Bridgeless PFC help - diyAudio
Go Back   Home > Forums > Amplifiers > Power Supplies

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 13th April 2010, 03:35 AM   #1
koosh75 is offline koosh75  United States
diyAudio Member
 
Join Date: Apr 2010
Location: California
Default Bridgeless PFC help

Hi, I need the eyes of an experienced PFC designer... Attached are schematics of the 750W semi-bridgless PFC I'm working on. L4981 controller, toroidal KoolMu boost inductor, CoolMOS transistors, 120VAC input. Series string of 6 100W incandescent bulbs as load. Testing it out w/ a variac, I haven't pushed it past 40% yet.

The diode bridge runs very hot, using a small heatsink, and eventually blew out. My understanding is that this bridge is used to bypass the boost diodes for inrush current during startup, and also to reduce common mode noise by linking the output ground to the AC line. Why is it carrying such current during normal operation? (The output voltage is higher than the peak rectified voltage from this bridge.)

How much capacitance is should I use on the AC-line side of the boost inductor? ST's appnote uses 1uF for an 800W converter, and I started with that value (C23). However, this doesn't attenuate the 50KHz switching noise on the AC line. I added 4.4 uF additional and the switching noise is reduced, but it's still significant. My variac is making a loud audible vibration, likely because of this? Do I need much more?

After removing the bridge temporarily, the fuse blows easily, even at 30% on the variac. Mosfets and diodes are running cool.

The attached photos include the current-sense voltage measured at the center tap of the current sense transformer L5. It looks as I expect, except that it looks a little distorted for the negative phase of the AC input.

The other photo is the AC line voltage measured before the EMI filter, with a total of 5.4 uF of capacitance across the boost inductor.

(The schematic specifies an ETD59 core, but I substituted two stacked KooMu toroids instead, for a total of 54,000mm^3. The measured value is about 440 uH. D19, D22 were changed to 8V each, rather than 5.1V).

(I'm powering the L4981 with an external supply, since there isn't enough current from the aux winding at 40%.)

Any ideas?
Thanks,
Brian
Attached Images
File Type: bmp F0039TEK.BMP (18.1 KB, 124 views)
File Type: bmp F0037TEK.BMP (23.5 KB, 64 views)
Attached Files
File Type: pdf 750W_PFC8.pdf (195.0 KB, 154 views)
  Reply With Quote
Old 13th April 2010, 11:47 AM   #2
ontoaba is offline ontoaba  Indonesia
diyAudio Member
 
ontoaba's Avatar
 
Join Date: Nov 2009
Location: Kudus, Malang, Dieng
Your diode working at high freq spiked voltage, you should use ultrafast and no need that big, and place some snubber if necessary.
Using MKP or equal caps is recommended for that 1uF caps to reduce the spike.
Your PCB layout is very bad. Bad tracing some high current line. Also you placed gate resistor too close to the drain, the other is better, you may look the differences between them at your current graph, and some other bad placed component.

I am not experienced well, and still have some problem with my own PFC too.
Just this what I know.
  Reply With Quote
Old 14th April 2010, 05:14 AM   #3
koosh75 is offline koosh75  United States
diyAudio Member
 
Join Date: Apr 2010
Location: California
Hi, thanks for your reply. I noticed that my PDF file wasn't showing the flood-fill on the PCB layout, so I've updated the file. You'll notice that the layout is actually not bad: the high-current loop area is very small, limited by the thickness of the board.

I think I disagree with you about the diode bridge having to use ultrafast diodes. This is not in the high-frequency current path. The top two diodes only conduct during startup, before the boost converter has charged the output caps. The lower two diodes I believe are also outside of the high-frequency path. The MOSFET body diode of the opposite phase conducts the 50KHz current. As a further example, see this semi-bridgeless example, which uses a standard-recovery GBU6J diode bridge. Choosing Between Semi-Bridgeless and Interleaved PFC Pre-Regulators | Jun 2009 Page of

Likewise, ON Semi Appnote AND8392/D uses a 15A 800V slow diode bridge in a semi-bridgeless PFC configuration:
http://www.onsemi.com/pub/Collateral/AND8392-D.PDF

I think something else is going on. Removing this bridge would turn my semi-bridgeless into a normal bridgeless PFC, nothing wrong with that. In this case, why is the current sense still distorted, and why does the fuse blow?

As for the input capacitor, I am using a 1uF Wima MKP, resulting in the waveforms shown. I'll take a closer look at the gate drive at the MOSFETs, but it looked ok when I checked it briefly. The snubbers on gate and drain are currently NOT stuffed, but I do have ferrite beads on the gates.
  Reply With Quote
Old 14th April 2010, 07:57 AM   #4
ontoaba is offline ontoaba  Indonesia
diyAudio Member
 
ontoaba's Avatar
 
Join Date: Nov 2009
Location: Kudus, Malang, Dieng
From that (voltage picture), your diode is working at high spiked place indeed. Placing it before the filter may solve the heating problem, but you will be another EMI problem to the input line. And your PCB layout is very bad, and from that (current picture), there may something disturbing the controller or may be sense faillure.
Lets see to the ON stuff. That PCB has (looks like) three ground plate area, Power stage, controller, and that small converter. You may search any PCB picture that may help you in designing your own.
With blown fuse, just remember that both your mosfet working at coupled transformer/inductor. If there some collision, your fuse will blow, seems this is what happening, from that (current picture) small collision frequently happened.
You need to analyze both gate voltage, and may be drain voltage if necessary.
good luck

Last edited by ontoaba; 14th April 2010 at 08:04 AM.
  Reply With Quote
Old 14th April 2010, 12:31 PM   #5
ontoaba is offline ontoaba  Indonesia
diyAudio Member
 
ontoaba's Avatar
 
Join Date: Nov 2009
Location: Kudus, Malang, Dieng
I am trying makes some PCB guide and found this PFC transformer floating, making your emi transformer added in series as PFC transformer.
This may solve some your problem.
Attached Images
File Type: png PCB.PNG (146.8 KB, 227 views)
  Reply With Quote
Old 14th April 2010, 05:54 PM   #6
diyAudio Member
 
megajocke's Avatar
 
Join Date: Jan 2003
Hmm...

Here's an equivalent circuit during the on time. (disregarding rds-on)

L4 is shorted for both input polarities by the lower bridge diodes. It is only the leakage inductance of L4 and circuit resistances that separates this from a hard short across the input. :O

I don't think a coupled inductor will work there either - look at how the input common mode potential changes between the switch on and off states.
Attached Images
File Type: png short.PNG (31.8 KB, 202 views)

Last edited by megajocke; 14th April 2010 at 05:58 PM.
  Reply With Quote
Old 14th April 2010, 06:49 PM   #7
koosh75 is offline koosh75  United States
diyAudio Member
 
Join Date: Apr 2010
Location: California
No, I don't think so. For the positive phase, during the on-time, the top diode you drew is reverse biased. The 50KHz return current flows through the opposite MOSFET body diode, which turns on quickly. The bottom diode you drew is forward biased, but its is slow diode, so it doesn't have time to turn on before the body diode begins conducting.

The lower bridge rectifier diodes only turn on and off at 120 Hz, to keep the output ground referenced to the line's low-side.

See the fig 4 example at http://powerelectronics.com/power_ma...efficiency.pdf

As for the coupled inductor, see the L4891 appnote Fig 9 and related verbage regarding the use of a coupled inductor rather than two separate inductors. http://www.st.com/stonline/products/...re/an/9119.pdf
  Reply With Quote
Old 15th April 2010, 11:34 AM   #8
ontoaba is offline ontoaba  Indonesia
diyAudio Member
 
ontoaba's Avatar
 
Join Date: Nov 2009
Location: Kudus, Malang, Dieng
Hi koosh,
I found another connector with High voltage line near the sensor wire on P13-P15 connection. Move away that line too.
And that bridge is needed on your circuit, why you remove it? Isn't all will be flying. I've writing about placing this bridge before the filter, it is not work, I am not recognize the function before, when you said that removing it makes your semi-bridgeless into a normal bridgeless PFC.
Also there is no grounding (small grounding caps) from ground to the earth.
Btw why you couple that inductor? Any reference?

Last edited by ontoaba; 15th April 2010 at 11:57 AM.
  Reply With Quote
Old 15th April 2010, 02:17 PM   #9
Eva is offline Eva  Spain
diyAudio Member
 
Eva's Avatar
 
Join Date: Oct 2003
Location: Near the sea
Send a message via MSN to Eva
Are you aware of the special common-mode filter requirements of that topology?

Half the switching waveform amplitude of square wave will appear as a common-mode component across the CM filter. The resulting CM interference may disturb the whole circuit.
__________________
I use to feel like the small child in The Emperor's New Clothes tale
  Reply With Quote
Old 15th April 2010, 08:00 PM   #10
diyAudio Member
 
megajocke's Avatar
 
Join Date: Jan 2003
Figure 9 in the ST appnote does not have a bridge fitted before the coupled inductor while figure 4 in the other document uses separate inductors. These features are mutually exclusive. No diode, no matter how slow, will be able to stand off the full line voltage, in the forward direction, for the full on-time of the switch without conducting a lot of current. This is what it would have had to do in a circuit with both a coupled inductor and a bridge before the inductor.

But in reality, the coupled inductor solution (without bridge by necessity) is practically unworkable too due to common-mode noise problems. The text regarding common mode noise referring to figure 9 in the ST document is wrong. The common mode rejection is better with separate inductors. Just look at the inductance seen by a common mode current. For coupled inductors it is just the leakage inductance, while it is the inductance of both filter inductors in parallell for non-coupled inductors. The latter will be much higher. It's nowhere near good enough though so that's why you'd want to fit the bridge like in figure 4 you are referring to, which is an approach that will not have extreme common-mode noise problems.

Last edited by megajocke; 15th April 2010 at 08:08 PM.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
1kW PFC-Boost ChocoHolic Power Supplies 75 23rd May 2011 07:55 AM
PFC voltage spikes tstitans Power Supplies 8 10th August 2009 06:34 PM
PFC inductor advice Andy F Power Supplies 48 21st January 2009 03:23 AM
Pfc + Ups = ??? ifrythings Power Supplies 9 30th December 2007 07:48 PM
PFC question medogrizli Power Supplies 12 3rd June 2007 03:49 PM


New To Site? Need Help?

All times are GMT. The time now is 12:06 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2