What is the problem with leakage inductance?

[tim9000]

The thread name is probably misleading, but I just want to make sure I'm clear in that: Are the only two problems caused by leakage inductance (stray magnetic flux causing imperfect mutual coupling) in a transformer, that of;
- poorer voltage regulation &
- EMI noise on other equipment
(?)

Also, on the note of voltage regulation, is that only caused by leakage flux?

Thanks

 

[JerryR]

The effects of leakage inductance depend on several factors. The largest one is the frequency of the signals passing through the transformer structure. For power transformers, which dominately operate with low frequency sine waves, regulation is definitely an issue.

EMI is less important due to the low frequencies. For transformers operating over a broad bandwidth, such as audio transformers, frequency distortion may be a problem. The magnitizing inductance must be large enough to pass the lowest frequencies.

The construction techniques needed for a large magnitizing inductance tend to increase the leakage inductance, particularly if a high isolation is needed between the primary and secondary. The higher leakage inductance may cause losses (poor regulation resulting in distortion) at higher frequencies.

Transformers used in applications such as switching power supplies are sensitive to voltage kickback from leakage inductance as the driving transistors turn off. This typically requires the use of snubber circuits to limit the voltage on the switching devices. This is also influenced by the construction techniques needed if isolation is required between the primary and secondary. These techniques tend to increase the leakage inductance.

EMI can be a problem in all cases. This is a function of the operating frequency of the transformer.

 

[berkeman]

Are the only two problems caused by leakage inductance (stray magnetic flux causing imperfect mutual coupling) in a transformer, that of;
- poorer voltage regulation &
- EMI noise on other equipment

Depending on the circuit, there can be other issues with leakage inductance (Lk).

For example, in communication circuits with limited supply voltages for the network drive circuit, if Lk is too large, it can cause headroom problems for the drive amp. Say you have a 3.3V power supply and you want to drive a large TX signal into a transmission line. If you have a large Lk, then extra voltage will develop on the inboard side of your comm transformer, and you will get distortion of the TX signal if the TX amp runs out of headroom. So you will usually try to minimize Lk in comm transformers, unless you have wide power supply rails to accommodate the extra drive voltage needed.

 

[berkeman]

Transformers used in applications such as switching power supplies are sensitive to voltage kickback from leakage inductance as the driving transistors turn off.

The kickback is generally due to both the magnetizing and leakage inductances, L = Lm + Lk.

 

[JerryR]

The kickback is generally due to both the magnetizing and leakage inductances, L = Lm + Lk.

In switching power supplies the kickback from the magnetizing inductance is generally clamped by the load. The kickback I am referring to will be the sum of the primary leakage inductance and the secondary leakage inductance as modified by the transformer turns ratio.

 

[berkeman]

In switching power supplies the kickback from the magnetizing inductance is generally clamped by the load. The kickback I am referring to will be the sum of the primary leakage inductance and the secondary leakage inductance as modified by the transformer turns ratio.

Yes, that's true for many switching topologies. In a CRT Flyback transformer, though, the kickback is part of the high-voltage pumping action.

 

[vintageplayer]

Another issue (mainly for larger power transformers) is efficiency loss. Lost power equates to lost dollars.

 

[tim9000]

Depending on the circuit, there can be other issues with leakage inductance (Lk).

For example, in communication circuits with limited supply voltages for the network drive circuit, if Lk is too large, it can cause headroom problems for the drive amp. Say you have a 3.3V power supply and you want to drive a large TX signal into a transmission line. If you have a large Lk, then extra voltage will develop on the inboard side of your comm transformer, and you will get distortion of the TX signal if the TX amp runs out of headroom. So you will usually try to minimize Lk in comm transformers, unless you have wide power supply rails to accommodate the extra drive voltage needed.

My lingo is lacking.
I'm really sorry but could you please elaborate on Headroom problems for drive? All I could find was regarding audio output.
What do you mean by 'comm transformer' (communication?) and 'inboard side'? I assume you just mean like the realistic transformer model (with the leakage inductance before the ideal transformer)
Also, what is kickback? Is it like if you open circuit an inductor? (Are you saying like if you open circuit a side of a big transformer, you'll still get a spike, in the open circuit coil, from the leakage?)
Thanks a lot

 

[tim9000]

Another issue (mainly for larger power transformers) is efficiency loss. Lost power equates to lost dollars.

Thanks for the reply, how does it translate into efficiency loss exactly though?

 

[vintageplayer]

Thanks for the reply, how does it translate into efficiency loss exactly though?

Leakage inductance is a result of leakage flux. This is the flux which does not perfectly couple both windings in a transformer and goes to waste.

https://commons.wikimedia.org/wiki/File:Flux_leakage.png

The more leakage inductance (leakage flux) you have, the less power you will see on the secondary side of the transformer. This is why power transformers always have iron cores (these help "direct" the flux and reduce leakage inductance). When it comes to transferring large amounts of power, even little power losses can add up and equate to large dollar losses. For smaller transformers (not used in large power supply), efficiency may not be as big a motivator for reducing leakage inductance.

 

[berkeman]

Also, what is kickback? Is it like if you open circuit an inductor? (Are you saying like if you open circuit a side of a big transformer, you'll still get a spike, in the open circuit coil, from the leakage?)

Yes, the kickback is the result of a switch opening in series with an inductor that is carrying current. Since the current through an inductor cannot change instantaneously when the switch opens, the voltage kickback generated can be quite large (limited only by the parasitic capacitance involved, and arcing breakdown voltages involved). Are you familiar with this differential equation relating the voltage and current for an inductor?

$$v(t) = L \frac{di(t)}{dt}$$

could you please elaborate on Headroom problems for drive? All I could find was regarding audio output.
What do you mean by 'comm transformer' (communication?) and 'inboard side'? I assume you just mean like the realistic transformer model (with the leakage inductance before the ideal transformer)

Some of the details are proprietary to the company that I work for, but I can speak in generalities. Yes, a comm transformer is a communication transformer, used in some communication network. You are probably familiar with Ethernet networks, right? Ethernet network physical connections ("Phy connections") use communication transformers tuned to the frequencies used for whatever bandwidths are needed.

The circuit that drives the "near side" of the comm transformer (the side of the transformer that faces the device, as opposed to the side that connects to the network wiring) has some power supply voltage. It is typically either 5V or 3.3V, although other voltages could be used. When you have a transmit (TX) amplifier circuit, it needs its power supply voltage to be larger than the output waveform that it is driving into the comm transformer. Depending on the type and topology of the TX amp, it can need on the order of 2V of "headroom" outside of the TX signal in order not to distort the signal. So if you have a 2V TX waveform that you want to get through a comm transformer onto the network, and you have a 5V power supply, your TX amp should be okay, since you have 5V-2V=3V of headroom, and you only need 2V of headroom.

But, if the comm transformer has a large leakage inductance, there will be an extra voltage drop across that leakage inductance (Lk) in addition to the voltage drop across the leakage magnetizing inductance (Lm), and that extra voltage drop erodes your headroom. If the extra voltage drop across the Lk increases the drive voltage into the near side of the comm transformer enough (like by 1V in this example), then you will start running out of headroom in your TX amp, and that will cause distortions in your TX waveform, which can cause comm errors in your network.

Hope that helps some. Ask more questions if things aren't clear yet.

Edit -- fixed a typo Leakage --> Magnetizing

 

[tim9000]

Leakage inductance is a result of leakage flux. This is the flux which does not perfectly couple both windings in a transformer and goes to waste.

https://commons.wikimedia.org/wiki/File:Flux_leakage.png

The more leakage inductance (leakage flux) you have, the less power you will see on the secondary side of the transformer. This is why power transformers always have iron cores (these help "direct" the flux and reduce leakage inductance). When it comes to transferring large amounts of power, even little power losses can add up and equate to large dollar losses. For smaller transformers (not used in large power supply), efficiency may not be as big a motivator for reducing leakage inductance.

I am, or atleast was very well acquainted with leakage flux, and mutual coupling.
The point I was getting at regarding efficiency is, yes I can see how eddy currents in the core create wasted heat, and copper losses in the coils too. But I'd have thought that since an inductor doesn't consume energy, that twice every cycle, that energy that is stored in the leakage inductance will just bounce back to the primary supply. So rather than it being: power out the secondary / power in the primary, its just like you'd have to deduct the leakage from the denominator, because it's not really going anywhere, just being stored for a bit. That's why I didn't see it affecting the efficiency, though I'm happy to be told otherwise if there is supporting pictures or equations in links.
To this end, I'd have thought maybe leakage flux would affect the power factor of the transformer?

 

[tim9000]

Yes, the kickback is the result of a switch opening in series with an inductor that is carrying current. Since the current through an inductor cannot change instantaneously when the switch opens, the voltage kickback generated can be quite large (limited only by the parasitic capacitance involved, and arcing breakdown voltages involved). Are you familiar with this differential equation relating the voltage and current for an inductor?

$$v(t) = L \frac{di(t)}{dt}$$Some of the details are proprietary to the company that I work for, but I can speak in generalities. Yes, a comm transformer is a communication transformer, used in some communication network. You are probably familiar with Ethernet networks, right? Ethernet network physical connections ("Phy connections") use communication transformers tuned to the frequencies used for whatever bandwidths are needed.

The circuit that drives the "near side" of the comm transformer (the side of the transformer that faces the device, as opposed to the side that connects to the network wiring) has some power supply voltage. It is typically either 5V or 3.3V, although other voltages could be used. When you have a transmit (TX) amplifier circuit, it needs its power supply voltage to be larger than the output waveform that it is driving into the comm transformer. Depending on the type and topology of the TX amp, it can need on the order of 2V of "headroom" outside of the TX signal in order not to distort the signal. So if you have a 2V TX waveform that you want to get through a comm transformer onto the network, and you have a 5V power supply, your TX amp should be okay, since you have 5V-2V=3V of headroom, and you only need 2V of headroom.

But, if the comm transformer has a large leakage inductance, there will be an extra voltage drop across that leakage inductance (Lk) in addition to the voltage drop across the leakage inductance (Lm), and that extra voltage drop erodes your headroom. If the extra voltage drop across the Lk increases the drive voltage into the near side of the comm transformer enough (like by 1V in this example), then you will start running out of headroom in your TX amp, and that will cause distortions in your TX waveform, which can cause comm errors in your network.

Hope that helps some. Ask more questions if things aren't clear yet.

Fascinating reply.

Well I'm struggling to get my head around this Ethernet business, as far as I was aware an ethernet cable was what you plugged into your router or switch then into your desktop, I mean I've seen heaps of them in server towers or whatever. But what you mean by 'Physical Connection' I'm not sure about, are you saying this is for Ethernet cables that travel long distances? And how do transformers come into play? Are they supplying the Ethernet cables with a high frequency, or power, or data, or both?
I think I get in principle what you're saying about some waveform being eroded from by a lack of power, but it's the aforementioned step in the reasoning that is letting me down, sorry.

Cheers

 

[berkeman]

No need to be sorry. Just use Google Images to look at Ethernet Phy circuits, and it will become more clear.

 

[tim9000]

No need to be sorry. Just use Google Images to look at Ethernet Phy circuits, and it will become more clear.

H'mm, https://www.google.com.au/url?sa=i&rct=j&q=&esrc=s&source=images&cd=&cad=rja&uact=8&ved=0ahUKEwjHxda778vKAhUnnqYKHb07DhMQjRwIBw&url=http://electronicdesign.com/boards/ethernet-port-protection-requires-smart-design-and-test-strategies&psig=AFQjCNETi_snsEc740leIdYJUYtEwP-IgA&ust=1454048407681234

So What's a TVS1/TVS2 chip for? See, those transformers just look like isolation transformers to me, how big are they? I would imagine not much bigger than the chips.
so this Phi chip, are they the last port of call before the signal gets transmitted down the Ethernet cable? Do they just facilitate the data that is already there, encoding and decoding the actual transmitted signal?

 

[vintageplayer]

I am, or atleast was very well acquainted with leakage flux, and mutual coupling.
The point I was getting at regarding efficiency is, yes I can see how eddy currents in the core create wasted heat, and copper losses in the coils too. But I'd have thought that since an inductor doesn't consume energy, that twice every cycle, that energy that is stored in the leakage inductance will just bounce back to the primary supply. So rather than it being: power out the secondary / power in the primary, its just like you'd have to deduct the leakage from the denominator, because it's not really going anywhere, just being stored for a bit. That's why I didn't see it affecting the efficiency, though I'm happy to be told otherwise if there is supporting pictures or equations in links.
To this end, I'd have thought maybe leakage flux would affect the power factor of the transformer?

Yes I agree, the only extra losses will be due to the larger circulating currents you will have to supply to achieve the same real output power.

 

[tim9000]

Yes I agree, the only extra losses will be due to the larger circulating currents you will have to supply to achieve the same real output power.

Right.
So I wonder, does leakage flux affect efficiency or functionality in any real term?
And does it affect power factor?