Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronic Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Offline Switch Mode EV chargers should be made of paralleled <3kW modules?

Status
Not open for further replies.

treez

Advanced Member level 5
Joined
Sep 22, 2008
Messages
7,519
Helped
547
Reputation
1,097
Reaction score
529
Trophy points
1,393
Location
cambridge
Activity points
75,593
Its very notable how High Power (3kW) Switch Mode Electric Vehicle chargers always use multiple paralleled modules, each individually of power less than 3kw, instead of using one single high power module.

The below chargers all testify to this….

Multi-Kw charger made up of 2kw modules in parallel….
http://www.eltek.com/detail_products.epl?id=1142907&cat=&k1=&k2=25515&k3=&k4=&close=1

Harmony EV charger, Multiple parallel modules…(look less than 2kw each)
http://www.supplydesign.com/technology.html

Single phase EV charger: 3.3kW
http://www.metricmind.com/products/brusa-nlg513-wx/
…………………………………………………………………………..
Is the reason for this because when you go above 3kw, off the shelf ferrite cores for the transformer get too large, and they stick out well above the components around them, meaning that gap-padding the other components to the lid of the case is not practical as the lid is too high due to it having to clear the transformer height?
………………………………………………………………………….
Also, the other reason I presume, is that Electric Vehicle chargers are often located outdoors, and so are susceptible to falling to mains transients…

Littelfuse transient protectors.
http://www.littelfuse.com/~/media/e...ghting_surge_protection_modules_flyer.pdf.pdf

There is obviously little point in making an expensive high power single module if there’s a chance of it just succumbing to mains transients…might as well make a load of cheap ones instead, then if any of then get taken out by a mains transient, you just replace the one that blows. Since MOVs have a tolerance, the module which has the lowest breakover voltage MOV will quench the transient and blow (eventually), leaving the others fairly unscathed.

This thread tells of the problem of mains transients, and how it means that no equipment should be made expensive (ie not a single high power module) , because it will simply not last if installed at a site where mains transients are severe…
https://www.edaboard.com/thread352771.html
...and with all these EV chargers coming on to the power network, and with them switching on and off regularly, the incidence of mains transients will get far worse.

So do you agree that Switch Mode EV chargers should be made up of multiple modules in parallel, all below 3kw each?
 

FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
46,801
Helped
13,894
Reputation
28,036
Reaction score
12,528
Trophy points
1,393
Location
Bochum, Germany
Activity points
273,097
From the three presented power supplies, the 5 kW (a bit more than 2 kW!) three phase "Harmony" modules looks interesting. Making a three phase power supply out of single phase modules doesn't look reasonable to me, except you want optional single phase operation as often required for E-Car chargers.
 
  • Like
Reactions: treez

    treez

    points: 2
    Helpful Answer Positive Rating

treez

Advanced Member level 5
Joined
Sep 22, 2008
Messages
7,519
Helped
547
Reputation
1,097
Reaction score
529
Trophy points
1,393
Location
cambridge
Activity points
75,593
ok thanks.
Making a three phase power supply out of single phase modules doesn't look reasonable to me
..surely its ideal, you put three equal power chargers, one on each phase and you have balance..surely?
 

FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
46,801
Helped
13,894
Reputation
28,036
Reaction score
12,528
Trophy points
1,393
Location
Bochum, Germany
Activity points
273,097
A three phase power supply gets along with almost no DC storage capacitors if it's not required to override mains interruption. Single phase supply has pulsating (0 to 200 %) input power and needs respective storage capacitors.
 
  • Like
Reactions: treez

    treez

    points: 2
    Helpful Answer Positive Rating

treez

Advanced Member level 5
Joined
Sep 22, 2008
Messages
7,519
Helped
547
Reputation
1,097
Reaction score
529
Trophy points
1,393
Location
cambridge
Activity points
75,593
The attached LTspice simulation of a three phase rectifier shows the problems that 3-phase operation has for an EV charger.
Whether you are star or delta connected, the phase currents are distorted compared to a single phase mains followed by active PFC stage.

I cannot see how the three phase input , via a three phase rectifier could be used as an input to an switch mode EV charger, the simulation bears this out. I am not saying it could not be done, just that the power factor would be too poor for an EV charger, where the shear number of them means that 99% PFC is required..surely?

By the way, many thanks for the LTspice sim, it was yourself FvM who corrected it for me some time back.
 

Attachments


FvM

Super Moderator
Staff member
Joined
Jan 22, 2008
Messages
46,801
Helped
13,894
Reputation
28,036
Reaction score
12,528
Trophy points
1,393
Location
Bochum, Germany
Activity points
273,097
The attached LTspice simulation of a three phase rectifier shows the problems that 3-phase operation has for an EV charger.
Whether you are star or delta connected, the phase currents are distorted compared to a single phase mains followed by active PFC stage.
I presumed you know how a three phase PFC works. It can't use a line-commutated DB6 rectifier. If not implemented as unidirectional converter, combining three single phase PFC outputs in a common DC bus, it will use a bidirectional active front end topology, as discussed in previous threads. https://www.edaboard.com/thread342659.html
 
  • Like
Reactions: treez

    treez

    points: 2
    Helpful Answer Positive Rating

cts_casemod

Member level 4
Joined
Dec 9, 2010
Messages
72
Helped
14
Reputation
28
Reaction score
14
Trophy points
1,288
Location
London
Activity points
1,950
When you go above a certain power level you have to either use multiple paralleled mosfets or a single large IGBT.
The problem with multi paralleled mosfets is that the current doesn't quite share very well, so you would have to considerably derate.

Using IGBT's one is limited by the slow switching frequencies and hence the size of the magnetics and filters for EMI increase considerably. Generally efficiency goes down too, which demands larger heatsinks.

For mosfets the idea is to use one or maybe two with current measurement. The controller IC turns off once the set current is reached. Since they can operate at high frequency the magnetics are small.

This is the reason for smaller modules. Its basically limited by the power of the individual mosfets. There are no 600 or 800V 100A mosfets commonly available on the market and even the rare exceptions perform worse than the smaller counterparts.

Three phase is a different story. I haven't come across many chargers that implement true three phase. A very rare number takes line to line and use two phases, but generally only a single phase at 32 or 63A is used.
 

treez

Advanced Member level 5
Joined
Sep 22, 2008
Messages
7,519
Helped
547
Reputation
1,097
Reaction score
529
Trophy points
1,393
Location
cambridge
Activity points
75,593
Re: Offline Switch Mode EV chargers should be made of paralleled &lt;3kW modules?

Basically what I am saying is, that there is utterly no point in making stuff like EV chargers using expensive methods…we should just use filthy cheap ways, (eg multiple dirt cheap, paralleled, low power modules) because for all you know, the equipment might be located in a region suffering horrendous mains transients…and will get blown in no time, and need replacing.

For example, I used to work in an electric drive company that used to make 20kw electric drives for newspaper production lines, which use many such drives along the length of the production line. Mains transients would regularly blow a drive up, so it was a total waste of time to make an expensive electric drive. In fact, the drives used to blow so regularly that the customer did not actually buy a quantity of drives, but instead, they bought say a 3 year contract during which time our company would keep them supplied with electric drives…replacing all the failed units as and when they failed.

This cements what I say, its simply a waste of time to make expensive flashy power supplies for connection to the mains, because they may just get blown by mains transients..especially EV car chargers which may be outdoors or at least near to the service entrance.
Even the transient protection modules made by Littelfuse eventually succumb to death by mains transient..

Littelfuse transient protectors.
http://www.littelfuse.com/~/media/e...ghting_surge_protection_modules_flyer.pdf.pdf

- - - Updated - - -

That just about says it all, even the Mains Transient Protection modules can’t last out against mains transients!!…eventually succumbing to them..as the above datasheet confesses.

This is the reason for smaller modules.
…I agree, indeed, using say a single high power module is not the way forward, because all that happens is that say if using an LLC converter, you would need to use paralleled FETs….but that doesn’t really work too well with an LLC, because then you end up needing even more magnetising current to flow in order to discharge/charge the FET Cds capacitances before the dead time elapses.
 

cts_casemod

Member level 4
Joined
Dec 9, 2010
Messages
72
Helped
14
Reputation
28
Reaction score
14
Trophy points
1,288
Location
London
Activity points
1,950
…I agree, indeed, using say a single high power module is not the way forward, because all that happens is that say if using an LLC converter, you would need to use paralleled FETs….but that doesn’t really work too well with an LLC, because then you end up needing even more magnetising current to flow in order to discharge/charge the FET Cds capacitances before the dead time elapses.
Spot on, how could I forget the device capacitance :)
 
Last edited:

Warpspeed

Advanced Member level 5
Joined
May 23, 2015
Messages
2,213
Helped
753
Reputation
1,508
Reaction score
731
Trophy points
113
Location
Melbourne, Australia
Activity points
17,766
Have to agree, expensive high tech solutions that blow up or have a very limited life with planned obsolescence in mind are definitely not in the best interests of the consumer.

Its not too bad if a ten dollar LED light or 12v dc wall pack needs replacing after a power surge, but its a different thing if a thousand dollar alternative energy battery charger goes bang, and is destroyed to an extent way beyond any economic repair.

I have been trying to build a reliable multi kilowatt three phase off line switching power supply for quite some time, (for personal home use) but it kept randomly blowing up for reasons I could never fully understand.

I am now using three honking great toroidal transformers with a six diode bridge, and large common mode inductor. It has survived several major power surges and seems to be immortal.

It must weigh 60Kg and is on wheels, but it should last forever.
 

FlapJack

Advanced Member level 1
Joined
Sep 16, 2010
Messages
400
Helped
101
Reputation
202
Reaction score
99
Trophy points
1,308
Location
USA
Activity points
3,822
I have been trying to build a reliable multi kilowatt three phase off line switching power supply for quite some time, (for personal home use) but it kept randomly blowing up for reasons I could never fully understand.

I am now using three honking great toroidal transformers with a six diode bridge, and large common mode inductor. It has survived several major power surges and seems to be immortal.

It must weigh 60Kg and is on wheels, but it should last forever.

To design for line spikes you have to know what to design for. Have you been monitoring the line and recording the transits.
 

Warpspeed

Advanced Member level 5
Joined
May 23, 2015
Messages
2,213
Helped
753
Reputation
1,508
Reaction score
731
Trophy points
113
Location
Melbourne, Australia
Activity points
17,766
This is a part time retirement home project, so a mains transient data logger is out of the question.
But I have used a Dranetz transient logger in the past, and the numbers are really scary.

Its one thing to knock off the microsecond duration several Kv spikes with a filter and some Transorbs. Lightning strike can be a bit more problematic.

But what if a truck flattens a power pole and the 6.6Kv line on top falls across the 240v lines directly underneath ?

You can only build in so much protection, then something a lot bigger comes along.... and its all over.
 

cts_casemod

Member level 4
Joined
Dec 9, 2010
Messages
72
Helped
14
Reputation
28
Reaction score
14
Trophy points
1,288
Location
London
Activity points
1,950
This is a part time retirement home project, so a mains transient data logger is out of the question.
But I have used a Dranetz transient logger in the past, and the numbers are really scary.

Its one thing to knock off the microsecond duration several Kv spikes with a filter and some Transorbs. Lightning strike can be a bit more problematic.

But what if a truck flattens a power pole and the 6.6Kv line on top falls across the 240v lines directly underneath ?

You can only build in so much protection, then something a lot bigger comes along.... and its all over.
Thinking that way what sort of electronic appliances have you got at home?

I much prefer a buck or boost topology to a isolated of some kind. If you need isolation, add a transformer (or three for three phase) before the converter. The transformer is the first step to clean transients.

The stresses seen by the switches are much smaller with a buck/boost converter than with a flyback/tapped inductor or even full bridge. By stresses I mean voltage transients due to inductive loads, since the designer has complete control over the current.

I have a number of non isolated buck converters operating from the mains and suffering constant abuse and none has failed yet. These range from 5W to 1KW and are used for battery charging or high power led driving. One was a custom built power supply to operate on my EV, since I was having a hard time finding a small 500VDC input power supply.

Boosts are more complicated as current can only be controlled above the input voltage, but if you use a diode to bypass the switching device chances are only that diode will be destructed in case of a short, before the input fuse is blown (always use a fast type). Very often I found that the diode survives as it is much tougher than the mosfet.
 

Warpspeed

Advanced Member level 5
Joined
May 23, 2015
Messages
2,213
Helped
753
Reputation
1,508
Reaction score
731
Trophy points
113
Location
Melbourne, Australia
Activity points
17,766
Thinking that way what sort of electronic appliances have you got at home?
Most typical "white goods" washing machines, driers, refrigerators, microvave ovens, stoves, etc.... are pretty immune, at least the motors and heating elements are.
The control electronics are usually very low power typically fed from a big resistor or capacitor, with a zener, or something equally crude which is practically bullet proof.

The more high end electronic stuff, computers, Hi fi, TV, and so on are usually pretty well protected and seem to survive quite well.

What I am having constantly fail here, are cheap and nasty switch mode wall packs and LED lights, where frequent voltage transients often blow out the reservoir electrolytic capacitor. As these capacitors are typically rated for 450v I wonder how high and wide these over voltage transients are.

My own attempts at getting +/- 210v dc non isolated supplies (dc with respect to mains neutral) were with a six diode bridge driving dual 15 amp buck regulators.

The raw rectified three phase waveforms never dip below about 260v, so a buck regulator down to a regulated 210v is entirely practical. There was no bulk dc bus storage capacitor used, to eliminate the inrush problem.
Just a very effective two stage low pass filter to prevent PWM ripple from going back into the mains. That was followed by heavy transorb over voltage clipping to keep over voltage spikes out of the 1Kv rated IGBTs.

In theory it should have worked fine. The combination of LPF and clipping should have been more than sufficient to stop mains spikes.
And it was effective for very long periods, but for no obvious reason it would fail randomly to the point of becoming a real annoyance.
I totally rebuilt it several times with different components and different features, but eventually gave it up.

When other parts of the whole system are perfected and completed I may revisit this problem as its become a real challenge..
 

cts_casemod

Member level 4
Joined
Dec 9, 2010
Messages
72
Helped
14
Reputation
28
Reaction score
14
Trophy points
1,288
Location
London
Activity points
1,950
Cheap LED bulbs are probably the worse thing one can get, same for low quality power supplies/chargers, etc. They dont have common mode choques so not only they burn out easily if hit by a spike they can also cause noise to other appliances.

I find it very unlikely that a peak would be able to travel trough the EMI filter, the input MOV, and charge the input capacitors high enough to cause damage. Remember a peak is absorbed by anything with a rectifier/capacitor, so TV's, computers, phone chargers, all absorb it, not just the converter. If they survived, something is wrong.

More likely your issue is coming from the converter itself.

If you're playing with high power IGBT's two things are essential. VCE Desat and active clamping

VCE desat turns off the IGBT in case of a large transient that takes the IGBT out of saturation. This is done by steps, not simply turning the device off.
Active clamping makes the IGBT conduct if the DC-LINK voltage is too high. In essence it commands the IGBT as a big linear regulator to absorb DC-LINK spikes.

I would break an arm if you still blow IGBT's after installing these. A good quality driver is expensive, but its worth every penny. Concept sells some good ones.http://www.ebay.com/itm/Original-factory-package-Concept-IGBT-Driver-2SC0108T2A0-17-/252181575869?hash=item3ab73178bd:g:IZEAAOSwAKxWUmiZ

Some care is also required with DC-LINK input capacitance. The IGBT switching on and off will generate spikes on the input. they need a low impedance to have where to go otherwise the high voltage shoots trough the IGBT insulation shorting it. The right amount of capacitance is needed, not just to protect the mains but to protect the IGBT itself. Generally film capacitors and snubbers across the IGBT. A large value is not needed. Electrolytic's WONT work.
 

Warpspeed

Advanced Member level 5
Joined
May 23, 2015
Messages
2,213
Helped
753
Reputation
1,508
Reaction score
731
Trophy points
113
Location
Melbourne, Australia
Activity points
17,766
The thing about a buck converter is there cannot be a sudden current spike that can pull the IGBTs out of saturation. There is always sufficient series inductance to slow the rate of current rise where the normal PWM current limit can handle it.

And I agree, an input filter with a cut off frequency around 1Khz should remove all the narrow high voltage spikes, leaving little for any hard clamping left to clamp.

The IGBTs were massively oversized as well.
There was nothing in the way of ringing or inherent problems with the switching waveform either.

I have no idea why it kept blowing up. It would go fine for months at high power then blow up twice within a few days under virtually no load.

I am running dc/dc kilowatt rated switching power supplies elsewhere in the same system without any problems. But anything connected to the mains seems to be cursed !
 

cts_casemod

Member level 4
Joined
Dec 9, 2010
Messages
72
Helped
14
Reputation
28
Reaction score
14
Trophy points
1,288
Location
London
Activity points
1,950
The thing about a buck converter is there cannot be a sudden current spike that can pull the IGBTs out of saturation. There is always sufficient series inductance to slow the rate of current rise where the normal PWM current limit can handle it.
Unless the device is latching, which one doesn't have great control over.

Have you contemplated using active rectifiers rather than a buck converter to get your desired output?
 

Warpspeed

Advanced Member level 5
Joined
May 23, 2015
Messages
2,213
Helped
753
Reputation
1,508
Reaction score
731
Trophy points
113
Location
Melbourne, Australia
Activity points
17,766
This buck converter used hysteric control for both output voltage, and an output current limit override.
The dynamic performance for step load change was spectacularly good.
The only problem with it was it occasionally had suicidal tendencies for reasons unknown.
 

Status
Not open for further replies.
Toggle Sidebar

Part and Inventory Search


Welcome to EDABoard.com

Sponsor

Sponsor

Design Fast


×
Top