your inductors air core? \$\endgroup\$ – WhatRoughBeast Oct 28 '16 at 23:22
  • \$\begingroup\$ You can see the resonant frequency is increasing with the higher current : in a half cycle of the gate waveform you're seeing nearly 3/4 cycle of the resonant waveform (a short half cycle and a negative going, i.e. the start of the opposite half cycle. This suggests the L or the C is decreasing : decreasing C is unlikely, but L will decrease if the inductor core saturates. At which point we need to know almost everything there is to know about the inductor : value, construction, core dimensions, core material etc. But mainly, what saturation current did you design it for? \$\endgroup\$ – Brian Drummond Oct 28 '16 at 23:44
  • \$\begingroup\$ My instinctual thought was to wonder if an air core inductor was being used. My quantitative thought was to marvel at the sheer level of \$\frac{\textrm{d}I}{\textrm{d}t}\$ here. \$\endgroup\$ – jonk Oct 29 '16 at 11:21

The vertical slope of the BH loop is proportional to inductance and higher impedance. Notice that have DC voltage offset thru the coil, where the average DC current rises by integration over time until it reaches the B limit or Bsat value and then L inductances drops to zero if very flat peaks and your AC waveform is now driving near zero ohms so high current peaks. Make sure the drive is balanced as the name suggests in "Alternating Current"

enter image description here

Something is obviously wrong with the design of the coil drive with DC offset or burden load resistor value. In order to detect 1000A peaks there must be no DC thru the windings.

Ensure there is no DC and load Burden Resistor is correctly chosen to ensure it is properly loaded.

Also read about Remanence and degaussing due to abrupt current shutoff in a transformer which also causes saturation. This is done in the lab with a Variac.

BTW your signal integrity is very poor on "The Good" side. This needs significant review and improvement with more details than you have offered.

| improve this answer | |

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.