On Tue, 12 Nov 2013 17:16:41 -0800, mike wrote:

On 11/12/2013 7:38 AM, Jim Thompson wrote:
On Tue, 12 Nov 2013 08:14:31 -0700, Jim Thompson
wrote:

On Tue, 12 Nov 2013 08:45:16 -0500, Jay wrote:

On Mon, 11 Nov 2013 08:16:16 -0700 Jim Thompson
wrote in Message id:
:

On Mon, 11 Nov 2013 07:38:52 -0500, Jay wrote:

Hello,

I am having a high failure rate on a 2N7000 FET that drives a relay coil.
Please see the attached jpeg. This gate drive is coming from a 5 volt
PIC18F452 pin 5 (RD7/PSP7) Any thoughts greatly appreciated. Note that it
will run for quite awhile, perhaps 20 to 30 thousand cycles before failing
entirely shorted Drain to Source.

Thanks.

The PIC18F452 pin 5 (RD7/PSP7), is that high-side-only drive? If so,
the 10K G-S pull-down is inadequate for quick turn-off, and the 2N7000
dissipates to death.

Or it could be the other way around and the PIC can't raise the gate
voltage quickly enough due to its large capacitance. Likewise a slow
death due to dissipation.

...Jim Thompson

Hi Jim,

Thanks. I think your second thought may be right. I took a scope capture
of the gate at turn on. (Attached)
Think that could be the culprit? Any ideas to get rid of it?

To address Tim's question:
The total trace length including wire is about 6" or so between the PIC
and the FET. 4" of that is on a ribbon cable between two connectors.
The diode is at the coil, not the FET.

Turn off looks fine, no ringing or anything...

Thanks everybody for your suggestions.

The time spent in that flat spot (gate charge plateau) is probably
what is getting you.

I'd probably rig up a bipolar driver between the micro and the gate.

More detail as to ON/OFF rate, etc?

...Jim Thompson

Sheeesh! It just occurred to me... the easy fix... One package of
74HC04 (or the small package equivalent). Parallel 5 of the six
inverters for directly driving the gate, then one inverter between
micro and the 5-paralleled ones. That ought to speed up things a bit
;-)

...Jim Thompson

Am I the only one old enough to remember the days when we used
to IDENTIFY a problem before we started throwing solutions at it?

You really need to plot the voltage AND the current in the FET
on BOTH edges.

I'm gonna make a big leap here cuz I don't know the current in the relay
relative to the capabilities of the FET. A quick look at the spec
suggests that, using the worst
case specs, there's plenty of room for the possibility that this is the
WRONG FET for the task.

If it's low-impedance pullup-only drive. Otherwise Miller can
eat your lunch at turn-on too.
If it really is Miller effect. And IF reducing the gate resistor
from 10K to 1K really does reduce the discharge time by a factor of
10.
Might that solve the problem?
Should be trivial to tell by looking at the CURRENT X Voltage profile
at the drain and gate.

I'm not saying you shouldn't fix it, but it is curious that you have
the remote alarm with mechanical relay going off 30,000 times.

Huh? Too many tokes ?:-}

...Jim Thompson
--
| James E.Thompson | mens |
| Analog Innovations | et |
| Analog/Mixed-Signal ASIC's and Discrete Systems | manus |
| San Tan Valley, AZ 85142 Skype: Contacts Only | |
| Voice480)460-2350 Fax: Available upon request | Brass Rat |
| E-mail Icon at http://www.analog-innovations.com | 1962 |

I love to cook with wine. Sometimes I even put it in the food.