odd result from SPICE, is this actually possible? / High Voltage / Forums

Forums

4hv.org :: Forums :: High Voltage

« Previous topic | Next topic »

odd result from SPICE, is this actually possible?

Move Thread

LAN_403

Crunchy Frog

Wed Oct 07 2009, 04:17AM

Registered Member #2422 Joined: Tue Oct 06 2009, 02:41AM
Location:
Posts: 85

By feeding two opposite unsmoothed doublers into either end of a transformer (the waveforms look like asin(x)+a and asin(x)-a, I get this output. This doesn't seem right, but if it's wrong why is it coming out of the simulator?

Myke

Wed Oct 07 2009, 04:32AM

Registered Member #540 Joined: Mon Feb 19 2007, 07:49PM
Location: MIT
Posts: 969

No, It's right. What it does in the first half cycle is charge the first capacitor. Then it charges the second capacitor in the second half of the cycle. The output voltage pulled from the two capacitors in a series arrangement. When they are both charged, the output voltage is twice the input voltage. Hope that made sense.

EDIT: whoops, didn't notice the measuring point was the transformer. Make sure you aren't measuring the voltage across the transformer primary? What is this for? The voltage multiplier only puts out DC so the transformer core would saturate and you would effectively short out the multiplier.

TwoSpoons

Wed Oct 07 2009, 04:48AM

Registered Member #1621 Joined: Tue Aug 05 2008, 05:26AM
Location:
Posts: 19

OK, I redrew your circuit to make it easy to see what is going on . Ran it in simterix and got tiny millivolt pulses on the output - transient result attached. No Idea how you got that graph out of your simulator - though its really hard to figure out what the x and y scales are.
]circuit.pdf[/file]

Though I get a very different result if I greatly increase the primary inductance (to 1 H, then I get 40 v pulses every half cycle) - what did you use in your circuit?

Crunchy Frog

Wed Oct 07 2009, 03:28PM

Registered Member #2422 Joined: Tue Oct 06 2009, 02:41AM
Location:
Posts: 85

The x scale is a few msec, the y scale is about 0-3kv

Edit: The voltage multiplier puts out sort-of-ac, in that its a sine wave varying from 0 to twice the original amplitude.

It looks sort of like this (both inputs on one graph)

/\    /\
--\--/--\--
   \/    \/

becomes

/\  /\  /\
__/  \/  \/  \
-------------
\  /\  /\  /\
 \/  \/  \/

So would the two put together be dc, or is there some current factor or something I'm missing?

Maybe I should switch simulators?

klugesmith

Thu Oct 08 2009, 09:51PM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

[quote]

/\  /\  /\
__/  \/  \/  \
-------------
\  /\  /\  /\
 \/  \/  \/

So would the two put together be dc, or is there some current factor or something I'm missing?
Maybe I should switch simulators?
[/quote1255037066]
I think you should switch transformer models.
The simulated output waveform in OP is consistent with your ASCII waveform drawings. Differential voltage between the two ends of primary winding rises to 300V in the first 2 half-cycles, then stays there (300 V DC) in the absence of a load. Then you have a 1:10 "perfect transformer" which works even at DC. It also transforms the current (from 1 pF capacitive load). So the current on primary side will be as if there were a 100 pF capacitor in place of the transformer: zero amps DC.

A good reason for a simulator to use a perfect transformer by default, is that you can't model any real transformer without knowing more parameters than the turns ratio. For example, how can a model "not transform DC" unless it can establish the minimum AC frequency that "does transform"? A convenient parameter is the primary magnetizing inductance. Remember that SPICE is not in the business of guessing ANY details of the user's circuit.

A very simple but more realistic transformer model for you would be 2 inductors, say 10 henries and 1000 henries, with a mutual inductance "K L1 L2 0.99" (gives M = 99 henries).

[edit] Was going to point out HV_wiki about transformers, but it doesn't address modeling. Here's a good and relevant tutorial about transformers in simulation:

There seems to be some disagreement about which model is the "perfect transformer" and which the "ideal transformer". Perhaps someone else would like to investigate & report on that convention.

Antonio

Fri Oct 09 2009, 02:09AM

Registered Member #834 Joined: Tue Jun 12 2007, 10:57PM
Location: Brazil
Posts: 644

An ideal transformer is completely characterized by the turns ratio. No real transformer behaves in this way, but it is an useful approximation for most cases.
The "perfect transformer" from the site mentioned is a linear transformer, two linear lossless inductors with linear coupling. This is a far better approximation, but is quite complicated to treat in designing circuits.
It is equivalent to an ideal transformer with a magnetizing inductance (the primary coil inductance L1) in parallel with the primary winding, a dispersion inductance in series with the secondary coil, with value L2*(1-k^2), and the turns ratio multiplied by k, the coupling coefficient.
An useful simplification is then to consider k=1, what keeps only the turns ratio and the primary inductance in the model.
The next approximation is to add resistances to both coils.
A real transformer has also distributed capacitances, and with a magnetic core the inductances can be significantly nonlinear.

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob