440V 3phase to 0-17V/1000A rectifier design help

Need help to design a 440V 3phase line to 0-17V DC variable rectifier, output Current: 1000A

I searched over internet, most of them are theoretical.

Which will be good using 6 pulses and 6SCR or using old magnetic reactor + Diodes?

If 6pulses is good, then how to generate 30 degree crossing of 3phase line as switching is done after 30 degree of each phase.

Which hardware are you talking about? Three phase transformer, three phase SCR bridge?

I can make transformer as per requirement. But how the control will be?

Output voltage can be varied according to the control angle, 0 to 90° for positive output voltage (1st quadrant/rectifier operation)


B6 circuit


45° control angle

FVM is correct - you can buy off the shelf SCR controllers for this, else you have to design it yourself - a lil bit tricky - or hire a power electronics engineer / consultant to design it for you ...

FvM said:

Output voltage can be varied according to the control angle, 0 to 90° for positive output voltage (1st quadrant/rectifier operation)

View attachment 143694
B6 circuit

View attachment 143695
45° control angle

Click to expand...

Is there any different way to sense the zero crossing and generating triggering pulse for three phase system? Or can be done what we do for single phase and 3 pulses for 3phase seperately?

you need to create a virtual neutral and three phase signals in phase with the sec side outputs...

Hi,

The total controllable phase angle is 120°.
But I'divide it into several sections.

With an "ON" angle of 0..60° the active voltage goes down to 0.
Example: 45°:
* phase1pos ON: 0..45°
* all off: 45° ... 60°
* phase3neg ON: 60..105°
* all off: 105° ... 120°
* phase2pos ON: 120..165°
* all off: 165° ...180°
* phase1neg ON: 180..225°
* all off: 225° ... 240°
* phase3pos ON: 240..285°
* all off: 285° ... 300°
* phase2neg ON: 300..345°
* all off: 345° ... 360°

With active angle from 60° to 120° there is no OFF state. The next ohase takes over before the voltage goes down to 0.

Now it depends on load how this involves regulatikn.
While with purely resistive loads you have 0...120° meaningful control angle.
But if you have some RC, LC, or you are driving current into a DC biased load (like charging a battery) you may have limited meaningful phase angle.
If you try to charge a battery, then the lower phase angles (maybe 0..30° is meaningless, as long as output voltage is smaller than battery voltage)
If you directely charge a capacitor, then the upper phase angles (maybe 90° ... 120° is meaningless, because the capacitor will charge to peak voltage at 90° anyway)

Klaus

Here's another diagram showing the relation of phase-to-phase voltages and currents for the above shown B6 circuit with filter inductor.



Due to the filter inductor working, the current never drops to zero.

The 6-phase pattern is achieved by triggering SCRs in the upper and lower branches alternatingly. Obviously there are different ways to derive the trigger timing.

This Rectifier will be used in Electroplating purpose. Is this resistive load?

Hi,

Maybe there are some minor DC remaining effects, some minior stray inductance, some minor stray capacitance...but I assume it is almost pure resistive load.

Therefore I expect that the full 120° range is usefull.

At least in Europe such a circuit is not allowed to operate at mains grid, because it generates a lot of overtones.
A clean circuit may use an adjustable transformer followed by a rectifier, or a 3 phase PFC circuit.

Klaus

You are talking very serious current levels here. Many years ago I worked for a company which manufactured such high power circuits.
Before I answer the question:
How much design effort do you want to spend on it? How much money are willing to spend? How much time do you have available?

These are not a rhetorical or sarcastic questions. Absolutely not, neither I want to dissuade you from your project. All I want to do is to really think about the project.

From experience, the biggest pitfalls of all, is that you will have to custom make or adapt most of your bussbars, heat management and other hardware.
You will have to have the means of designing and winding large custom transformers.

You really need to think all of your protection circuitry: which includes semiconductor-grade fuses on each leg of the three phase rectifier, AC fuses on each phase input, DC-grade output fuses and thermo-magnetic breakers. And to prevent replacing too may of the very expensive output fuses (A Littelfuse 17T900 is US$129) a very desirable feature is "electronic fuses", which via output current sensing and monitoring circuitry, and perhaps over temperature sensing, to control a contactor.

If this rectifier is to be connected to a battery or to a motor which can produce Back-EMF, a reverse blocking diode or even a DC contactor is required.

In the end, after you have designed, costed out and built all these hardware items, one realizes that there is little money and time left to design and troubleshoot the electronics portion.

So how did we do it back then? With motorized three-phase Variacs which servo controlled the applied voltage to the power transformer.
Definitively a low-tech approach, but it worked and was extremely rugged. And we always delivered on time and on budget.

So if I use a 3phase Variac in the input side of my hand made transformer (note1) then rectify it using high ampere Diodes will be the best for this range of current?

note1: I've made a transformer of this ratings: Input: 440V,Y and Output: 17V,Y. Kept output current: 1200A based on a rough calculation.

I've not enough time to make R&D with this. Maybe 15days I can hold. I've studied some papers and application notes. Summarizing them, it is really costly and too much complex to design in this current range.

15 days is too little time to design, build and test this project of this power level.

Following a lo- tech approach will shorten the project significantly

17kVA output equates to 8 amps per phase input on a 440Vac line - so yes your 3 phase variac should handle this OK, for a 6 diode rectifier, each diode sees the 1kA for 33.3% of the time, so Iave = 334A, so 300A diodes would do, it is a matter of heatsinking, ~ 350 watts per diode x6 = 2.1kW in the heatsink. It is possible you could use 45V shottkies with lower losses.

The variac can then provide coarse control - which may well be good enough for electro-plating.

Good luck.

OK, I've made a rough design outline. Here it is. If you have any suggestion please let me know. I've started sourcing equipment.
First I'm making a trial for 250A. If it works fine then I'll start for 1000A.

a 6 diode rectifier would be better ...

That means using Delta instead of Y and using 6 diodes will be better?

Hi,

a 6 diode rectifier would be better ...

Click to expand...

"Better" in what manner?
* current waveform, overtones.. --> only a small amount better. The current waveform will be far away from being sinusoidal, but it will become more symmetrical.
(In either case the given circuit is not allowed to be operated on european (and other regions) mains grid)
* efficiency: --> I don't think so.
* cost, part count,... --> I don't think so

Klaus

Three versus six pulse rectifier has a large impact on transformer rms current and respectively transformer losses and minimal size.

Regarding harmonic currents, the design misses yet the expected output choke. I presume that you didn't make actual calculations for imagined filter capacitor or estimated current waveforms?