Purpose of SCRs in Power Factor Circuit

It was built by Zytec. Code name Tazmo. Does anyone have schematics for it?

Hello Greenia, Your close on what the SCRs do.

My take on the circuit is this. There is no filter capacitor on the full wave bridge rectifier but an active filter circuit using the two SCRs. The full wave bridge rectifier generates a series of positive going sine waves based on the AC frequency. D6 acts as block to keep any voltage feeding back to the full wave rectifier. The circuit fed by R8 is the firing circuit for the SCRs to fill in the gaps in the full wave bridge rectifier sine waves. The ic MC34262 is the controller to detect the zero crossing points in the sine wave which then supplies the delayed signal and power to the collector of Q3 and Q5 which then turns D2/D4 on/off. The SCRs act as a full wave rectifier filling in the gaps from the bridge rectifier. This makes the power supply appears resistive to the AC line, thus significantly reducing the harmonic current content .

Power factor controller TC from ON Semiconductor MC34262 PDF.

This may not be the exact way the circuit works, but based on your print and reading about the power factor controller IC and no filter caps to speak of it makes the most sense as to what the SCRs are doing.

I hope this helps Charles.

Hello MetalSmiths, thanks for the reply. Perhaps a little more of the schematic will help. I found the original patent for it. It looks like what the MC34262 is based on. Bucher, II James D; Zytec Corporation; United States 4683529 - Switching power supply with automatic power factor correction 1987.

Here is some more of the schematic, tell me if you still think the function of the SCRs is the same. The EMI filter and other components have been included. There is a slight change in the circuit with the cathodes of the SCRs going in to the inductor and feeding back through D10. The IC at the bottom is the MC34262.

Thanks, Greenja

Greenja, Thanks for the second drawing. The Filter caps on the ac input side of the full wave bridge rectifier are to help with noise and harmonics. What I was talking about is a large or several electrolytic capacitor on the dc output side of the full wave bridge rectifier and/or inductor to help filter the dc. This is typical in linear power supply. But with this unit with PF correction I think what I described is more or less how it works.

Charles

MetalSmiths, I think I've almost got it.

You weren't referring to the output Caps were you. There are 3 470uf 450V caps. I believe C3, L2 and R3 provide filtering of the DC output voltage(these are not the correct values as the components are missing). I understand what the RFI-EMI circuit is doing on the AC side. The SCRs look like they are turned on to maintain power output at the desired voltage level. I just can't seem to get the simulation to run using Multisim to get a visual picture of what's happening so I can have a clear understanding.

I am reading over the App Note AND8123/D PFC Using Critical Conduction Mode.

Here's the DC side of the circuit. After passing through D16 and then back through D14, I believe the DC output gets filtered.

Here's an image of the PFC board. The SCRs are in the lower right, in front of the bridge. The inductor and some caps are missing. It has another inductor at the top left with supplies isolated 12V and 5 volts. The PFC is in the middle, comparators for the 12 & 5 Vdc top center right.

MetalSmiths, I do believe you are right about this. There is however another part of the circuit that I have to go through. There is an inductor and 2 caps. They were removed from the board and I forgot to include them in the diagram. Back to the old drawing board. I will send an update with these components.

Thanks,

Greenja

Greenja,

Without seeing the whole picture I am thinking that the bridge provides for the power supplies and its output also provides for regulation by the amplitude being sensed by the triggering circuit to trigger the SCRs. The SCRs drive the transformer that provides high voltage to the 470uF 450V capacitor bank. The transformer circuit also feeds back to the trigger control circuit. Thus the amplitude of the High voltage is keep at a fixed level when the line amplitude and/or load varies.

Jon

Yes, that does seem to be what it's doing. I am just starting to use MC34262 PFC chip and from all the circuits I've seen and from the literature, the SCRs are not needed. The chip maintains the 400 VDC output by varying the frequency on the Mosfets.

This is a power supply from a server, so perhaps there is a special condition that requires the SCRs. I was working on the board 1 section at a time, but the SCRs left me a little puzzled and has taken up too much time. I'll move onto the other sections and post the complete schematic. Thanks for the input.

Greenja

Greenja,

The SCRs and associated trigger circuit drive the transformer that provides the High Voltage supply used by the Mosfets.

The SCRs trigger circuit is controlled by voltage feedback from the high voltage side of the transformer to regulate the Voltage on the Big filter Caps to keep them from launching off the board.

The bridge rectifier supplies the low Voltage system and synchronizes the SCR triggering.

On the DC side of the supply the MC34262 PFC chip works with the stored power to maintain the output Voltage.

According to the document for the MC34262 PFC chip it serves as a Regulating Pulse Width Modulator with specialized features for Pulse Repitition Rate variation for Power Factor Control.

Excellent for Flourescent lighting.

Jon

Kudukdweller9, You got it!

I was looking at the Bridge and it high voltage and amps rating 600V, 25A thinking that it was supplying all the power to the MOSFETs. That's how they have it set up in the design examples from the data sheet.

Thanks Jon for the help on that one.

Greenja,

Welcome.

Looks like an interesting project.

Data sheets do not account for the engineers who find interesting ways of incoporating things.

Your application is using an an SCR driven transformer inverter for the raw DC power.

The bridge would be most likely be on a heat sink if it was doing anything that averages near its rated capacity. Perhaps it is for handling inrush and startup surges as well as lower current operation.

It makes me wonder about the loads of the 12 and 5 Volt supplies.

Jon

Well, since you asked...

3.3 V .... 47A
5.0 V .... 65A
12.0 V .... 20A (24A pk @ 25% duty cycle)
-12 V .... 1 A
5.0 V....... 2.0A (Standby)

Total Power 605W...max power 650W for 1 minute.

Here is the fully populated and unpopulated boards.

Zytec actually patented the Heat sink design for the SCR's and Mosfets. As you can see there is a large heat sink on the Bridge. I want to replace the four IRFP460 with FDL100N50 Mosfets that can deliver 2500 Watts ($9.00 USD each). I'll have to upgrade the SCRs and increase the inductor size. It's an attempt at a motor VSD/Solar Inverter/High PF power supply. The Sun Microsystem Ultra450 has 3 power supplies in them. I was able to buy one for $100.00 a few years back hoping to get a chance to do something with the power supplies.

Greenja

Greenja,

Thanks much.

It all makes sense now.

Hard working server supply.

Jon

Hi,

Not sure if the schematic shown is complete. The source voltage seemingly is floating and the return path for any load current seems to be the GND only. The source voltage can therefore swing to both the posistive and the negative side of the GND. If this is true, there is need for D2 and D4 to cater to full cycle control. For this the cathodes should be tied together.

Hello All,

I'm just following this thread and we have used the FDL100N50 product before. I'm just wondering if the PCB proper has sufficient meat in the traces or if you are planning on some hard wiring. Don't think those PS boards were ever made with more than 2 oz. copper... .. just an FYI

Best on ya'

GLB

GLB, very good point. As for the previous post about the schematic, it is not complete.

The PFC board itself will not be used, a new board based on this design and some components will be made. I had the power supplies and it gave me an excellent opportunity to understand power factor correction with working boards.

As for the FDL100N50, the data sheet specs seem quite impressive. They seem to recommend it for PFC. What was the highest highest wattage you got up to and at what temperature? How realistic is the W/C de-rating data? Did your parallel them up? Have you used them in motor speed control?

All the info on them would be greatly appreciated. It would be nice to use only one type of MOSFET for both PFC and motor control.

Greenja

Power factor my b*t. This cirquit auto switches between voltage doubling rectification for ~110VAC and plain full bridge rectification for ~220VAC so you have about 320 VDC to the Mosfet both cases. Get real please guys!
--64 6E 69 4D 65 6C 70 6D 69 53--

Guest,

Bridge rectified circuit for 220V mains and SCRs operate for 110V mains. Trigger circuit senses lower voltage and brings the SCRs on line.

You seem to be familiar with this design and could provide some details.

As you can see the info provided was pretty weak.

What else do you have?

This circuit auto switches between voltage doubling rectification for ~110VAC and plain full bridge rectification for ~220VAC so you have about 320 VDC to the Mosfet both cases. Get real please guys!
--64 6E 69 4D 65 6C 70 6D 69 53--

--d n i m e l p m i s -- Simple mind. Clever!

Kuduk

Ghost, er, Guest,

Since this supply is only running off of ~120 Volts 50 Hz, what if one rectifier circuit was producing primary power by driving an LC circuit that, at the same time, provides phase shifted dc pulses to the SCRs and trigger circuit to provide a boost via another winding of the transformer to fill in during the dropout periods of the first rectifier? Having the two rectifier circuits work in concert can provide regulation of the power applied to the FETs.

I have seen this done using the rectifier filter inductor to produce current in adjacent diode coupled windings when the field collapses to send juice to the output supplies during the off state of the inverter switchers. It fills in the power gaps nicely and reduces the need for larger fiter components

We are very interested in this project so if there is some documentary evidence that you are willing to share it would be greatly appereciated.

Kuduk

Here another look at the circuit.

Complete PFC board schematic providing 400 VDC bus.

Power Mosfets, Output Caps and MC34262 PFC chip

Kuduk, I think you are right on the money with your answer. There is inductor L2 that would provide power when the Mosfets are switched off. The output is then fed to a TOPSwitch controller TOP202YA1 and transformer to provide and isolated supply for driving the mosfets on the second board. The second board provides the 12, 5, -12 and 3.3 volts from the DC bus.

BTW, who's we?

Output Caps with Mosfets.

Power for second board MOSFETS and drivers.

Greenja,

Looking good. You certainly have put in a lot of work on this.

Q8 takes care of the power factor function.

BTW, who's we?

We is me and those folks looking in and not posting.

Kuduk

Q4, Q6, Q7 & Q8 (IRFP460) take care of the power factor correction, they are in parallel providing the 220W x 4 = 880W at 25 deg C. The derating is 1.7 W/C which gives a some room from the 650W rating for 1 minute. The switching Diode is actually a STT12a06 (now obsolete) 600V 12 Amps with a 22W dissipation.

The Drv output from the MC34262 feeds Q11 which drives all the Mosfets on at the same time. Q12 drives them all off hard. C17 is connected to VCC (not shown).

Lots of work...yes, but it will make the math required to doing modifications on it a lot easier. Plus the supplies are going for cheap now. Check out ebay. They are going for $9.00!!!! I'm not in the US, so if you (or anyone else) buys them up, I'll take 2 from you at a later date!

Greenja,

Yep, Q4, Q6, Q7 & Q8. I was looking at the drawing with only Q8 showing.

Kuduk

Some theory now. Charging a capacitor with rectified AC does not give any significant AC current to AC voltage phase shift since capacitors are supposed to keep most of their voltage till the next cycle (that's the whole idea about using them and of course the bigger capacitor per Iout the better. Limiting factors only cap cost, Initial high charging current and fast enough discharge at shut off). So switching cirquits themselves have very good PF at AC no need to correct. The only possible significant phase shift is done by the EMI filter coils implemented at AC input to prevent switching frequencies back to grid, but still not worth the trouble to correct. SCRs can of cource be used sometimes for PF correction in combination with coils and caps exploiting the fact that a pulse turns them on and they automaticaly turn off at current zero (not voltage zero) but nowdays IGBTs and dedicated ICs or microcontrollers are used instead.

Ghost, er, Guest

Ghost, er, Guest,

Thanks for you excellent input.

The EMI filters dont have much effect on line frequencies. They are looking for stuff like noise from the grid and from the supply. They look like a straight wire to line frequencies. No phase issues there.

Each half cycle of ac from the mains presented to an inductive or capacitive input high power supply causes enought distortion of the waveform of the power grid and the internal workings of the supply to need a Power Factor Correction circuit compensates for it.

A room full of servers could cause some grief on the power grid if it were not for each server having a phase correcting power supply.

I mentioned before that I think the active part of the rectifier circuit (SCRs) is to regulate the high Voltage to a set point around 360 to 400 Volts? Using are brute force rectifier alone would put the system at the mercy of power grid fluctuations.

Even my cheap UPS compensates power grid fluctuations. (such fluctuations are rare where I live)

At higher line Voltage the SCRs would not have to work much. At lower line Voltage they would be working hard to hold up the load.

IGBTs and dedicated ICs or microcontrollers are used instead.

That PFC board must be kind of old. I can imagine there are a lot of them going to recyclers now.

Kuduk

Kuduk, Ghost, er, Guest,

Yes these boards are old, pre internet bubble. That's why they're so cheap! As for a dedicated IC, well that's what the MC34262 PFC controller is for.

Now on the theory side of it, the PFC is trying to make the current waveform as sinusoidal as possible and in phase with the voltage. This diagram is from On Semi AND8123/D.

What the other side of the power supply would have seen is a huge input spike from the inductor switching. The whole purpose of this type of PFC is to average out this spike, hence smoothing out the huge spike over the entire cycle. This figure was taken from Fairchild AN-42047.

Now I haven't come across any literature that shows an IGBT providing PFC. If there is, please post a link to it. On Semi, PIC and others have PFC algorithms for their DSPs with schematics and software to build complete boards. What I found was that this board does everything their boards and software would do. I haven't calculated all the losses inherent in this design, but the fact that it provides PFC and the inverter function saves a lot of time. Simply unclip PFC board and plug it into your (test) design.

So the first test was to building a VFD using space vector algorithms, similar to the On Semi 3PH AC board. It was only after going over the ON Semi design that I realized the 300-1359-xx was pretty much the same. I found a PS21254-E IGBT 3 phase module from an old washing machine and added that to the 300-1359-xx to make a VFD for a fractional HP 3PH AC induction motor. By 'borrowing' from the ON Semi design the brake, back EMF, current sensing, phase sensing, UV & OV protection circuits, their entire board was reproduced.

The next design using the same 300-1359 will provide 600 Watts (x3) of power from solar panels and battery bank. What I need to figure out is how to effectively turn 'off' the PFC function while still maintaining the boost function with as little modification as possible.

So, the original question of the purpose of the SCR was to ensure that I didn't inadvertently cook something on the boards and to figure out which components could simply be re-sized to meet a design change (taking the existing inductor & trace sizing into consideration).

In the Sun Ultra450 server that these power supplies were taken from, they are paralleled, which means there should be no limit to the number of supplies I can use for an inverter, just keep adding to the system. Sure there are other ways of doing the same thing, but none that I know of for under $10.00! Plus you get a very nice case that can be hot swapped!

Hi what this IC supposed to do is get the ABS(V0*sinωt) (unfiltered rectified AC) voltage and stepping it up by BOOST chopping it, charges the bulk capacitor in a controllable manner, so the load grid sees appriximates a more steady load comparing with normal charging current, that starts the moment the rectified voltage goes over capacitor's lost voltage since last cycle and stops shortly after Vpeak. And since this is another full DC/DC conversion before your "main" one you just double your costs and losses. Bigger capacitors and/or a simple DC reactor would be more effective if you asked me. The actual phase shift with big enough capacitors is insignificant and the "PFCorrection" is all about harmonics from short high charging currents to the caps, nothing else. It even is at the limits of "power factor" definition this aproach but since I've seen worse won't argue on that now.
But I do have a question. Assuming the main DC switching (after "PFC") is done at double the AC peak voltage (>310VDC) when working with 110VAC as you said so doubling rectification is required and that cannot be done without 2 separate oposite connected capacitors one charged at positive peak ant the other at negative peak from grid (single rectification each that without capacitors would give 0VDC please study the relevant rectification cirquit), does "PFC" happen after bulk capacitors? That would be enough for me to throw the board away and design from scratch
And a short answer about IGBTs Since a few years that their effective working frequencies went over audio band they can be used for most switching apps. ST for example sells IGBTs working over 130Khz for PFC. Google it. And about the SCRs on this cirquit can't see anything to do with "PFC" from drawings. It's common to use 2 SCRs or 1 TRIAC to autoswitch between different grid voltages and avoid an external switch so if using single voltage just discard and use "fixed" rectification. Cheers Ghost, er, Guest.

Ghost,er,Guest,

Congratulations you have provided me with the information I need to find out what the SCRs are really doing. You are correct, SCRs are used as 'Pre-regulators' and for voltage switching. They allow for switching with little losses at mains frequency and do not play a part in the PFC function. Here is the circuit from Linear Technologies LT1083 that shows the SCR configuration. Here is another circuit from another forum that gives a little more detail.

Power factor correction takes place prior to Bulk capacitors.

Thanks for the info on the IGBTs. I'll be sticking to MOSFETs for PFC though.

Didn't quite understand the some of the first part of your comment. I believe the whole point of the original design was to provide for active PFC so that you could use smaller components and reduce your cost, rather than just throwing big bulky components on your board.

As it stands, I now feel confident about the SCRs and their autoswitching function thanks to you.

Greenja