zener voltage drop is weird / General Science and Electronics / Forums

Forums

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

« Previous topic | Next topic »

zener voltage drop is weird

1 2 next

Move Thread

LAN_403

IamSmooth

Thu Jan 07 2010, 02:47AM

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

Please look at the following pictues

voltage measured at A: approx +15 to -15v

voltage measured at B: approx +15v to -20v

voltages at A and B with 150R resistor replacing mosfet GS junction: A +15 to -15v, B +10v to -15v

The schematic is is a resistor followed by a 5v zener going to the gate of a mosfet. A transformer going from +15 to -15 is driving the signal. The mosfet source is connected to ground. Now I expect the voltage to be 5v more at point A compared to point B. I also expect the positive voltage to show a difference and not the negative voltage. I have put voltage markers on the tracing so you can compare one with the other. Measurements were made with the positive probe at point A and B; the negative probe was connected to ground.

What I see is the voltage at A is 5v more positive than B. However, this is occurring on the negative cycle. A is -15v and B is -20v. The upper tracing of the square-wave is only affected by the standard diode voltage drop of 0.6v. What I expected to see was Apeak to be 15v, B to be 10v, and both troughs to be -15v. Another puzzling thing is how am I now getting the signal at point B going from +15v to -20v? My drive is only +15/-15v, and I have a -20v reading?

In the last picture I removed the mosfet and put a 150R resistor from the gate to source (ground). Now, the voltage at B is registering as 10v, A is 15v, and both are at -15v.

It seems to me the mosfet has properties that I am not appreciating. Can anyone straighten this out for me?

Myke

Thu Jan 07 2010, 03:50AM

Registered Member #540 Joined: Mon Feb 19 2007, 07:49PM
Location: MIT
Posts: 969

I'd think the zener isn't functioning in the way that you want because you need some current flowing through the zener to have a 5V drop appear across it in the reverse direction. The gate of the MOSFET is pretty good insulator so you can't get a constant voltage drop across the body of the diode. I might have to draw pictures since I don't know if I explained it well enough...
I'm not sure about why the -20V is there.
I'm wondering how you are getting a square wave off the end of a GDT...

In A, I would expect the zener to drop 5V across it.
In B, when the cap is discharged, it would behave similarly to A.
C is what the equivalent would be once the cap has fully charged (if it has enough time to). I would expect to see 0V across the zener as there is no current flowing through it.

Connect your scope probes across the zener to see what voltages appear across it. The voltage across the zener in the reverse direction should only vary between 0 and 5 volts.

IamSmooth

Thu Jan 07 2010, 04:21AM

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

The diodes work fine. I made the measurement across points A and B, and I got the expected result. It is a square wave going from 5v to 0 (or real close) with A being 5v more positive than B.

The -20v must have something to do with the gate capacitance getting charged. Anyone want to chime in?

radiotech

Thu Jan 07 2010, 05:41AM

Registered Member #2463 Joined: Wed Nov 11 2009, 03:49AM
Location:
Posts: 1546

I looked at the datasheet for the FDH44N50 and it appears to be insulated gate device with low leakage and capacitance. How can the current flow through the Zener? They will have forward drop of about 0.6 V and reverse breakdown at the Zener's rated voltage, provided current is flowing. I have never seen a zener that works on nA.

Dr. Dark Current

Thu Jan 07 2010, 12:12PM

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

Is the MOSFET's drain connected to a powered up switched circuit?
If so, the gate is probably being charged through the reverse transfer capacitance (edit: but anyway this is weird because the reverse transfer current should 'act' opposite the charging current). Try to power up just the gate driver and disconnect the drain/turn off the drain supply.

radiotech: The currents during gate charging and discharging are pretty big and should be enough to cause the zener drop.

IamSmooth

Thu Jan 07 2010, 12:53PM

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

Dr. Kilovolt wrote ...

Is the MOSFET's drain connected to a powered up switched circuit?
If so, the gate is probably being charged through the reverse transfer capacitance (edit: but anyway this is weird because the reverse transfer current should 'act' opposite the charging current). Try to power up just the gate driver and disconnect the drain/turn off the drain supply.
.

I thought of this, too. The tracing is only the inverter drive circuit. The HV rail going to the drain is not powered. I am wondering if I can model the mosfet gate with a small capacitance and diode and see what happens.

I wonder if anyone can try to reproduce this on a breadboard. I am going to try to reproduce it with a function generator in order to take the gate drive out of the equation.

radiotech

Thu Jan 07 2010, 08:53PM

Registered Member #2463 Joined: Wed Nov 11 2009, 03:49AM
Location:
Posts: 1546

Re gate capacitance; The data sheet says 5535 pFd. Ifthe frequency that diagram with the mosfet is operating t was 30 kHz gate-source AC current would be 0.016 amp. Dr. Kilovolt had a good point about gate/source C. Here are some tests using
Mykes figure "B" with a 1n4723,5.6 volt in series with 4700 pF fed with Sine 30 kHz, and scope across cap, wave was distorted sin. With a 1K in place of cap, wave looked exactly like zener acting. with lower part down from 0 about .5 V and upper above about 5 V. However with signal changed to square, the zener acted like a 680 ohm R with the cap load producing a bit of rounding of the leading edge on top, and on bottom with diode reversed.No zener knee was seen with square wave.

The next thing is to find out why the 1N4723 acted this way, or if this experiment was flawed. The only thing is the original pictures seem to support the zener turning into a resistor with square wave
current. I tried to magnify the image to better see the scope numeric readout.

I do hope to get the new digital cam into the shop and be able to post pics soon. As well there is a digital scope there needing some minor repair that can produce image files in CPM or DOS which might work on this notebook.

Dr. Dark Current

Thu Jan 07 2010, 10:08PM

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

Another thought, maybe the gate voltage "rings up" because of the GDT's leakage inductance. Bypass the GDT and see if the waveform changes.

GeordieBoy

Thu Jan 07 2010, 10:29PM

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

I think Dr kV is right. What you are seeing is probably due to the fact that the gate capacitance is effectively being resonantly charged through the GDT's leakage inductance.

When you add a 5V zener diode in series you are essentially letting the voltage on the gate wander by up to 5V relative to the output from the GDT before current will flow to correct it.

I'm guess this behaviour is what you want? ...otherwise you wouldn't have taken the rather unusual action of driving a MOSFET's gate through a series zener diode in the first place?

Regarding MOSFET gate capacitance: It's value varies massively depending on the instantaneous drain-source voltage, and also to some extent depending on the gate-source voltage. There is also capacitance between the drain and the gate that tends to cause apparent extra gate current to flow when the drain voltage is rapidly changing.

Modern Switched-mode MOSFETs are highly non-linear devices!

-Richie,

IamSmooth

Thu Jan 07 2010, 11:01PM

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

These are interesting thoughts KV and Georgie because I have been in my basement experimenting by adding capacitance and inductance. I took the basic setup: resistor, diode, mosfet

and found the tracing to be as expected. The voltage behind the zener was 5v less.

I then added a 125uH inductor in front of the resistor. I drove the test circuit with 45khz square-wave from a signal generator.

I found that this reproduced the excessive negative voltage that exceeded the generator voltage.

On the bottom tracing the most negative voltage is AFTER the zener.

I had tried combinations of capcitors on the breadboard, but this did not create the same effect, although removing the mosfet on the circuit board (not the breadboard) and putting a 0.01uf capacitor across the gate-source junction did result in the negative voltage after the zener to exceed in magnitude the negative generator voltage.

I saw this effect even after bypassing the GDT, so this effect must be from the stray inductance on the mosfet.

Is this behavior that I want? The voltage "ringing" isn't what I was after; it was the timing delay to ensure both mosfets weren't on at the same time.

1 2 next

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