EmbeddedRelated.com
Forums
Memfault Beyond the Launch

EMI: which controller is 'immune'

Started by Herbert Demmel March 21, 2006
Hello everbody,

I used ST6 microcontrollers in some projects without any problems 
regarding ElectroMagnetic Inference. ST6 family was designed to operate 
properly in bad environment (like car)

Now I have to switch to another micro (ST7, AVRmega, PIC, ARM7)

My customer told me of problems (hanging software, reset, latch-up or 
similar) in an application with some electrostatic discharge.

Which family do you suggest?

Thank you, Herbert
Much more important than which processor you choose, is the design
around it - good interfaces & filtering to prevent EMI issues getting
to the processor, and causing these "glitches".

"Herbert Demmel" <programmer@demmel-m.de> wrote in message
news:Xns978DD6386349Cprogrammerdemmelmde@ID-41558.user.uni-berlin.de...
> Hello everbody, > > I used ST6 microcontrollers in some projects without any problems > regarding ElectroMagnetic Inference. ST6 family was designed to operate > properly in bad environment (like car) > > Now I have to switch to another micro (ST7, AVRmega, PIC, ARM7) > > My customer told me of problems (hanging software, reset, latch-up or > similar) in an application with some electrostatic discharge. > > Which family do you suggest?
Any family will do. All the problems you mention are caused by improper hardware design (PCB layout, filtering). I have made a small board (90 x 66mm) with an ATmega162. Four inputs with optocouplers, three outputs (RS422, RS232) with rc filters (10 Ohm, 10nF) and one 4.7uH common mode filter on the power supply. The PCB is double layer and most of the bottom layer is groundplane. Where the groundplane is open, there is no power supply trace on the top layer and the larger openings are never under an IC. This board passed all IEC945 tests, which included RF field 10V/m from 80MHz to 1GHz, conducted RF of 3Vrms from 150kHz to 80MHz on all ports, fast transients on all ports (1.2kV, 50ns pulses, at 5kHz, 15ms bursts every 300ms) and ESD at 8kV contact discharge and air discharge. No problem detected. Meindert
On Tue, 21 Mar 2006 22:35:50 +0100, in comp.arch.embedded "Meindert
Sprang" <ms@NOJUNKcustomORSPAMware.nl> wrote:

>"Herbert Demmel" <programmer@demmel-m.de> wrote in message >news:Xns978DD6386349Cprogrammerdemmelmde@ID-41558.user.uni-berlin.de... >> Hello everbody, >> >> I used ST6 microcontrollers in some projects without any problems >> regarding ElectroMagnetic Inference. ST6 family was designed to operate >> properly in bad environment (like car) >> >> Now I have to switch to another micro (ST7, AVRmega, PIC, ARM7) >> >> My customer told me of problems (hanging software, reset, latch-up or >> similar) in an application with some electrostatic discharge. >> >> Which family do you suggest? > >Any family will do. All the problems you mention are caused by improper >hardware design (PCB layout, filtering). I have made a small board (90 x >66mm) with an ATmega162. Four inputs with optocouplers, three outputs >(RS422, RS232) with rc filters (10 Ohm, 10nF) and one 4.7uH common mode >filter on the power supply. The PCB is double layer and most of the bottom >layer is groundplane. Where the groundplane is open, there is no power >supply trace on the top layer and the larger openings are never under an IC. >This board passed all IEC945 tests, which included RF field 10V/m from 80MHz >to 1GHz, conducted RF of 3Vrms from 150kHz to 80MHz on all ports, fast >transients on all ports (1.2kV, 50ns pulses, at 5kHz, 15ms bursts every >300ms) and ESD at 8kV contact discharge and air discharge. No problem >detected. > >Meindert >
Excellent. Would it be possible to see the layout of the board, possibly on a.b.s.e? How many versions did it take you to get it correct? martin
"martin griffith" <mart_in_medina@yahoo.esXXX> wrote in message
news:e60122tf96nimev5aq2ujhgcds1c37mgse@4ax.com...
> Would it be possible to see the layout of the board, possibly on > a.b.s.e?
Yep, it's there as "Board MPX-SB"
> How many versions did it take you to get it correct?
Only 2. It was a design we already had, which complied with EN61000-6-1 and EN61000-6-3 (EMC, domestic/light industrial). One large customer required EN60945/IEC945, so they tested it and we added one W&#4294967295;rth common mode filter (rectangle marked L1 in the bottom right corner) in the power supply to meet the requirements. EN60945 requires very high surpression of RF emissions on certain marine frequencies. Meindert
On Wed, 22 Mar 2006 07:33:54 +0100, in comp.arch.embedded "Meindert
Sprang" <ms@NOJUNKcustomORSPAMware.nl> wrote:

>"martin griffith" <mart_in_medina@yahoo.esXXX> wrote in message >news:e60122tf96nimev5aq2ujhgcds1c37mgse@4ax.com... >> Would it be possible to see the layout of the board, possibly on >> a.b.s.e? > >Yep, it's there as "Board MPX-SB" > >> How many versions did it take you to get it correct? > >Only 2. It was a design we already had, which complied with EN61000-6-1 and >EN61000-6-3 (EMC, domestic/light industrial). One large customer required >EN60945/IEC945, so they tested it and we added one W&#4294967295;rth common mode filter >(rectangle marked L1 in the bottom right corner) in the power supply to meet >the requirements. EN60945 requires very high surpression of RF emissions on >certain marine frequencies. > >Meindert >
Thanks. It hasnt turned up on http://www.usenet-replayer.com/cgi/content/archive yet. I'll keep trying martin
"martin griffith" <mart_in_medina@yahoo.esXXX> wrote in message
news:5r2222l8lilsk8po8fbhd93mvk87ua4asf@4ax.com...
> Thanks. It hasnt turned up on > http://www.usenet-replayer.com/cgi/content/archive > > yet. > I'll keep trying
I think you'd better access a newsserver directly. Much faster. Meindert

Memfault Beyond the Launch