PWM Ringing at high Freq

Any idea how to remove ringing when PWMing? PWMing my laser diode in a low side driver configuration. Meaning my laser diode is connected to voltage+resistor in series and then the switch. (Done so I can avoid the almighty evil of a Pfet or the complexity of using a N channel mosfet as a high side driver). PWMing at about 700khz... could go higher but the ringing gets worse. Heres my scope output from my rigol DS1052SE.


CH1 = voltage across diode.



Second image:

CH1 is the gate of the mosfet. CH2 is the + terminal of the diode with respect to ground.



Any ideas? My source is a DC battery with an extra 20mF of capacitance added in parallel to help filter out any noise. Cabling is pretty short so it shouldn't be inductive kick back. Perhaps I should try a low side driver configuration? Ringing is less apparent at lower frequencies but i'm trying to build a high feq driver.

Basic schematic representation. To lazy to rewdraw... Pretend thats a nchannel mosfet and frequency source says 700khz...

What is the measured diode current during the on cycle?

I wonder if the oscillator is causing it.

CH2 in image 2 shows a tiny bit of wiggle at the end of the on cycle across the gate. That small voltage change will result in a large dV at the diode, which will result in a huge dI through the diode.

Maybe you should add a little capacitance to the gate, with a cap in parallel.

Maybe start with a 2nF cap and go up from there?

For such high frequencies, I wonder if a little more complexity is worth the trouble.

Maybe try a crystal oscillator source?

Is your current signal source a simple oscillator? Or is it something partial to square wave outputs like a (very) high speed 555?

How about a picture of the setup....
In that Schematic I don't see any filtering at all...
In the 2nd SS of the scope... the signal driving the
gate of the mosfet is pretty dirty.... IMO...
You know what they say... Garbage in... Garbage out..

I would start with a clean input signal and work from there.
BTW... a 20uf cap across a battery is kinda small... unless
your battery can't easily supply the current required by
your circuit.. In that case I would use a bigger battery instead
of a small 20uF cap.


Jerry

What you have there is ringing as denoted by the dampend sine wave when the signal goes thru transition. These are cause by stray RC, RL, and LC tank circuts. Without a complete schematic and pictures of your setup it is hard to tell exactly where the problem lays. The reason it get worse with an increase in frequency is that the small values in stray circuts have more of an effect as the frequency is increased.

This may help: http://en.wikipedia.org/wiki/Ringing_(signal)

Also some understanding of capacitive reactence, and inductive reactence in relation to
frequencies may help too.

LasersBee, as stated in the original post I was using 20mF of capacitance which is a decent amount. Was using overkiill so people wouldn't just tell me to add capacitance which tends to be all fix for many people.

After further consideration of the circuit i'm pretty sure one of the killing factors is the frequency. At 700khz+ speed you need a pretty darn good driver in order to be able to drive a mosfet. Issue is mosfet gates are actually small capacitors and that fact is not normally an issue. However, if your trying to drive it at insane speeds, that cap acts as a short to source and you need a driver that can provide .5A + of inrush current in order to be able to charge the gate in time (not something many ICs can do).
Oddly enough though, just the addition of the diode seemed to reak hell. Heres an example of my output across a purely resistive load.


Heres an example of the frequency source without connection to the gate.




For those who were wondering... heres the circuit. Note this is not current limited. This circuit is designed for high frequencies. You cannot current limit at 1Mhz... although its pretty easy to do at around 20khz. 1 pot adj frequency, 1 adj duty cycle, the pullup pot on the output of the op amps needs to be adjusted based on the current draw.



^^ mistake, cntrl pin is connected to a cap and then gnd.

I wonder what the capacitance on the LD is?

Roxx said:

I wonder what the capacitance on the LD is?

Click to expand...

Good question! I have no idea. I'm also curious as to how much the inductance of my long wires to the laser diode affect the circuit as well. Every small bit of capacitance and inductance is massively critical at high switching speeds. Sadly, I'm just a student now. Hopefully when I take optics, they cover the finer points of a light emitting diode model.