Maker Pro
Maker Pro

Need forward resonant control chip.

Y

Yzordderex

Jan 1, 1970
0
I am looking for a control chip for a forward resonant converter. It
will run around 500kHz and drive a rf deck. Output voltage 0-120vdc
into about 10ohm load. I'm thinking fixed off time to allow drain
voltage to resonate back to zero for soft turn-on. This is for a
friend of mine to play with. I am using a full bridge UC3875n
controller in my own design. Friend is hoping to simplify.

regards,
Bob

http://mysite.verizon.net/vze6qsvc/coolradiostuff22/
 
W

Winfield Hill

Jan 1, 1970
0
Yzordderex wrote...
I am looking for a control chip for a forward resonant converter.
It will run around 500kHz and drive a rf deck. Output voltage
0-120vdc into about 10ohm load. I'm thinking fixed off time to
allow drain voltage to resonate back to zero for soft turn-on.
This is for a friend of mine to play with. I am using a full
bridge UC3875n controller in my own design. Friend is hoping
to simplify.

Check out the ucc3895 resonant ZVS phase-shift controller,
http://focus.ti.com/docs/prod/folders/print/ucc3895.html

In a proper design it can handle your needs, although in my
estimation engineering with this chip has the capability of
separating the men from the boys, if you catch my drift. :>)
TI provides a half-dozen app notes to help you get started.

Multiple types of bench measurements can help you supplement
the data sheet's sometimes sparse details, or decode confusing
information. The end result can be impressive. I made a 500W
600kHz controller that I'm fairly happy with, even though it
ballooned to over 150 components, including safety shutoffs.
So I'm not so sure about your simplification goal. Improved
performance is what you'll be enjoying.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)
 
W

Winfield Hill

Jan 1, 1970
0
Winfield Hill wrote...
Check out the ucc3895 resonant ZVS phase-shift controller,
http://focus.ti.com/docs/prod/folders/print/ucc3895.html

[ snip ] I made a 500W 600kHz controller that I'm fairly
happy with ...

A primary claim to fame for my 600kHz 500W system is its
smooth 10,000:1 output-voltage operating range, namely from
1V minimum (with fine control at 1V) to 10kV maximum (with
precise control there). TI's ucc3895 phase-shift controller
was one critical part of the design, a 10:1 pre-controller
supply was another. My 60Vdc 500W 600kHz H-bridge uses my
favorite HIP4081A to drive four Fairchild FQP33N10 MOSFETs.
The latter are big enough, but not too big (this is a very
important issue at high frequencies). I haven't measured
the system's overall efficiency, but when it continuously
delivers 400W at 300kHz, it does so with just one small
external heat sink and no fan. I'd guess the total losses
(dissipation) are under 15W, which is about 96% efficient.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)
 
H

Harry Dellamano

Jan 1, 1970
0
Winfield Hill said:
Winfield Hill wrote...
Check out the ucc3895 resonant ZVS phase-shift controller,
http://focus.ti.com/docs/prod/folders/print/ucc3895.html

[ snip ] I made a 500W 600kHz controller that I'm fairly
happy with ...

A primary claim to fame for my 600kHz 500W system is its
smooth 10,000:1 output-voltage operating range, namely from
1V minimum (with fine control at 1V) to 10kV maximum (with
precise control there). TI's ucc3895 phase-shift controller
was one critical part of the design, a 10:1 pre-controller
supply was another. My 60Vdc 500W 600kHz H-bridge uses my
favorite HIP4081A to drive four Fairchild FQP33N10 MOSFETs.
The latter are big enough, but not too big (this is a very
important issue at high frequencies). I haven't measured
the system's overall efficiency, but when it continuously
delivers 400W at 300kHz, it does so with just one small
external heat sink and no fan. I'd guess the total losses
(dissipation) are under 15W, which is about 96% efficient.

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)

Sounds impressive Win.
Were you able to attain soft switching over the entire load range? Did you
have any noise problems if/when hard switching? The choice of MOSFETs is
important. Look at the STW45NM50FD, it is a top performer at higher
voltages.
When I use this configuration, the method of current sensing is key to
avoid charge dumping if/when hard switching. There are configurations that
attain soft switching for all line load ranges that perform nicely.
Regards
Harry
 
W

Winfield Hill

Jan 1, 1970
0
Harry Dellamano wrote...
Sounds impressive Win.
Were you able to attain soft switching over the entire load range?

No, of course not!
Did you have any noise problems if/when hard switching?

I had some difficulty, which I solved with snubbers and paralleled
mbr10100 10A 100V Schottky diodes (after blowing out a few FETs).
A major issue to prevent excessive damaging spikes was completely
eliminating shoot-through. Turning the FETs completely off rapidly
was an issue. I used a 10-ohm gate resistor to slow down turn-on,
but paralleled it with a 1n5817 Schottky diode for fast turn-off.
The choice of MOSFETs is important. Look at the STW45NM50FD, it
is a top performer at higher voltages.

Yes, but way too big for my 500W 600kHz bridge. In fact I went
down a factor-of-two step when I realized I had much less Ron
loss than switching losses. Compare the gate capacitances for
my FQP33N10 (1150pF) with your STW45NM50 (3600pF). OTOH, I was
operating at 70V max, rather than 360V, etc.
There are configurations that attain soft switching for all line
load ranges that perform nicely.

Yes, the ucc3895 has an adaptive delay for this purpose, but I
didn't use it. What methods have you used?

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)
 
H

Harry Dellamano

Jan 1, 1970
0
Winfield Hill said:
Harry Dellamano wrote...

No, of course not!


I had some difficulty, which I solved with snubbers and paralleled
mbr10100 10A 100V Schottky diodes (after blowing out a few FETs).
A major issue to prevent excessive damaging spikes was completely
eliminating shoot-through. Turning the FETs completely off rapidly
was an issue. I used a 10-ohm gate resistor to slow down turn-on,
but paralleled it with a 1n5817 Schottky diode for fast turn-off.


Yes, but way too big for my 500W 600kHz bridge. In fact I went
down a factor-of-two step when I realized I had much less Ron
loss than switching losses. Compare the gate capacitances for
my FQP33N10 (1150pF) with your STW45NM50 (3600pF). OTOH, I was
operating at 70V max, rather than 360V, etc.


Yes, the ucc3895 has an adaptive delay for this purpose, but I
didn't use it. What methods have you used?

Thanks,
- Win
As you know, gate drive and dead time is the key. Take a look at the
ZETEX's ZX6TDB2E6 and some of their newer versions coming out. Rsat<100m
and fast. They make great gate buffer drivers, able to push/pull 6 amps
<30nS and also protect the gate driver IC.
When ZVS, the recovery time of the parasitic diode is not an issues but
when chaos breaks out there is not a full half cycle to recover and shoot-
through currents thru the ceiling. That is why you use the parallel Schottky
diode and I use the SWT45NM50FD. The FD denotes Fast Diode recovery of the
parasitic diode and parallel Schottkys are not necessary. Must be some 100V
FD FETs someplace.
I am working on a ZVS PSFB topology that allows soft switchings for all
line and loads. Sure looks great in SPICE. Bench testing later in the week.
Cheers
Harry
 
Y

Yzordderex

Jan 1, 1970
0
Sounds like some impressive work all the way around! I built a phase
shift controller to run a Class F push-pull rf deck. I think I went
over this tho. The deck runs 30vdc for a 75w carrier, and 60v for
modulation peaks of 300w.

I am using the UC3875 part. I know it's limited in it's performance,
but works quite nice in the application. It has built in gate drivers
which seem to work well. I used voltage mode control, as the load is
well defined. I used a set of irfi644 parts in a plastic package. I
wanted small gates to drive. I modulated the delay pins with the
audio, and achieved zvs all up and down the envelope.

I suspect it's going to be king modulator for a long time, and can't
really think of an easier way to do it at this point. Till now most
have been using a monster bank of caps and running open loop at 80kHz.
A multipole filter on the output to get rid of the switching trash.
(FCC only wants you transmitting on one freq at a time.) My friend is
just bored and needs something to do. Maybe I'll let him in on one of
my new projects.

regards,
bob
 
W

Winfield Hill

Jan 1, 1970
0
Harry Dellamano wrote ...
As you know, gate drive and dead time is the key. Take a look
at ZETEX's ZX6TDB2E6 and some of their newer versions coming out.
Rsat <100m and fast. They make great gate buffer drivers, able
to push/pull 6 amps <30nS and also protect the gate driver IC.

Dinky little SOT23 transistors? Advanced info only... Hmm.
When ZVS, the recovery time of the parasitic diode is not an issues
but when chaos breaks out there is not a full half cycle to recover
and shoot-through currents thru the ceiling. That is why you use the
parallel Schottky diode and I use the SWT45NM50FD.

You mean the STW45NM50FD, right?
The FD denotes Fast Diode recovery of the parasitic diode and parallel
Schottkys are not necessary. Must be some 100V FD FETs someplace.

Yes, perhaps from ST.
I am working on a ZVS PSFB topology that allows soft switchings for
all line and loads. Sure looks great in SPICE. Bench testing later
in the week.

That sounds like fun!

Thanks,
- Win

(email: use hill_at_rowland-dot-org for now)
 
Top