diyAudio

diyAudio (http://www.diyaudio.com/forums/)
-   Power Supplies (http://www.diyaudio.com/forums/power-supplies/)
-   -   DIY SMPS - PFC/DC-DC (http://www.diyaudio.com/forums/power-supplies/177222-diy-smps-pfc-dc-dc.html)

crazifunguy 14th November 2010 01:13 AM

DIY SMPS - PFC/DC-DC
 
Hello SMPS Community

I am working on a SMPS that I would like to use for some amps I want to build. I have some schematics for a CCM PFC and a DC-DC converter. The design could meet class D and 80+ specs according simulation tests. I do however have some questions. I am well rounded in electronics but I need some design help. Here are my specs so far.

PFC Input Voltage
85V - 265V AC 50/60HZ

PFC Output Voltage
400V DC
Current Unknown

DC-DC Converter Output Voltage
+70V 10A -70V 10A

Total Power ~1400W


I am stuck calculating the PFC Output current going to the DC-DC converter. I am thinking I would need 4A that would be 1600W. But such a design requires HUGE inductors for the input choke. Can anyone help me with the required current output from the PFC to the DC-DC converter. I have been working with poweresim to assist in the design.

Topology
CCM PFC with Half Bridge DC-DC

luka 15th November 2010 05:05 AM

I don't think you need THAT bit of a inductor, probably the same size as for the half bridge. And yes you would need only 4A, peak current in PFC would be higher..

What calculations have you already done?

dtproff 16th November 2010 06:34 AM

1 Attachment(s)
For 1400W you will probably need a bi-phase CCM PFC. That puts about 700W/phase. Here are the calcs I did for a 500W (1KW bi-phase) using a PQ3535 core.

Anyway.. adapted as needed.

Tony

dtproff 16th November 2010 06:35 AM

I should have mentioned that I modified it for 700W.

crazifunguy 16th November 2010 09:10 AM

1 Attachment(s)
Quote:

Originally Posted by dtproff (Post 2366734)
For 1400W you will probably need a bi-phase CCM PFC. That puts about 700W/phase. Here are the calcs I did for a 500W (1KW bi-phase) using a PQ3535 core.

Anyway.. adapted as needed.

Tony


Impressive.....Is that a program that you have saved or did you manually re-calculate all of that?

Just so everyone has an idea of what I am playing with.
PowerEsim - Free SMPS Switching Power Supply / Transformer Design Software

My project file is attached below. Just open it up with no password.

Take a close look at how rediculous the inductor specs are. 8awg wire????
This type of circuit design is new to me so the learning curve is massive. It will be fun and challenging for me.

LETS MAKE SOME POWER!!!!!

Workhorse 16th November 2010 09:18 AM

Quote:

Originally Posted by crazifunguy (Post 2364371)
Hello SMPS Community

I am working on a SMPS that I would like to use for some amps I want to build. I have some schematics for a CCM PFC and a DC-DC converter. The design could meet class D and 80+ specs according simulation tests. I do however have some questions. I am well rounded in electronics but I need some design help. Here are my specs so far.

PFC Input Voltage
85V - 265V AC 50/60HZ

PFC Output Voltage
400V DC
Current Unknown

DC-DC Converter Output Voltage
+70V 10A -70V 10A

Total Power ~1400W


I am stuck calculating the PFC Output current going to the DC-DC converter. I am thinking I would need 4A that would be 1600W. But such a design requires HUGE inductors for the input choke. Can anyone help me with the required current output from the PFC to the DC-DC converter. I have been working with poweresim to assist in the design.

Topology
CCM PFC with Half Bridge DC-DC


If you need 1600W from your PFC then the Peak Inductor current which will be drawn from mains AC should be calculated at worst case which is:
1600/[85X1.414]=13A approx.

You need an inductor which should be capable of handling 12-15A of current easily.

The total PFC output current will be then 4A at 400V output

dtproff 18th November 2010 07:09 AM

This is my standard MathCAD file for CCM operation.

For 85Vac and 400Vout you need about a 70% duty cycle. On top of that you will have a ripple current of somewhere between 25-50% of Idc (I used 40% and in my calcs 1400W). So that will be Idc+Delta I/2 where Delta I is the ripple current.

(1600W/(85*1.414))/0/7)+0.4*(1600W/(85*1.414))/0/7)/2=22.8A peak. This is the amount of peak current you have to use for the flux density calculation of Bmax.

The equations from workhorse are correct for the wire diameter as far as estimating what wire gauge would be appropriate.

After that you have to gap the core if you are using Ferrite.

I split your design up into 2 cores running 180 out of phase. You can use a single core running 1 phase but you will probably need a larger core.

As to the wire thickness, For CCM operation you don't have as much of a problem with the harmonics (now as much need for Litz wire).

One more item... AC Inlets (IEC320) are rated for 15A so be careful how much further you push this. Most Server sypplies are derated for low line operation above 1500W.

Anyway. Good luck.

crazifunguy 18th November 2010 06:35 PM

Quote:

Originally Posted by dtproff (Post 2368953)
This is my standard MathCAD file for CCM operation.

For 85Vac and 400Vout you need about a 70% duty cycle. On top of that you will have a ripple current of somewhere between 25-50% of Idc (I used 40% and in my calcs 1400W). So that will be Idc+Delta I/2 where Delta I is the ripple current.

(1600W/(85*1.414))/0/7)+0.4*(1600W/(85*1.414))/0/7)/2=22.8A peak. This is the amount of peak current you have to use for the flux density calculation of Bmax.

The equations from workhorse are correct for the wire diameter as far as estimating what wire gauge would be appropriate.

After that you have to gap the core if you are using Ferrite.

I split your design up into 2 cores running 180 out of phase. You can use a single core running 1 phase but you will probably need a larger core.

As to the wire thickness, For CCM operation you don't have as much of a problem with the harmonics (now as much need for Litz wire).

One more item... AC Inlets (IEC320) are rated for 15A so be careful how much further you push this. Most Server sypplies are derated for low line operation above 1500W.

Anyway. Good luck.

Keep in mind I choose a very wide input voltage range. Typically It will be used in the 108-132Vac range. I choose 85-265 to make it universal for other countries. As far as the current draw....This supply will never be used at the max output I would say at the most it will see a 65% load.

Other Notes:
I find it very hard to locate cores and bobbins for specific sizes. Also if the recomended wire size for example is 8AWG using 4 16AWG in parallel should meet the specs failry close.

Is there any advantage of using a transformer style choke vs a toridal?

TechGuy 19th November 2010 01:13 AM

" find it very hard to locate cores and bobbins for specific sizes. Also if the recomended wire size for example is 8AWG using 4 16AWG in parallel should meet the specs failry close"

Switching freq. is more important than wire gauge. Don't forget about the skin effect.
I suspect the your PFC switch freq. is arround 100 Khz. So a 16 AWG wire will not be fully utilized. That said you can probably utilitize a thinner gauge wire to hande 1600Watts. Remember your not transmitting current over long distances. the wire used in your inductor will be pretty short, especially if your switching frequency is very high. You dont need much inductance if your switching frequency is high.

I would search for a skin effect calculator on the net the provides wire resistance give wire gauge and freq. Then calc power loss and see how small gauge you can get away with.

One recommendation is to use flat magnetic wire since it has more circular mils and will have a lower AC resistance for the given material used. Another method would to use copper foil instead of wire, but you would need to use a e-core, RM8, or PQ type core since it would be very difficuilt to wind foil on a toroid.

Wire Gauge table with skin effect column:
American Wire Gauge table and AWG Electrical Current Load Limits with skin depth frequencies

Vendor that sells flat magnetic wire (online purchases):
Flat Magnet / Specialty Wire

I suspect that a single wire of 15 AWG flat wire will be sufficient to handle 1600W (but I haven't done the math to back it up).

"Is there any advantage of using a transformer style choke vs a toridal?"

Toroids will have better thermal disappation than other types, but will leak more EMI than other designs.

By the way, Winding a transformer with multiple parallel wires becomes difficult. At most I can do is about 4 parallel wires of 19 AWG at a time without running into problems.

crazifunguy 19th November 2010 03:33 PM

1 Attachment(s)
Attached is the schematic
Freq 85 KHz
L1 - 1.76 mH (EE55 7 turns)


All times are GMT. The time now is 11:09 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright 1999-2014 diyAudio


Content Relevant URLs by vBSEO 3.3.2