Not sure, that this is the best I have, but try http://www.smps.us
There you can find some theory and useful application notes from semiconductor suppliers.
There you can find some theory and useful application notes from semiconductor suppliers.
Some of the best tutorials are on National Semi's website -- they have an interactive smps design tool. Their discussion of the physics is probably the most comprehensive.
TI is the largest (in the U.S. at least), having aquired Unitrode a few years back.
Linear Technology probably has a very good explanation of noise in SMPS.
OnSemi also has a good interactive tool designed by Ridley Engineering: http://www.ridleyengineering.com/software.htm
here's a good power-point presentation on Power Factor Correction:
http://www.ftn.ns.ac.yu/ieee/jch-ns/arch_2001/10.pps
TI is the largest (in the U.S. at least), having aquired Unitrode a few years back.
Linear Technology probably has a very good explanation of noise in SMPS.
OnSemi also has a good interactive tool designed by Ridley Engineering: http://www.ridleyengineering.com/software.htm
here's a good power-point presentation on Power Factor Correction:
http://www.ftn.ns.ac.yu/ieee/jch-ns/arch_2001/10.pps
I just played around a little with that SMPS online calc tool - it strikes me how large cores it suggests, for an SMPS capable of 30 volts, 18 amps@50kHz, it is suggesting ETD59 core, isnt this a bit large? I thought the ETD59 core could handle power in the kW range ( I recently disassembled a 1,2kW 24V telecom PSU, it had a single ETD59 core )
That may be because you have chosen the wrong topology. Such a big transformer would be actually required in order to get 540W output only with non-push-pull topologies.
For push-pull topologies, even the E34 core from a conventional 200W ATX psu can easily acomodate enough magnet wire for 500W output (I've done it). Also, currently I'm using E42/21/20 for 900W at 32Khz and the winding area is far from filled. I think that even an ETD29 may be suitable for your needs.
For push-pull topologies, even the E34 core from a conventional 200W ATX psu can easily acomodate enough magnet wire for 500W output (I've done it). Also, currently I'm using E42/21/20 for 900W at 32Khz and the winding area is far from filled. I think that even an ETD29 may be suitable for your needs.
Yupp, i chose push-pull, full bridge topology. Try it yourself, and take a look =) input voltage 10 to 16 volts, output 30 volts, 18 amps. 50 kHz.... and it recommends the ETD59.... weird...
I'm currently building a 4*100 watt car amp for a friend, with high current capabilities, I'm going to use an ETD34 I got, i think it will be more than enough - but then i tried that online calculator...
I'm currently building a 4*100 watt car amp for a friend, with high current capabilities, I'm going to use an ETD34 I got, i think it will be more than enough - but then i tried that online calculator...
For a simgle unregulated push-pull 12V to 30V design with an ETD34 core and a 100Khz clock (50Khz effective), I would use 3+3 turns for the primaries and 8+8 turns for the secondaries. Worst case peak flux density would be 300mT with 16V input, which would be easily handled by any decent power ferrite.
You can calculate peak flux as follows :
B= (V/N) * (T/S)
Where V is total voltage applied, N is the number of turns (so V/N is just the amound of volts per turn), T is half the period of a cycle (5uS for 100Khz clock and 50Khz effective, that's because the flux swings from negative to positive) and S is the minimum cross sectional area of the core in m^2 (0.000091 for ETD34).
I think that I would place the primaries in the center, each one in a layer, and I would split the secondaries in two 4 turn halves that should be connected in series and placed each in a layer, under and over the primaries respectively. That would create a little asymmetry in the primary windings due to slightly different magnet wire lenghts, but a bit of dead time and some compensation in form of PCB track resistance will avoid any transformer saturation issue.
Such windings should be made with several magnet wires connected in paralell, so that the entire width of the winding window is filled. I have not made any calculation, but I think that there is enough room for windings capable of whitstanding in excess of 500W. Copper foil could also be used, but it's a mess to isolate and connect.
You can calculate peak flux as follows :
B= (V/N) * (T/S)
Where V is total voltage applied, N is the number of turns (so V/N is just the amound of volts per turn), T is half the period of a cycle (5uS for 100Khz clock and 50Khz effective, that's because the flux swings from negative to positive) and S is the minimum cross sectional area of the core in m^2 (0.000091 for ETD34).
I think that I would place the primaries in the center, each one in a layer, and I would split the secondaries in two 4 turn halves that should be connected in series and placed each in a layer, under and over the primaries respectively. That would create a little asymmetry in the primary windings due to slightly different magnet wire lenghts, but a bit of dead time and some compensation in form of PCB track resistance will avoid any transformer saturation issue.
Such windings should be made with several magnet wires connected in paralell, so that the entire width of the winding window is filled. I have not made any calculation, but I think that there is enough room for windings capable of whitstanding in excess of 500W. Copper foil could also be used, but it's a mess to isolate and connect.
hello dear friend
my name is Hadi from iran
i want to creat my own car amplifier but can not to creat smps for that.
i see in diy site you do that in past.
can you please help me to creat a good smps?
sory for my bad english
my email adress is Hadi_ghr_e_61@hotmail.com
very thanks for your helping.
have a nise day dear friend.
my name is Hadi from iran
i want to creat my own car amplifier but can not to creat smps for that.
i see in diy site you do that in past.
can you please help me to creat a good smps?
sory for my bad english
my email adress is Hadi_ghr_e_61@hotmail.com
very thanks for your helping.
have a nise day dear friend.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.