Voltage single phase active front end rectifier - simulation problem

[FreeRiderCz]

Hello, im trying to run this schematic :



with this regulation:



In OrCad 16.6 , i used PI regulator and PWM, but my simulation is making errors... and im unable to solve them. Can anyone pinpoint what am i doing wrong here? I ´m kinda self learning about these and i try to run some simulations to get some results, but i guess i did mistake i did not see...

Simulation in rar - orcad 16.6 files:

View attachment Active Front end rectifier single phase.rar

 

[FvM]

This is the same voltage DC bus topology as in your previous thread https://www.edaboard.com/showthread.php?382191-Active-front-end-rectifier-single-phase-current. Why are you naming it differently, why starting separate threads?

If the naming refers to the controller topology, it doesn't change the inverter type. Most active front end systems are using both current and voltage control loops.

 

[FreeRiderCz]

Samo topology but different type of regulation. I´m just looking for a help thats all.

 

[FvM]

I don't use Orcad16, unfortunately.

 

[FreeRiderCz]

Well i hope somebody will see this and gime some pointers

 

[BradtheRad]

* Is your schematic a bidirectional converter? Because it is not obvious what is the path of power flow.

* I think you should start with the simplest possible circuit, and see if the simulator has any problem. Try a single transistor/ mosfet. Then a half-bridge. Etc.

* The inductor may cause power factor error, and it also produces current lag. These effects make it difficult to apply PWM. By shutting current off abruptly it may create high-voltage spikes which can destroy devices.

* Your H-bridge is made solely from N devices. To turn them on-and-off requires bias volt levels which are sufficiently high or low at proper times during the cycle. I have seen a simulator have trouble when bias voltage is outside the supply rails.

 

[FreeRiderCz]

Thanks for a reply BradtheRad,

i was trying to simulate single wave and 2 wave rectifiers and they were working,

active front end rectifiers can be bidirectional but this scheme is focused only in one way to get higher voltage on the output than on the input

To your last 2 points, my english is not that good, so inductor behind power source is needed by the design in theory of this rectifier you advice me to get rid of it? Im not sure what exactly N devices mean in english.

Any suggestion how to get rid of that bias voltage?

thanks for reply

 

[FvM]

* The inductor may cause power factor error, and it also produces current lag. These effects make it difficult to apply PWM. By shutting current off abruptly it may create high-voltage spikes which can destroy devices.

It's the purpose of the AFE controller to correct the power factor and generate a sinusoidal mains current. Voltage spikes are clamped by design of the H bridge.

* Your H-bridge is made solely from N devices. To turn them on-and-off requires bias volt levels which are sufficiently high or low at proper times during the cycle. I have seen a simulator have trouble when bias voltage is outside the supply rails.

The schematic in post #1 is a principle circuit, abstracting from details like gate drivers. There a different abstraction levels in the simulation of power electronic circuits, we don't know which is used in FreeRiderCz's PSpice design. We should see pdf prints of the simulation schematic.

 

[FreeRiderCz]

It's the purpose of the AFE controller to correct the power factor and generate a sinusoidal mains current. Voltage spikes are clamped by design of the H bridge.



The schematic in post #1 is a principle circuit, abstracting from details like gate drivers. There a different abstraction levels in the simulation of power electronic circuits, we don't know which is used in FreeRiderCz's PSpice design. We should see pdf prints of the simulation schematic.

I didnt have simulation schematics - i just simply try to simulate this scheme in post #1 just as it is (dunno if thats the correct attitude ? ) but i didnt find anwhere on internet someone simulating AFE controlers in OrCad.

 

[FvM]

In OrCad 16.6 , i used PI regulator and PWM, but my simulation is making errors.

The control schematic involves a bit more than PI and PWM, there's also sine generation. Not clear how you implemented it. It's definitely possible with PSpice or Ltspice, using behavioral modelling elements.

 

[FreeRiderCz]

The control schematic involves a bit more than PI and PWM, there's also sine generation. Not clear how you implemented it. It's definitely possible with PSpice or Ltspice, using behavioral modelling elements.

Well i followed mathematical desrcibe when i make this regulation schematics, if by sin generation you mean this SIN block then i have it there:


i followed this mathematic explanation for epsilon regulation:

https://drive.google.com/file/d/1iZlgO7sVXG4VazvnvSmmwGzvQ2nlHobF/view?usp=sharing

 

[FvM]

Theta is a continuously increasing phase, not a constant. In other words, you have a sine generator and its phase is shifted according to the PI output. In a real AFE, theta has to be phase locked to the mains voltage.

 

[FreeRiderCz]

Theta is a continuously increasing phase, not a constant. In other words, you have a sine generator and its phase is shifted according to the PI output. In a real AFE, theta has to be phase locked to the mains voltage.

Oh i know what you mean, i just think that theta (phase shift of voltage and current on input) is still 30 or 60 angle or 90 im not sure now (thats why i used constat for theta), also i didnt know how to get continuous measuring of that phase shift (theta), i was solving how to get "probe" to get continuous measurement of Uc on condensator.

Simply i dont know how to get measurement of phase :-D, does anybody know?

 

[FvM]

i just think that theta (phase shift of voltage and current on input) is still 30 or 60 angle or 90

Theta is not a phase shift, it's the running phase of the mains voltage, e.g. theta = ω*t = 2*pi*50*t

 

[FreeRiderCz]

Theta is not a phase shift, it's the running phase of the mains voltage, e.g. theta = ω*t = 2*pi*50*t

Oh so its the (theta = ω*t = 2*pi*50*t) omega *t of time base of that input voltage, am i right?

 

[BradtheRad]

inductor behind power source is needed by the design in theory of this rectifier you advice me to get rid of it?

Im not sure what exactly N devices mean in english.

Any suggestion how to get rid of that bias voltage?

Your schematic has icons for NPN transistors although N-mosfets may be intended on account of neighboring diodes (body diodes).

The aim is not to get rid of bias voltage but to apply sufficiently high and low bias voltage (PWM signals), in order to turn the transistors completely on and completely off.

Your schematic has no ground icons. Some simulators need to have a ground icon somewhere in a circuit, otherwise you get an error.

While you are in the experimental stage it is easier if you omit components such as the inductor that create unpredictable effects. Perhaps it is needed for regulation (or smoothing, or choke effect, etc.). You can add it later when the simulation runs without errors.

 

[FreeRiderCz]

Your schematic has icons for NPN transistors although N-mosfets may be intended on account of neighboring diodes (body diodes).

The aim is not to get rid of bias voltage but to apply sufficiently high and low bias voltage (PWM signals), in order to turn the transistors completely on and completely off.

Your schematic has no ground icons. Some simulators need to have a ground icon somewhere in a circuit, otherwise you get an error.

While you are in the experimental stage it is easier if you omit components such as the inductor that create unpredictable effects. Perhaps it is needed for regulation (or smoothing, or choke effect, etc.). You can add it later when the simulation runs without errors.

I understand, i try to avoid inductor,

about that ground icon, i use one at input and i thought its enough, should i add ground also on output side and to regulatory part?

Could you suggest any N-mosfet instead parts i was using?

thank you wery much

edit: i actually got 2 grounds 1 on input side and one on output side...

 

[FvM]

i actually got 2 grounds 1 on input side and one on output side...

No, there can be only one ground node (node 0) in SPICE simulation circuits. Its choice isn't particularly critical, but the circuit must not be floating. I would choose the DC- node as ground.

In some cases, circuit parts that are separated by ideal diodes and switches need to be connected by a Megohm resistor to get faster convergence.

 

[FreeRiderCz]

Well only one node didnt help, still getting same error ....

 

[FvM]

Unfortunately we neither know which error you get nor which exact circuit you are simulating.