Recommendation for Power factor correction control Buck or Boost?

Go To Last Post
14 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have a 3 phase 400Vac supply rectified to 540VDC.

 

A full bridge phase shift converter converts the 540VDC @10A to 10Vdc @500A.

 

I try to figure out a power factor correction frontend for this converter.

 

The main question is : Buck or Boost?

 

I tend to use the buck topology rather than the populated Boost

 

My Reasons for Buck Topolgy:

 

Option 1

-No Inrush current. (soft start is possible)

-Possiblity to step down the 540VDC to 320VDC so that the phase shift converter design will be simpler and cheaper. Semiconductor Voltage rating is 600V instead 1200V which saves cost.

 

Option 2

There is also a possiblity that I convert 540VDC to 50VDC through the Power factor correction converter. The semiconductor Voltage rating reduces to 100V instead of 1200V which gives me the possibility to reduce semiconductor cost and increase the frequency and reduce system volume.

 

Any recommendation between Option 1 and Option 2 , Or maybe some remarks for Boost Topolgy? (Boost topolgy has some drawbacks in my case. The Bus voltage must be 40-50V greater than 540VDC bus which makes the design and implementation harder and cost of the semiconductors increases. The inrush current is also a problem for boost topolgy which needs extra effort and circuit and increase costs. 

 

 Thanks for your input.

 

This topic has a solution.

________________________________ We dream of a world where current does not need the voltage to flow.

Last Edited: Wed. Mar 13, 2019 - 06:59 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Is there not a trade off between higher current vs higher voltage?

and it might be ( I am not a High voltage specialist) that higher current is a higher danger level. Voltage will jump across air, but high currents can be deadly, so it might be that for safety reasons people go for high voltage and low current(and limited) instead of lower voltage but more current.

just my 2ct

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The converter needs a power factor correction state. So I must use at least 1 high voltage igbt and a high voltage diode for both cases.

 

Assuming the following power stage ( phase shift converter after the PFC) ,a 1200V 10A semiconductor with its 1200V insulated gate driver is much more expensive than a 600V 15A semiconductor + gate driver. The 1200V design of the post converter is also much more complex and the danger level ist much much higher.

 

But I am not sure what are the main drawbacks of PFC Buck topology. As I am working always on the industrial supply network(400Vac), there will be no need for a wide range input Voltage. The High Voltage Bus will be between 540VDC and 450VDC.  Because of this reason I tend to use the buck topology.

 

Any ideas?

________________________________ We dream of a world where current does not need the voltage to flow.

Last Edited: Fri. Mar 8, 2019 - 10:07 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

3 phase 400Vac supply rectified to 540VDC

how is that rectified ?

if it's just 6 diodes, you will need to change it! (for the grid it needs to look as 3 resistors, diodes will only load in the tops.)

 

This reply has been marked as the solution. 
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

incal wrote:
There is also a possiblity that I convert 540VDC to 50VDC through the Power factor correction converter.

 

An electric car battery is made of many small elements connected in series and parallel in such a way that it ends up to about 400 - 500V. Must be a reason for why they decided this voltage and not 100 or 48V or any other low voltage. 

 

A three phase PFC is not just one, are three in fact where you face multiple isolated gate drivers and inductors. My opinion is that you do not need it. 5kW is not something that challenges a three phase 400VAC grid connection.

 

You need to go back to your project requirements. About PFC, look at local regulations and start from there. The electric power company is worried about a bad power factor, not the consumer. This is because a bad power factor increases the current which in turn increases the loses on the distribution network. If not mandatory, avoid it to save cost. 

 

The crest factor is also three times lower because there are six picks for a three phase bridge versus two picks for one phase.

full wave three phase rectifier conduction waveform

 

I did a Google search for "VFD input block diagram" and all what I saw was the regular 6 diodes three phase bridge rectifier maybe followed by an inductor to smooth the AC picks for a reduced crest factor and the following capacitors. This is what I would do. Add that inrush relay - resistor and get done with it. You have enough to worry about the full active bridge.

 

If you are still keen about PFC, draw a rough schematic and post it here. I think you will get more comments. So far is a bit not clear how exactly you want to build the PFC.

 

A single PFC be it buck or boost, will not help much if you keep the three phase bridge because you only can load a phase for 60° (see above figure). You can't take any advantage of a lower voltage from a buck PFC because overall, the efficiency will be lower. The PFC cost will be much higher that the difference in semiconductors cost rated 1200V vs 600V.

 

Again, keep it simple.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I plan to use a dedicated power factor controller IC from TI. (Ucc29910) .

 

This is the basic PFC-Buck schemtic:

 

 

 

 

 

This is the datasheet:

 

http://www.ti.com/lit/ds/symlink/ucc29910a.pdf

 

 

As far as I investigated the datasheet, this Ic deals only with the rectified Line. I predict that it shapes the input current modulating the chopper so that it will come close to a sinusoidal form. 

 

I plan to cascade a phase shift converter (full bridge) after this stage. So that the needle form of the input current of the phase shift converter will be delivered by the PFC state. But the input of the PFC-Buck converter will be a sinus as far as the algorithm of the PFC-IC suceeds.

 

I think that this is not a regular Buck converter but it has also the ability to shape the input current regardless of the load current by exchanging and storing some buffer energy in the PFC inductor. (I am not %100 sure this is my prediction)

 

The rectification of the 400Vac line is through a B6 diode bridge.(six diodes) (Yes it has 300hz ripple just as you showed).

 

The phase shift full bridge converter has really sharp input current just like a needle. This results in very poor PFC. I hope that this arrangement improves power factor. As it is a buck configuration and the switch is series connected to the line, it will have an inrush current management. (soft start) . It will also be possible to use 600V semiconductors.

 

Do you think that adding a passive PFC is more effective for 3 phase networks. A big inductor and an overdimensioned capacitor bank may smoothen the input current of the system and can deliver the high peak current demand of the cascading phase shift converter.

 

What do you think about this setup? (Active and passive variants)

________________________________ We dream of a world where current does not need the voltage to flow.

Last Edited: Sat. Mar 9, 2019 - 07:46 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The 3 phase 400Vac Line will be rectified through a B6 bridge. (Yes six diodes). I plan to use a dedicated PFC controller (buck type). Interestingly I noticed that the PFC-IC deals only with the rectified DC bus. 

 

I sent a basic schematics of the PFC stage above. The above illustration has monophase 230Vac input and a B4 rectifier. In my case I have 3 phase 400Vac and a B6 bridge rectifier.

 

Do you think that this IC is not effective enough to improve the power factor.

 

(As it is shown in the schemetics, The PFC-IC has no information of the Ac line. It samples only the rectified Bus. It is a possiblity that the algoritm of the IC works best for 1 phase 230VAC. But fails for 3 phase network if the angle between the phases is less.

 

Thanks for your input. 

________________________________ We dream of a world where current does not need the voltage to flow.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

There is also a modified version of the schematics where the high side switch is eliminated but there is a need for an isolated measurement of the PFC output voltage.

 

I notice that this IC sample the output of the rectifier (also after smoothing capacitor) (see linesns input of the IC). It also samples the output of the PFC. (Vbulk pin of the IC).

 

If the capacitor after the rectifier is a smoothening capacitor, the IC will not have any information abaout the phase angle. But if it is a small EMC filter capacitor(such as a film capacitor) then it may be that the IC has also information over the phase angle so that it can adjust the current synchronisation. If it is just a small film capacitor then it may be that the phase information is important for the IC to work properly. 

 

Unfortunatelly the the 3 phase rectifier decreases the ripple and the phase information is less accurate. Just 60 degree and very little ripple on the top. I am not sure if the IC can handle that if it must extract phase information from the rectified bus to implement the its algorithm

 

________________________________ We dream of a world where current does not need the voltage to flow.

Last Edited: Sat. Mar 9, 2019 - 08:47 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

angelu wrote:
. The electric power company is worried about a bad power factor, not the consumer.

Not strictly true. For non-domestic customers some power companies install kVA meters as opposed to kW. Then the worry of bad power factor is transferred to the customer.

It is, after all, he who has the ability to measure and do something about it such as install power factor correction capacitors on the incoming such as I mentioned in the OP previous thread about his low voltage monster converter.

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Another issue is that maybe you attempt to claim that your device is more energy efficient than the competitior.

You claim you deliver 10V @500A with an efficiency of %98. (for example using active rectification on the secondary side with ultra low rdson mosfets).

 

At the end you end up demonstrating it with a clamp ammeter. And if your power factor is poor, it may be so that your converter efficiency of %98 can not be demonstrated on the display of the ammeter.

 

Because the ammeter adds up the whole fluctuating reactive power and its current components on its display.

 

You have build an efficient converter but you can not demonstrate it through an ammeter.

 

It will be very hard to explain to the customer that the ammeter is wrong.

 

And at the end if you want to sell it to someone (local or abroad) ,you may have to satisfy with the rules of EN 61000-3-2 or something similar.

 

I am also not a fan of PFC, I also believe that it is an extra cost for the developer. But if it can be constructed as a buck topology to save further costs of the following circuit (lower voltage semiconductors,being able to use cheap IPM modules available for mass production instead of much smaller and expensive industrial markets (1200V market), cheaper capacitors (caps are only available up to 450V), maybe getting rid of the balancing resistors of the high voltage cap-bank for 540VDC (only 1 400V cap instead of 2xseries connected 400V caps) etc., the cost of PFC may diminish.

________________________________ We dream of a world where current does not need the voltage to flow.

Last Edited: Sat. Mar 9, 2019 - 01:53 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

incal wrote:
Unfortunatelly the the 3 phase rectifier decreases the ripple and the phase information is less accurate. Just 60 degree and very little ripple on the top. I am not sure if the IC can handle that if it must extract phase information from the rectified bus to implement the its algorithm

With only 60° per half cycle you cannot claim this is a PFC. And in the end a proper inductor between bridge and caps will do almost the same thing.

 

Without PFC I don't think you get  more than 95%. With PFC, subtract at least 2%, depends how much you lower the voltage.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

angelu wrote:

incal wrote:
Unfortunatelly the the 3 phase rectifier decreases the ripple and the phase information is less accurate. Just 60 degree and very little ripple on the top. I am not sure if the IC can handle that if it must extract phase information from the rectified bus to implement the its algorithm

With only 60° per half cycle you cannot claim this is a PFC. And in the end a proper inductor between bridge and caps will do almost the same thing.

 

Without PFC I don't think you get  more than 95%. With PFC, subtract at least 2%, depends how much you lower the voltage.

 

I investigated the IC application note and noticed that the capacitor just after the rectifier is a very small filter capacitor (nF range , 470nF).

 

This means this IC needs to measure the deep grooves , the zero crossing/touching points of the AC line just after the rectifier.

 

So I just as recommended , I will use 3 seperate PFC regulators for each phase. There will be also 3 seperate isolated phase shift converter for each PFC leg. Finally I will combine the outputs of each 1700W converter to reach 5000W.

 

I send the schematics . Do you think that a big inductor (1mH..2mmH) at the input can satisfy a PF of 0.95 without using any active PFC.

 

What do you think abaout the following schematics? the following schematics has also an advantage that i can use cheaper semiconductors (600V) and capacitors for the phase shift converter.

 

 

 

 

 

 

________________________________ We dream of a world where current does not need the voltage to flow.

Last Edited: Sun. Mar 10, 2019 - 04:25 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

    350 / 540 = 0.6481. arcsin(0.6481) = 40.4°. This means that your buck PFC will work for two times 90 - 40.4 which is almost 100° out of 180°. Better than 60° when using a 6 diode bridge rectifier.

 

    If you could use the neutral, then you could use boost PFC which would regulate over entire AC cycle and give you a safe 400VDC. Using 450V caps and 600V semiconductors would work nicely. However, some prefer not to wire the neutral.

 

    One thing that you need to worry is how to balance all three sections so you load the three phases equally. With a 6 diode bridge, this comes naturally.

 

    I do not know much about passive PFC.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I also asked the same question on TI forum. TI says that buck topology is not suitable for high power and high voltages. Adding that this topology is very sensitive to failure of the semiconductor switches during a line surge or lightning strike.   

 

They recommend the boost topology. That means my dc bus voltage should be 40-50V higher then the rectified bus voltage. It is around 590V-600V. This means lower frequencies and more complex design (gate drives etc.). Or I take an advanture to the world of SiC Mosfets.

 

I think that boost topology is easy to paralell because it has a diode at its output. The buck topology does not have a diode at its output, so that extra diode may be needed to paralel buck pfcs from different phases. 

I may combine the outputs of each PFC-boost stage to a common DC-Bus. It can be possible to build only 1 full bridge converter after this combined common PFC-Output.

 

If one conveter is fed through 3 seperate PFC-buck regulators, the issue with the balancing is also solved.But I am not sure if the 3 seperate PFC-buck regulators are in paralell , will there be a problem when regulating the output voltage. 

 

But I am not sure if paralelling 3 PFC-boost stage for each seperate phase have any down sides. I also dont like to use the neutral line.

 

If I have 3 phase supply, I always use the rectification between phases without using the neutral line. I think it is better that the converter itself creates its own neutral line regardless of the outer neutral line.

 

It would be very instersting that there exists an intelligent IC that can sample directly the output of B6 rectifier and make assumptions through 60 degree valleys. It should be theoretically possible to extract all phase information through the output of the 3 phase rectifier.

A PFC for the 3 phase network with only 1 switch could be possible.

 

I am confused. I have 2 options:

 

1- 3 PFC-Buck for each Phase combined to a common DC-Bus + 1 full bridge converter. PF 0.96-0.98

 

2- A big inductor after the B6 bridge. PF 0.85-0.9?! (I am not sure). What may be the disadvantages of the big inductor. (2mH..5mH). I may need bigger DC bus capacitors or the slope, voltage and the duty cycle of the cascading converter is limited.  

 

 

 

 

 

________________________________ We dream of a world where current does not need the voltage to flow.

Last Edited: Tue. Mar 12, 2019 - 06:01 AM