How to separate two GND in Proteus

[Chriss]

Hi!
I run into an issue with Proteus 8 on my simulation when I try to use the scope.

I designing a circuit where I have a AC main high voltage side separated from a low voltage site.
These two circuit's are separated with an optocoupler.

When I use my scope I have some wrong measurement's.

My major question is:

When I have a simple circuit with an AC source, do I have to connect it to a GND?
When I disconnect GND from the high voltage side, Proteus can't simulate, running into error.

Actually, how can I tell the scope not to use the DC/low voltage side GND as a reference when I scoping the High voltage?

thanks.

[PlainName]

It needs to be referenced against something, which is typically gnd. Unlike real life, you won't blow the scope up if you shove 240V up the ground pin

[Chriss]

Yea, you are right, but if I put a optocoupler MOC3021 and the low side is a DC circuit, the high side ( where the diac is ) is an AC circuit.
If I didn't put a gnd to the high side and try to simulate, I get always an error like some timestep error.

If I put a gnd to the low and high side, I have a bad scoping signal.
I assume the two gnd's are act as the same ref.point.
Like they are in real the same gnd line.

In real life these two gnds have to be separated.
How can I achieve this in Proteus 8?

Thanks.

[PlainName]

I am not sure how you are doing this: are you treating AC1 and AC2 as the signal and ground to the scope? Or are you having AC zero crossing at scope ground and the signal either AC1 or AC2? It sounds to me like the former, and that won't work because the AC signal isn't referenced to anything.

The latter, where AC1 and AC2 both swing positive and negative (alternately, of course) relative to ground, is what you want, I think. So far as I know, you can't have a floating signal because, well, it's floating and not referenced to anything. If you don't want the AC to swing below reference ground you can add a DC offset to it which shouldn't affect how your circuit operates if the two sides are really separate.

Can you post your circuit? A printout would be best since not everyone has your version of Proteus (or even any version).

[Chriss]

So, here is a really basic schematic where I also have problem to simulate with Proteus 8.

I posted the "project" if we can call this so. 
It is just to show one of the problem I have when I try to use AC and DC side in the same schematic in Proteus 8.

When I try to simulate the circuit without the GND connected on the AC side, then if I open the switch I get an error
like on the screen. I dont really understand what is that error meaning.

If I connect the GND on the AC side, then everything is working, no error, but on a more complexe schematic I got a signal distorsion too.

So, I'm not sure what is wrong and how to really separate the AC and DC ground's.
Or how to take measurement with the scope but to be suere the scope is referencing to the AC GND if I measure the AC side etc.

Thanks for any advice.

[PlainName]

You need the ground for reference. If you remove the DC side and replace the opto with the switch, you get the same issue, so it's not a DC-AC thing, just an unreferenced AC thing.

The bottom side of the AC generator isn't a reference - it's just the inverted version of the other side. A signal, in other words. You can connect the top side to ground instead, and the circuit works fine, which shows that both sides are technically signals. So without ground the circuit is floating and the simulator doesn't know what to do with it.

But that's my view. I may be wrong

[Chriss]

Yea, I come also acros that thinking, but how can we simulate a real AC DC circuit in Proteus?

What power source is a correct AC power source in Proteus?

This is all so confusing for me in Proteus...

[PlainName]

With it ground referenced, you are doing a proper DC/AC simulation. It's just that the AC isn't floating. You can add a DC offset if you want to have it going actually 120V negative as well as 120V positive.

[Chriss]

It is completely crazy in some meaner. 

Here is probably the stupidest circuit in the world and the simulation is a mess.

I put a bridge rectifier grounded and without ground.
The result is totally crap.

Ok, I can live with this cos I know why this is but I didn't expect from a simulation app this kind of behavior on this under the basic level.

I'm asking myself, what kind of error are there when we try to simulate more complex circuits?

Or is there a way how to do this circuit setup and measuring in a correct way?

What power source do you use in your circuit when you deal with AC current?
I saw people use the VSINE part, but other use the ALTERNATOR part.

Look the picture how Proteus simulate a bridge rectifier...

[PlainName]

I think those traces are appropriate for the circuits.

In the first (grounded) the scope reference is the lower AC line, so chB will show a sine wave (the AC waveform). The bridge is grounded on the negative side, so you will get a half-wave on chC. However, you might normally expect that be a rectified wave of double the AC frequency, but recall that here the lower AC line is grounded, so only the upper line is driving the rectifier and it never goes negative. Thus the chopped half-wave.

In the second (un-grounded) the scope reference is the bridge negative terminal. Only the AC upper line can drive it, and only when it is positive with reference to both the lower line and ground (because the lower line is shorted through the bridge to ground. Again, a chopped half-wave is appropriate, I think. The scope chC will be a rectified sinewave, double the AC frequency, which is what we see.

What we can't see, possibly because of the range settings, is the effect the bridge diodes have. But that's not surprising since the AC is 240V and the diodes are 0.7V or so, and pretty much invisible unless you zoom in.

[Chriss]

Here I did a similar circuit testing but in Multisim.
Here is the result, I can not achieve this in Proteus 8.

I think, the Multisim simulation is more accurate to the real life situation.

Check the picture I made from Multisim.

[PlainName]

Can't comment on relative accuracy, but I note the multisim circuit has the ground you complain about in the Proteus circuit.

[Doctorandus_P]

It's a common limitation of simulation packages.
A simple workaround, is to have only one "real GND", and give the other one some other name. Also connect them with a 1GOhm resistor or similar, so there are no "floating nodes".

[Chriss]

Thank for this tip!
I like the idea.

Btw.
I saw a video on the net, where a guy used two types of GND in his circuit in Proteus.

Are they insulated?
Where can I find that other EARTH sign in Proteus like in this circuit?

[woofy]

Where can I find that other EARTH sign in Proteus like in this circuit?

[Chriss]

WoooW!
I discovered something new in my Proteus 8!
Look into the pic.

I do not have all the terminals in my list like you guys!

what is wrong on my side?

[Chriss]

Nobody has any idea why in my Proteus is missing the Chassie terminal simbol? 

[PlainName]

Sorry, not a clue. Wonder if it depends on which level of Proteus you have. In any case, you can add a new one yourself - it only wants a name and net label.

[Chriss]

Can you pls help me how to add a ne one?
If I add a new GND simbol to my Terminal list, will that gnd be separated from any other gnd in the schematic?
Would that gnd be a reference gnd for only in the circuit where it is connected?

I use Proteus 8 Professional.

[PlainName]

From the help file:

ISIS permits the definition of user defined symbols for use as logical or physical terminals. These are made in the same way as ordinary symbols except that you must place a Node marker to specify the position of the terminal's connection point, and a Label marker to specify the position and orientation of its net label.

To make a user defined terminal

* Select an appropriate 2D graphics icon - typically the Line icon - from the Mode Selector toolbar

* In the Object Selector select an appropriate graphics style. This will nearly always be the TERMINAL style though the BUS WIRE style may also be appropriate for small parts of the symbol.

*Select and place graphics objects as required to form the body of the terminal. Any graphic that needs to have a fixed appearance should be edited, the appropriate Follow Global? checkboxes unchecked and the graphics style attribute changed.

*Select the Markers icon.  Place a Node or Busnode marker where you want the wire or bus to connect the terminal, and a Label marker where you want its net label to be. You can also place an Origin marker to define where its origin will be.

* Tag the objects that will comprise the port by by entering selection mode and dragging a tag-box around them with the left mouse button.

* Invoke the Make Symbol command, set the type to Terminal and then select a name and library for the new terminal.
Click OK to complete the operation.

Or, to put it another way, go through the motions of creating a symbol (just draw it with a line of style TERMINAL), add a MARKER where you want the wire to connect, select them all, right click and select "Make Symbol" and make sure it's of type TERMINAL. Very similar to making a component except you use the TERMINAL line style (I imagine you could use something else if you wanted, but TERMINAL has specific colour and thickness properties) and don't have any pins.

Strangley, when I looked at this yesterday I could right click in the terminals tray and select 'create'. No idea why that doesn't work for me now.