Trying to find out what topology this SMPS is?

[treez]

Hello,

We have had a 3kw offline (85-265VAC) battery charger made for us by a consultancy. They did not give us the schematic. The chargers are failing and we are wondering why. First of all we are trying to assess what topology it is.

The schematic that we have determined so far, by using the DMM on the PCB is as in the pdf. However, when we simulate this in LTspice , it doesn’t give as much output voltage as we see in the actual circuit on the bench. We are wondering if we have got the position of L7 correct?

We are running it into a 28 Ohm load and have set it to 250V output voltage. It is delivering obviously 8.9A into the load.

Do you know if there is some version of an LLC resonant converter that can give high gain so as to give the output voltage that we are seeing? By designing our own LLC resonant converter we find it is not possible to achieve 250V across a 28 Ohm load from a 415VDC input voltage. –But that is the output voltage that we are seeing on the bench with the circuit that they designed for us.

As such, do you know if there is a “type” of LLC resonant converter that can achieve what we are seeing here on the bench? I am not too sure if I have the power stage schematic exactly right.
The switching frequency is definitely 100KHz, and the resonant frequency of L1 and C4//C1 is 110KHz, so it seems as if that is the resonant inductor and capacitance respectively.

(pdf schematic and LTspice simulation of it attached.)

 

[treez]

Sorry here is the actual circuit now. ( I was given a wrong version before). It does the job, but do you know what are the voltage spikes across L1? (they don't seem to cause any dissipation so I guess they are ok?) Also, why do they use L7 across the primary? Is it to make it capable of doing the entire output voltage range? Also, why have such large resonant caps and such a small resonant inductor? The external resonant inductor (L1) is on the order of typical leakage inductance, and leakage can be loosely toleranced, so why have the resonant inductor so small in value?, such as its value can vary over a big percentage and mess up the operation.

The actual spec is
vin = 415vdc
vout = 150-300vdc
pout = 3000w

 

[FvM]

The topology is LLC converter. It's variable voltage conversion factor relies on varying the switching frequency (or tuning the resonant circuit, which is obviously not possible with this circuit). It's effectively impossible that circuit works with 100 kHz fixed switching frequency.

 

[Velkarn]

L7 looks like trying to adjust magnetizing inductance value

 

[Easy peasy]

It could well be a series resonant converter with phase shift control.

 

[Easy peasy]

Sorry it is an LLC, the small L and large C are a direct result of tuning for 3kW, and the fact that the applied volts are halved by the half bridge topology. The tuning may not be perfect, this can be ascertained by the freq range from no load to full load at a given input voltage, e.g. 300kHz no load to 180kHz full load is about OK for an LLC. Full bridge would be better at this power level...

 

[Easy peasy]

As to failing, this is a bit more common for the LLC than people let on, for dynamic response to load changes and overloads if the freq goes too low, hard switching results and most mosfets are not up to it (applied dv/dt when supposedly off) and go bang due to internal diode recovery causing the internal xtor structure to come on and effectively latch. Some makers are more immune than others, IXYS being of notable quality in this regard. There are hundreds of LLC designs out there that have failed due to this effect...

 

[treez]

it is an LLC, the small L and large C are a direct result of tuning for 3kW

Would you say that the heavy load here (3kw) makes the LLC less favourable?...since at heavy load, the Q factor of the LLC stage gets higher, and as we know about LLC converters, when the Q factor gets higher, the gain (vout/vin) tends to get lower. -Then in order to raise the gain up again one has to either reduce the resonant inductor or increase the resonant capacitor, or reduce the magnetising inductor, none of these is particularly welcome.

At some heavy load, it appears almost impossible to get the switching frequency at which you operate, far enough away from the frequency below which capacitive operation starts, which as you know, is disastrous. So do you believe that for a LLC converter, there is realistically an upper load range?....eg 7kw?

After all, to get decent behaviour at 7kw, one ends up with a ridiculously low magnetising inductance, and loads of circulating current?

 

[Easy peasy]

The LLC tuned for its load is efficient enough from 40% to 100% load, raising the Q raises the circulating current, and the gain of the power stage actually goes up, meaning you have to reduce the freq less for a change in power.
There is no upper power limit for LLC, they have been made at 1MW.
The trick is in the tuning of the power stage for Vin range, Vout range, power and desired freq of operation.

 

[Easy peasy]

here is a link to a paper on a 1MW bi-directional LLC converter:
**broken link removed**

 

[treez]

thanks, SunnySkyGuy linked that one, it doesn't step up or down in both directions.

 

[FvM]

thanks, SunnySkyGuy linked that one, it doesn't step up or down in both directions.

Yes, the topology is fixed frequency, fixed voltage ratio series resonant converter.

 

[Easy peasy]

It processes power both ways, it is an LLC with a high-ish parallel L, or series resonant converter, it demonstrates LLC can be used to any power.
On a more useful power level I have seen it done at 3kW from 24VDC to 380VDC, bi-directionally, as an LLC with reduced Lmag in the Tx (a multi gap toroid).

 

[FvM]

It processes power both ways, it is an LLC with a high-ish parallel L, or series resonant converter, it demonstrates LLC can be used to any power.

Maybe you have a different design in mind. As far as you are referring to the linked "1 Megawatt 20 kHz Converter", it's clearly designated "Series Resonant Converter with Constant Frequency Modulation" in the paper. And there's no parallel inductor or intentional reduction of transformer inductance involved. Conventional fixed voltage series resonant converters with bidirectional energy flow are pretty standard in high power applications. They can be considered as "DC transformers".

It's a different question under which circumstances the design could be supplemented to a LLC and operated with variable frequency or variable inductance to achieve voltage variation.