Eng-Tips is the largest engineering community on the Internet

Intelligent Work Forums for Engineering Professionals

  • Congratulations waross on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!

Line inductance and PFC

Status
Not open for further replies.

spinach

Electrical
Aug 5, 2008
11
GB
Hello,

I am designing an offline Power factor corrected SMPS (continuous mode boost converter at front).

Since it is mains powered, i am worried how i will be able to draw the PFC's high frquency pulse currents through the Line Inductance. (That line inductace being the inductance of the power wires and the Electricity supply company's transformer secondary coil)

I am thinking of putting a big AC capacitor on the line (AC) side of the diode bridge.....however, that worries me because then that will ruin the power factor.

....What i could do is put a capacitor of sufficient value across the input AC line...such that it "cancls" the inductive reactance of the line inductance and makes my power supply look resistive....however, i am worried that such a capacitor would have to be of a value and voltage rating that would make it very expensive.

When i look at circuit diagrams of EMC filters...they confuse me because the capacitor values are always less than 4n7.....is it worth putting these in at all?....
......surely they won't give sufficint energy storage to supply the pulse currents to the switch mode power supply?

Then there's the common mode choke.....that appears to be the only other option for providing the energy storage for the pulse current input to my Switched mode power supply........but i thought that all the common mode choke did was stop currents leaking to earth, and not providing energy storage?

And with the common mode choke, an enginer told me they never saturate and never need a gap in their core, -but surely if they conduct enough current that the magntising current puts a flux in the core which will saturate it, -then it saturates i would have thought.

I have never had this problem before because bfore i always drew the SMPS currents from the caps after the diode bridge.
 
Replies continue below

Recommended for you

The high-frequency switching components are shorted by capacitors in the 22 - 47 nF range. No probs, just do it. A 47 nF capacitor will not affect your PF very much.

The line is usually, correctly or not, regarded as a transmission line with the characteristic impedance being 50 ohms resistive. At least, that is what the LISNs are in most cases.

Gunnar Englund
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
 
Keep in mind that CM chokes should deal with AC current pulses of some 3ms width. They have opposite polarity windings for cutting down any noise coming with the same phase through both AC lines (common mode), while your shunt cap will only short differential harmonics between mains wires.
They may be simulated as part of the same filter, though they are meant for different types of noise. I mean, CM choke will clear common mode noise but current harmonics are absorbed only by their winding inductances.

CM chokes and caps show some self-resonance due to its construction. For example, a 1mH choke will turn into a capacitive reactance over 1MHz, so any further harmonic at those frequencies and up is absorbed mainly by your shunt cap.
 
Thankyou for these replies.

If i could expand a little, i have recently been (trying to) play with a PWM controller chip. -Using variable frequency as the (non-capacitively smoothed) rectified input line went up and down.

-But i have been working off of a 36V AC RMS Mains transformer just to make things easier.

The problem is that when i start pulsing the Mosfet with higher on times, the rectified DC bus goes mad....it just becomes a wacky load of pulses going down to zero. -The fuse (1Amp) on the secondary does not blow during this.

I am wondering if my problem is that because i am using the step down transormer (240V TO 36v) the line impedance (which i believe is mostly inductive) as seen looking back through the transformer is magnified about 9 times.....this then is a huge indutance through which i am trying to draw current pulses and its just not happening and the inductance is pushing my DC bus (and the secondary output volts of the transformer) down to zero.

I am thinking of getting some 440V AC RMS 30uF motor run capacitors and putting them on the primary side...hoping that they will cancel out the inductance of the line...(i believe line inductance is about 1.4mH/mile and 14nF/mile (?))

Any ideas on wy my DC bus is dropping to zero much appreciated......when i am not switching the MOSFET the DC bus is fine.

This has really puzzled me and now i am wondering that when i go direct off line i will still be somewhat affected by the input line inductance, which of course will include leakage inductance of the supply company's transformers.

I feel certain that this line inductance will have to be mitigated in some way, ...possibly by adding large AC capacitance ....but then maybe that will have large leakage current which i thought is illegal in SMPS's...in which case i won't be able to get enough capacitance and this won't work.

I notice on the datasheet of one (won't mention the name) off-line PFC contrller they put a 1uF AC capacitor at the input..i thought this was illegal?
 
Well, there really should be no need for a 30uF capacitor on the AC side of the rectifier. There will typically be X1 rated caps L-L or L-N and Y2 caps L-G and N-G for noise suppression.

Seriously, we have no idea what your circuit looks like or what type of PWM controller chip you're playing with or how you are varying the duty cycle so it's really hard to try and help you.

 
"line inductance, which of course will include leakage inductance of the supply company's transformers"

You are way off beaten paths and accepted practice here. I doubt seriously that your (say around 1 A maximum) load has any influence at all on - or is influenced by - the transformer's inductance. No, I don't doubt it - I am positive.

Such transformers are at least in the 100 kVA range - much more likely at 300 - 800 kVA, or larger.

You really need to start with a basic understanding of your circuit. Looking for the reason your circuit behaves badly in other places than your circuit will not help. Especially not if you think your company's distribution transformer or the line inductance is the reason for that bad performance.

Gunnar Englund
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
 
I am puzzled by your references to the mostly inductive lines. I was always under the impression that the action of utility lines was mostly resistive and that furthermore, the characteristics of the load determined the relative reactance of a circuit.
Under short circuit conditions the highly inductive component of the transformer impedance dominates, but under short circuit conditions there is no voltage to the load anyway.
When you consider the complete circuit, that is load, utility supply including transformers, the impedance of the load dominates the "inductance" of the total circuit.
The resistance and inductive reactance of a power transformer are several orders of magnitude less than the resistance and inductive reactance of a power supply.
For the design of a power supply,you may want to consider the inductance of the connection cable and branch circuit conductors, but you are wasting your time worrying about the "Inductance of the utility supply". For the purpose of a power supply design, consider the utility supply to be infinite and non-reactive.

Bill
--------------------
"Why not the best?"
Jimmy Carter
 
Hello

I just put a 30uF AC capacitor on the output of the secondary of my 36V transformer and now my circuit works.

-I think this is proof that it was the line inductance...as seen looking back through my transformer's secondary which was causing me the problem.

....BTW, when i put just 1uF on the secondary it still didn't work...showing that this was not enough capacitance to mitigate the supply mains line inductance, including the leakage inductance of my own transformer.

Please remember.....my circuit switches the unsmoothed rectified mains......no DC capacitors after the diode bridge to draw currents from.

I know (especially now) that Line inductance is a very real problem for my type of power supply (which also means for PFC corrected power supplies as they too have no caps after the diode bridge from which to draw current).

..in college, we were told about "overlap" problem in mains connected diode bridges.....these mean losses and its due to the line inductance.....we were not told that there was any fix for this "overlap" problem....we were just told that it was a scourge of which we should be aware.

...if you scope a DC bus after a mains connected diode bridge you will see the sine "hump" flatten at the top.....this is directly the result of the supply line inductance including the leakage inductance of the power supply company's transformers.....as you well know it's because you can't suddenly brake in inductive current and both top diodes are on together for a period of time.

The problem i now face is that off line SMPS's aren't allowed to have 30uF caps on the input AC line...as the leakage would be too great......i just hope that my transformer is making the line inductance appear greater and that when i go off line the seen line inductance will be less.......

.....its a matter of how much less?.....i can't remember whether its Np^2/Ns^2 * Line inductance or just Np/Ns * Line inductance....

First of all i tried a one transistor isolated Forward converter "pseudo PFC" idea (just switching the DC bus when it rose up above the certain level...that didn't work at all and i now know never would have because a forward converter draws "transformer" current for the load as opposed to "magnetisation" current as with a boost or flyback etc...."transformer" current is far too high-rise pulsey for the line inductance to allow through.
 
What can I say. I've certainly learned something today.

Bill
--------------------
"Why not the best?"
Jimmy Carter
 
No Bill, you haven't learned anything.

The inductance of the transformer is not the distribution transformer but the transformer used to get 36 V. It has nothing to do with line inductance or distribution transformer inductance. This guy still hasn't got it and I am not going to try and tell him what is going on. He will learn eventually - as have all of us.

Referring to diode overlap (the commutating angle) tells us that he has read about high-current rectifiers. But that is not the problem here.

His newbie status is all too obvious. He has to learn from his own mistakes. Will not listen to those in the know.

Gunnar Englund
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
 
Gunnar, What can I say. I've certainly learned something today -on tech support :)
 
Thanks Gunnar, sincerely. I though that he was working on a power supply to be used directly on the mains. As soon as you reminded me about his 36 v transformer, I had a quick flash of embarrassment.

Bill
--------------------
"Why not the best?"
Jimmy Carter
 
Any appliance sharing the mains wires is adding a new shunt reactance to your line, so I think you'll get mad trying to approach the problem that way.

A well-designed PFC stage should mask your PS as a resistive load to the mains, so inductance issues shouldn't affect the output seriously.
BTW, the bulk cap is meant to act like a battery supply. A Low-ESR, 2-3uF/[Output Watts] size is all your MOSFET needs to switch safely without ugly DC ripple.

Transformer impedance should match your DC rail voltage, MOSFET rating and PWM frequency. Too high and your MOSFET will be fighting with low current slopes and high voltages. Too low (or PWM too wide) and your MOSFET will be pulling current through a saturated primary -a wire jumper, literally, with the obvious MOSFET overheat and DC ripple issues.

Finally, if your feedback loop is fast enough, it will be able to follow line or DC rail 'humps', getting a better output regulation.
 
It is very simple to calculate the approcimate leakage inductance. The transformer's regulation is typically around 10 %. That means that voltage drops 3.6 V when load is increased from 0 to rated current. Almost all of this voltage drop is due to reactance (OK, some resistance also) so an approximate leakage inductance value would be L=(3.6V/2.8A)/(2*PI*line frequency). Or a few millihenries. That's a huge inductance if you compare with the distribution transformer's inductance, which usually is in the microhenry range.

I really cannot understand your attitude - do you not have a principal you can ask? And, why on earth are you designing with an obsolete part?



Gunnar Englund
--------------------------------------
100 % recycled posting: Electrons, ideas, finger-tips have been used over and over again...
 

Care to share the wattage of this supply?

I know you can find complete examples of PFC front ends up to likely 500W from companies that sell the IC's. Not one will require a 30uF capacitor before the rectifier.

...if you scope a DC bus after a mains connected diode bridge you will see the sine "hump" flatten at the top.....this is directly the result of the supply line inductance including the leakage inductance of the power supply company's transformers.....as you well know it's because you can't suddenly brake in inductive current and both top diodes are on together for a period of time.

Huh? Diode overlap or commutation would occur at the zero crossing as one diode is still turning off while the other diode has already turned on.

 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top