1. Home
  2. Electronics Design
  3. IGBT bridge, ringing, inductance -- surprise

IGBT bridge, ringing, inductance -- surprise

That's all about my IGBT inverting bridge project.

I dragged out my bigger DC power supply and tried to play with higher loads and higher inductances in series with inputs.

I used this power supply

formatting link

and two 100W, 0.2 ohm resistors in series to provide 0.4 ohm AC load.

I used currents of up to 20 amps.

In the snubber circuit, for now I used simply one 4.7 uF Cornell-Dubilier 940C capacitor, connected with 4 inch 10 gauge wire leads. No resistor yet.

Again, this is a shorting bridge that does not ever open in normal operation.

For inductor, I used a Home Depot spool of 12 gauge solid wire, my guesstimate is that there are 300 feet of wire in it. So, it made an air cored inductor.

I watched voltage across the DC rail with an oscilloscope.

Here's my finding: adding the inductor in series with input did NOT change the ratio of ringing to voltage. Since the inductor has resistance, it caused voltage drop and current drop and decrease in ringing.

Ringing was about 2x voltage both with inductior as well as without. Since current is proportional to voltage, I suppose that ringing is a linear function of current.

This is explainable and supports the hope that bridge is truly shorting. (it can be proven in another way, since the bridge conducts a little bit of current without load).

So, if my observation is correct, then, the ringing is in no way related to inductance of the input.

That is very good news. That means that the big inductor of my welder, in series with the bridge, would not be a problem.

The question is, what causes the ringing?

Now, here are some numbers. At 8 amps, the ringing was about 10 volts. That means that at 200 amps, assuming linear relationship, the ringing would be 250 volts.

That's not acceptable. I would like it to be under 150 volts. I can easily add more capacitance, someone gave me for free about eight 2 uF low ESR capacitors.

i
Reply to
Ignoramus1487
Loading thread data ...

Now I understand what is going on.

I was wrong.

When I plug the air cored inductor in series into the DC line, the ripple goes down a lot MORE than would be warranted by voltage drop and reduction in current. (I incorrectly assumed proportionality previously)

The current drops by perhaps 30% (1/3) and the ripple drops by 66% (3 times).

I finally understand why it happens.

WHAT HAPPENS WITHOUT THE INDUCTOR:

The low inductance, capacitor enhanced DC power supply replies to shorting of the bridge by momentarily increasing the current a great deal.

When the bridge finally un-shorts, that current suddenly is stopped and that causes the ripple.

WHAT HAPPENS WITH THE INDUCTOR:

Despite shorting of the bridge, the time when it is shorted is not enough to increase current, so, the snubber capacitor is able to absorb that current.

Maybe I am wrong. I will do some more experimenting tomorrow.

i
Reply to
Ignoramus1487

What happens is, when the bridge shorts, input-terminal voltage drops to approximately zero, and field starts building up in the inductance. When the bridge unshorts, this energy has to go somewhere. The common solution in switchmode power supplies is to add a highspeed diode (to draw current from the peak) and dump that current surge into a capacitor. This charges the capacitor, which is drained back to say, +V, to dissipate the excess power.

In my bridge, I have approximately 50-100A (peak design value) going through two transistors, with maybe 20-100nH inductance until the nearest snubber capacitance (if any), or the opposing transistors (since I'm using the internal diodes to handle the somewhat slower main flyback pulse, which is also accounted for in the design). Since the transistors switch off in say

100ns (I may run as long as 0.5-1us off-switching time for this reason), that's 100A switching off in 0.000 000 1 second, or dI/dt * L = V = 100V, which is actually pretty good. (A few more inches of wire, giving up to say 0.3uH, would be deadly, pushing peak + flyback voltage above avalanche voltage and making a $50 mess of magic smoke.)

Tim

-- Deep Fryer: a very philosophical monk. Website:

formatting link

Reply to
Tim Williams

Yep.

Yes. That's exactly what I did, after some screwing around, however the diode I used is a regular speed rectifier diode (simply because I had one). I ordered a ultrafast diode on DigiKey, item 497-4408-5-ND, and will use it instead when I get it in the mail.

At first, I tried a simple RC snubber, but it was dissipating too much energy and made the capacitor emit a lot of audio noise. After I did not like it, I switched to a RCD snubber, you can see it here:

formatting link

Note that that URL tells a story how I tested my bridge under full power. I got all the way up to that point. So, what I will be doing now is icing on the cake. The cake itself is already cooked.

That's very interesting. I am sorry if I missed something, but what is your power supply for?

Sounds exciting... This power electronics is a fun thing to think about.

I feel really good myself about having a working (so far) inverter. I hope that if my son goes eary to bed tonight, I would be able to weld some aluminum this weekend.

i
Reply to
Ignoramus12834

Heh, that'll happen. Electrostatic forces causing elastic deformation of the dielectric... you think that's something, you should listen to a ceramic disc some time!

RC snubbers take a bit of fidgeting. Chances are your capacitor was too large (causing the resistor to dissipate too much of the fundamental frequency), or the resistor too small (though that depends on the Thevenin equivalent of the power source!).

Yep, fun stuff, huh?

Induction heating. I have a load which is inductive as far as the square wave output is concerned, so although the bridge always has to be conducting, I have diodes (internal to the transistors) that automatically clamp the voltage to the rails, and the rails are low impedance, keeping voltage within limits.

Yes, it is. And now I'm drawing schematics to completely tubify by project. Nah, I probably won't build it, but sooner or later I'll

Yep, me too. Just need to add control circuits, then transistor drivers, then kick up the power side a notch from MOSFETs and 1.2kVA to the IGBTs and

12kVA offline source.
formatting link
(Two small strips of titanium inside my 2 x 2" coil (1uH) heated to yellow (this picture is exactly as I saw it, the color temperature is correct this time!) in under a minute.)

(OBTW, for anyone else watching my posts, here's the pile of capacitors that will become the tank capacitor. Digikey 495-1333-ND.

formatting link
More pictures and blurb coming soon as things come together.)

Tim

-- Deep Fryer: a very philosophical monk. Website:

formatting link

Reply to
Tim Williams

Yes.

The pictures are great! That pile of caps ought to be quite expensive... Keep us posted on your progress...

i
Reply to
Ignoramus18456

I agree.

What happened a couple days ago gives the phrase "cake is cooked" a wholly different meaning. :)

Looks very exciting. I have a thought to make an arc furnace...

i
Reply to
Ignoramus20878

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.