Why don't the clamping diodes clamp everything? / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Why don't the clamping diodes clamp everything?

1 2 next

Move Thread

LAN_403

IamSmooth

Mon Dec 21 2009, 01:36AM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

I am feeding the capacitor tank voltage into my PLL chip. On end is to ground; the other goes through a 20k resistor. Right before it goes to the chip I have two MUR4100 (4A, 1000kv ultra fast) diodes connected. One is connected to the 15v supply and the other to ground. I would expect if the signal shoots below 0v the diode will clamp the voltage close to zero; if the voltage goes above 15v the other diode will act as a short. Here is my tracing:

Voltage range is 0 to 15v (1.5 divisions)

How come there is a significant overshoot? Why doesn't the diode clamp this, and how can I fix the problem so I get a nice flat tracing?

Z28Fistergod

Mon Dec 21 2009, 02:50AM

Registered Member #2040 Joined: Fri Mar 20 2009, 10:13PM
Location: Fairfax VA
Posts: 180

Series inductance and the speed of the diodes will keep them from clamping everything. Some small signal diodes would probably work better than the 4A monsters you have, those are kind of overkill for PLL feedback.

If you check the data sheet for the diodes you'll probably find they don't work well at whatever frequency those oscillations are going at.

IamSmooth

Mon Dec 21 2009, 03:33AM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

what about the 1n4148 or 1n4937? What if I added a ferrite bead?

ragnar

Mon Dec 21 2009, 11:39AM

Registered Member #63 Joined: Thu Feb 09 2006, 06:18AM
Location:
Posts: 1425

Z28F is right, MUR4100s have longer reverse-recovery times and a larger junction capacitance.

Try 1N4148 signal diodes or 1N5819 schottky diodes and compare the results with your oscilloscope.

Don't add any ferrite beads

Dr. Dark Current

Mon Dec 21 2009, 12:07PM

Registered Member #152 Joined: Sun Feb 12 2006, 03:36PM
Location: Czech Rep.
Posts: 3384

Maybe your scope is picking up interference from the switching transitions of the bridge. I have had this happen.

GeordieBoy

Mon Dec 21 2009, 01:32PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

It is true that forward recovery time of clamping diodes depends on the inductance in series with the clamping loop. That is why you want to keep your switches, clamping diodes and DC bus cap all as close as possible together to minimise the clamping loop's inductance.

Dr Kv is also right, that it could be switching hash being picked up by the measurement loop. Connect the flying ground clip to the tip of the scope probe and check that you get a flat line on the scope screen. Then touch this scope probe (with the ground clip still on) onto the 0v line of the circuit you are probing. You should still get a flat line with no glitches. If you get spikes when you do this test then the probe is most likely picking up common-mode interference.

If interference is magnetically coupled the spikes will start to appear when you get close to the circuit under test, and before you touch the probe tip to the circuit. In this case winding the earth lead around the probe tip is a good fix to minimise the loop area in the measurement circuit.

If the noise only appears when the grounded probe tip touches the circuit under test, then big ferrite cores cliped over the scope probe cable are the best solution. These form a common-mode choke and prevent HF currents flowing from the DUT down the screen of the scope probe.

Those are just a few tricks that are often used when probing power electronics circuits.

-Richie,

IamSmooth

Mon Dec 21 2009, 02:54PM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

The timing coincides with the inverter switches, but is the noise really in the circuit under test, or is it artifact that the probe is picking up? I am using a differential probe for measuring this circuit.

GeordieBoy

Mon Dec 21 2009, 03:59PM

Registered Member #1232 Joined: Wed Jan 16 2008, 10:53PM
Location: Doon tha Toon!
Posts: 881

> but is the noise really in the circuit under test, or is it artifact that the probe is picking up?

See my previous post.

teravolt

Tue Dec 22 2009, 03:01AM

Registered Member #195 Joined: Fri Feb 17 2006, 08:27PM
Location: Berkeley, ca.
Posts: 1111

1: Can you float it and create a common for your probe with a ground?
2: If your spike is real are your fets are cross conducting? Do you have any dead time? what type of setup do you have

IamSmooth

Tue Dec 22 2009, 04:12AM

Registered Member #190 Joined: Fri Feb 17 2006, 12:00AM
Location:
Posts: 1567

The 1N4148s worked better. When I viewed it with the current through the inverter the tracing was basically clean.

1 2 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob