Toroid or EE Ferrite core, which is Best for 240V mains SMPS

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi SMPS Masters,

In my country I have access to both Toroid as well as EE ferrite cores, which core could be best suited for Mains AC SMPS work....
I want to switch my SMPS at 25-45KHZ, using IGBT's...But Struck in choosing the right shape for core...

I Hope Eva as always must have the wonderful answer for this too...


:) K a n w a r
 
Toroids in SMPS are not as wonderful as most people think. There is too much "repeat what you heard without worrying whether it's true or not".

With toroids, stray magnetic fields only cancel out in the far field, while in the near field they produce stray fields in all planes. On the other hand, E cores only produce fields in a single plane. Thus a toroid is likely to disturb a SMPS control circuit (or even a current sense resistor!!) placed in close proximity to it despite the plane in which it is placed, but an E core will be completely harmless if the plane of the windings is perpendicular to the plane of the sensitive PCB tracks (daughterboards come handy here).

Also, achieving 4KV primary-to-secondary safety insulation in a toroid is quite difficult, while with a E core it's a matter of adding the traditional 3 layers of mylar tape, leaving some (>3mm) margins between the windings and the sides of the coil former, and between primary and secondary side magnet wires themselves where they come out of the bobbin to the pins (>5mm).

I recommend E or ETD cores for the main transformer and iron powder for the output inductors.

The following picture shows the PFC converter in which I'm now working (450V output, 1.7KW@88V AC input, 4.6KW@230V AC input). I had to place the boost inductors that way in order to reduce the amount of leakage flux disturbing the 0.01ohm current sense resistor (that shiny thing with the shape of a bridge) and the control circuit daughterboard. I'm considering two current sense resistors with the current flowing through them in opposite directions in order to fully cancel the effects of stray flux from the toroids...
 

Attachments

  • pfc3k_1.jpg
    pfc3k_1.jpg
    81.2 KB · Views: 1,519
Thanks Eva, for the wonderful and helpful Suggestions...

Could i also use Iron powder cores for Common-mode Inductors for Mains Filtering.......

So its clearly evident from your detailed description that Toroid cores as Transformers are a bad thing in Mains SMPS....

I am currently working on 5KW SMPS to power my 3.2KW Class-H amp...

regards,
K a n w a r
 
An iorn powder core is the last thing that you would want in a common mode filter. Permeability is low so resulting inductnce is low, and resulting impedance in the Mhz range is also low. Standard power or signal ferrites should produce far better results, particularly when winding capacitances (though which RF leaks) are low.

BTW: A common-mode choke wound on a 3E25 toroid core is shown behind the capacitors in the picture from my previous post.
 
Hi Eva,

So what you are saying is that Standard Power Ferrites are also much better than toroids when it comes to common-mode Filtering...is it right.....

Then again what suited best to the Secondary side after the rectification for the LC filter for the Rails...
I have seen QSC PLX amps uses Ferrite beads on secondary side after the diodes....Whereas Crest PRO series uses a power ferrite based common-mode inductor on the rails itself.....[rails of power amp]

K a n w a r
 
Workhorse said:


Then again what suited best to the Secondary side after the rectification for the LC filter for the Rails...
I have seen QSC PLX amps uses Ferrite beads on secondary side after the diodes....Whereas Crest PRO series uses a power ferrite based common-mode inductor on the rails itself.....[rails of power amp]

K a n w a r
You have to make the distinction between the main power filter and the suppression components: for the power choke, a metal powder toroid (iron or molypermalloy) is quite suitable, as is a gapped ferrite core.
Ferrite beads on the diodes serve to remove the RF caused by the switching or the hard recovery.
And common mode inductors on the output rail are there to remove ...well, common mode interferences, or to decouple groups of outputs from one another.
LV
 
Workhorse said:
Hi Eva,

So what you are saying is that Standard Power Ferrites are also much better than toroids when it comes to common-mode Filtering...is it right.....

Then again what suited best to the Secondary side after the rectification for the LC filter for the Rails...
I have seen QSC PLX amps uses Ferrite beads on secondary side after the diodes....Whereas Crest PRO series uses a power ferrite based common-mode inductor on the rails itself.....[rails of power amp]

K a n w a r

You have to distinguish between regulated and unregulated supplies. In a regulated circuit, after the diodes you are going to find either a big iron powder toroid or a big gapped ferrite core, usually with two or more coupled windings (those are differential mode, not common moe). In an unregulated circuit you are going to find small smoothing pi filters made with ferrite beads, which sometimes are also present after the main inductor in regulated supplies.

Note that QSC uses a particular SMPS topology in their highest power models, which involves a PFC stage feeding directly the primary side of a transformer through a full bridge. Don't try to understand how that works at first because it's a bit complex.
 
it hasn't dawned on me before this thread but is it possible to get good enough regulation going with an unregulated design right after a PFC module (astec, 380VDC output at 950W)?

and also, I can't find locally a big enough EE core, will a 2.4" dia toroid (amidon FT 240-77 from what I recall)but to increase separation, I would wind half the circle with the primary and the other half of the circle the secondary? I guess that would result to EMI nightmare?
 
Ferrite Toroid

DJ-

YES, that would be am EMI nightmare. You do NOT want to wind your toroid that way, due to the poor magnetic coupling from windings not wound over each other. Since you have a nice BIG toroid (I have several FT-240-77s myself), room on the core won't be too much of an issue. At 40kHz switching frequency, from Amidon's datasheets, your '240 should be good for ~1.5kW throughput, topping out at ~1.7kW @ 100kHz.

EVA is right when she says that EE cores are preferred for Mains-powered SMPSs, because EEs have better primary-to-secondary 4kV isolation (hi-pot) and better primary-to-secondary magnetic coupling, but if a ferrite toroid is wound and taped correctly, you can have just as good isolation. This is not to disagree with EVA, as she is rarely (if ever) in error on this stuff, but I am trying to point out that you can have reasonably good results with a line-operated SMPS using s toroid if it is done correctly. I have had good success with toroids of many sizes in both DC-DC and AC-DC supplies.

Toroids are much better suited to DC-DC SMPSs (like in almost any car amp), where the windings ratio can exceed 7:1 (secondary to primary), running anywhere from +/-33-55VDC out. Also, because all of the voltages are low (L.T. 60VDC) hi-pot isolation is not so much an issue.

For offline SPMSs, toroids are not well suited for applications where the primary-to-secondary windings ratio is greater than, say, 7:1 because of the poor coupling that would result from the low number of secondary turns on the core. Thus, a toroid for a 5V, 100A mains-powered SMPS would be a poor choice.

The key to success with toroids in AC-DC SMPSs is having a primary-to-secondary windings ratio of LT 7:1. LT 4:1 is even better. For example, a half-bridge ferrite toroid running off a 400V PFC front end (200V on the primary windings) with outputs of, say, +/- 45V would work well because the primary-to-secondary ratio is something like (200/90) or ~2.2:1, providing good coupling, due to the increased number of secondary turns covering the entire 360 degrees. Proper hi-pot siolation can be achieved if sufficient layers of transformer tape (I use teflon tape) are sandwiched in between the windings.

Even better coupling on the toroid can be achieved if you wind half of the primary, then 3 layers of tape, then both secondaries, then 3 layers of tape, then the other half of the primary, followed by a final 3 layers of tape. As for the plane of radiation EVA mentions, I did not even stop to think of that, but she's right (again). However, if good coupling is achieved, and good component layout on the board is employed, then magnetic strayfields and their resultant interference can be minimized.

As for powdered-iron toroids on the output, you will definitely need it, if your supply is constant-frequency PWM regulated, as it performs the same function as it would in a classic non-isolated buck conveter.

If I ever ressurect any of my dead or disassembled AC-DC converters, I will post pics of them here.

EVA, question for you: Since powdered-iorn cores are not very good for AC line filters in common-mode, can they be used for differential-mode?

Steve
 
I'm planning to use a full bridge on the primary side.

I was thinking of winding the primary first then the seondary over it (with enough layers of tape).

since I will be using it in a high power amp (someday!), it will have outputs higher than +/-40V.

I plan it to be unregulated since the PFC module that I have already puts out 380VDC regulated......
 
DJ-

Actually, the output from the PFC is what manufacturers call "semi-regulated", expecting that the PWM, or whatever the DC-DC section of the converter is, to take care of the remaining regulation to make it tighter for line- & load variations. As far as winding the primary first, I have done this, and it's OK, but splitting the primary up so it fully envelops the secondary, both inside and outside, your magnetic coupling will be maximized. Strike what I said before about winding half the pri outside & half inside. Even better, to minimize leakage inductance, since your primary will undoubtedly be either bifilar or quadrifilar, depending on what gauge wire you're using, you could wind two strands inside the secondary, and the other two strands on the outside of the secondary.

Jimbo,

Sorry, dude. :( Didn't mean to do that. Don't know what to say about that. Perhaps you could wind it for 24V, and then buck that down to 5V, synch'ing it to the main PWM's oscillator. :xeye:
Just a WAG.
 
ok,
i will wind as many turns as i can on the secondary probably 6-8 turns.

i have 40 turns on a ft240-77 of 6 number 22 wires fed with 320v (doubles line) the secondary is 3 turns of 30 number 22 wires to get about 18 volts at 50 amps. but the fet keep vapourizing the opposite pairs fry.... full bridge.....

so i will change the secondary to 6-8 turn or what will fit over the primary of the toroid. with about 15-20 wires.. then buck .........


i wonder if i can get an etd49 and try that with the 320 to 18v.

i might try the two half bridges in parallel. so..... what would the best way to balance ? two coils around a common toroid?


jimbo............................................:smash:
 
Eva said:


You have to distinguish between regulated and unregulated supplies. In a regulated circuit, after the diodes you are going to find either a big iron powder toroid or a big gapped ferrite core, usually with two or more coupled windings (those are differential mode, not common moe). In an unregulated circuit you are going to find small smoothing pi filters made with ferrite beads, which sometimes are also present after the main inductor in regulated supplies.

Note that QSC uses a particular SMPS topology in their highest power models, which involves a PFC stage feeding directly the primary side of a transformer through a full bridge. Don't try to understand how that works at first because it's a bit complex.

Thanks Eva,

My Design Goal is to develop an 5KW unregulated supply and its without PFC....
I am going to use Full-Bridge Topology
Output Transformer Core is EE80
Gate Driver with EE25 core
SwFq=35KHz
Now for Pi Filters with ferrite beads....
Input Filtering with Iron powder Toroids....
PWM controller=SG3525 + Sync with Crystal Oscillator using CD4060 for eliminating BFO effects

K a n w a r





:)
 
DJ-

How are U PFC'ing it? Constant-frequency? If so, then synch your SG3525 to the PFC t oavoid beat frequencies. Do a search to find threads covering this.

A switching frequency of ~60kHz would be good for ~1.5kW thru this core before core loss becomes an issue. Also, remember, while higher frequency will mean less turns on the pri & sec., a higher V(in) will mean more turns. So for

F(sw) = 60kHz, F(osc) = 120kHz

N(pri) = [Vin(min) e8]/[K x F(sw) x A(e) x B(max)],

Where Vin(min) = 380V
K = A constant (4.0 for sq wv and 4.44 for sine wv),
A(e) = Effective crossectional area of the core,
B(max) = maximum flux density in Gauss @ given frequency

So, N(pri) = (380 x e8) / (4 x 60000 x 1.57 x 1000) = 100T

While this might seem high at first, considering the frequency is 60kHz, and the input voltage to the primary is 380V (because of PFC), instead of (180 x 1.414) - 2.5V (diode drops) = 252V giving 28T, you have 100T.

Now go on and calculate your secondaires using classical methods.

N(sec) = N(pri) x [V(out)/V(in)] = 100 x (50 / 380) = ~13T.

So, primary will be 100T, and secondaries will be 13T + 13T. I would do 14T +14T to accommodate the diode drops of ~2.5V for ultrafasts, and ~2.0V for Schottkys.

Eva, Thanks for the advice. Would a #77 ferrite toroid make a good common-mode choke for the AC Line filter?
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.