hi power smps

Status
Not open for further replies.
Jump to Latest
#62
well,
the diodes seem to balance the dead time level of the transformer, and reduce some of the overshoot, but that was before i changed transformers from a toriod to a double e core.

i was trying to find the source for where i saw the diodes all over the place, but i have not found it yet..

i have calculated the switching freq at about 63khz and the osc freq at about 126khz.

i am doing testing tonite or tomorrow and we will. see....................

thanks......
 
#64
Ultima Thule said:
Jimbo,

I have not read this thread too thoroughly so forgive me.

Anyhow I looked at your last schematic, why do you have lots of diodes around all switching FET's??

If the bounce is MORE than the rails voltage of 320 Volts it wont be clamped because your diodes prevent the intrinsinc FET diode to clamp the overshoot.

Also you have more switching losses with these diodes and they cost money, make no sense to me.

Cheers Michael
Click to expand...

michael,
untill i can find the source for the diodes and why i guess i'll remove them...
 
#65
Hi Jimbo,

what you brought up regarding those diodes is entirely something new for me.

Though, I would like to stress that any small imbalance in your core will of course make it saturate easilly, I hope you are going to use pin 9 on MC33/34025 in some way as I can't see it being connected in your schematics.

btw, I went back and read other posts in this thread and think Eva made good comments about FET's and the body diode, was it in post #13...

Cheers Michael
 
#67
Hi guys (& gals) 😀,

I haven't posted here in a while, so I thought I'd drop by. 😉

It has been my experience with RC snubbers on the secondary side of AT/ATX supplies, for the 5V winding (low voltage), the RC goes across rectifier bridge, and for the 12V winding (high voltage), one RC combo goes in parallel with each rectifier.

Extrapolating this to our +/- 35-40V outputs, I would place an RC snubber in parallel with each diode of the full-wave bridge, just like the 12V for the pc supply. Obviously, the math comes grom Chryssis/Pressman/Brown's books.

Thoughts?

Steve
 
#69
The purpose of these RC snubbers is to smmoth a bit the abrupt impedance changes that diodes exhibit both when they turn on and when they turn off. They are best placed in parallel with each diode, although that may result sometimes prohibitive in cost and space terms. These impedance changes have to be smoothed (and the resulting ringing damped) because they are a source of high frequency excitation (usually up to 50Mhz) for any tank circuit in the nearhood, which are usually formed by transformer winding parasitistics, PCB track inductance and capacitor parasitistics.
 
#71
ok,
i can tak care of the ringing on the rectifiers and the transformer when the fets come on.

i have rinning also in the deadtime of the primary when the full bridge is off.

i can suppress that but it sucks a lot of power off of the transiston between the a and b on and off modes.

i ahve seen other wave orms where thois ringin is allowed to go on until the next fet comes on, and the rining goes from ail to rail
 
#74
i removed all of the series diodes around the fets, i cannont find the source info that i found.

so 🙂 🙂 🙂 🙂 🙂 😉
the pnp gate drive circuit that EVA has shared sems to work. no more instananeous explosion here in nebraska now.


i have successfully run at 48 amps at 17 volts.

however.

i have noted that when the output voltage starts regulating (shorening the pulses) mechanical noise appears and the pulses are all over the place....

assuming that my voltage feedback is not stable...

also i am working on the current limiter, colicraft, full bridge resistor network fed to pin 9 of 33025...
 
#76
jamesrnz said:
ok,
i can tak care of the ringing on the rectifiers and the transformer when the fets come on.

i have rinning also in the deadtime of the primary when the full bridge is off.

i can suppress that but it sucks a lot of power off of the transiston between the a and b on and off modes.

i ahve seen other wave orms where thois ringin is allowed to go on until the next fet comes on, and the rining goes from ail to rail
Click to expand...


It depends on the frequency and on how much current is involved in that ringing. If the transformer leakage inductance wants to ring at, say, 2Mhz or 4Mhz with MOSFET parasitistic capacitances, that will not cause much trouble. However, other forms of ringing above, say, 10Mhz where PCB parasitistics, transistor and diode lead parasitistics and capacitor parasitistics are involved must be tamed, preferably by improving layout. For example, I have recently found out that TO-247 packaged MOSFETs and IGBTs tend to self resonate at 20Mhz to 40Mhz (25nH lead inductance and 250pF to 1000pF C-E or D-S internal capacitance) , which is indeed a big pitfall and a powerful reason to use parallel TO-220 transistors instead and avoid big packages. I had to slow down turn on, add RF decoupling with ferrite beads in the gates and strong RC snubbers to prevent such RF ringing (that would be happily radiated by magnetics otherwise).
 
#77
eva,

i believe you are correct i have some high frquency ringing that is probably due to poor layout, my prototype is fairly ugly and sloppy.

but i have been able to make power now, i have to tame the voltage feedback instability when the voltage feedback kicks in at higher currents.

i have been reading about that guy 'bode' and poles and zeros... neat:xeye:
 
#78
luka said:
Hi jamesrnz,

I'm sorry, I didn't find your out voltage, what is it ? Are U using halfbridge for this tests? What will U use this supply for?

Maybe you should read my thread, since we have the same problem.

THX, Luka
Click to expand...

Hey,

i am building a battery charger for a really big battery output about 17 volts at 50 amps continuous.

i am using full bridge.

i am just about ready to post a schematic of what i have so far.
 
#80
There are few things that are possible:
1.)your 330uf is way to low for 50A current, so there must be large ripple voltage/current that is heating your cap.
2.)your inductor could be to small, but I think that if you will put bigger problem won't go away, but it could help.Don't use to big one because of the responce time to any changes.
3.) is the most important: Feedback is most likely to couse trouble, it is responsible for mech. noise, but there could be another thing.If your input cap. is to small, 50Hz rectified (100Hz)will be travelling from primary to secondary side.
What is your input cap size?
 
Status
Not open for further replies.
Top Bottom