Forums

AC spike go pass XFRM and regulator to reset CPU [Part 2]: It is irradiated?

Started by Rodo July 1, 2008
Hi all,

A while back I posted a problem I was having with a spike (see original text 
below). Someone suggested that the spike would be in the ground line as 
well. So I connected the scope probe and ground to the ground of the circuit 
and yeap, there it was. I felt rather silly connecting the scope probe and 
ground together but oh well ...
This week I disconnected the probe but left it next to the circuit and I 
noticed that I also got the spike. So I thought it must be getting in the 
scope from the AC ground. So ... for some reason I move the probe on top of 
the scope which sits in a higher shelf and repeated the test. I got no 
spike. Several tries show the spike if the scope probe is close (1 or 2 
inches) to the circuit but it goes away if I place the probe further from 
it. So it looks like the spike is being generated and transmitted (for lack 
of a better word) everywhere to the circuit?  Is this something that would 
require more ground to shield the PIC?

Comment/ suggestions welcome .... this is driving me nuts :-(.

Thanks in advance.



---------------- Original message
I have a device that is powered by 240VAC (also works from 120VAC). The AC 
drives the primary of a transformer (with a fuse and a MOV). There is also a 
fan connected to the 240VAC with a switch. The secondary of the transformer 
has a bridge, a large electrolytic cap (1000uF/25v) and an LM7805 regulator. 
There is a smaller cap (10uF/10v and 01uF/25v) at the output of the LM7805. 
Then the 5vdc connects to a PIC.

Almost every time I flip the fan switch a spike shows up at the output of 
the regulator. This is the part that I mostly do not get: even if I 
disconnect the fan and I flip the switch... the spike shows up ... and it is 
a bit larger in amplitude. I grabbed a picture of the spike ... you can see 
it at: http://mysite.verizon.net/rodo/ds0000.bmp

The spike is obviously not always exactly the same. Sometimes is a bigger, 
or smaller more or less oscillations. But the general timming is the same. 
By this I mean I do not have to change the scope's setting to see it.

I tried to: add caps, common mode choke, remove ground (from circuit and 
scope), add larger cap at the output of LM7805, etc., to find why is there a 
spike when there is no load (I removed the fan remember). The ultimate thing 
is that I need to avoid the spike from occuring (with or without load) 
because it is reseting the CPU.

Could someone enlighten me or point me in the right direction please?

Thank you
----------------- End original message


Rodo wrote:

> Hi all, > > A while back I posted a problem I was having with a spike (see original > text below). Someone suggested that the spike would be in the ground > line as well. So I connected the scope probe and ground to the ground > of the circuit and yeap, there it was. I felt rather silly connecting > the scope probe and ground together but oh well ... > This week I disconnected the probe but left it next to the circuit and > I noticed that I also got the spike. So I thought it must be getting in > the scope from the AC ground. So ... for some reason I move the probe > on top of the scope which sits in a higher shelf and repeated the test. > I got no spike. Several tries show the spike if the scope probe is > close (1 or 2 inches) to the circuit but it goes away if I place the > probe further from it. So it looks like the spike is being generated > and transmitted (for lack > of a better word) everywhere to the circuit? Is this something that > would require more ground to shield the PIC? > > Comment/ suggestions welcome .... this is driving me nuts :-(. > > Thanks in advance.
Scope probes are usually very high impedance and it is not unusual for an open probe to pick up noise from everywhere. You need a sound approach to exploring this phenomena. Are you trying to do this exploratory work in a lab or is the environment you are in the shop floor? This may have a bearing on the noise you see. If you are working in the lab is the ground in the lab a very low impedance. I have had substantial copper busbar fitted around all the sockets in a lab to ensure a very low impedance ground network. Check your scope's grounding is secure (for safety reasons) and as low impedance as possible. You should also probably use a LISN (see http://en.wikipedia.org/wiki/LISN) to help with performing measurements. If you connect the probe to a small coil (just a few turns of stiff wire wrapped around a 4mm diameter former) you could hunt for the localised source of the spikes (in a non-contact way). -- ******************************************************************** Paul E. Bennett...............<email://Paul_E.Bennett@topmail.co.uk> Forth based HIDECS Consultancy Mob: +44 (0)7811-639972 Tel: +44 (0)1235-811095 Going Forth Safely ..... EBA. www.electric-boat-association.org.uk.. ********************************************************************
"Paul E. Bennett" <Paul_E.Bennett@topmail.co.uk> wrote in message 
news:6d1sntFe5u3U1@mid.individual.net...
> Rodo wrote: > >> Hi all, >> >> A while back I posted a problem I was having with a spike (see original >> text below). Someone suggested that the spike would be in the ground >> line as well. So I connected the scope probe and ground to the ground >> of the circuit and yeap, there it was. I felt rather silly connecting >> the scope probe and ground together but oh well ... >> This week I disconnected the probe but left it next to the circuit and >> I noticed that I also got the spike. So I thought it must be getting in >> the scope from the AC ground. So ... for some reason I move the probe >> on top of the scope which sits in a higher shelf and repeated the test. >> I got no spike. Several tries show the spike if the scope probe is >> close (1 or 2 inches) to the circuit but it goes away if I place the >> probe further from it. So it looks like the spike is being generated >> and transmitted (for lack >> of a better word) everywhere to the circuit? Is this something that >> would require more ground to shield the PIC? >> >> Comment/ suggestions welcome .... this is driving me nuts :-(. >> >> Thanks in advance. > > Scope probes are usually very high impedance and it is not unusual for an > open probe to pick up noise from everywhere. You need a sound approach > to exploring this phenomena. > > Are you trying to do this exploratory work in a lab or is the environment > you are in the shop floor? This may have a bearing on the noise you see. > > If you are working in the lab is the ground in the lab a very low > impedance. I have had substantial copper busbar fitted around all the > sockets in a lab to ensure a very low impedance ground network. Check > your scope's grounding is secure (for safety reasons) and as low > impedance as possible. You should also probably use a LISN (see > http://en.wikipedia.org/wiki/LISN) to help with performing measurements. > > If you connect the probe to a small coil (just a few turns of stiff wire > wrapped around a 4mm diameter former) you could hunt for the localised > source of the spikes (in a non-contact way).
I'm in the production area but the spike only happens when I hit the sw. I'll try to hunt the spike with a coil as you suggested next week. Thanks
On Thu, 03 Jul 2008 06:44:01 GMT, "Rodo" <noway@youwish.com> wrote:

> >"Paul E. Bennett" <Paul_E.Bennett@topmail.co.uk> wrote in message >news:6d1sntFe5u3U1@mid.individual.net...
>> Scope probes are usually very high impedance and it is not unusual for an >> open probe to pick up noise from everywhere. You need a sound approach >> to exploring this phenomena. >> >> Are you trying to do this exploratory work in a lab or is the environment >> you are in the shop floor? This may have a bearing on the noise you see. >> >> If you are working in the lab is the ground in the lab a very low >> impedance. I have had substantial copper busbar fitted around all the >> sockets in a lab to ensure a very low impedance ground network. Check >> your scope's grounding is secure (for safety reasons) and as low >> impedance as possible. You should also probably use a LISN (see >> http://en.wikipedia.org/wiki/LISN) to help with performing measurements. >> >> If you connect the probe to a small coil (just a few turns of stiff wire >> wrapped around a 4mm diameter former) you could hunt for the localised >> source of the spikes (in a non-contact way). > >I'm in the production area but the spike only happens when I hit the sw. >I'll try to hunt the spike with a coil as you suggested next week.
It sounds more that you have some kind of capacitively coupled interference. A mains transformer might have a 10-100 pF stray capacitance between primary and secondary and when you touch tho fan switch with your hand or let current flow into the wiring to the fan (even if the actual fan is disconnected), you are introducing similar capacitive coupling. Also the probe 3-5 cm from the PCB might cause some capacitive coupling. Such hum problems were common with audio gears in the past (especially with tube gear) or with very high impedance measurement systems. In a bad audio setup you could get a bad hum problem when touching any metallic part of an electric guitar, which was in contact with the cable shield. The capacitively coupling between the stage lights and your body allowed sufficiently current to flow and cause hum problems. Using proper cables and connectors this problem could be rectified. I have never heard of anyone having similar problems with some microprocessor card, but the layout must be very bad :-). A solid ground plane usually solve these kind of problems. The old trick to reduce mains hum and other higher frequency noise from entering from the primary side to he secondary side on a transformer was to use a static shield between primary and secondary and connect it to ground. When making a static shield it is very important that you do not create a single turn short circuit :-). Paul