It is currently Sat May 30, 2020 10:13 pm



Post new topic Reply to topic
Author Message
ekstra   ara
 Post subject: Does output waveform change if load varies?
PostPosted: Fri Jan 19, 2018 2:32 am 

Joined: Fri Jan 19, 2018 2:19 am
Posts: 1
I know about inverters is as follows(correct me if i am wrong thank you) Dc source such as battery fed current into transformer by controlled switches such as mosfet or IGBT. This controlled current fed into primary side of step up transformer. While the control circuit modulate the high frequency carrier wave and referance sine wave. By modulating we get gate signals as follow.
Image

Using fullbridge we converte dc waveform into single sine wave pwm signal. And when fedding it to primary winding of inverter we use LC low pass filter and use filter also on output(i read on research paper but not sure).

Image

Ok my question is :

How to filter this pwm sine wave into sine wave? Does LC filter do the job?

This inverter circuit i explain here is seems much simple than commercial inverter. Do I missing anything?

Does output waveform change if load varies?

Correction and suggestions are welcome for question. Thank you.


Top
 Profile Send private message  
Reply with quote  
 Post subject: Re: Does output waveform change if load varies?
PostPosted: Fri Jan 19, 2018 6:54 am 
Plasma Level
User avatar

Joined: Tue Oct 26, 2010 9:08 am
Posts: 2174
Location: North Carolina
Way outside of the topics of the forum, but here goes...

With ANY power electronic converter, there are always two components in the system. There is the actual power electronics chain, and there is a control system which is essentially a specialized computer. The development history of power electronics is closely tied to development of computers. So keep in mind that especially on the smaller ones you will see all the electronics associated with the power supply and sensors with the controller as well as the power electronics. Further, you cannot drive the gate side of a power transistor with typical 3-5 VDC transistor outputs of a microprocessor so there is at a bare minimum a bigger power transistor and associated bias electronics to boost the power output, or more often a pulse transformer and capacitor and associated electronics to boost the 3-5 VDC signals up to perhaps 20-30 V or maybe even around 100 V to drive the gates. Eventually on higher voltage systems, particularly medium voltage where isolation is a necessity, this gives way to on board gate drivers operating from fiber optic signalling. Your drawings do not show any of those details.

With this in mind, typically the output would be either voltage controlled or current controlled. In other words as the load (never mind the power semiconductors themselves) fluctuate, the processor will adjust the pulse waveform to either maintain a constant voltage or a constant current, depending on the design. And this is often an inner control loop also fed by an outer much slower controller that gives a command signal for the desired voltage/current waveforms.

Also not shown is that some very cheap AC/AC systems have a simple diode rectifier on the front end but that's almost outdated technology these days. The theoretical best THD (harmonic distortion) you can get with these is 6% without filtering. By the time you put a big heavy LC filter on it you almost may as well just switch out the diodes for active devices (SCR's, MOSFET's, IGBT's) and run this from the same control system. This has the added advantage of regeneration if you need it.

Also almost nobody actually does PWM modulation anymore. That's VERY old school. These days most people are doing some form of lattice-style control where we look at the output on an AC system as a vector and modulate the output accordingly. PWM drives are much noisier and less efficient, and switch way too much. But sticking with PWM modulation the percentage of on time determines the output voltage (or current if we are doing current control). So as the load fluctuates the controller should be reading the output voltage/current and modulating the PWM modulation with the outer control loop I mentioned earlier. Usually we read voltage and/or current though because it's a heck of a lot easier and faster. With voltage inputs you have an inherently high impedance loop so by keeping capacitance low bandwidth for the sensor is in the kilohertz or even megahertz range. A simple shunt (controlled resistor usually <<1 ohm) converts current into voltage in DC and drives are operating at the waveform level so everything is effectively DC, so that gives us a very fast current sensor, too but at a consequence of generating some heat. CT's are so slow they are used sparingly such as for the outer loop inputs while the inner loop does voltage control.

Finally answering the LC/RC filter question, yep that's how it's done although inductors are so much larger and somewhat lossy that RC filters are pretty common if we can manage to pick values where the resistor stays small. Plus the property that inductors tend to resist changes to current works against a drive that is attempting to switch things on and off quickly. In addition, don't forget that if the load is highly inductive such as a motor or firing into a transformer, the drive can be designed with the assumption that there is already a certain amount of inductance in the load (or line) which helps to reduce or eliminate inductors, particularly in VFD's and electronic lighting ballasts.

Remember that a filter has a finite bandwidth and Q (roll off). As we make the filter specifications tighter, the size and number of components grows as do the filter losses. So most power electronics systems have a "carrier" frequency (with a PWM design) of a minimum of around 3 KHz and go up to around 10-20 KHz in some designs. The consequence of low carrier frequency is that it makes a lot of audible noise and causes unacceptable pulsations in the load. As the carrier frequency goes up this stuff disappears but new problems arise because it excites resonances in the system (cable and load reactance). When resonances occur (ringing) it easily boosts localized voltages to the point that it overcomes and destroys the cable and/or load insulation very quickly. Also power electronics has finite switching bandwidths and power electronics consumes very little power in an "on" or "off" state but consumes a great deal of power while switching. Increased carrier frequency increases thermal load as efficiency falls off so it has to be derated to compensate for the increased carrier frequency particularly above about 5 KHz.

Unless you are designing your own electronics this stuff has become so cheap that there's no point. You can buy premade modules called IPM's (integrated power modules) that contain everything in one piece with well built thermal designs for connecting to a heat sink along with tons of documentation. At larger sizes you get away from that and start going to more discretized but still modularized construction. That's for something special. Drives are so cheap these days that if you don't have to build one, don't do it. Parallax, Texas Instruments, and many others give away free drive control software that you just have to customize for your application. IPM's and bigger stuff is available from the usual sources such as Vishay or Semikron, and the processor vendors also give plenty of pointers. I think for smaller ones you can also get IP (Verilog code) from Lattice, And for servo systems, servo drives for years are/were written in "C". The servo drives are sold as basically a controller (computer) with a compiler and libraries of code for it, and an "amplifier" which is just the power electronics end of things that runs the tight inner control loops under control of the controller. Again you can't beat these guys on price and availability unless you are going into the drives market

As a repair guy the stuff has become so inexpensive and so highly integrated that realistically unless you get to very large drives (500 HP+) except for very small items and soft starts (because they are so simple), the cost of repairs usually exceeds the cost of outright replacement so all of this stuff has become disposable/"module level" replacement. I'm only aware of all of the little idiosyncrasies and design details because as a troubleshooter/repair person I need to know how the thing works to be able to fix it and to be able to set it up. Everything I described above is typically taught in a 2 week "dives" class up to some point. Usually the drives classes stop at the point of getting down to component level things. That's where a university-level "power electronics" class comes in where you get into the engineering and design aspects of drives.


Top
 Profile Send private message  
Reply with quote  
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 2 posts ] 

All times are UTC - 7 hours


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Jump to:  
cron
© 2019 Arcflash Forum / Brainfiller, Inc. | P.O. Box 12024 | Scottsdale, AZ 85267 USA | 800-874-8883