FET not turning fully on with AC on drain / General Science and Electronics / Forums

Forums

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

« Previous topic | Next topic »

FET not turning fully on with AC on drain

1 2 next

Move Thread

LAN_403

the_anomaly

Wed Jun 27 2018, 06:08PM

Registered Member #19 Joined: Thu Feb 02 2006, 03:19PM
Location: Jacksonville, FL
Posts: 168

Hello all,

I built a simple circuit to drive a nichrome wire heating element. Unfortunately, the FET is overheating withing a few minutes of power being applied. When I scope the drain, I find it several volts above ground with an AC waveform on it. The gate signal looks good to me. I control the FET with PWM but for the moment I am driving it with 100% on time. Power is supplied from a variac into a 1000VA isolation transformer to the circuit. Currently running around 20-30VAC. I have used colors in coordination to the scope screen shot to show where I am probing; blue is the gate, green is the drain, purple is the high side.

Schematic:

Scope waveforms:

Any ideas on what I am doing wrong? I added a snubber to the FET between the drain and source terminals, 2nF cap in series with 20ohm resistor. This is not shown on the schematic.

Blackcurrant

Wed Jun 27 2018, 09:26PM

Registered Member #2989 Joined: Sun Jul 11 2010, 12:01AM
Location: UK
Posts: 94

D1 ?
or a short load

also

the_anomaly

Thu Jun 28 2018, 12:47PM

Registered Member #19 Joined: Thu Feb 02 2006, 03:19PM
Location: Jacksonville, FL
Posts: 168

I think I figured it out. Heating coils resistance 7.1 ohms, FET is 1-2ohms as it heats up which accounts for the voltage I am seeing across it.

Sulaiman

Thu Jun 28 2018, 12:50PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Is it necessary to use d.c. for the heater/control ?
It would be much easier and much more efficient to use mains plus a triac
(often fired by an opto-triac)
if you need rapid control.

If burst fire operation (on/off/on over several seconds to minutes) is acceptable
(it usually is due to thermal inertia)
then just an ssr should suffice

the_anomaly

Tue Jul 03 2018, 02:26PM

Registered Member #19 Joined: Thu Feb 02 2006, 03:19PM
Location: Jacksonville, FL
Posts: 168

I think I solved the overheating problem. I switched to R6030KNZ1C9 for the power FET which has about a 1/10th the RDS on as the STW12NK80Z I was using.

Is it necessary to use d.c. for the heater/control ?
It would be much easier and much more efficient to use mains plus a triac
(often fired by an opto-triac)
if you need rapid control.

It's not necessary but I chose DC as this is what I am most familiar with. My full system is computer software that runs a PID loop to control the temperature with a microprocessor running the circuit. The computer asks the micro for current temperature, runs the PID loop, then adjusts the microprocessors PWM output. With basic tuning I get good accuracy and response time with almost no over/undershoot of desired temperature.

I have a new problem, ringing. I've been trying to use snubbers with limited success. Here is what the ringing looks like with a typical snubber:

The green trace is the FET drain, the blue is the FET gate.

So far I have tried RC snubbers composed of 1200pF, 680pF, and 330pF in all combinations with 47ohm, 20ohm, and 4.7ohm resistors. They all vary in performance but not by much; the best I get is a few volts off the drain ringing and a small reduction in the gate ringing.

Question, is the ringing on the gate the main source of the problem? I have been focused on reducing the ringing on the load, I am wondering if I should be trying to clean up the gate signal instead.

hen918

Tue Jul 03 2018, 02:51PM

Registered Member #11591 Joined: Wed Mar 20 2013, 08:20PM
Location: UK
Posts: 556

Yes, the gate signal is the first one that needs to be cleaned up (ringing that huge could destroy the oxide layer in the FET, and will be causing the ringing on the drain), once that looks nice, if you still have problems, then you turn your attention to the drain. I would recommend adding / increasing the gate resistor, this should solve your problem.

Sulaiman

Tue Jul 03 2018, 05:16PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

I too would look at the gate drive,
I think that I'd try R4 as in the original diagram but 100R to 10R.

Choose your pwm frequencies wisely and filter the load wiring against radiating emi.

Unless you must start with a d.c. supply,
I think that you are making your system unnecessarily complex, expensive and difficult.

Consider the humble but excellent REX-C100 ssr controller as a reference design.

the_anomaly

Tue Jul 03 2018, 06:26PM

Registered Member #19 Joined: Thu Feb 02 2006, 03:19PM
Location: Jacksonville, FL
Posts: 168

I would recommend adding / increasing the gate resistor, this should solve your problem.

I have seen a lot of people recommend small resistances like 4.7-10 ohms and I had been under the impression we want a very small resistance so lots of current can flow. Now I see that's not entirely true. I found this application note from Fairchild: Driving and Layout Design for Fast Switching
Super-Junction MOSFETs. On page 5 they discuss using a ferrite bead in series with the gate resistor. I took out both gate resistors and put a ferrite bead in their place. I did not leave myself enough space under the heatsink to do much soldering otherwise I would have put the bead in series with a gate resistor... The new waveform looks much better although still needs improvement:

I'll pull the heatsink off and up the gate resistance with and without the ferrite bead and hopefully post the conclusion soon.

Unless you must start with a d.c. supply,
I think that you are making your system unnecessarily complex, expensive and difficult.

True, but there is more to the project than I have talked about. I plan to control pressures, fan speeds, in addition to the temperature ect. so for that purpose I do need computer/micro control. I'm doing everything incrementally so for this small project it is overkill but it is needed in the long run.

Plasma

Tue Jul 03 2018, 08:28PM

Registered Member #61406 Joined: Thu Jan 05 2017, 11:31PM
Location:
Posts: 268

Hi, your right your underdapene the signal, the were meaning higher resistance, a inductor is charging up a capacitor, so use high resistance, as the capacitor is the domanitor high resistance is best.

Sulaiman

Tue Jul 03 2018, 08:44PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

I suspect that along the way you will discover that for each power source / load there is an 'optimal' solution,
and there is no 'universal' drive/controller other than possibly at the signal level,
but for

a one-off hobby project it does not matter - the weirder the better, it's all learning,
and hopefully fun,
but it is unlikely to be a shareable/collaborative approach.

Lots of my own hobby circuits are 'odd' - because I can.

1 2 next

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