It's hard to filter a spike,

G

George Herold

Guest
This comes out of my weird 60 Hz pickup thread on sed.
It seems like a basic thing, hence SEB.
I've got a Fet gate drive that couples,
via a common cable into a signal line.
(C_couple ~= 10-100pF, the source impedance drivibng the
signal line is kinda high ~2k ohm.
A fast switching spike on the gate makes a pulse
on my signal. As I try to filter it out...
say by slowing down the switch edges.
I decrease dV/dt, so less voltage, but the
pulse get's spread out in time... by the
same slope. The result being basically no change
in the amount of fundamental. (the frequency creating the spikes)

grumble, it took me part of a day to "discover"
this fairly obvious fact in retrospect.
I don't expect any answers.
(putting the gate drive inside coax works,
but has other issues.)

George H.
 
On Wed, 18 Jan 2017 16:41:50 -0800, George Herold wrote:

This comes out of my weird 60 Hz pickup thread on sed.
It seems like a basic thing, hence SEB.
I've got a Fet gate drive that couples,
via a common cable into a signal line.
(C_couple ~= 10-100pF, the source impedance drivibng the signal line is
kinda high ~2k ohm.
A fast switching spike on the gate makes a pulse on my signal. As I try
to filter it out...
say by slowing down the switch edges.
I decrease dV/dt, so less voltage, but the pulse get's spread out in
time... by the same slope. The result being basically no change in the
amount of fundamental. (the frequency creating the spikes)

grumble, it took me part of a day to "discover"
this fairly obvious fact in retrospect.
I don't expect any answers.
(putting the gate drive inside coax works,
but has other issues.)

George H.
Click to expand...

That's why I was babbling about lots of diodes in your other thread: if
you limit the spike first, and then low-pass, you'll actually reduce its
fundamental content.

Better would be to have some circuit that limits the slope -- basically a
low-pass filter that ignores any excursion more than X volts above or
below the output voltage.

--
Tim Wescott
Control systems, embedded software and circuit design
I'm looking for work! See my website if you're interested
http://www.wescottdesign.com
 
On Wed, 18 Jan 2017 22:26:16 -0600, Tim Wescott <tim@seemywebsite.com>
wrote:

On Wed, 18 Jan 2017 16:41:50 -0800, George Herold wrote:

This comes out of my weird 60 Hz pickup thread on sed.
It seems like a basic thing, hence SEB.
I've got a Fet gate drive that couples,
via a common cable into a signal line.
(C_couple ~= 10-100pF, the source impedance drivibng the signal line is
kinda high ~2k ohm.
A fast switching spike on the gate makes a pulse on my signal. As I try
to filter it out...
say by slowing down the switch edges.
I decrease dV/dt, so less voltage, but the pulse get's spread out in
time... by the same slope. The result being basically no change in the
amount of fundamental. (the frequency creating the spikes)

grumble, it took me part of a day to "discover"
this fairly obvious fact in retrospect.
I don't expect any answers.
(putting the gate drive inside coax works,
but has other issues.)

George H.

That's why I was babbling about lots of diodes in your other thread: if
you limit the spike first, and then low-pass, you'll actually reduce its
fundamental content.

Better would be to have some circuit that limits the slope -- basically a
low-pass filter that ignores any excursion more than X volts above or
below the output voltage.
Click to expand...

Still better is to eleiminate the source of spikes.


w.
 
On Wednesday, January 18, 2017 at 11:26:23 PM UTC-5, Tim Wescott wrote:
On Wed, 18 Jan 2017 16:41:50 -0800, George Herold wrote:

This comes out of my weird 60 Hz pickup thread on sed.
It seems like a basic thing, hence SEB.
I've got a Fet gate drive that couples,
via a common cable into a signal line.
(C_couple ~= 10-100pF, the source impedance drivibng the signal line is
kinda high ~2k ohm.
A fast switching spike on the gate makes a pulse on my signal. As I try
to filter it out...
say by slowing down the switch edges.
I decrease dV/dt, so less voltage, but the pulse get's spread out in
time... by the same slope. The result being basically no change in the
amount of fundamental. (the frequency creating the spikes)

grumble, it took me part of a day to "discover"
this fairly obvious fact in retrospect.
I don't expect any answers.
(putting the gate drive inside coax works,
but has other issues.)

George H.

That's why I was babbling about lots of diodes in your other thread: if
you limit the spike first, and then low-pass, you'll actually reduce its
fundamental content.
Click to expand...
I thought about limiting the spike at it's source, but there is not much to be
gained. (the gate swings -15* to +15 V.. I could cut that down to 0 to +10,
but only a factor of three.) I hadn't thought about limiting on the input line.
Clamped at zero and ~1.5V... three PN diodes and a Schottky going the other way for
the zero?
Better would be to have some circuit that limits the slope -- basically a
low-pass filter that ignores any excursion more than X volts above or
below the output voltage.
Are you talking about on the input line?
Click to expand...

George H.
* (-15 V on the gate! I'm abusing the maximum GS voltage!)
--
Tim Wescott
Control systems, embedded software and circuit design
I'm looking for work! See my website if you're interested
http://www.wescottdesign.com
Click to expand...
 
On Thursday, January 19, 2017 at 7:45:31 AM UTC-5, Martin Riddle wrote:
On Wed, 18 Jan 2017 16:41:50 -0800 (PST), George Herold
gherold@teachspin.com> wrote:

This comes out of my weird 60 Hz pickup thread on sed.
It seems like a basic thing, hence SEB.
I've got a Fet gate drive that couples,
via a common cable into a signal line.
(C_couple ~= 10-100pF, the source impedance drivibng the
signal line is kinda high ~2k ohm.
A fast switching spike on the gate makes a pulse
on my signal. As I try to filter it out...
say by slowing down the switch edges.
I decrease dV/dt, so less voltage, but the
pulse get's spread out in time... by the
same slope. The result being basically no change
in the amount of fundamental. (the frequency creating the spikes)

grumble, it took me part of a day to "discover"
this fairly obvious fact in retrospect.
I don't expect any answers.
(putting the gate drive inside coax works,
but has other issues.)

George H.





Did you try to shore up the ground return to the driver?
Shorten its path, much the same as the coax did?
Click to expand...

Hmm I hadn't thought about the source return line.
I'll have to look at it.
Thanks.

(On second thought I don't mind if you reply :^)

George H.
Cheers
Click to expand...
 
On Thursday, January 19, 2017 at 9:20:28 AM UTC-5, George Herold wrote:
On Wednesday, January 18, 2017 at 11:26:23 PM UTC-5, Tim Wescott wrote:
On Wed, 18 Jan 2017 16:41:50 -0800, George Herold wrote:

This comes out of my weird 60 Hz pickup thread on sed.
It seems like a basic thing, hence SEB.
I've got a Fet gate drive that couples,
via a common cable into a signal line.
(C_couple ~= 10-100pF, the source impedance drivibng the signal line is
kinda high ~2k ohm.
A fast switching spike on the gate makes a pulse on my signal. As I try
to filter it out...
say by slowing down the switch edges.
I decrease dV/dt, so less voltage, but the pulse get's spread out in
time... by the same slope. The result being basically no change in the
amount of fundamental. (the frequency creating the spikes)

grumble, it took me part of a day to "discover"
this fairly obvious fact in retrospect.
I don't expect any answers.
(putting the gate drive inside coax works,
but has other issues.)

George H.

That's why I was babbling about lots of diodes in your other thread: if
you limit the spike first, and then low-pass, you'll actually reduce its
fundamental content.
I thought about limiting the spike at it's source, but there is not much to be
gained. (the gate swings -15* to +15 V.. I could cut that down to 0 to +10,
but only a factor of three.) I hadn't thought about limiting on the input line.
Clamped at zero and ~1.5V... three PN diodes and a Schottky going the other way for
the zero?

Better would be to have some circuit that limits the slope -- basically a
low-pass filter that ignores any excursion more than X volts above or
below the output voltage.
Are you talking about on the input line?
Click to expand...

OK I screwed up again. Yesterday I tried slowing down the gate drive on the FET.
I don't know what I did wrong, but I thought it screwed up the slow feedback loop.

I did it again today, and it worked fine??
Anyway I've now got the gate transitioning very slowly.. ~100ms TC and
everything seems happy.

George H.

George H.
* (-15 V on the gate! I'm abusing the maximum GS voltage!)

--
Tim Wescott
Control systems, embedded software and circuit design
I'm looking for work! See my website if you're interested
http://www.wescottdesign.com
Click to expand...
 
On Wed, 18 Jan 2017 16:41:50 -0800 (PST), George Herold
<gherold@teachspin.com> wrote:

This comes out of my weird 60 Hz pickup thread on sed.
It seems like a basic thing, hence SEB.
I've got a Fet gate drive that couples,
via a common cable into a signal line.
(C_couple ~= 10-100pF, the source impedance drivibng the
signal line is kinda high ~2k ohm.
A fast switching spike on the gate makes a pulse
on my signal. As I try to filter it out...
say by slowing down the switch edges.
I decrease dV/dt, so less voltage, but the
pulse get's spread out in time... by the
same slope. The result being basically no change
in the amount of fundamental. (the frequency creating the spikes)

grumble, it took me part of a day to "discover"
this fairly obvious fact in retrospect.
I don't expect any answers.
(putting the gate drive inside coax works,
but has other issues.)

George H.
Click to expand...
Did you try to shore up the ground return to the driver?
Shorten its path, much the same as the coax did?

Cheers
 
On Thu, 19 Jan 2017 09:49:54 +0100, Helmut Wabnig wrote:

On Wed, 18 Jan 2017 22:26:16 -0600, Tim Wescott <tim@seemywebsite.com
wrote:

On Wed, 18 Jan 2017 16:41:50 -0800, George Herold wrote:

This comes out of my weird 60 Hz pickup thread on sed.
It seems like a basic thing, hence SEB.
I've got a Fet gate drive that couples,
via a common cable into a signal line.
(C_couple ~= 10-100pF, the source impedance drivibng the signal line
is kinda high ~2k ohm.
A fast switching spike on the gate makes a pulse on my signal. As I
try to filter it out...
say by slowing down the switch edges.
I decrease dV/dt, so less voltage, but the pulse get's spread out in
time... by the same slope. The result being basically no change in
the amount of fundamental. (the frequency creating the spikes)

grumble, it took me part of a day to "discover"
this fairly obvious fact in retrospect.
I don't expect any answers.
(putting the gate drive inside coax works,
but has other issues.)

George H.

That's why I was babbling about lots of diodes in your other thread: if
you limit the spike first, and then low-pass, you'll actually reduce its
fundamental content.

Better would be to have some circuit that limits the slope -- basically
a low-pass filter that ignores any excursion more than X volts above or
below the output voltage.

Still better is to eleiminate the source of spikes.


w.
Click to expand...

Yes!



--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

I'm looking for work -- see my website!
 
On Thu, 19 Jan 2017 09:49:54 +0100, Helmut Wabnig wrote:
On Wed, 18 Jan 2017 22:26:16 -0600, Tim Wescott <t...@seemywebsite.com
wrote:

On Wed, 18 Jan 2017 16:41:50 -0800, George Herold wrote:

This comes out of my weird 60 Hz pickup thread on sed.
It seems like a basic thing, hence SEB.
I've got a Fet gate drive that couples,
via a common cable into a signal line.
(C_couple ~= 10-100pF, the source impedance drivibng the signal line
is kinda high ~2k ohm.
A fast switching spike on the gate makes a pulse on my signal. As I
try to filter it out...
say by slowing down the switch edges.
I decrease dV/dt, so less voltage, but the pulse get's spread out in
time... by the same slope. The result being basically no change in
the amount of fundamental. (the frequency creating the spikes)

grumble, it took me part of a day to "discover"
this fairly obvious fact in retrospect.
I don't expect any answers.
(putting the gate drive inside coax works,
but has other issues.)

George H.

That's why I was babbling about lots of diodes in your other thread: if
you limit the spike first, and then low-pass, you'll actually reduce its
fundamental content.

Better would be to have some circuit that limits the slope -- basically
a low-pass filter that ignores any excursion more than X volts above or
below the output voltage.

Still better is to eleiminate the source of spikes.
Click to expand...

Fat chance if you can't control the weather, what happens off of the property, etc ...
 
You must log in or register to reply here.

Welcome to EDABoard.com

Sponsor

Back
Top