Searching \ for '[EE] Approach to noice problems' 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=approach+noice+problems
Search entire site for: 'Approach to noice problems'.

Exact match. Not showing close matches.
PICList Thread
'[EE] Approach to noice problems'
2005\02\03@123608 by Ake Hedman

flavicon
face
I have two identical boards on my desk  and have some noise problems with one of them. Both have a sensor input and a cable of about two meters attached to the input plus are connected to a CAN bus. The sensors are decouples on the board with a 0.1uF cap. Above my bench I have a fluorescent lamp and one of the boards get very upset (watchdog reset) when this lamp is turned on/off.

First it is strange that only one board does this but my question is how do one approach a problem like this?
I have also tested a piezo gas lightning device close to the boards but it does not disturb the systems.

/Ake

--  ---
Ake Hedman (YAP - Yet Another Programmer)
eurosource, Brattbergavägen 17, 820 50 LOS, Sweden
Phone: (46) 657 413430 Cellular: (46) 73 84 84 102
Company home: http://www.eurosource.se      Kryddor/Te/Kaffe: http://www.brattberg.com
Personal homepage: http://www.eurosource.se/akhe
Automated home: http://www.vscp.org

2005\02\03@124643 by Paul James E.

picon face

Ake,

In my opinion, it sounds like a bad ground somwehere.   Maybe a cold
solder joint on a component, or a dirty connection to ground, but a bad
gound somehow.   Another thing I would try is replacing the .1 uf  decoupling cap.  It's not entirely unlikely that it could be bad.
I have had this happen a few times.  Using a magnifying glass, inspect  all the solder joints on the board, and reflow the solder on any joints
that look suspect.  I'll bet you find a bad ground somewhere on the board.
Let us know what you find out.

                                              Regards,

                                                Jim
             

{Quote hidden}

> -

2005\02\03@132252 by Jose Da Silva

flavicon
face
On Thursday 03 February 2005 09:35 am, Ake Hedman wrote:
> I have two identical boards on my desk  and have some noise problems
> with one of them. Both have a sensor input and a cable of about two
> meters attached to the input plus are connected to a CAN bus. The
> sensors are decouples on the board with a 0.1uF cap. Above my bench I
> have a fluorescent lamp and one of the boards get very upset (watchdog
> reset) when this lamp is turned on/off.

If it is not a shielded cable, then you have a "nice" antenna to collect all
that RF noise. The 0.1uF caps by themselves are not enough. You need to
think of adding either a resistance or an inductor on the sensor line before
the cap and, preferably you add them to both lines (assuming 1 pin for
gnd/vcc and 2nd pin for sensor input).

> First it is strange that only one board does this but my question is how
> do one approach a problem like this?
> I have also tested a piezo gas lightning device close to the boards but
> it does not disturb the systems.

As you have found out, not all frequencies appear to affect it, or affect
every board the same way however, you are still dealing with an RF source
and an antenna. If you look at some of the tests done to pass FCC or EU or
other electrical/electronics agencies, you will note that the tests have
some very high voltages which will affect cables like yours of only 2 meters
length.
1st, you specify that the sensor should be on a shielded cable to make your
circuit more immune from noise, 2nd you design your board so that people who
seem to know more than they think and still use plain wire won't affect your
circuit.... so basically, you make your sensor input a low-pass filter
input. RC might be cheap, but LC may do nicer.

2005\02\03@164649 by Ake Hedman

flavicon
face
Jose,

my main problem appeared to be a ground problem just as Jim pointed out. I still need to follow you  advices though for he final card.

I have no decoupling on the sensor lines today. Only on the power to the sensors. The sensor in this project gives out a square wave of 1 kHz - 4 kHz and having a cap. on it will ruin the signal.

Would something like Murata BLM21AG102SN1 (http://www.murata.com/catalog/c31/es0100.pdf) do this job? I know I can test this but just now I feel I need to hold the hand of someone who has done this more than once....

Another path in on the board is the CAN bus which has a very long cable. Should I do the same thing there?

Regards
/Ake

Jose Da Silva wrote:

{Quote hidden}

--  ---
Ake Hedman (YAP - Yet Another Programmer)
eurosource, Brattbergavägen 17, 820 50 LOS, Sweden
Phone: (46) 657 413430 Cellular: (46) 73 84 84 102
Company home: http://www.eurosource.se      Kryddor/Te/Kaffe: http://www.brattberg.com
Personal homepage: http://www.eurosource.se/akhe
Automated home: http://www.vscp.org

2005\02\03@165641 by Ake Hedman

flavicon
face
Thanks Jim that was it! I could not find a bad solder joint but  connected an extra wire form a point on the board to the power supply and that made a big difference. Still have to find out why this is but have an approach.

Also need to look at nice in general but thats another (and later)  question.

Thanks again.
/Ake

Paul James E. wrote:

{Quote hidden}

>>-

2005\02\04@033424 by SO-8859-1?Q?Ruben_J=F6nsson?=

flavicon
face
Hi Ake,

Identify the current path of the noise through your board and either block it or make it as easy for it as possible to pass through your board. It only becomes a problem if the current is converted to a voltage as seen by your circuits (through an impedance of some sort, perhaps a bad ground plane).

Remember that the noise (or transient or surge) wants to go back to it's source and it probably does so through earth. Either in your powersupply (directly or capacitively coupled) or capacitively coupled directly from your board to an earthed groundplane.

One effective way to keep the noise currents away from your board is to put it in a metallic enclosure and have shielded cables where the shield is directly connected to the enclosure through a very good connection such as an EMC cable gland. Connect the shield at both ends. Block the power supply with inductors and decouple with capacitors to the enclosure (not the ground plane) in order to make the impedance for the noise much lower through the enclosure than through your board. You could also decouple the sensor signal to the enclosure.

If you don't have a metallic enclosure, mounting the board on a chassis plate where the shield is connected to the plate and the powersupply decoupled to it, could have the same effect.

If you want to filter the signal directly you have to do this through a good groundplane. Knowing what frequency your board is most sensitive to the noise will help you pick the right filter components. An EMC test will tell you that. Either way, a good very ground plane is always helpful.

Good luck / Ruben

{Quote hidden}

> -

2005\02\04@042028 by Ake Hedman

flavicon
face
Thanks Ruben,

for taking the time to write this information down. Being mostly a software guy myself this is very hard subject  to get a path in to. Most people you ask  However I have decide to learn this now. Your info was  *very* valuable.

Thanks again.
/Ake


Ruben Jönsson wrote:

{Quote hidden}

>>-

2005\02\04@042438 by Jose Da Silva

flavicon
face
On Thursday 03 February 2005 01:46 pm, Ake Hedman wrote:
> my main problem appeared to be a ground problem just as Jim pointed out.
> I still need to follow you  advices though for he final card.

Good to hear.

> I have no decoupling on the sensor lines today. Only on the power to the
> sensors. The sensor in this project gives out a square wave of 1 kHz - 4
> kHz and having a cap. on it will ruin the signal.

A)does it have to be a perfect square wave?   Not likely
B)is it okay to degrade the square wave? Probably. Is 10% okay? Probably.

...shoot for an RC or LC at about 40khz. That shouldn't affect your signal by
very much.  If you can make your RC or LC closer to 20khz, then even better.
Go for a combination of value RC or LC that only affects your circuit up to
1/10 at let's say 8 or 12 khz  (for example, the fluorescent lights are
running at approximately 90khz).

> Would something like Murata BLM21AG102SN1
> (http://www.murata.com/catalog/c31/es0100.pdf) do this job?

Well, look at the values on the spec sheet. You have an inductor of about
10ohms at 100mhz, or 1 ohm at 10mhz or 0.1 ohm at 1mhz. in other words, when
you are dealing with maybe 100khz or 50khz, those beads really aren't doing
more than costing you money. Leave the beads for cell-phone circuits at
900mhz or higher. What you want will be better served with 10 ohm resistors,
easier to find, cheaper too.

Can your circuit handle 100 ohms instead of 10 ohm resistors? Then you will
have better immunity.
What type of circuit do you want to aim for? You mentioned you had to send
vcc to the sensor too, so you may want to aim for something like this:

vcc--100R-+-100R-->to sensor vcc+
         |
        0.5u
         |
     chassisgnd

gnd--100R-+-100R-->to sensor gnd-
         |
        0.5u
         |
     chassisgnd

input-100R-+-100R--<input
          |
         0.25u
          |
      chassisgnd

chassis---------->shield connection(only connect one side, not both)

chassis is the metal box you put your circuit inside of.

Perhaps the above is way overboard versus what you plan on doing, but gives
you a general idea of things.

note, if we substituted a 390uH inductor, it would be roughly 100ohms at
40khz yet neglible at 4khz and far higher at higher frequencies.


> I know I can
> test this but just now I feel I need to hold the hand of someone who has
> done this more than once....

There is no hard/fast/rules I can point you towards.
I don't know anything about your sensor or your circuit.
I don't know if you are running on a battery circuit or an AC circuit.
I don't know if you are going to run this in the middle of a forest with no
local noise, or if you are going to run it within a setting full of
hi-voltage motors generating lots of spark-noise spikes.
I don't know if you will be running this for your personal use, or for dry
arid lighting-frequented areas.
I don't know if you will run it in a noisy environment such as a car
ignition, boat ignition or general house noise or industrial factory
location or close to a radio tower of one sort or another.
Far easier for you to work it out on your end versus me multiple guessing
here.  ;-)
Since you are in Europe (I think), then one or more of these will give you an
idea of what you are guarding your circuits against:
----------------------
Immunity: IEC 1000
EN55024 per EMC Directive 89/336/EEC, including
IEC 61000-4-2 Electrostatic discharge immunity test
IEC 61000-4-3 Radiated, radio-frequency, electromagnetic field immunity test
IEC 61000-4-4 Electrical fast transient/burst immunity test
IEC 61000-4-5 Surge immunity test
IEC 61000-4-6 Immunity to conducted disturbances, induced by radio frequency
fields
IEC 61000-4-8 Power frequency magnetic field immunity test
IEC 61000-4-11 Voltage dips, short interruptions and voltage variations
immunity tests

Line Distortion: EN 61000-3-2 per EMC Directive 89/336/EEC

Voltage Fluctuations and Flicker: EN 61000-3-3 per EMC Directive 89/336/EEC
----------------------
These people would like to have their components installed in your circuit.
Some of the info is useful (TThe IEC61000 info is real-world problems to
consider... perhaps not useful where you live, but will certainly be useful
in a dry-static-lightning environment) this doc seems to give you a general
feel of IEC61000
http://www.semtech.com/pdf/an96-07.pdf

(they also have example circuits).
perhaps a page to start with.... look at the circuits here:
http://www.semtech.com/products/products_line_familydisplay.asp_Q_FamilyID_E_7_A_LineID_E_1

here is another circuit (see top of page 4):
http://www.semtech.com/pdf/stf701.pdf


> Another path in on the board is the CAN bus which has a very long cable.
> Should I do the same thing there?

I have not looked at CAN, however, I would believe some CAN chip
manufacturers will supply real sample circuits which you can model your
circuit on while others may have sample circuits which have never gone
beyond a "paper" example (in other words, not tested in real life).
You'll have to ask someone else with experience in this area then.

Maybe the above is more info than you want, but hopefully useful :-)

cheers!

2005\02\04@043056 by Jose Da Silva

flavicon
face
On Friday 04 February 2005 12:35 am, Ruben Jönsson wrote:

Thanks Ruben, you put it in basic common sense language, far better to understand than what I sent.   :-)

2005\02\04@050441 by Ake Hedman

flavicon
face
Jose,

This was very useful and will get me going. I really needed this because I don't yet have the self confidence when it comes to hardware to think this out for my self but I can see the logic in it now when Ruben and you put me on the right track.

The stuff I am working on is some open hardware /software project for home automation and control  ( http://www.vscp.org/vscp/modules/index.html).  I have been writing hardware near  software since the 80-th  but this is my first public project where  I also do some hardware without having a hardware engineer to blame for bad designed hardware. ;-)  I will use plastic boxes http://www.vscp.org/vscp/modules/kelvin_sht/images/camden.jpg  for the modules. So I have to come up with something else then a chassis for the chassis gnd or maybe ground to the DIN rail will do?.

Thanks again for much appreciated help and info.

Cheers
/Ake


Jose Da Silva wrote:

{Quote hidden}

--  ---
Ake Hedman (YAP - Yet Another Programmer)
eurosource, Brattbergavägen 17, 820 50 LOS, Sweden
Phone: (46) 657 413430 Cellular: (46) 73 84 84 102
Company home: http://www.eurosource.se      Kryddor/Te/Kaffe: http://www.brattberg.com
Personal homepage: http://www.eurosource.se/akhe
Automated home: http://www.vscp.org

2005\02\04@134449 by Peter L. Peres

picon face


On Thu, 3 Feb 2005, Ake Hedman wrote:

{Quote hidden}

Are the boards new (never worked) ? Imho you are looking for a cold
solder joint or cracked part (cap, resistor), likely in the grounds or
decoupling areas. If you have a fox (from a fox&hound set) you can try
injecting signal into the ground at various chips and see if it is
sensitive.

Peter

2005\02\04@173720 by Ake Hedman

flavicon
face
Peter L. Peres wrote:

{Quote hidden}

Yes it looks like that was the problem. at least part of it.

But Pete... a fox, from a fox&hound set???  Can I use one of my cats instead... ;-) What is it?

Regards
/Ake

--  ---
Ake Hedman (YAP - Yet Another Programmer)
eurosource, Brattbergavägen 17, 820 50 LOS, Sweden
Phone: (46) 657 413430 Cellular: (46) 73 84 84 102
Company home: http://www.eurosource.se      Kryddor/Te/Kaffe: http://www.brattberg.com
Personal homepage: http://www.eurosource.se/akhe
Automated home: http://www.vscp.org

2005\02\04@180457 by Dave VanHorn

flavicon
face

>
>But Pete... a fox, from a fox&hound set???  Can I use one of my cats
>instead... ;-) What is it?

No, that would be a cat-scan of course.  :)


2005\02\04@203634 by Jose Da Silva

flavicon
face
On Friday 04 February 2005 03:01 pm, Dave VanHorn wrote:
> >But Pete... a fox, from a fox&hound set???  Can I use one of my cats
> >instead... ;-) What is it?
>
> No, that would be a cat-scan of course.  :)

Did somebody say cat-scan?  ;-)

2005\02\05@065754 by Lee Jones

flavicon
face
> But Pete... a fox, from a fox&hound set???  Can I use one
> of my cats instead... ;-) What is it?

I haven't seen a serious reply on-list.

In United States, fox & hound set is a telephone service term
for a signal generator and inductive amplifier/speaker set.

Fox is sometimes called a flicker set (because the signal has
two tone components that sounds like  wedo wedo wedo wedo  so
it flickers between the two frequencies) or a tone generator.

The hound may also be called a banana, as one major manufacturer
of them makes them in bright yellow and the shape is similar.
It's also called a tone tracer.

Tone generator is clipped to one end of pair of wires in a
bundle of dozens or hundreds of pairs.  At far end, you use
the inductive amplifier to figure out which cable contains
that wire pair by just getting the amplifier's tip close to
the cable containing the pair.

For example, you've got dozens of Ethernet UTP cables coming
from wallplates all over an office.  All the cables appear in
a central wiring closet.  Previous installer did not label
anything, just punched the cables up on a patch panel.  Inject
signal at a wallplate jack; find the patch panel cable using
the inductive amp inside the wiring closet.

In a perfect world, you would not need the set at all since
all of the cables would be properly labeled during installation
and there would never be shorts/cuts/breaks/etc in your cable
plant.  In the real world, I always carry a tone generator and
tracer set with me when going to a customer site.

                                               Lee Jones

2005\02\06@040213 by Peter L. Peres

picon face


On Fri, 4 Feb 2005, Ake Hedman wrote:

> But Pete... a fox, from a fox&hound set???  Can I use one of my cats
> instead... ;-) What is it?

You can use the cat only if you are cruel. The fox&hound set is a set of
2 devices, one of which injects a signal into a circuit and the other
acts as a detector for it. They exist in many variations. The simplest
would be an audio generator as fox and a high impedance headphone for
hound, to debug audio circuits. Equivalents exist for audio, pulse
circuits, transmission line circuits and even for tracing buried pipes.

Peter

2005\02\06@062805 by Ake Hedman

flavicon
face
Peter L. Peres wrote:

>
>
> On Fri, 4 Feb 2005, Ake Hedman wrote:
>
>> But Pete... a fox, from a fox&hound set???  Can I use one of my cats
>> instead... ;-) What is it?
>
>
> You can use the cat only if you are cruel. The fox&hound set is a set
> of 2 devices, one of which injects a signal into a circuit and the
> other acts as a detector for it. They exist in many variations. The
> simplest would be an audio generator as fox and a high impedance
> headphone for hound, to debug audio circuits. Equivalents exist for
> audio, pulse circuits, transmission line circuits and even for tracing
> buried pipes.
>
> Peter

OK. Pete I got it explained to me off list also. Had never heard the device being named that before even is seeing theme here as well.

Do you say a (my)  signal generator will be suitable for this?

Regards
/Ake

--  ---
Ake Hedman (YAP - Yet Another Programmer)
eurosource, Brattbergavägen 17, 820 50 LOS, Sweden
Phone: (46) 657 413430 Cellular: (46) 73 84 84 102
Company home: http://www.eurosource.se      Kryddor/Te/Kaffe: http://www.brattberg.com
Personal homepage: http://www.eurosource.se/akhe
Automated home: http://www.vscp.org

2005\02\07@020314 by Peter L. Peres

picon face

On Sun, 6 Feb 2005, Ake Hedman wrote:

{Quote hidden}

Yes, you make a high impedance injector (1kohm resistor in your case)
and inject signal into the circuit. You know where the cicruit is not
supposed to be sensitive at all (f.ex. supply rails) so if you inject
there and the circuit fails then you know something is amiss, like bad
grounds etc. Alternatively, you inject into the signal path and look at
the signal at the output with suitable tools (scope etc).

Peter

More... (looser matching)
- Last day of these posts
- In 2005 , 2006 only
- Today
- New search...