I cant really find anywhere on how to wind a toroidal.
soo, im going to use a 4-0-4 and a 12-0-12 or 1:3 turns ratio for a 12V to 36v SMPS.
well i need to see if im doing this right. the centertap just dont look right on my toroid. but it may just be me.
but i have 4 windings, centertap, then another 4 windings. im using 4 strands of 26AWG wire. going to run somewhere between 35 to 50khz. and a couple of LM3886 amps.
check and see if my primary looks right, if not, how do i fix?
also when i go to put on the secondary. how should i start the secondary, and which direction do i wind in respect to my primary? (i wound clockwise).
here is the pic:
soo, im going to use a 4-0-4 and a 12-0-12 or 1:3 turns ratio for a 12V to 36v SMPS.
well i need to see if im doing this right. the centertap just dont look right on my toroid. but it may just be me.
but i have 4 windings, centertap, then another 4 windings. im using 4 strands of 26AWG wire. going to run somewhere between 35 to 50khz. and a couple of LM3886 amps.
check and see if my primary looks right, if not, how do i fix?
also when i go to put on the secondary. how should i start the secondary, and which direction do i wind in respect to my primary? (i wound clockwise).
here is the pic:
An externally hosted image should be here but it was not working when we last tested it.
keep primaries close to each other, not opposite sides of toroid (use bifilar winding for example)
4 turns can be very dfficult to distribute evenly accross entire toroid, but one possbility is to make more than one primary, for example wind 4 identical sections of 4+4 bifilar to toroid with single wire and connect them parallei on PCB.
4 turns can be very dfficult to distribute evenly accross entire toroid, but one possbility is to make more than one primary, for example wind 4 identical sections of 4+4 bifilar to toroid with single wire and connect them parallei on PCB.
well i didnt do something right.
did 4 - 0 - 4 turns
and 12 - 0 - 12 turns.
with an input voltage of about 12V, and input frequency of about 32khz, i get an unregulated output of about 82v DC.
so i built a regulator circuit. im not using an SG3525. i decided to do something different. i used an AVR ATmega8 CPU to control it, using dual PWMs.
use the ADC and a 10-1 divider, 5V max reso, about 3.2v being normal voltage, times 10, is 32v.
thats about an ADC value of 655.
i wrote a little program that if above 655, i decrease the PWM. if below 655, i increase the PWM. if 655, i leave the PWM duty setting alone.
works like a charm.
did 4 - 0 - 4 turns
and 12 - 0 - 12 turns.
with an input voltage of about 12V, and input frequency of about 32khz, i get an unregulated output of about 82v DC.
so i built a regulator circuit. im not using an SG3525. i decided to do something different. i used an AVR ATmega8 CPU to control it, using dual PWMs.
use the ADC and a 10-1 divider, 5V max reso, about 3.2v being normal voltage, times 10, is 32v.
thats about an ADC value of 655.
i wrote a little program that if above 655, i decrease the PWM. if below 655, i increase the PWM. if 655, i leave the PWM duty setting alone.
works like a charm.
mbates14 said:
Incorrect. This assumes that your 8 turns around the core result in the center-tap being on the opposite side of the core. This is not the best approach from a component layout perspective. All primary connections should be on the same side of the core, and same thing for all secondaries. So, your primary winding should be wound as: 4T around the entire core, and another 4T wound right next to the first four. Both halves of the primary should be wound as closely together as possible. A great way to insure this is to wind them in bi-filar fashion. Assuming each primary half is a single conductor, cut two pieces of equal length, and wind them together on the corre. Take the "dot" end of one winding half and tie it to the "no dot" end of the other winding half. This is your center-tap. Repeat this procedure for the 12T+12T secondaries.
Also, if high voltages are involved in the design (>60VDC), a layer of transformer tape should separate the primary and secondary windings.
Steve
ok i understand. i had to google up bi-filar windings, and i understand what your saying.
by tearing apart a blown up commecial car amp, its transformer is wound the same way, but the windings are toast.
so i just used that core, and re-wound it using the way you describe.
still i get around 85V unregulated from each half. but that could be the way i have my test circuit build and all that stray inductance and capacitance.
here is my test circuit AFTER i built the regulation circuit and wrote the regulator software:
by tearing apart a blown up commecial car amp, its transformer is wound the same way, but the windings are toast.
so i just used that core, and re-wound it using the way you describe.
still i get around 85V unregulated from each half. but that could be the way i have my test circuit build and all that stray inductance and capacitance.
here is my test circuit AFTER i built the regulation circuit and wrote the regulator software:
An externally hosted image should be here but it was not working when we last tested it.
well that didnt wrok too well.
i hooked the amp and did a test run, well it went
rails dropped between 12 to 21 volts, and had intermittent hissing/distorted audio, and sometimes it was fine. when it was fine, and you turned it up, it would just start hissing and the rails would drop, everything would get hot.
then it just blew the FETs.
i hooked the amp and did a test run, well it went
rails dropped between 12 to 21 volts, and had intermittent hissing/distorted audio, and sometimes it was fine. when it was fine, and you turned it up, it would just start hissing and the rails would drop, everything would get hot.
then it just blew the FETs.
That could be caused by other factors than your transformer:
A. Your not driving your mosfets into saturation (or not quickly enough).Are you using a Gate driver? If the transistors are not fully turned on (Saturated) than the transistor operates like a resistor causing excessive heat and limits current input into the transformer (output voltage saging).
Another cause of blow Mosfet is if voltage spikes are exceeding the max voltage rating of your Mosfet. Note that your Mosfets need to have at least twice the voltage rating of your input voltage with Push-Pull designs. You also need to accomidate for voltage spikes caused by inductance leakage. Usually when I test a new SMPS design I use higher voltage MOSFETS until I I've got the voltage spikes under control, and then switch to the higher current/ Lower voltage MOSFET. Note that voltage spikes don't usually appear unless you have a load connected to the secondary winding
B. The core is saturating or you have flux walking. If it's just the case of Core saturation, you may be able to fix it by switching at a higher frequency or by increasing the number of turns on your transformer. As for for Flux walking, Check the primary inductance of both push-pull halfs. If they are unbalanced, its can cause an inbalance where the transformer walks into saturation as the current input is higher on one half of the push-pull cycle, than the other (ie two steps forward, one step back). You either need to balance the two primary halfs or use a controller than can adjust for imbalances. The issue is with a low number of primary turns, it can become tricky to achive balance since you'll probably have to add or subtract a fraction of a turn to achive balance. Personally, I don't like Push-Pull because of these problems.
FWIW: During testing you can include an inductor and fuse\breaker in series with your primary in order to mitigate excessive current if your core saturates resulting in explosing transistors. The inductor will of course limit the amount of output current, but it can help avoid blowing transistors when you do some initial testing. Once you checked that you don't have core saturation or issues with unsaturated transistor switching, you can remove the inductor (or if you wish, decrease the size of the inductor to increase your output current for higher output current testing). Its also a good idea to include current monitoring using your SMPS controller, since even with a full working design you could still destroy your transistors with excessive output loads.
As for the excessive output voltage it could be caused by inductance leakage spike noise leaking into the secondary winding. Check the output with an oscilliscope to see if you have large voltage spikes in the waveform. If the main body of the waveform is well above your target voltage, then your primary to secondary ratio is wrong. If its just voltage spikes, you can probably filter them out with snubbers on the primary and an output filter. BTW, You are using an output Inductor, right? You will need output inductor for your SMPS to operate correctly.
I would probably test with a controlled output device (ie a power resistor) rather than a stereo amp or some other finished product that has dynamic current loads. You need a static current load to check your design for faults. I use a programmably load bank for testing, but there is nothing wrong with using a set of power resistors to setup various static loads. Once you got all the bugs worked out with static loads, then you can move on to dynamic load testing. This is to work out bugs in your feedback loop which can destabilize with dynamic loads.
i fixed it. the problem was the voltage from my AVR was only 0 to 5V to 0v on PWM, and wasnt pushing the transistors into saturation.
i fixed that using optoisolators.
anyway, the next problem is the spikes in the primary. if i use a 12v 5amp power supply. everythings all hunkey dorey. if i use a 12v battery, it works for a few minutes, blows the fets.
i fixed that using optoisolators.
anyway, the next problem is the spikes in the primary. if i use a 12v 5amp power supply. everythings all hunkey dorey. if i use a 12v battery, it works for a few minutes, blows the fets.
You can use a gate driver like IR4427 to perform the level shift and current gain, or you can go the discrete way with a minimum of 3 transistors. Wrong gate drive will cause the MOSFETs to fail. You should adjust turn-on and turn-off times by means of series gate resistors in order to get clean drain switching transitions in the 50ns to 500ns range, while providing a low enough turn-off impedance to prevent each gate from being pulled high by drain-to-gate capacitance when the opposite MOSFET is turned on. Faster turn-off than turn-on is recommended.
You have to follow some basic circuit layout guidelines for 12V push-pull converters. An input CLC filter is required with the last capacitor bank near the MOSFETs and the transformer. This CLC filter should be sized to handle the primary peak current and should resonate well below twice the switching frequency, Fr=1/(2*pi*sqr(L*C). RC snubbers across the primaries are also required for reducing EMI and they are important (together with some dead time) to achieve self flux-balancing in the transformer and prevent saturation. Use f=1/(R*C) as a guideline to chose snubber values, where f is the drain ringing frequency. RC snubbers across the secondaries are usually required too to damp ringing after diodes turn off.
The transformer has to be wound in such a way that each winding is spread around the entire core. Primaries should be bifilar-like to maximize coupling between them, and secondaries too, but never ever twist the wires of different windings because this would result in too much capacitance and resonance. Tight coupling will reduce leakage inductance, which is the source of primary drain spikes.
All this information is available through the search engine, and you are a very lazy person if you haven't already found it. 
You have to follow some basic circuit layout guidelines for 12V push-pull converters. An input CLC filter is required with the last capacitor bank near the MOSFETs and the transformer. This CLC filter should be sized to handle the primary peak current and should resonate well below twice the switching frequency, Fr=1/(2*pi*sqr(L*C). RC snubbers across the primaries are also required for reducing EMI and they are important (together with some dead time) to achieve self flux-balancing in the transformer and prevent saturation. Use f=1/(R*C) as a guideline to chose snubber values, where f is the drain ringing frequency. RC snubbers across the secondaries are usually required too to damp ringing after diodes turn off.
The transformer has to be wound in such a way that each winding is spread around the entire core. Primaries should be bifilar-like to maximize coupling between them, and secondaries too, but never ever twist the wires of different windings because this would result in too much capacitance and resonance. Tight coupling will reduce leakage inductance, which is the source of primary drain spikes.
All this information is available through the search engine, and you are a very lazy person if you haven't already found it.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.