GND connection can reset MCU? How?

gnd computer

Hello, I have an MCU that controls many (20) relays. Most of times that I connect GND of multimter or osciloscope(both have 220V supplies) to GND of my circuit, I hear lots of opening and closing in relays suggesting that MCU is reseted in short intervals. This happens even before that I connect the other lead (+ lead) to measure voltage of a point. after about a seconds this opening and closing will end. Another clue is that when I connect one multimeter and after opening and closing of relays end, then connecting GND of another multimeter won't cause such resets. can anybody explain this.
Thank you.

usb surge earth connection

That could simply be noise.
I mean these are things that might need careful study of the situation.sadly true is these conditions cant have a theroritical proof..
you know random conditions ??
like the multimeter can act like a capacitor bwn the lines.. so puttin another one will not have any more effects...
its taking time to settle because it might be charging the cap...a deliberate capacitor might help.. and again it might have some other effects also..

gnd+computer

To avoid situation like this you should have two separate voltages: one (can be "dirty", say, 12V for relays and second - "clean and regulated" for MCU ..
Also, it is good idea to have optocouplers between the MCU and relay drivers ..
This just isolates grounds ..

You may try to connect EARTH to the MCU's ground ..

Regards,
IanP

mcu relay reset

Thanks cameo_2007 and IanP, The ground and earth are separated, Earth is driven by an OP27 that has a 100Ohm series resistor to GND voltage. Because also the circuit has sensitive analog section, I'm not sure about effect of conecting earth to digital ground(though analog ground and digital ground are separated connecting at just one point). About capacitor well it could be the cause. Also one imporatant fact about the circuit that I forgot to mention in the original post is that computer GND also is connected to metal enclosure of the circuit(they are connected via USB) but GNDs of circuit and computer(USB) are optically isolated. The computer usually has no earth connection(because end user ususally do not connect computer earth). Though multimeter and osciloscope also do not have earth connection. The circuit metal enclosure has earth connection though earth is not connected to circuit GND but circuit drives the earth via OP-AMP as explained. If the problem is that MCU is resetted upon connection of multimeter GND, how optoisolation or having separate voltage for relays and MCU may prevent such events, because resetting MCU will open and close relays anyhow.

reset mcu

I have seen things like this before. In the real world, there is no such thing as totally isolated. Every surface has an electrical potential to every other surface in the universe. When you attach the ground wire, you rapidly bring the two surfaces to approximately the same potential. The rapid shift causes a current spike across your ground plane. When spikes like this flow under your microcontroller or your reset circuit, the processor can get scrambled and will often run away and reset.

How does one prove this is what is happening? Before connecting the scope ground, connect the scope ground with a high value resistor like 10 MEG. This allows potential to equalize slowly without spikes. Then after the potentials are the same, jumper out the high value resistor to make a direct ground connection. If this prevents the relays from clicking then you know the reason.

Hello Banjo, Thanks. I suspected that surge current could be the cause. It is reasonable. I remember once I worked with a circuit that had this problem. In that circuit computer had not been made optically isolated and computer had not been connceted to earth. Once, for a moment I accidentally connected earth to GND of circuit. The result was catastrophic, apart from a few burnt ICs a 50mil track to computer GND had been disrupted! I will test your suggestion. Thanks.

surge current is different.It happens only when the power supply is first applied.
one possible reason for that earth-ground incident,may be 'coz the earth must not be properly earthed and there must've been some voltage in that and screwed the ckt gnd.

Hello cameo_2007, about the incident that I explained, I think the reason was because there was substantial voltage between earth and circuit GND and when I accidentally connected earth to circuit GND, computer drew huge current that led to disruption of a thick track. I'm not still 100% sure that surge current could be the cause of MCU resets but it is a candidate.One of the users of this circuit claims that even touching the circuit GND with bare hand occasionally can cause relay switching(possibly because of MCU reset) but I have not been there and can't verify it for sure but if we assume that the user is right then the reason can't be surge current because amount of current that can be injected by touching circuit GND is very low and can't be considered surge and lead to MCU reset.

the compueter doesnt draw huge amount of current just like that.if it had,then your computer hardwae shud also fail.here i think there mustve been some amount of voltage,maybe more than abt 100 or 200 and the track couldnt withstand and burnt off.could be the reason to my knowledge,i'm not claiming it.
and i already said,surge current is entirely different,it comes only while starting the board,and now way the reason for resetting the micro.
the reason i said,could be simply be noise and comes in random conditions.i donno any more details on that.try getting expert helps there study the conditions practically. and one way to solve this problem is simple,,just dont touch the micro or the board while its on