Searching \ for '[EE] Do multi-track oscilloscopes create ground lo' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/index.htm?key=multi+track+oscilloscopes
Search entire site for: 'Do multi-track oscilloscopes create ground lo'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Do multi-track oscilloscopes create ground lo'
2011\11\17@135202 by Electron

flavicon
face

Hi again :o
a title that resembles the one of Bladerunner's original novel. :D

I have a 4 traces oscilloscope, actually I need all of the 4 traces
to correlate various parts of the circuit I'm working on (designing),
which is battery operated and thus isolated from mains' ground.

The oscilloscope's 4 channels' grounds are not isolated each other,
and now I developed the concern that the oscilloscope may itself
cause ground loops.

Should I use only one probes' ground, and let disconnected all the
other 3?

By the way, it would be cool if LTSpice "wire" let specify the
thickness (AWG) and length, so to simulate parasitic inductance and
resistance of the wire, which means a lot (we were talking about
ground loops, weren't we? :) ).

With kind regards,
Mario

2011\11\17@143733 by Sean Breheny

face picon face
You are right to consider possible problems but in many cases there is
no real problem once you consider the situation.

I don't like the term "ground loop" because too many people
misunderstand it. I prefer to say something like "unintended ground
voltage differences"

There are two possible problems which the scope grounds can have: one
is that you connect various points of the circuit which are at
different voltages to the scope common ground. The other is that you
affect your scope measurements by introducing extra noise.

I often use differential probes in sensitive circuits. If your circuit
has one big ground plane and you simply want to make general
measurements, then connecting the several scope grounds to the plane
should be fine. If you are dealing with signals below 100kHz, then
connecting only one of the grounds is OK. However, as you go above
this frequency, transmission line effects in the scope probe cable
begin to appear and you need both ends of each cable to be grounded to
prevent this.

Sean


On Thu, Nov 17, 2011 at 1:51 PM, Electron <spam_OUTelectron2k4TakeThisOuTspaminfinito.it> wrote:
{Quote hidden}

>

2011\11\17@152113 by Dwayne Reid

flavicon
face
At 11:51 AM 11/17/2011, Electron wrote:

>I have a 4 traces oscilloscope, actually I need all of the 4 traces
>to correlate various parts of the circuit I'm working on (designing),
>which is battery operated and thus isolated from mains' ground.
>
>The oscilloscope's 4 channels' grounds are not isolated each other,
>and now I developed the concern that the oscilloscope may itself
>cause ground loops.
>
>Should I use only one probes' ground, and let disconnected all the
>other 3?

As usual, "It depends".

You can get by with a single ground connection if you working at low speeds (DC - hundreds of KHz).  However, as soon as you need wide bandwidth, each probe NEEDS is own ground connection - this ground connection needs to approach zero length as the bandwidth increases towards hundreds of MHz.

I use a scope much more than I use a multi-meter, even when I'm just looking at general voltage levels.  If I'm looking at a high-level audio or a slow digital signal, I just use a single ground connection for multiple channels.

All that changes as soon as I need to look at high-speed or low-level audio signals.  My favorite probe ground connection is shown by Bob Pease in several of his articles and in his book "Troubleshooting Analog Circuits" - its a length of bare wire wrapped around the ground barrel at the front of the scope probe, then soldered to the closest ground point.  The effective ground lead length is somewhere between 1/4" - 1/2", depending on how close you can get the probe tip to the node that you are trying to observe.

Experience is the best judge.  If you see a waveform that just doesn't look right, change the ground connection technique and see if that improves things.

dwayne

-- Dwayne Reid   <.....dwaynerKILLspamspam@spam@planet.eon.net>
Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
(780) 489-3199 voice          (780) 487-6397 fax
http://www.trinity-electronics.com
Custom Electronics Design and Manufacturing

2011\11\17@165123 by Electron

flavicon
face

Thank You for all the interesting replies.. but changing slightly topic
and entering into PCB design, let me ask one thing:

a ground plane with a big hole in the center is a ground plane or is a ground loop?

I'm tempted to use star-like 0V connections instead of a ground plane, as I cannot
guarantee it to be a "plane", due to non sporadic signal tracks immersed into it.

I understand the need to keep the resistance and nonetheless the inductance as low
as possible, but I have seen things like "flooded" ground tracks that go nowhere,
then thin tracks that connect large ground zones, etc.. and now I know this is very
bad design. Perhaps designing with a star topology in mind will produce a better
and more *aware* result, what do you think?

As for the 'scope issues, I understood. Thank You very much for this reply and for
the ones in the other thread, about simulating isolated circuits on LTSpice (I write
this here as a note to not inflate the post count too much with "just" thanks).

Greets,
Mario

2011\11\17@175839 by Sean Breheny

face picon face
You are welcome!

You are illustrating the reason why I hate the term "ground loop" - it
has nothing inherently to do with loops!

It simply means any situation where you expect two points in the
circuit to be at the same voltage, so you can use them as a common
reference, but they are not close enough to each other in voltage for
your purposes.

This situation may be caused by a hole in the middle of a ground plane
but it can also be caused by other things and a big hole does not
necessarily cause it.

On the other hand, large holes in ground planes can cause other
problems, such as unintentional antennas. If you have a device which
is putting high frequency current through ground and you have a cutout
in the ground plane so that the current has to take "the long path
around", this structure will radiate quite a bit of RF.

A poorly-done star ground (i.e., one which is done without
understanding of where high currents are flowing and which point is
the appropriate common point) can be far worse than a ground plane. A
well done star ground can also be far better than a ground plane for
low frequency signals - especially when there are also large currents
in the system. For RF and digital, though, usually a ground plane
which is as continuous as possible is the best solution.

Even a ground plane, though, has a certain minimum inductance and
resistance (especially AC resistance) between any two points. Some
experiments I've done indicate that, in the 10MHz frequency range, the
inductance dominates and there is about 5 nanoHenries of inductance
per inch of distance between two points on a ground plane. I looked
into this when I was designing a motor drive circuit which had very
fast rising and falling PWM edges. I made a circuit which charged a
capacitor and then dumped it through a pair of points about 5 inches
apart on a continuous copper plane on a PCB within about 100
nanoseconds. I used a high-frequency current probe to monitor the
current and a differential scope probe to look at the voltage drop
directly between various points on the plane.

Sean


On Thu, Nov 17, 2011 at 4:51 PM, Electron <electron2k4spamKILLspaminfinito.it> wrote:
{Quote hidden}

>

2011\11\18@085801 by Electron

flavicon
face
At 23.58 2011.11.17, you wrote:
>You are welcome!
>
>You are illustrating the reason why I hate the term "ground loop" - it
>has nothing inherently to do with loops!

True, perhaps (please tell me if I'm wrong, as I'm ignorant in these
matter which I'm beginning to study only now) with only one exception:
if a wire acts like a RF antenna, would a "ground loop" act like a
magnetic pickup? (assuming there are changing magnetic fields in the
room, of course).

This situation excluded, I agree that ground loop is an improper term,
and all the matter (as I just simulated in LTSpice, using resistors and
inductors in place of wires) in the end is that different supposedly-gnd
points are at different potential, thus current will flow if you connect
them somewhere. That made me understand the importance of star topology,
although that has its own problems too (namely the extra RF picked up).


{Quote hidden}

Is it a matter of total length of the conductor? (of course sharp turns,
length, etc.. will matter too, but I mean You're referring here only to
the fact that at high frequences tracks must be as short as possible?).
i.e. transmission lines' effects.


{Quote hidden}

Cool rig, I wish someday I'll have it too. :-)

Cheers,
MarI/O

{Quote hidden}

>> -

2011\11\18@091234 by alan.b.pearce

face picon face
> >You are illustrating the reason why I hate the term "ground loop" - it
> >has nothing inherently to do with loops!
>
> True, perhaps (please tell me if I'm wrong, as I'm ignorant in these matter which
> I'm beginning to study only now) with only one exception:
> if a wire acts like a RF antenna, would a "ground loop" act like a magnetic pickup?
> (assuming there are changing magnetic fields in the room, of course).

Ground Loop is quite a reasonable term to use at mains frequency, I agree it is a bit of a misnomer at RF.
-- Scanned by iCritical.

2011\11\18@095831 by Sean Breheny

face picon face
Hi Alan,

What is the specific situation at mains frequency that you would call
a "ground loop"?

Sean


On Fri, Nov 18, 2011 at 9:11 AM,  <EraseMEalan.b.pearcespam_OUTspamTakeThisOuTstfc.ac.uk> wrote:
{Quote hidden}

>

2011\11\18@110954 by Dwayne Reid

flavicon
face
At 07:58 AM 11/18/2011, Sean Breheny wrote:

>What is the specific situation at mains frequency that you would call
>a "ground loop"?
>
>Sean
>
>
>On Fri, Nov 18, 2011 at 9:11 AM,  <alan.b.pearcespamspam_OUTstfc.ac.uk> wrote:
> >
> > Ground Loop is quite a reasonable term to use at mains frequency,
> I agree it is a bit of a misnomer at RF.

I'm not Alan, but I'll jump in here.

For me, a 'ground loop' is a situation that occurs when an unbalanced signal becomes influenced by a current that flows in the ground conductor that is associated with that signal feeding the input of a device.  Phew! Long, run-on sentence there.

But it does describe the situation, whether that signal is an audio signal feeding a preamplifier or any signal feeding the input of a 'scope.

dwayne

-- Dwayne Reid   <@spam@dwaynerKILLspamspamplanet.eon.net>
Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
(780) 489-3199 voice          (780) 487-6397 fax
http://www.trinity-electronics.com
Custom Electronics Design and Manufacturing

2011\11\21@064904 by alan.b.pearce

face picon face
> Hi Alan,
>
> What is the specific situation at mains frequency that you would call a
> "ground loop"?
>
> Sean

The hum you get in a PA system when connecting multiple items together, and then breaking the ground at one point removes the hum.
-- Scanned by iCritical.

2011\11\21@085950 by Electron

flavicon
face
At 12.48 2011.11.21, you wrote:
>> Hi Alan,
>>
>> What is the specific situation at mains frequency that you would call a
>> "ground loop"?
>>
>> Sean
>
>The hum you get in a PA system when connecting multiple items
>together, and then breaking the ground at one point removes the hum.

By the way, is this hum picked from nearby magnetic 50/60Hz fields?

Because on an amp the hum caused by improper ground connection may be just
power supply hum injected into the other circuits (thus load dependent),
which (due to rectification) will be 100/120Hz.

But I suspect that a "ground loop" may act also like a magnetic "antenna" (a
reluctor pickup), thus picking up magnetic fields leaked from nearby high
tension electric company's line transformers, etc..

Effectively breaking the loop is like cutting one of the two wires of a
reluctor pickup..

More... (looser matching)
- Last day of these posts
- In 2011 , 2012 only
- Today
- New search...