# When building a crude full wave rectifier how to improve it?

tim9000
Say you were building a full-wave rectifier. How many simple ways (if any) are there to reduce the THD on the source caused by using a full-bridge rectifier load?

Also, If you used a low pass filter and then an isolation TX (or the other way around?), how much would that help to this end (if at all)?

I'm sure you would get odd harmonics and that the TX would act as a LP filter in it's own right, but what else do you need to consider, and how can you combat the non-linearity of this load?

Thanks!

There are many " steps" in how to do this. AC line filter, then DC Choke, then many versions of active filter, or active rectifier, PFC controller.

VERY easy to build in LTSice and "play" with the passives.

berkeman
tim9000
Thanks heaps for the document, it is a good place to start. I did have old notes on this stuff and other power electronic topologies and theory, but it will be some time before they turn up again.

When I said simple I suppose I was meaning 'non-active' methods. But I am still interested in anything do-able.

I'll hopefully get a chance to take a better look at the document over the weekend. However, I did take a quick look and I can see that there is the least reactive current is situation C). Which is interesting, presumably it would have lower THD too. [I expect that the sizes of choke were selected to be approximately the same over AC and DC application (maybe taking extra size necessary to compensate for the tendency for DC to saturate)]. From memory the DC inductors keep the diodes forward biased for longer once they try to snap-off. But I wonder if there would be a diminishing return by using both AC line reactors on the supply side AND the LC filter on the DC load side?...

Actually - on a recent drive project we saw the best performance with the mixed approach - but also due to cost and losses, as the DC choke is bigger = $, but we have 1% DC bus voltage improvement with AC and DC inductors. 1% is at 400KW... so it adds up! Science Advisor Do you have one phase or three phase AC? How much power are you considering? Are you trying to get a PF closer to 1 on the supply side? If the load on the bridge rectifier is resistive, then there is no PF problem. What is the load? Windadct Again -- great exercise for LTSpice or PSIM demo. tim9000 Do you have one phase or three phase AC? How much power are you considering? Are you trying to get a PF closer to 1 on the supply side? If the load on the bridge rectifier is resistive, then there is no PF problem. What is the load? I actually was thinking about PF on supply side. Although, I suppose there may be a need for paying attention to the load side too. I always assumed that the power to the load was of PF = 1. If you had a DC load that was only resistive, could the PF of the load not be 1? (With or without a DC capacitor) I was just thinking about a resistive load, but I should consider loads in general, because I started this thread with just the notion of building the best DC rectifier I could, for fun. Actually - on a recent drive project we saw the best performance with the mixed approach - but also due to cost and losses, as the DC choke is bigger =$, but we have 1% DC bus voltage improvement with AC and DC inductors. 1% is at 400KW... so it adds up!

Very interesting, I understand that a DC choke would cost more than an AC one. But I am surprised that the best performance result was a mix, did you have any idea why? And what sort of component sizes are we talking about for this VSD project? Also, since you found it was still useful to use the AC choke, do you think it would also be useful to have a delta configured AC capacitors on the supply before the rectifier? And if so, would it make any difference if this was before or after the AC choke?

Yes, I will simulate on LTSpice before I begin constructing anything.

Thanks

The DC choke helps with the DC ripple as well, and reduces the load on the DC Caps, no one component is chosen based on one parameter.

We only do the Power Electronics - but the selection of the passives affect the sizing of the devices we provide. This project has a 880A DC rectifier, ( half controlled for soft start) and 850A AC inverter.

If you had a DC load that was only resistive, could the PF of the load not be 1? (With or without a DC capacitor)
A rectified resistive load without a capacitor will have an input PF very close to 1.

Since the storage capacitor is in parallel with the load, the power factor will be effected by the presence of the reactive capacitance, even if the final load is resistive.

If you have a diode bridge feeding a capacitor then, for low load currents, you will have a narrow pulse of input current just before the voltage peak of each half cycle. For heavy loads the current will flow mainly on the rising edge as the capacitor is recharged.

HowlerMonkey
Don't ask for more than the rectifier can deliver.

In guitar amps, it is common to push a tube rectifier circuit beyond it's rating in the quest for a certain effect on the amplifier.

They are actually looking for distortion as opposed to having none at all.

I know my old dynaco amps brag about being able to pass square waves without ringing meaning a very low amount of distortion.

AlexCaledin
Say you were building a full-wave rectifier. How many simple ways (if any) are there to reduce the THD on the source...?

- the simplest way is only one - AC motor + DC generator! (Sounds nice, right? )

It can work as the cooler, so it's the coolest way too.

And think of the superb surge protection!

Last edited:
HowlerMonkey
Old DC welders did the same thing.

tim9000
Did anyone have any response to this:
since you found it was still useful to use the AC choke, do you think it would also be useful to have a delta configured AC capacitors on the supply before the rectifier? And if so, would it make any difference if this was before or after the AC choke?

The DC choke helps with the DC ripple as well, and reduces the load on the DC Caps, no one component is chosen based on one parameter.
I understand a DC choke would help with the ripple (as long as it isn't saturated). But I didn't consider how it would reduce the load on the caps.

Did you have any more information on the list of considerations you used in arriving at the sizes of your passive components?

That is a fair amount of DC current, is this to drive a motor, or some sort of system?

When you say 'half controlled for soft start' do you mean that you just ramp up the voltage slowely?

Since the storage capacitor is in parallel with the load, the power factor will be effected by the presence of the reactive capacitance, even if the final load is resistive.

I suppose thinking about it, the voltage on the load comes purely from the current when the diodes are conducting. Excluding the DC capacitor --> So it doesn't matter if the load has leading or lagging VArs the voltage of the load is based on the diode bridge rectifier conduction current.

- the simplest way is only one - AC motor + DC generator! (Sounds nice, right? )
Yeah that would be the best, but I might have space, cost and power constraints. Also, I might not have three phase and I don't like the idea of single phase induction motors or the carbon brushes of DC motors.

Old DC welders did the same thing.
Did the same thing as what, sorry?

Thanks all

HowlerMonkey
Old DC welders produced DC by using an AC motor to turn a DC generator.

The current in the DC link capacitors is dependent on the DC ripple, reduce the ripple reduce the current in the caps. For this one they are using 6mF of FIlm caps, and we calculate the lifetime to be 85k Hrs based on full load heating due to the ripple.

This is an AC motor drive.

I do not know if they are using a full LC AC filter scheme, pretty sure just AC chokes. They are building an add-on active filter now, for new and retrofit applications.

Yes the Half-Controlled rectifier is to allow the ramp up of V on the DC link. There are a few ways to do this, with a relay contact and resistor, plus a main, full current rated, contactor for example, but that is mechanical items vs a "solid-state" solution.

tim9000
The current in the DC link capacitors is dependent on the DC ripple, reduce the ripple reduce the current in the caps. For this one they are using 6mF of FIlm caps, and we calculate the lifetime to be 85k Hrs based on full load heating due to the ripple.

This is an AC motor drive.

I do not know if they are using a full LC AC filter scheme, pretty sure just AC chokes. They are building an add-on active filter now, for new and retrofit applications.

Yes the Half-Controlled rectifier is to allow the ramp up of V on the DC link. There are a few ways to do this, with a relay contact and resistor, plus a main, full current rated, contactor for example, but that is mechanical items vs a "solid-state" solution.
Sorry for the late reply, I thought I already replied and was waiting.
Talking about adding AC capacitors, so when you make this addition to the AC chokes, you said LC filter, is an L and a C 'greater than the sum of their parts' so to speak? This is to say, if you tried just L, and just C, would the result of combining them into an LC filter be more attenuation than the sum of each attenuation?

So in summary, the non 'active' methods are really-really basic, pretty much just caps and inductors, that is a shame.
Also, to double check, so what if your supply had a really poor power factor, how would the rectifier diodes respond to that: the voltage accross the load would still be in phase with the current wouldn't it? As opposed to the supply voltage?

Cheers

Not sure what you mean by "supply had a really poor power factor", it is the load current (in this case the rectifier) that determines the Power Factor. The whole paragraph has lost me.

tim9000
Not sure what you mean by "supply had a really poor power factor", it is the load current (in this case the rectifier) that determines the Power Factor. The whole paragraph has lost me.
Apologies, you've got me worried now, maybe what I'm asking isn't possible. I'll rephrase: say you were using a synchronous generator which is under-excited, which I believe means that the current is leading the voltage, It could just as easily be over-excited for the example. Anyway, you're using this generator to supply your rectifier. So the the power factor of the supply isn't perfect (I'm taking this as axiomatic, please correct me if I'm mistaken). Now, assuming your DC load is purely real, will the DC load voltage and the current both be in-phase with the current of the synchronous generator?

The other point I was making was to lament that it seems to me based on this thread: if you want to remove the harmonics of the rectifier you're either stuck with the complexity of active solutions, or the brute simplicity of L and C passive components. There is no middle ground. (a statement for your confirmation)

The other-other point on this note which I was inquiring was: Is an LC filter more effective than 'the sum of it's parts'? I.e. if you took the attenuation of a choke (L) and separately considered the attenuation of using just a line capacitor (C). And superimposed those attenuations of the harmonics together, would the theoretical total attenuation of adding the attenuation of L-setup (without capacitor) and the attenuation of C-setup (without inductor) together, be the same as the attenuation seen from using them in conjunction together as an LC filter? Or does it happen that when you combine them the attenuation is even greater? (I wouldn't expect so, but I'm asking just in case)

I hope this clarifies my last post.

Thanks

Last edited:
Throwing a generator parameters into this discussion only makes it more complicated, rarely is the rectifier so closely coupled to a generator that the two issues are interacting- we look at the grid as an ideal or static in both cases.

For rectifiers, there are theoretical limits for each filter passive type. Worth pointing out that it is also very common to use a phase shifting source transformer in higher power to help in 3 phase. (Y : D/Y, and D : Y/D would provide 4 different phase shift angles)

I guess the point is there IS NO ONE BEST SOLUTION. The DC LC circuit also has critical impact on the quality of the DC. In real products we start with a specification that defines the objective, AC supply, THD, PF, DC ripple, lifetime, cost, size, complexity, etc...

IMO it is time to start simulating / building, compare the quality / impact of the different elements and configurations.

tim9000
IMO it is time to start simulating / building, compare the quality / impact of the different elements and configurations.
Yes, totally. I'll figure out how to supply the rectifier with a voltage and current out of phase and see what happens.