Reset winding sizing for forward converters / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Reset winding sizing for forward converters

Move Thread

LAN_403

ConKbot of Doom

Mon Jun 20 2011, 08:49PM

Registered Member #509 Joined: Sat Feb 10 2007, 07:02AM
Location:
Posts: 329

I'm putting together a battery balancer using the principal shown here

Basically just unregulated bi-directional DC-DC converters with a common bus to balance the charge of seriesed cells to ensure that their voltages are all the same.

Looking at page 9, this seems like it would be best to address each converter as a forward converter without the rectifier and smoothing inductor.

I have a few toroids of "77" material, put some windings on one. Ive been poking at it with a transistor I had handy, power supply, and sig-gen for a bit, and have mostly made sense of the forward converter. Without a snubber or reset winding, the big current spike as it went into saturation was obvious. With a reset winding and diode, the waveforms make sense, and match up with what ive seen elsewhere.

Since it is more efficient to use a reset winding vs a RCD snubber, I'm going to go with one of those, but trying to wind a toroid with bunches of different windings all tightly coupled becomes problematic after a while, so I'd like to use thinner wire for the reset winding if possible.

I would hope to not have to do more then 3-5A RMS on the converter under worse case (one cell sagging under load, and the balancing converters running to try and maintain voltage on the weak cell, but I wouldnt really expect a converter to handle the full current of the pack) So I can probably step down to 22AWG rather then 20AWG I'm using now, but the reset winding is only carrying the energy stored in the core, so RMS current should be a lot less then the primary correct? So I could use something like 24-28 AWG for that?

How tight does the coupling on the reset winding need to be? Currently I'm wrapping the 3 windings flat, next to each other the whole way, spread evenly around the toroid. Is it ok to wrap the reset over the primary and secondary? Or should I strive to keep them all flat in a single layer?

Is there a trick that would let me wind the secondaries on, then put a thin, conforming layer of insulating tape on, and wind a primary and reset winding on a toroid, all in a single flat layer? Only thing I could see would be to wind all 4 on to get the right spacing, unwind/cut off the primary and reset, add tape, rewind primaries. This is only ~10" of wire for each one, so thats not that unplausible, but just seems silly none the less.

Thanks!

Mattski

Tue Jun 21 2011, 08:45AM

Registered Member #1792 Joined: Fri Oct 31 2008, 08:12PM
Location: University of California
Posts: 527

ConKbot of Doom wrote ...
Since it is more efficient to use a reset winding vs a RCD snubber, I'm going to go with one of those, but trying to wind a toroid with bunches of different windings all tightly coupled becomes problematic after a while, so I'd like to use thinner wire for the reset winding if possible.

I would hope to not have to do more then 3-5A RMS on the converter under worse case (one cell sagging under load, and the balancing converters running to try and maintain voltage on the weak cell, but I wouldnt really expect a converter to handle the full current of the pack) So I can probably step down to 22AWG rather then 20AWG I'm using now, but the reset winding is only carrying the energy stored in the core, so RMS current should be a lot less then the primary correct? So I could use something like 24-28 AWG for that?

Yes, RMS current in the reset will always be less than in the primary winding. The peak current in the reset winding is equal to the magnetizing current in the primary (assuming equal # turns in primary and reset) which will be Vin*D/f_switch where D is your duty cycle. Figure out what your maximum duty cycle is going to be so you can find the peak current. The reset winding should conduct for D/f_switch seconds (again assuming Nprimary=Nreset) linearly decreasing from Ipeak to zero, giving an RMS current if I recall my math correctly of I_reset,rms=sqrt(D/3)*I_peak. Decreasing the reset:primary turns ratio increases transistor voltage stress, decreases reset time, and increases peak reset current.

wrote ...
How tight does the coupling on the reset winding need to be? Currently I'm wrapping the 3 windings flat, next to each other the whole way, spread evenly around the toroid. Is it ok to wrap the reset over the primary and secondary? Or should I strive to keep them all flat in a single layer?

To get the lowest leakage inductance (highest coupling) in a toroidal transformer my understanding is that you should wind all windings all the way around the core, and bifilar/trifilar type windings are even better. I'm not really sure offhand what the impact would be of a high leakage inductance in the reset winding.

ConKbot of Doom

Thu Jul 07 2011, 01:21AM

Registered Member #509 Joined: Sat Feb 10 2007, 07:02AM
Location:
Posts: 329

Thanks, thats what I thought. :)
I re-wound a test transformer and its all working as good as it could be given the parasitics of my test setup. A mosfet gate with 5' or RG-58 between it and the signal generator rings quite a bit.

I need to take the time and solder up a breadboard with my gate drive circuit on it. I take it for a very low quiescent current draw gate driver (or at least when off/low) a small mosfet totem pole to drive my power mosfet would be the easiest way? I cant seem to find any optocouplers that dont have a supply current of a few mA (other then plain phototransistor output ones of course)

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob