They are small psu from china,labeled '400 watts' and '550 watts'.......honestly i dont think they are anywhere close to 200W. Their main formers are much smaller compared to a genuine 215W psu from my cpu.
I just want to grind the outer legs then smoothen them with a sandpaper to see how it'll turn out. Am a little busy(in college) and i dont know how much time i can have for this project. Am not in a hurry to build it,i want the knowledge and experience first. Maybe i'll use the small cores first.
I just bought a KA3525A, i haven't read the datasheet yet; is it an sg3525 equivalent?
Regards,Xeclipse
Last edited:
Don't grind the legs,the only part of a core set that should be is the center leg,you can add gaps in the outer legs with a business card(1mm).Don't let the size of the core fool you,it's the frequency that you use it that determines the power up to a point,taking in the saturation properties.The size of the center leg will give you some idea,small less power while large more power...Beware of gapping cores unless necessary.
Am not trying to gap the core.......am trying to eliminate the gap in the center leg by grinding the outer legs.
now with +/-66v it is over 400w into 4r load, far more then my speaker can handle and/or I need... still thinking if I would connected it to +/-75
Just remember to take some photos when the magic smoke come out of that speaker!
Hey Guys,
The only way a high-side N-Channel MOSFET can be driven w/o a floating-gate driver IC (IR2113 or equivalent) or gate-drive xfmr is by using a discrete boot-strap circuit. I forgot the circuit layout, but if I come across it, I will post it here.
Cheers,
Steve
Refer to classd amplifiers in this forum,used in quite a few designs.My question is why bother,a gate drive transformer can be realized with a few turns on a ferrite core and ic's are cheap,fewer parts less costs,improve mtf.There are quite a few app notes on this subject.A gate drive transformer is cheaper and consumes less pcb area.
A good app note from TI which you can find here http://www.ti.com/lit/ml/slup083/slup083.pdf has a step by step design of gate drive design as well as other info for the proper design of a HB converter,yes it current mode but applies to voltage mode as well.
Last edited:
Ok lets break this down.
everything stays the same. As you can see +VIN is connected to Vref (pin16) via R9. -VIN is connected to COMP (pin 9).
To explain more: To get max duty cycle, as it is wire now, you have to have 5v on +VIN, since OP inside is wired as voltage follower (if I remember right, didn't check in a long time, but I think it is like that)
^^
This is for unregulated, if regulated, opto would pull down the voltage, to reduce the duty cycle
In other words, 5v = max D.C., 0v = min D.C. = 0 D.C.
If you creck the pictures, you can see that I have 2nd board for feecback
All you have to do to go from regulated >> unregulated is to pull that card out, everything else stays the same. My was unregulated most of the time, I only had card in (or use those pins) is when I played with regulation, which I might have to again, unless I get bigger core for IRS amps, current ones won't hold up with full smps + pfc combo voltage
that is right. you can bring any source of clock, inpulses form anywhere, even signal generator to this pin. Idea was back in the day, if I would need it, I would get this clock of the Damp board, which had 4060 on it, one of its pins provided the clock for the amp, and other would be/ could be for the smps. This way smps would no work at this freq. as it does now with standart RC setup.
What happens to deadtime at max DC?
What if i set deadtime resistor to 0?
This is where i got lost,can i leave it floating or do i ground it?
Regards,Xeclipse.
There is always some dead time (as you set it), D.C. in not 50% for this same reason. Resistor needed is in range from 0 to 500r I think
Floating, since RC take over... only if you would bring pulses (and what you need to do, and how should pulses look like is in datasheet), it would sync to that freq.
Last edited: