
Home  Forums  Rules  Articles  diyAudio Store  Gallery  Blogs  Register  Donations  FAQ  Calendar  Search  Today's Posts  Mark Forums Read  Search 

Please consider donating to help us continue to serve you.
Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving 

Thread Tools  Search this Thread 
27th May 2011, 09:45 AM  #21 
diyAudio Member
Join Date: Jul 2009
Location: West bengal/siliguri

guys i did the pain staking calculation ................ from the book luka referred me to .......
calculation: skin depth = 6.62/ root of frequency = .0.0209cm wire diameter = 2 X skin depth = 0.209 X 2 = 0.48cm ~ 26SWG. bare wire diameter Aw = 3.14D^2/4 = 0.00137 I/P power Pin = Po/n = 400/0.95 = 421W Ipk = pin X 1.414 /Vin(min) = 421 X 1.414 / 90 = 6.61A Ripple current = 0.2Ipk = 1.323 A 
27th May 2011, 09:58 AM  #22 
diyAudio Member
Join Date: Jul 2009
Location: West bengal/siliguri

max duty cycle Dmax = (Vo  (Vin X 1.414))/Vo = 0.68
Inductance L = (Vin(min) X 1.414)Dmax/ ripple current X frequency = .000685H energy Required Eng = LIpk^2/2 = 0.0149 (wattsecond). electrical cofficient Ke= 0.145PoBm^2/10000 = .0003625. core geometry coefficient Kg = (Eng)^2/ Ke X a = 0.615{cm^5} ok here is a slight problem the Kg value suggets me use a bigger core say ETD49 or so wat happens if I use a smaller core say ETD34. Last edited by sekhar; 27th May 2011 at 10:20 AM. 
27th May 2011, 10:05 AM  #23 
diyAudio Member
Join Date: Jul 2009
Location: West bengal/siliguri

lets proceed with ETD34 .....................
current density J =2(Eng)(10000)/Bin X Ap X Ku =1233A/cm^2 this also seems very high........... neways lets proceed............. Irms = Ipk/1.414 = 4.67A bare wire area Aw(n) = 4.67/1233 = .00394 cm^2. number of strands Sn = Aw(n)/ area of 26SWG wie = 3.07 ~ 3 number of turns = Wa X Ku /Sn X (area of 26SWG wire) = 129. 
27th May 2011, 10:11 AM  #24 
diyAudio Member
Join Date: Jul 2009
Location: West bengal/siliguri

calculate air gap Lg = (0.4 X 3.14 X N^2 X Ae)/L X 100000000 = 0.2971cm or 117mils...
fringe flux F =(1 + (Lg/root of Ae)ln(2G/Lg)) = 1.692.... little mistake due to log ...... new number of turns = root of{(Lg X L)/0.4 X 3.14 X Ae X F X (10^8)} = 99.36 ~ 99 
27th May 2011, 10:20 AM  #25 
diyAudio Member
Join Date: Jul 2009
Location: West bengal/siliguri

cal Bpk = F[{(0.4 X 3.14 X N X Ipk(10^4)}/Lg] = 0.468
new resistance per cm = 1345/Sn = 448.33 winding resistance R = (MLT) N 448.33 (10^6) = 0.315E copper loss Pcu = Irms^2 X R = 6.86W. Regulation a = Pcu/Po =1.717% Ac flux density Bac = 0.02768T core loss Pfc=0.018W total loss =Pcu +Pfc = 6.878W Temperature Rise = Tr = 82.35C these things dont seem to be in place ................... tell me if i had some mistakes........... regards sekhar 
27th May 2011, 02:40 PM  #26 
diyAudio Member
Join Date: Dec 2003
Location: Munich

...ermh..., I guess you did not mean 0.48cm, but 0.48mm...
Anyway, please note that the skin effect as calculated in your formulas is valid for a single wire in free space only. Practically this calculation does still fit in single layer windings, and more or less in two layer windings. But not for a massive bunch of conductors. You will need to consider the proximity effect, which is the extended consequence from the skin effect, when going for multiple conductors. The skin effect only takes into account the self induced effects of one wire. The proximity effect additionally takes into account the fields which are caused by the other conductors close by. You may search the web for graphs, which indicate Rac/Rdc for typical winding arrangements depending on frequency. Furtheron you are neglecting the eddy current losses, which are caused in the winding by the fringing field of the core. Sorry, I do not have a practical calculation model on hand, without finite elements. You can be sure that you will have more losses than your calculation indicates. What happens if you chose an etd34 instead of etd49? More losses than with an etd 49, and consequently much more temperature rise than with an etd49. New calculation of turns: Forget it. a) Your wild formula with a mixture of factors and units seems to be intended for pure AC flux situations with pos an neg flux. But your PFC will have only one flux polarity in the core, no sign reversal. ==> Giving you a factortwoerror for the number of turns. a) Simplified fringing field calculations tend to indicate an accuracy, which is simply not there. 
27th May 2011, 03:50 PM  #27  
diyAudio Member
Join Date: Jul 2009
Location: West bengal/siliguri

Quote:
So should i go forward with this design or should i sit back do some more studying................. regards sekhar 

27th May 2011, 04:04 PM  #28 
diyAudio Member
Join Date: Jul 2009
Location: West bengal/siliguri

@choholic
ok lets start the calculation all over again ........................ u tell me wat to do i will give all my input data, and lets see if can get this thing right atleast on paper......... then I will try making it .............. and thanks again for pointing out my mistakes................. and here is the schematic for the pfc regards sekhar 
28th May 2011, 04:16 PM  #29  
diyAudio Member
Join Date: Dec 2003
Location: Munich

Quote:
Sorry, I am out of the game. 

28th May 2011, 05:01 PM  #30  
diyAudio Member
Join Date: Jul 2009
Location: West bengal/siliguri

Quote:
regards sekhar 

Thread Tools  Search this Thread 


Similar Threads  
Thread  Thread Starter  Forum  Replies  Last Post 
New PFC Handbook  Krisfr  Power Supplies  0  19th February 2011 05:37 PM 
Bridgeless PFC help  koosh75  Power Supplies  17  22nd September 2010 04:13 PM 
Pfc + Ups = ???  ifrythings  Power Supplies  9  30th December 2007 08:48 PM 
PFC question  medogrizli  Power Supplies  12  3rd June 2007 04:49 PM 
New To Site?  Need Help? 