3-phase rectifier efficiency

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

Here is the Description of the issue: We have a 3-phase generator that in its standard working conditions, turning speed is around 3000 to 8500 rpm (which means that the output frequency is around 50 Hz to about 140 Hz). The manufacturer states that the output voltage (phase to phase) is between 35 and 95 V AC depending on speed and 300 W output power is guaranteed even in 3000 rpm without decreasing in voltage level. I want to build a regulator for this generator that be able to give me 30 V DC - 10 A (300 W). The simple solution is using a 3-phase rectifier and a buck regulator. But this solution is two main drawbacks:

  • The first is for 10 A I would have a 10 A * 1.4 V = 14 W loss just in rectifying circuit and for 90% efficiency I can have total of 30 W in losses. So 14 W its just a lot.
  • The second is in the case of 95 V AC I would have almost 127 V DC after rectifing! And no ready to use ICs like LM2596 and common buck controllers can handle such voltages.

My primary solution is using 6 MOSFETs (instead of a 3-phase-diode-bridge) like inverters used to control speed of induction motors. And get 30 V DC right after them.

But the problem is "ready to made controllers" like LTC3722 have just a lot of functions and protections that i should use dosens of circuits to manually create them. such a big work! and Of course the Propability to failure in such circuits would be much more than Single-Chip-Solutions like "LM2596".So if you know any IC(buck regulators...Buck controllers... 3phase bridge controllers and...)or have any discreet circuit solution in mind, I'll be grateful if share it with me here.

Last Edited: Wed. Aug 16, 2017 - 06:58 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

What about Schottky diodes? I think you could reduce the voltage drop to half.

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

How much knowledge do you have about electronics?

Designing a circuit like this is doable, but quite a challenge, especially when you need a very high  efficiency.

As you've noticed, once power levels get higher the amount of ready to use single chip solutions get less.

 

Maybe browse around a bit in the Ali / Ebay specials ?

For example: this one goes up to 108V input and claims to be able to handle up to 20 A.

https://www.aliexpress.com/item/...

 

A 1:3 input voltage range is also quite big.

Most off the shelf solutions do not reach that.

 

MOSfets are not really fit for switching from an AC input source.

They always have a body diode which starts conducting if the polarity reverses.

But you can use MOS fets in those active rectifiers, but then they are used "reversed".

Explaining it is quite simple.

First you temporarily forget the gate and use the body diode of the MOSfet as a regular diode.

Then you turn on the fet during the time the body diode conducts.

RdsOn of you fet should be so low that it takes all the current.

But be sure to turn the MOSfet off before the polarity reverses...

https://duckduckgo.com/html?q=+s...

https://en.wikipedia.org/wiki/Ac...

 

esmaeil6514 wrote:
... primary solution = MOSfets ... And get 30 V DC right after them.

So this won't work because of above.

(Or prove me wrong with a schematic :)

 

And maybe look into Silicon Carbide.

This semiconductor has some nice properties and is gaining a maket share in especially high power electronics.

 

And this is probably not the best forum for serious analog electronics. Maybe try the forum @ eevblog?

Have fun.

Paul van der Hoeven.
Bunch of old projects with AVR's:
http://www.hoevendesign.com

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

esmaeil6514 wrote:
And no ready to use ICs like LM2596 and common buck controllers can handle such voltages.

The controllers don't need to handle the high voltages, they just need to control the switching element. How many switchmode supplies do we do in everyday equipment that are running off the mains voltage? Controllers such as the UC3842 are common in such applications. They run off low voltage at a moderately low current. They don't need to handle the high voltages - all they need to do is control the mosfet that requires are 12V and thr feedback voltages are dropped by a resistor divider. So, I'd suggest a 300W buck converter to run at those voltages is rather straightforward.

 

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

turning speed is around 3000 to 8500 rpm

What drives the generator? (fuel engine, wind, water ??) and can the speed ever go below or above that? If this is the case then you will have bigger problems than what is stated above.

John Samperi

Ampertronics Pty. Ltd.

www.ampertronics.com.au

* Electronic Design * Custom Products * Contract Assembly

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

Kartman wrote:
How many switchmode supplies do we do in everyday equipment that are running off the mains voltage?

Quite a lot actually.

Lots of (very small) SMPS circuits have a DIP8 running directly on 300Vdc and a 400V switch.

Used often in those wall warts or USB SMPS's, but they are usually limited to a couple of watts.

 

But for Higher power SMPS's your absolutely right. All I've seen use external switching elements.

Paul van der Hoeven.
Bunch of old projects with AVR's:
http://www.hoevendesign.com

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

perhaps take a look at something like this:

 

http://www.linear.com/product/LT...

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

Upvoted sparrow2

short answer, excellent link to the LTC3895 - 150V Low IQ, Synchronous Step-Down DC/DC Controller.

It claims to do 5A with an efficiency of 95%. Nice.

But to reach 10A you need to at least use different MOSfet's.

Paul van der Hoeven.
Bunch of old projects with AVR's:
http://www.hoevendesign.com

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

esmaeil6514 wrote:

Here is the Description of the issue: We have a 3-phase generator that in its standard working conditions, turning speed is around 3000 to 8500 rpm (which means that the output frequency is around 50 Hz to about 140 Hz). The manufacturer states that the output voltage (phase to phase) is between 35 and 95 V AC depending on speed and 300 W output power is guaranteed even in 3000 rpm without decreasing in voltage level. I want to build a regulator for this generator that be able to give me 30 V DC - 10 A (300 W). The simple solution is using a 3-phase rectifier and a buck regulator. But this solution is two main drawbacks:

  • The first is for 10 A I would have a 10 A * 1.4 V = 14 W loss just in rectifying circuit and for 90% efficiency I can have total of 30 W in losses. So 14 W its just a lot.
  • The second is in the case of 95 V AC I would have almost 127 V DC after rectifing! And no ready to use ICs like LM2596 and common buck controllers can handle such voltages.

My primary solution is using 6 MOSFETs (instead of a 3-phase-diode-bridge) like inverters used to control speed of induction motors. And get 30 V DC right after them.

 

The rectifier current applies at the 35~95VAC level, so it will be less than 10A. Schottky diodes could be one solution.

You could also make a 3 phase  diode bridge, using MOSFETS, and isolated gate drive all those FETS carefully, to turn ON the FET when the Diode is forward biased, to make a 'better diode'.

You can get 200V MOSFETS in the 25mOhms to sub 10mOhms region, so that's 2.5 ~ 1W series loss at a conservative 10A average (plus gate drive power budgets)

 

However, that's a lot of parts, and a Schottky is 200v  15A at 890mV, and quite a lot cheaper/simpler. 

 

Comes down to how much you are prepared to pay for that last 4~6 watts of power saving :)  Probably there are other places you can gain 4~6 watts for less $ ?

 

Last Edited: Sun. Jul 30, 2017 - 10:39 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hi and thanks for considering my problem…

To El Tangas: About schottky diodes, as the forward voltage drops, the allowable reverse bias voltage also drops! As an example “STPS3045C” is power schottky with 0.5v forward drop and 45v allowable reverse bias. The best diode I found was “MBR 20200”  with 200v reverse bias capability and 0.9v forward drop which means 18W loss in 10A “just in rectification circuit”.

To Paulvdh: thanks for searching for me. The module you found is DC-DC and I need a  AC-DC converter. The best I found was “EA15A3H” that is Generator Automatic Voltage Regulator. “I think” it  exactly did what I told you about(using inverter bridge). I can’t use that module cause it designed for 380v line voltage and  my generator produse 35-95VAC…

And for your next idea you gave me(sensing diode conduction time) its just a great idea! Thanks and I’ll think about it… it also has drawbacks like I cant use PWM technics for making my desired voltage right after bridge…

To Kartman: you are right but all buck regulators need a DC voltage to work and will need the  bad ‘rectifier’ circuit! And of course buck regulators has efficiency of “tipically” 90%... adding 10% to losses…

To js:  its driven from an engine . No, the speed in most cases is around 6000rpm…

About How many switchmode supplies… most of them (if not say all) are fly backs that uses isolated transes. for 300W the trans became big and I don’t have space for it. Alse I think I can solve the problem without trans(I also don’t need isolation…).

To sparrow2: I saw LTC3722 but this one is just better. Still need DC voltage…

To Who-me: 2W compared to 18W  means more than 5% better efficiency and it worth trying… and about the price I’m richsmiley)

If I drove any circuits I’ll post it here for you. The best solution till know I think would be using mosfets as diodes and something like LTC3895 after it…

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

My point regarding switched mode supplies is that many of them use a uc3842 ic and use a mosfet to switch the high voltage - thus you don't need a high voltage ic. So, using one of these very cheap ics you can make your buck converter easily and cheaply.

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

esmaeil6514 wrote:
About schottky diodes, as the forward voltage drops, the allowable reverse bias voltage also drops!

Obvisouly you have to find a suitably-rated schottky rectifier - Mouser has plenty, for a start:

 

http://www.mouser.co.uk/Semicond...

 

 0.9v forward drop which means 18W loss in 10A “just in rectification circuit”.

You've missed the point, again, that the rectifier is on the "primary" side - so it does not see the 10A output current.

 

You say the worst-case input is 35V - so 300W is then 8.6A

Top input is 95V - when 300W is only 3.2A

 

I cant use PWM technics for making my desired voltage right after bridge…

Why not? That is how most SMPS work!

 

for 300W the trans became big and I don’t have space for it.

Look at the size of an off-the-shelf 300W SMPS; eg, a PC power supply

 

One of the key features of SMPS is that they switch at high frequencies - so the transformers (if used) can be much smaller ...

 

EDIT

 

remove spurious line

 

Last Edited: Wed. Aug 16, 2017 - 09:36 AM