EmbeddedRelated.com
Forums

Ethernet transformwer with 4kV isolation

Started by Joe G (Home) March 28, 2011
On Tue, 29 Mar 2011 18:09:10 -0700, Joerg <invalid@invalid.invalid>
wrote:

>But I am not sure your 240VAC floating setup is legit in the first place.
What so special about that ? AC/DC radios for 220/240 V mains had series capacitors on the antenna and antenna ground terminals rated at 3750 Vdc. Tube televisions for 220/240 V mains used a half wave rectifier and hence the internal chassis was either at the neutral potential or riding at full phase voltage (220/240 V). For this reason, the number of external interfaces was very limited. On the antenna side it was quite easy to have sufficient isolation with heavily insulated wires on different holes on the ferrite core. The external loudspeaker was also quite easy to arrange, since an audio output transformer was needed anyway. Feeding early computer games or teletext into the video chain of such AC/DC receiver required some optoisolators (4N25 etc.). Various standards require 1500 V or 2500 V (typically DC for at least a minute). The reason for such high voltages (compared to the RMS voltage) is that such peak voltages really exists in a typical mains powered circuit. A steep pulse current in the live conductor will generate also a significant voltage drop in the neutral connector due to the neutral wire inductane/impedance and hence the voltage can be significantly higher than any grounded potential. For the OP, the simplest thing would be to use some standard Ethernet/fibre converters and use 0.1 - 1 m of fiber between the grounded electronics and the other system riding on 240 V mains.
<upsidedown@downunder.com> wrote in message 
news:4dv6p61ld7c9agc8j9i5kgsptef3m61be0@4ax.com...
> On Tue, 29 Mar 2011 18:09:10 -0700, Joerg <invalid@invalid.invalid> > wrote: > >>But I am not sure your 240VAC floating setup is legit in the first place. > > What so special about that ? > > AC/DC radios for 220/240 V mains had series capacitors on the antenna > and antenna ground terminals rated at 3750 Vdc. > > Tube televisions for 220/240 V mains used a half wave rectifier and > hence the internal chassis was either at the neutral potential or > riding at full phase voltage (220/240 V). For this reason, the number > of external interfaces was very limited. On the antenna side it was > quite easy to have sufficient isolation with heavily insulated wires > on different holes on the ferrite core. The external loudspeaker was > also quite easy to arrange, since an audio output transformer was > needed anyway. > > Feeding early computer games or teletext into the video chain of such > AC/DC receiver required some optoisolators (4N25 etc.). > > Various standards require 1500 V or 2500 V (typically DC for at least > a minute). The reason for such high voltages (compared to the RMS > voltage) is that such peak voltages really exists in a typical mains > powered circuit. > > A steep pulse current in the live conductor will generate also a > significant voltage drop in the neutral connector due to the neutral > wire inductane/impedance and hence the voltage can be significantly > higher than any grounded potential. > > For the OP, the simplest thing would be to use some standard > Ethernet/fibre converters and use 0.1 - 1 m of fiber between the > grounded electronics and the other system riding on 240 V mains.
Yep, It's all about how you interpret the safety standard. I was attacking it from the point of view that there is already a transformer so lets just a high isolated one. There other places in the circuit where the safety barrier can be out. By the way are you in OZ Joe from Melb.
On Thu, 31 Mar 2011 21:41:56 +1100, "Joe G \(Home\)"
<joe.g@optusnet.com.au> wrote:

> ><upsidedown@downunder.com> wrote in message >news:4dv6p61ld7c9agc8j9i5kgsptef3m61be0@4ax.com...
>> For the OP, the simplest thing would be to use some standard >> Ethernet/fibre converters and use 0.1 - 1 m of fiber between the >> grounded electronics and the other system riding on 240 V mains. > >Yep, It's all about how you interpret the safety standard. > >I was attacking it from the point of view that there is already a >transformer so lets just a high isolated one. > >There other places in the circuit where the safety barrier can be out.
If you are making a low volume product and intend to sell it all over the world, in practice, the simplest thing is to do some "overenginering" and in this case use some fiber, even if you could get rid of simple optoisolators in a particular market.
>By the way are you in OZ
Nope, sorry too disappoint you, but your guess is off by nearly 20000 km :-).
Jim Stewart wrote:
> Joerg wrote: > >> You can get medical grade isolation transformers for Ethernet. Not sure >> how much this one costs but they are usually well north of $100: >> >> http://www.baaske-medical.de/product_info.php/info/p478_Netzwerk-Isolator-MED-MI-1005.html/XTCsid/uclhpdj0bctjeufg6sf7pkjkn4 >> >> >> You'd have to ask them where they can be bought in Australia. Or maybe >> you guys have local suppliers for such transformers. The IT department >> of a large hospital might know. >> >> But I am not sure your 240VAC floating setup is legit in the first place. > > IMHO, I don't think it would pass UL, CE or CSA without > approved high side components and creepage distances. >
Obviously, it has passed: http://www.baaske-medical.us/media/content/Productinfo_MI1005_US.pdf?xa7ec4=1770b39dc5d6e4a07a37064d3877c62f Quote: "The MI 1005 is approved to following standards: DIN EN / IEC 60601-1:2007 (3rd Ed), DIN EN / IEC 60601-1-2:2007, UL 60601-1" If in doubt about the cert of a part ask the vendor for a copy. I always got that upon request.
> Then there's the issue of the cable and connector insulation > and touch distances...
Well, yeah, of course your installation is not supposed to compromise the isolation provided by the device. The same goes for just about any part with a test certificate. -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.
upsidedown@downunder.com wrote:
> On Tue, 29 Mar 2011 18:09:10 -0700, Joerg <invalid@invalid.invalid> > wrote: > >> But I am not sure your 240VAC floating setup is legit in the first place. > > What so special about that ? >
The part needs to be rated for that purpose. It probably is, but if it goes into a product I would not release it without written confirmation by the vendor that it's ok. This goes for all parts, including the supply to that device.
> AC/DC radios for 220/240 V mains had series capacitors on the antenna > and antenna ground terminals rated at 3750 Vdc. >
I've repaired a lot of those. Not all had that rating on the caps. But I bet the mfgs all had copies of the certs for all parts that crossed that barrier, locked away in steel cabinets :-)
> Tube televisions for 220/240 V mains used a half wave rectifier and > hence the internal chassis was either at the neutral potential or > riding at full phase voltage (220/240 V). For this reason, the number > of external interfaces was very limited. On the antenna side it was > quite easy to have sufficient isolation with heavily insulated wires > on different holes on the ferrite core. The external loudspeaker was > also quite easy to arrange, since an audio output transformer was > needed anyway. > > Feeding early computer games or teletext into the video chain of such > AC/DC receiver required some optoisolators (4N25 etc.). > > Various standards require 1500 V or 2500 V (typically DC for at least > a minute). The reason for such high voltages (compared to the RMS > voltage) is that such peak voltages really exists in a typical mains > powered circuit. >
There you said it, for one minute. Huge difference, as HV barriers can "age" with application of HV over time. Some do, some don't. Therefore, I can only repeat my advice: Get it in writing.
> A steep pulse current in the live conductor will generate also a > significant voltage drop in the neutral connector due to the neutral > wire inductane/impedance and hence the voltage can be significantly > higher than any grounded potential. > > For the OP, the simplest thing would be to use some standard > Ethernet/fibre converters and use 0.1 - 1 m of fiber between the > grounded electronics and the other system riding on 240 V mains. >
Fiber is one of the safest method. But can get expensive. -- Regards, Joerg http://www.analogconsultants.com/ "gmail" domain blocked because of excessive spam. Use another domain or send PM.