• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

Three phase active rectifier.

Status
Not open for further replies.

Pommie

Well-Known Member
Most Helpful Member
I currently have a project which takes a 12V 3ph supply and rectifies it with a Schottky bridge. Even with Schottky diodes, I'm loosing about 2V (edit, my mistake - should be 0.6V) and would therefore like to use Mosfets in an active bridge configuration to improve efficiency. Has anyone got experience in this area? Knows of any control chips? Found any examples? I found a few single phase examples but not 3 phase. I do not have access to the neutral(?) connection, only the phases but could make a neutral with 3 resistors in a star formation.

All suggestions welcome.

Mike.
 
Last edited:

tcmtech

Banned
Most Helpful Member
If you're losing ~2 volts in a full wave bridge rectifier they are either not Schottky or else they are either very poorly made ones.
 

crutschow

Well-Known Member
Most Helpful Member
The voltage drop depends upon the current.
How much current is the bridge providing?
Is the output filtered to get smooth DC?
Perhaps this paper will help.
 

Pommie

Well-Known Member
Most Helpful Member
The diodes are MBR1645 and the power used is 125W. From the datasheet they will drop about 0.3V each and so my 2V above was completely wrong. I was thinking of a latter stage of development where the power will be around 2kW and the diodes that will be required then have considerably larger drops. This is just a proof of concept and working out how to build an active bridge would be advantageous.

Edit, the output will be filtered to DC but with PFC to look resistive.
Edit2, thanks for the paper. I'll read later when I've more time. Gotta go shopping (hangs head in shame).

Mike.
 

spec

Well-Known Member
Most Helpful Member
Hi Pommie,

What is the source of the 12V? Is it a transformer?

If you are using a transformer as the voltage/current source, the output voltage will also drop as a result of the current loading on the transformer.

What value and type of reservoir capacitor are you using?

As you no doubt know, in a rectifier/reservoir capacitor circuit, the current is supplied to the reservoir capacitor in huge gulps, at the peaks of the rectified voltage camel-humps. The size of the current depends, to a first approximation, on the source impedance, band-gap voltage of the rectifier diodes, dynamic impedance of the rectifier diodes, and the impedance of the reservoir capacitor.

By the way, the band-gap voltage of silicon semiconductor drops by approximately 2mV/Deg C. So, with a diode operating at a typical junction temperature of 125 Deg C, the band-gap voltage would be reduced by 200mV from the band-gap voltage at 25 Def C.

The peak current, rather than the reservoir capacitor drain current, should be used to determine the rectifier diode forward drop, or MOSFETs Vds, rather than the reservoir load current. A good working estimate of the peak current is ten times the reservoir load current, 100A in this case.

If you use MOSFET active rectifiers, this high current needs to be taken into account. Using MOSFETs, would give you a forward voltage drop (Vds) of,

Vds= Ipeak * Rdss * 2

Where,
Vds : Voltage between drain and source of the MOSFET in Volts.
Rds : Resistance between drain and source of the MOSFET in Ohms.
2 : MOSFET channel temperature/Rds factor

Take a 0.010 Ohm Rds MOSFET,

Vds =100A * 0.01 * 2 = 2V.

Then multiply that by 2, because you would have a minimum of two MOSFETs in series to form a bridge rectifier (I think this is correct, but have not checked), and you get 4V, so MOSFET active rectifiers may not be the answer.

If you do decide to use schottky rectifiers, I would advise going for as higher current rectifiers as you can reasonably afford/fit, say 100A upwards. The next best thing is to parallel diodes to get a lower forward drop.

The ripple voltage across the reservoir capacitor will also subtract from the average output voltage. To a first approximation, the peak-to-peak ripple voltage in Australia (240V, 50Hz) can be calculated by,

Vripple = (I * 0.01)/ Cres

Where,
Vripple: Peak to peak ripple voltage across the reservoir capacitor in Volts
I: Load current flowing out of the reservoir capacitor in Amps
Cres: Reservoir capacitor capacitance in Farads

spec

DATASHEETS
http://www.onsemi.com/pub_link/Collateral/MBR1635-D.PDF
 
Last edited:

MrAl

Well-Known Member
Most Helpful Member
I currently have a project which takes a 12V 3ph supply and rectifies it with a Schottky bridge. Even with Schottky diodes, I'm loosing about 2V (edit, my mistake - should be 0.6V) and would therefore like to use Mosfets in an active bridge configuration to improve efficiency. Has anyone got experience in this area? Knows of any control chips? Found any examples? I found a few single phase examples but not 3 phase. I do not have access to the neutral(?) connection, only the phases but could make a neutral with 3 resistors in a star formation.

All suggestions welcome.

Mike.

Hi there,

The way i understand it is relatively simple. There are four mosfets wired up as a bridge considering their internal body diode polarities only. That means that even without turning even one on, it forms a full wave bridge rectifier.
Of course the body diodes in mosfets are not that good, so there will still be voltage drops, but once the mosfets are turned on the voltage drops are much lower. The turn on is determined by measuring the drain to source voltage and working the gate signal like would be done in a spice model of a zero voltage diode. The gate signals have to be differential of course because the source voltage is always changing relative to 'ground'.

To find out more, check out the IRF1166 or IRF1167. That should provide some practical information.
 

schmitt trigger

Well-Known Member
The key with Pommie's request is that he want to rectify a three phase voltage

The problem with synchronous rectifiers is that they conduct equally well in both polarities. In a single phase the gate drive is perfectly synchronized with the drain voltage so cross conduction does not occur.
In a three phase rectifier however, unless care is taken to fully synchronize the gate drive timing, there will be an overlap where the Mosfet is still being driven and an opposite phase's voltage has increased which will create a cross conduction path.

The cross conduction, being very low impedance, will create very large fault currents.
 

spec

Well-Known Member
Most Helpful Member
The key with Pommie's request is that he want to rectify a three phase voltage

The problem with synchronous rectifiers is that they conduct equally well in both polarities. In a single phase the gate drive is perfectly synchronized with the drain voltage so cross conduction does not occur.
In a three phase rectifier however, unless care is taken to fully synchronize the gate drive timing, there will be an overlap where the Mosfet is still being driven and an opposite phase's voltage has increased which will create a cross conduction path.

The cross conduction, being very low impedance, will create very large fault currents.
Hmm, not too sure what you mean- to me, 'synchronous rectifier' is a generic term for a rectifier, normally a MOSFET, that is turned on/off synchronously in relation to the power line and are used on N phases.

With MOSFET synchronous rectifiers the conduction is in the forward or reverse direction, depending on application. In a normal bridge application the conduction would be in the reverse direction, in parallel with the forward conduction of the intrinsic D/S schottky diode.

spec
 

Pommie

Well-Known Member
Most Helpful Member
I think I'm going to stick with schottky diodes for now. I'm going to concentrate on the power factor correction aspect of the project first.

Thanks for the help so far. Will return to this some time in the future.

Mike.
 

spec

Well-Known Member
Most Helpful Member
I think I'm going to stick with schottky diodes for now. I'm going to concentrate on the power factor correction aspect of the project first.

Thanks for the help so far. Will return to this some time in the future.

Mike.
Good move Pommie.:)

spec
 

MrAl

Well-Known Member
Most Helpful Member
The key with Pommie's request is that he want to rectify a three phase voltage

The problem with synchronous rectifiers is that they conduct equally well in both polarities. In a single phase the gate drive is perfectly synchronized with the drain voltage so cross conduction does not occur.
In a three phase rectifier however, unless care is taken to fully synchronize the gate drive timing, there will be an overlap where the Mosfet is still being driven and an opposite phase's voltage has increased which will create a cross conduction path.

The cross conduction, being very low impedance, will create very large fault currents.
Hi there,

You do realize this has been done ad infinitum with three phase systems using SCR's right?

Also, a simple model for a diode would be a switch with a very small resistance in series, and when the voltage on one terminal (call it the anode) becomes greater than the voltage on the other terminal (call it the cathode) the switch turns on, and that means the diode model conducts. It stays 'on' until the voltage polarity goes to zero.
Thus i can not see how this would not work somehow with MOSFET's where we measure that voltage and generate a drive signal based on that voltage just like in the model.

However, if you have found some definite fault condition that occurs with the MOSFET bridge then perhaps you can illustrate either with an exact description of the fault or a simulation of some kind showing the fault. Describing the problem would at least let others follow your thoughts on this.

The main reason for wanting to do something like this is to either increase efficiency or to reduce voltage drop. When the system voltage is low (like 12vac) losing 0.5v is much more significant than when the system voltage is 120vac.

There are better Schottky's too. The company formerly known as Zetex made some really good Schottky's.
 

tcmtech

Banned
Most Helpful Member
As for finding good quality high current Shottky or Avalanche diodes they are standard components in most every high current 12 - 24 volt commercial application alternator now which makes them pretty cheap to find as prebuilt bridge rectifier setups.
 

schmitt trigger

Well-Known Member
Hi there,

You do realize this has been done ad infinitum with three phase systems using SCR's right?


Thus i can not see how this would not work somehow with MOSFET's where we measure that voltage and generate a drive signal based on that voltage just like in the model.

.
Of course I do, I've been involved with power electronics since the late 1970s.

Your second sentence is exactly what is meant by "synchronizing". One has to synchronize the Mosfet's gate drive to a conduction angle identical to what a diode would have, because the Mosfet does not have natural commutation.
The SCR will naturally commutate even though the gate may be still driven.

In plain English....an SCR only requires synchronized timing of the turn-on pulse.
On a Mosfet, additionally the width of that pulse must also be controlled to prevent the channel from conducting while the current reverses its flow.
 

Pommie

Well-Known Member
Most Helpful Member
I didn't realize that you could use those Linear chips in a 3ph design. Nice find.

Thanks,

Mike.
 

MrAl

Well-Known Member
Most Helpful Member
Of course I do, I've been involved with power electronics since the late 1970s.

Your second sentence is exactly what is meant by "synchronizing". One has to synchronize the Mosfet's gate drive to a conduction angle identical to what a diode would have, because the Mosfet does not have natural commutation.
The SCR will naturally commutate even though the gate may be still driven.

In plain English....an SCR only requires synchronized timing of the turn-on pulse.
On a Mosfet, additionally the width of that pulse must also be controlled to prevent the channel from conducting while the current reverses its flow.
Hi again,

The MOSFET driver variety accomplish that by measuring the voltage from drain to source. Knowing that voltage (and it's polarity) is the key to knowing when to switch the MOSFET on and when to switch it off.

This is different than when doing say a synchronous buck converter where the gate signals are generated a priori.
 

sbkenn

New Member
Hmm, not too sure what you mean- to me, 'synchronous rectifier' is a generic term for a rectifier, normally a MOSFET, that is turned on/off synchronously in relation to the power line and are used on N phases.

With MOSFET synchronous rectifiers the conduction is in the forward or reverse direction, depending on application. In a normal bridge application the conduction would be in the reverse direction, in parallel with the forward conduction of the intrinsic D/S schottky diode.

spec
Actually, in this situation, the MosFET is turned on to conduct in the SAME direction as the diode.
 
Status
Not open for further replies.

Latest threads

EE World Online Articles

Loading
Top