Saturation of Current Transformer

What is the difference between "partial" and "complete"?

A quick look at the saturation curve for your particular CT will answer the question completely. Contact the manufacturer for further assistance.

Unsaturated output current (blue) is sinusoidal, a copy of the input current being measured. If the current exceeds the saturation current, the core becomes completely magnetized, the flux does not increase and the induced secondary current returns to zero until the flux begins decreasing again (black curve). Of course, the higher the input current, the narrower the "shark fin" becomes.

https://ideology.atlassian.net/wiki/display/AP/High+Impedance,+Merz-Price,+Circulating+Current+Differential

So a really crappy transformer (that easily goes into saturation) and is way over-driven could be useful as a voltage spike generator? I had never really thought about how useful that might be. Put a full-bridge rectifier on that, and all the pulses go positive, correct? I have a silly reason to be interested in making something like that for my research.

You might want a flyback transformer. I've used one to make a high voltage power supply (many years ago). There ought to be tons of them on the surplus market since televisions have gone flat screen.

https://en.wikipedia.org/wiki/Flyback_transformer

If you want to create very short rise time peaky spikes, a spiral vector inversion generator is a pretty easy pulse generator to build and simple to control with spark gaps.

James Stuart note off-topic....

On-topic..... Initially, CT saturation flattens the peaks of the CT output. If the saturation is so much that the output falls to zero when it should be at maximum peak then it is fully saturated. There are degrees of "full" saturation, since a massive overcurrent will leave just + and - spikes close to zero input current with almost half cycle length periods of zero output between.

Off-topic....

You get alternate + and - spikes as in the graph. When I was a young engineer, telecom multiplex systems were analog and needed lots of harmonics of base frequencies. The way to get harmonics of 4 kHz was to feed a "square loop" saturation curve ferrite core via a series-tuned L-C resonant circuit (over 1 kV rms across inductor) to give a sinusoidal current drive. The short spikes from the saturable core gave a very uniform harmonic spectrum despite environmental variations. Good magnetic materials are actually very stable with temperature and time etc - I have read that magnetic amplifiers buried in nuclear radiation environments can maintain levels within a fraction of a percent for years.

67model

I will try playing around with this idea. There will be hydrogen-oxygen mixture nearby (not necessarily in the same space), so I am somewhat cautious about any potential source that needs to spark, or would typically produce a spark (unless engineered correctly). I like your idea. Will explore further.

Thanks for the input on how saturation effects the output.

Thank you all. Now I get a little idea on this. In short, Partial saturation means still the CT produces secondary current (which is non linear to the primary current) and complete saturation means there is zero secondary currrent. Is this right?

Referring to comment #6 by "Original Poster" - Please note comment #8.

I think you would be right if you wrote that there is zero secondary current for part of a half cycle at times when secondary current would be large if it were a linear multiple of primary current. Also, although the secondary current may fall to "practically" zero, there will always be some current due to the "air cored" flux of the primary bar linking the toroidal coil of the CT.

Some CTs, often called Rogowski Coils, are entirely air cored and can avoid saturation problems, but need special integrating amplifiers.

You should realise that "metering" CTs have high permeability cores and no air gap, while "protection" relay CTs have less easily saturated cores and air gaps to reduce the flux, achieving a greater overcurrent tolerance with usefull accuracy by sacrificing high accuracy at normal currents.

Finally, if I have remembered correctly, the secondary rms current over a whole cycle may still be usefully proportional to primary rms current despite saturation.

67model

So it really comes down to the design of the CT as to what characteristics will be produced in over current driven mode....

True.

Protection CTs are usually described as e.g. 5P10, meaning 5% error at 10 times rated primary current - with the rated load ohms (burden).

Metering CTs usually described as e.g. "Class 1", meaning 1% error at rated primary current and rated load ohms.

Standards like ANSI or IEC for CTs give error expectations in phase as well as magnitude & effect of secondary load (burden). Historically, meters and relays were electromechanical & the burden was partly inductive while digital instruments are likely to be resistive.

To be accurate, Rogowski Coils are not CTs. They take the e.m.f. from the coil, which is fundamentally proportional to rate of change of flux,flux is proportional to current without saturation, so integration gives current.