My ferrite core SMPS
rp181, Sun Jan 11 2009, 04:33AMHello all,
So i got a U126/91/20 core from ferroxcube awhile back, and I am actually doing stuff with it now. I plan on using a L298 for a H-bridge design. I plan to start with hard switching, and then later move on to SLR topology. Input will be voltage will be 330v, and a maximum of 10A (limit of my fuse right now). I will be using STGW30NC60W IGBT's. I made a bobbin with 1/8" acrylic, and i am using 30 AWG wire. the insulation thickness is.005 inches. Here are some pics:
A picture of old PVC bobbin vs. the new one.
The bobbin look horrible, but i say its pretty good for only using a acrylic sheet and a Xacto knife to make it (lots of scoring and breaking).
Any comment/questions/suggestion's for the driver (like frequency)?
Re: My ferrite core SMPS
Dr. Dark Current, Sun Jan 11 2009, 10:28AM
Cool core. Be careful when handling it, I got two monster cores from old TVs and even with normal handling, they developed some chips and dents around the area where the core halves touch.
I would go with ~50kHz for a first try.
Remember that the square wave drive puts out a lot of harmonics so you can excite the transformer's self resonant frequency even if it's magnitudes higher than your drive frequency.
For this reason I tune the driver at low power with unloaded transformer to a frequency close to the intended one, at which it draws the least current.
Dr. Dark Current, Sun Jan 11 2009, 10:28AM
Cool core. Be careful when handling it, I got two monster cores from old TVs and even with normal handling, they developed some chips and dents around the area where the core halves touch.
I would go with ~50kHz for a first try.
Remember that the square wave drive puts out a lot of harmonics so you can excite the transformer's self resonant frequency even if it's magnitudes higher than your drive frequency.
For this reason I tune the driver at low power with unloaded transformer to a frequency close to the intended one, at which it draws the least current.
Re: My ferrite core SMPS
rp181, Sun Jan 11 2009, 03:19PM
Yep, i already got small chips, and one bigger chip.
I heard somewhere (can't recall where) that due to some reason, the voltage output will be double than what your turns ratio will tell you. If i have a 10 turn primary (enough turns?) then estimated output would be around 5kV. Would the out put be 10kV?
For SLR topologies, everyone uses either a external inductor or with high leakage winding's. Is there anything wrong with gapping the core?
rp181, Sun Jan 11 2009, 03:19PM
Yep, i already got small chips, and one bigger chip.
I heard somewhere (can't recall where) that due to some reason, the voltage output will be double than what your turns ratio will tell you. If i have a 10 turn primary (enough turns?) then estimated output would be around 5kV. Would the out put be 10kV?
For SLR topologies, everyone uses either a external inductor or with high leakage winding's. Is there anything wrong with gapping the core?
Re: My ferrite core SMPS
uzzors2k, Sun Jan 11 2009, 04:42PM
Nice core. You probably heard it here.
And yes, the 2 * ideal ratio thing seems consistent.
Gapping the core ruins open circuit performance because the open circuit inductance will be close to the short-circuit inductance. With an ungapped core the open circuit resonant frequency will be so low that almost no current is drawn, which is good. Really, gapping the core is only useful in SLR as a last ditch effort to reduce the value of leakage inductance. If you fear that the leakage inductance will be too high then wind the transformer with good coupling, ie windings underneath eachother, and use an external inductor to get the required inductance.
uzzors2k, Sun Jan 11 2009, 04:42PM
Nice core. You probably heard it here.
And yes, the 2 * ideal ratio thing seems consistent.Gapping the core ruins open circuit performance because the open circuit inductance will be close to the short-circuit inductance. With an ungapped core the open circuit resonant frequency will be so low that almost no current is drawn, which is good. Really, gapping the core is only useful in SLR as a last ditch effort to reduce the value of leakage inductance. If you fear that the leakage inductance will be too high then wind the transformer with good coupling, ie windings underneath eachother, and use an external inductor to get the required inductance.
Re: My ferrite core SMPS
rp181, Sun Jan 11 2009, 06:00PM
Thats good, i ran short of wire so there are ~150 turns. If you look at the first picture, it wound to about a cursor width to the left of the white tape.
rp181, Sun Jan 11 2009, 06:00PM
Thats good, i ran short of wire so there are ~150 turns. If you look at the first picture, it wound to about a cursor width to the left of the white tape.
Re: My ferrite core SMPS
Dr. Dark Current, Sun Jan 11 2009, 06:02PM
AFAIK the gap does not do much with leakage inductance (if something then increase it)...
Dr. Dark Current, Sun Jan 11 2009, 06:02PM
Uzzors wrote ...
Gapping the core ruins open circuit performance because the open circuit inductance will be close to the short-circuit inductance. With an ungapped core the open circuit resonant frequency will be so low that almost no current is drawn, which is good. Really, gapping the core is only useful in SLR as a last ditch effort to reduce the value of leakage inductance. If you fear that the leakage inductance will be too high then wind the transformer with good coupling, ie windings underneath eachother, and use an external inductor to get the required inductance.
Well, you don't want any gap with bridge inverters, especially with the SLR. Even a very little gap often alters open circuit performance greatly.Gapping the core ruins open circuit performance because the open circuit inductance will be close to the short-circuit inductance. With an ungapped core the open circuit resonant frequency will be so low that almost no current is drawn, which is good. Really, gapping the core is only useful in SLR as a last ditch effort to reduce the value of leakage inductance. If you fear that the leakage inductance will be too high then wind the transformer with good coupling, ie windings underneath eachother, and use an external inductor to get the required inductance.
AFAIK the gap does not do much with leakage inductance (if something then increase it)...
Re: My ferrite core SMPS
rp181, Sun Jan 11 2009, 08:38PM
I will probably go with a external inductor on a wire spool (ID about 1inch).
Wont the inductor filter some harmonics? How do you think those IGBT's i posted will stand up? I can't find anything on there current rating vs. frequency.
On richard burnet's(?) site, i saw some thing about putting 2 or more H bridges together to distribute the power for induction heating. Would this same setup work if i split the inductor between them?
I finished the secondary. I ran out of wire so you can see the extra space on one side.
rp181, Sun Jan 11 2009, 08:38PM
I will probably go with a external inductor on a wire spool (ID about 1inch).
Wont the inductor filter some harmonics? How do you think those IGBT's i posted will stand up? I can't find anything on there current rating vs. frequency.
On richard burnet's(?) site, i saw some thing about putting 2 or more H bridges together to distribute the power for induction heating. Would this same setup work if i split the inductor between them?
I finished the secondary. I ran out of wire so you can see the extra space on one side.
Re: My ferrite core SMPS
Dr. Dark Current, Sun Jan 11 2009, 10:14PM
If its too big, spread the primary turns. If its too small, wind the primary in layers (so that the effective "core surface" under the winding is smaller).
Edit: Changing the leakage also changes the resonant frequency.
Dr. Dark Current, Sun Jan 11 2009, 10:14PM
rp181 wrote ...
I will probably go with a external inductor on a wire spool (ID about 1inch).
Why? Just wind the primary opposite the secondary, this should provide just about right leakage for experimental arc drawing. I will probably go with a external inductor on a wire spool (ID about 1inch).
If its too big, spread the primary turns. If its too small, wind the primary in layers (so that the effective "core surface" under the winding is smaller).
Edit: Changing the leakage also changes the resonant frequency.
rp181 wrote ...
Wont the inductor filter some harmonics?
no, but it adds inductance decreasing the system's resonant frequency.Wont the inductor filter some harmonics?
rp181 wrote ...
How do you think those IGBT's i posted will stand up? I can't find anything on there current rating vs. frequency.
They should be good to at least 100kHz or a bit more. Be sure to include clamping diodes on them (at least 30A high-speed diodes).How do you think those IGBT's i posted will stand up? I can't find anything on there current rating vs. frequency.
rp181 wrote ...
On richard burnet's(?) site, i saw some thing about putting 2 or more H bridges together to distribute the power for induction heating. Would this same setup work if i split the inductor between them?
Maybe, but just paralelling the transistors should have similar effect. I don't know exactly how big is your core, but these IGBTs in a full bridge should be able to supply ~4-8kVA powered from rectified 230V (or doubled 120V).On richard burnet's(?) site, i saw some thing about putting 2 or more H bridges together to distribute the power for induction heating. Would this same setup work if i split the inductor between them?
Re: My ferrite core SMPS
rp181, Mon Jan 12 2009, 12:34AM
I didn't think winding on opposite legs would be enough, but it sounds good.
If you go to ferroxcube's website, core sets, U, its the last one on the list.
EDIT: I plan on making a final bobbin much like finn hammer's so i can tap it for different voltages. What gauge wire to you guys recommend? I wan't to keep it in air, but i will put it under oil if i have to.
rp181, Mon Jan 12 2009, 12:34AM
I didn't think winding on opposite legs would be enough, but it sounds good.
If you go to ferroxcube's website, core sets, U, its the last one on the list.
EDIT: I plan on making a final bobbin much like finn hammer's so i can tap it for different voltages. What gauge wire to you guys recommend? I wan't to keep it in air, but i will put it under oil if i have to.
Re: My ferrite core SMPS
uzzors2k, Mon Jan 12 2009, 05:48PM
You're right, my bad.
uzzors2k, Mon Jan 12 2009, 05:48PM
Dr. Kilovolt wrote ...
AFAIK the gap does not do much with leakage inductance (if something then increase it)...
AFAIK the gap does not do much with leakage inductance (if something then increase it)...
You're right, my bad.
Re: My ferrite core SMPS
rp181, Tue Jan 13 2009, 03:27AM
I am winding a new secondary after i ran out off wire on the other. On this one, i epoxied the acrylic together (rather than tape) so i could use it under oil. I am winding the secondary in two sections, of about 80 turns each, for a total of 160. On the primary side, I was planning to use 330v with 10 turns, for 33 v/turn. I think this core can handle more (based on finn's measurements), so i think i could go a 5 turn primary. That gives me a output of 5kV a section, for 10kV total. After the harmonics doubling, that gives me 20kV. Anyone see any flaw in my logic?
rp181, Tue Jan 13 2009, 03:27AM
I am winding a new secondary after i ran out off wire on the other. On this one, i epoxied the acrylic together (rather than tape) so i could use it under oil. I am winding the secondary in two sections, of about 80 turns each, for a total of 160. On the primary side, I was planning to use 330v with 10 turns, for 33 v/turn. I think this core can handle more (based on finn's measurements), so i think i could go a 5 turn primary. That gives me a output of 5kV a section, for 10kV total. After the harmonics doubling, that gives me 20kV. Anyone see any flaw in my logic?
Print this page