Ringing in a boost converter output

Status
Not open for further replies.
A

Artlav

Full Member level 2
Joined
Nov 26, 2010
Messages
144
Helped
2
Reputation
4
Reaction score
2
Trophy points
1,298
Visit site
Activity points
2,723
Hello.

I'm making a boost converter to power HID lamps from lead-acid/LiPo batteries.
The schematic is quite straightforward:


However, there is quite a lot of ringing on the output.
Converter frequency is 100KHz.
On each turn-on and turn-off of the MOSFET there is a high frequency ring.

+-2V ringing, at 20V, 2A total output into a resistor load, ~20% duty cycle:


Zoom in (1..5 us/cell, same 2V/cell), apparently out of baud for the scope or something:


At higher power it gets worse.

What could be causing it?
Can it be what makes the lamp (150W xenon short arc) flicker like TV static?
 

1.

Question: Do you still get the ringing when you disconnect the transformer labelled 'HV igniter'?

Because it is an inductor, it may contribute to the ringing.

2.

Your smoothing capacitor is a high value. It will absorb heavy current bursts. It will take a longer time to charge on power-up.

My simulation shows it can be as low as 68 uF, and do a reasonable job of smoothing. It may lessen the tendency to ring.
 

BradtheRad said:
Do you still get the ringing when you disconnect the transformer labelled 'HV igniter'?
Click to expand...
Elimination by shorting the leads together on the transformer have no measurable effect.

BradtheRad said:
Your smoothing capacitor is a high value. My simulation shows it can be as low as 68 uF, and do a reasonable job of smoothing.
Click to expand...
I had 100uF initially - it heated up like crazy and there were several V of square wave ripple, at 150W. The thing is intended to go up to about a KW peak.
So i choose the capacitor based on heat dissipation capability/ESR rather than minimal sufficient smoothing size.

- - - Updated - - -

Correction - with the transformer shorted the ringing's amplitude increase noticeably when measured at the load.
The ringing seems to be somewhere in 10s of MHz range.
 

The ringing is mostly an effect of parasitic circuit impedance and to a smaller part caused by capacitor internal inductance. I don't believe that the ringing is actually present across the capacitors, but it's surely present somewhere in your circuit and "seen" by the oscillsocope probe.

All in all it's normal behaviour of switched mode converters with unsuitable circuit layout.
 

Got rid of most of the ringing by adding an RC snubber across D-S of the MOSFET. 1000pF+5Ω.
Unfortunately, the arc is still flickering like TV static, so that wasn't it.
However, i never seen such a lamp work before, so that might even be normal.

FvM said:
All in all it's normal behaviour of switched mode converters with unsuitable circuit layout.
Click to expand...
Quite likely, since it's all bolts and wires.
 

HID lamp powered at 100 Khz?
What sort of HID? Sodium, Metal Halide, Mercury?

I've read somewhere that HID lamps cannot be driven so high in frequency.....because of arc "dancing".
The frequency is more on the order of about 400 Hz. Yes, that is Hertz.

But I could be wrong.

- - - Updated - - -

I actually went to a commercial HID ballast website, from General Electric, and the frequency is actually even lower, only from 75 to 140Hz....actual frequency depends on the lamp type..
 

You answered before I updated my message...see my reply above.

I am not familiar the characteristics of the Xenon lamp, I ignore whether a DC arc is suitable or not.
 

It's supposed to be driven with DC - small sharp cathode, fat heat dissipating anode, + and - marks, anode-up orientation, and so on and on.
 

Xenon arc lamps are DC operated. And the power supply must have a constant current characteristic because the arc discharge has a negative differential impedance. I'm not sure about the required time constant, but 6800 µF output capacitance may be a way too high to achieve stable lamp operation.

I suggest to review known working Xenon lamp power supplies.
 

Status
Not open for further replies.

Similar threads

Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…