A stupid question about Sine Generator in Spice/LTSpice

Does your transformer model have any DC resistance? If not, a non-zero sine angle can make persistant DC circulating current. That's not usually a problem in real life.

Add a big series capacitor?

Den mandag den 1. august 2016 kl. 23.43.51 UTC+2 skrev M Philbrook:

for 90 deg you want the sine to start a quarter cycle later instead of starting at 0 with peak voltage?

-Lasse

For the case I was doing, driving a 3 phase transformer and having DC in there just does not cut it. It seems that I've done this before with out suffering from this while driving transformers.

In the past I've used the DC offset to correct for this and it seem to work out ok but not in this case. Also, I just did an update on LTSpice and the engine seems to be very cranky now, I switch to the other solver.

Currently using the "Delay" parameter seems to be the only way but it puts a hacky startup in there, not that it really matters much.

I suppose large caps may have got me by, but putting those on the sheet just adds components that are not there in reality.

I think we need a model made up for a 3 phase supply..

Thanks. Jamie

yes but in this case it would be 0, 120 and 240 degrees

The problem being is, spice starts the sine wave at the 0 volt line no matter the phase shift and thus we no long have a symmatrical sine wave with 0 phase. Jamie.

Repeat: do your transformer winding models have any DC resistance?

Is it possible to close a switch, between source and load, after an appropriate time into the simulation?

Looks fine to me, in two simulator environments...

Tim

-- Seven Transistor Labs, LLC Electrical Engineering Consultation and Contract Design Website:

Given that in real life your contactors won't energize all three windings s imultaneously and won't sweetly switch at the zero-crossings, it's probably worth investigating just why that screws up your simulation. I second JL 's suggestion to check how realistic your transformer model is. Zero resist ance and k=1.000000 probably isn't what you want.

Cheers

Phil Hobbs

In LTspice there is a function...

stp(x), u(x), unit step, 1 if x > 0, else 0

Multiply your sine function by stp(time-delay) ...Jim Thompson

Den tirsdag den 2. august 2016 kl. 02.28.14 UTC+2 skrev M Philbrook:

but that's how it would likely be in the real world when you switch on three phase power and if you turn off after an integer number of cycles it would average out

-Lasse

Can you post your simulation for us to look at?

You want all three sines to start at zero amplitude at T0?

yes it does and i wish it would work that way here.

If I don't do anything other than set the phase,voltage and freq, the sine starts at the given phase I entered at 0 volts, not where it should be.

Yours is starting where it should be.

I did note that I just down loaded an updated version of ltspice and running it on a Win10 machine, I don't know if that has any bearing but i am going to test that sim on an older PC with an older install of LTspice to see if there is any difference.

I was able to correct the issue using the delay parameter which did start at the proper level and used a flat line to finally connect to the starting angle of the sinewave.

Maybe I have a bad copy of spice on my pc. jamie

no. I wanted 0 phase to start at 0 which it does and the 120, 240 angle to start at the proper levels.

V = Sin (120) * Level; V = Sin (240) * Level; those are around ~ 0.866 and - 0.866 * level, which is where I would expext the sinewave to start on the left side.

Other posters have shown that it does do that with their test but not here.

I'll try a 3 phase non induciton load this time.. I can't have any DC in it.

Jamie

Just another data point - it works fine for me using the Phi(deg) phase parameter on an AC voltage source. As you find the engine 'very cranky', I'd suggest looking somewhere completely different. Do all your inductors have a little series resistance, for example? Any floating nodes which need a few gigohms to ground?

Cheers

Ok, this is the issue, I just experimented using R's and L's tied into a Y network to ground. That worked as it should have.

The problem is doing a Y primary transformer,the secondary having a delta. I have the K properties set so that only each phase is coupled with one secondary as it normally is. So I have 3 K directives using this K1 L1 L2 0.99 L1 being a primary of one of the Y's and L2 being the secondary of one of the delta's and that is repeated for the other two sets

What is happening is the delta on the secondary side is effecting the starting point back to the primary side. It is basically offseting the starting point in the simulation, but this does not happen in the real world! Its adding an offset due to the feed back if you want to call it that. Using the Delay parameter of 5.5ms for the 120 and 11 ms for the 240 fixes that because it starts at 0 volts anyway but delays the actual starting of the sinewave, for which the interpolation is correct.

its just a simple 3 phase transformer, Y primary side and Delta on the secondary side. I have the Star connections bonded to common on the Y side but I guess you could run it unbounded too.

I'll see how large the file is, maybe I can post it here. jamie

Hand-operated breakers are asynchronous to the line, so the random sine phase can squirt some DC into inductive loads. But the circuit resistance will make that DC decay pretty soon. Or eventually.

I don't know if any big breakars are line synchronized somehow. It would seem difficult.

Might I suggest that all should join the LTspice List...

I've learned quite a lot about LTspice from just reading the E-mailed digests. ...Jim Thompson

g the

That's very likely an underestimate of k. I've measured power transformers in the 0.9989ish range.

Cheers

Phil Hobbs