RC snubber on flyback primary
in [Design]
|
Prev: 89C51ED2
Next: IR motion sensor
From: seegoon99 on 10 Oct 2008 08:26 On Oct 10, 8:21 am, Eeyore <rabbitsfriendsandrelati...(a)hotmail.com> wrote: > Jim Thompson wrote: > > Back almost 30 years ago, when I was designing off-line switchers, I > > used a Tek current probe along with a voltage probe to plot the > > operating SOA on the 'scope screen. Learned a lot about designing > > snubbers that way, and made a lot of high efficiency stuff. > > Lots of app notes around IIRC on snubber design too. > > Graham You may need a voltage clamp across your primary as well as a snubber. Try a diode in series with R//C. Diode Anode to Drain of fet and R//C to +supply. This will clamp the drain spike and stop your fet blowing. The r-c cct across your primary will help kill high frequency ringing , but does not help much with the leading edge spike. Which can be big!! I have attached a simple flyback type cct which demo's the "clamping" cct. After 300us the "clamp " is enabled. Cheers Rob. Version 4 SHEET 1 1128 692 WIRE -112 -16 -256 -16 WIRE -32 -16 -112 -16 WIRE 96 -16 -32 -16 WIRE 208 -16 96 -16 WIRE -32 16 -32 -16 WIRE 208 16 208 -16 WIRE -112 32 -112 -16 WIRE 400 32 336 32 WIRE 480 32 464 32 WIRE 688 32 480 32 WIRE 480 48 480 32 WIRE 688 48 688 32 WIRE -256 112 -256 64 WIRE 96 112 96 64 WIRE -112 128 -112 96 WIRE -80 128 -112 128 WIRE -32 128 -32 96 WIRE -32 128 -80 128 WIRE 336 128 336 112 WIRE 416 128 336 128 WIRE 480 128 480 112 WIRE 480 128 416 128 WIRE 688 128 688 112 WIRE 688 128 480 128 WIRE 416 144 416 128 WIRE 208 176 208 96 WIRE 208 176 96 176 WIRE -80 256 -80 128 WIRE -64 256 -80 256 WIRE 48 256 0 256 WIRE 208 256 208 176 WIRE 208 256 128 256 WIRE 64 304 48 304 WIRE 128 304 112 304 WIRE 48 336 48 304 WIRE 48 336 16 336 WIRE 208 336 208 256 WIRE 0 352 -192 352 WIRE 48 352 48 336 WIRE 128 352 128 304 WIRE 0 416 0 352 WIRE 160 416 0 416 WIRE 208 432 208 416 WIRE -192 448 -192 352 WIRE -192 640 -192 528 WIRE 208 640 208 432 WIRE 208 640 -192 640 WIRE 208 672 208 640 FLAG 208 672 0 FLAG 416 144 0 FLAG -256 112 0 FLAG 16 336 0 SYMBOL voltage -192 432 R0 WINDOW 123 0 0 Left 0 WINDOW 39 -142 141 Left 0 WINDOW 3 -137 170 Left 0 SYMATTR SpiceLine Rser=12 SYMATTR Value PULSE(0 12 0 10n 10n 3.5u 10u) SYMATTR InstName V1 SYMBOL ind2 192 0 R0 SYMATTR InstName L1 SYMATTR Value 1m SYMATTR Type ind SYMATTR SpiceLine Rser=1 Cpar=100p SYMBOL voltage -256 -32 R0 WINDOW 123 0 0 Left 0 WINDOW 39 24 132 Left 0 SYMATTR SpiceLine Rser=0.1 SYMATTR InstName V2 SYMATTR Value 320 SYMBOL ind2 352 128 R180 WINDOW 0 36 80 Left 0 WINDOW 3 36 40 Left 0 SYMATTR InstName L2 SYMATTR Value 50µ SYMATTR Type ind SYMATTR SpiceLine Rser=0.5 Cpar=30p SYMBOL cap 464 48 R0 SYMATTR InstName C5 SYMATTR Value 100µ SYMBOL res 672 16 R0 SYMATTR InstName R4 SYMATTR Value 10 SYMBOL cap -128 32 R0 SYMATTR InstName C1 SYMATTR Value 22n SYMBOL res -48 0 R0 SYMATTR InstName R2 SYMATTR Value 56k SYMBOL diode 0 240 R90 WINDOW 0 0 32 VBottom 0 WINDOW 3 32 32 VTop 0 SYMATTR InstName D1 SYMBOL res 80 -32 R0 SYMATTR InstName R1 SYMATTR Value 1k SYMBOL cap 80 112 R0 SYMATTR InstName C2 SYMATTR Value 100p SYMBOL diode 400 48 R270 WINDOW 0 32 32 VTop 0 WINDOW 3 0 32 VBottom 0 SYMATTR InstName D2 SYMBOL nmos 160 336 R0 SYMATTR InstName M1 SYMATTR Value STW11NM80 SYMBOL sw 32 256 R270 WINDOW 3 -42 120 Left 0 SYMATTR InstName S1 SYMBOL voltage 144 352 R90 WINDOW 0 -32 56 VBottom 0 WINDOW 3 32 56 VTop 0 WINDOW 123 0 0 Left 0 WINDOW 39 0 0 Left 0 SYMATTR InstName V3 SYMATTR Value PULSE(0 10 300u) TEXT 264 536 Left 0 !.tran 500u TEXT 264 -24 Left 0 !k l1 l2 0.999 TEXT -136 552 Left 0 !.ic V(Vsence = 0) TEXT 248 616 Left 0 !.model SW SW(Ron=0.01 Roff=10Meg Vt=0.6 Vh=1) TEXT 296 232 Left 0 ;Set C2 to 1pF effectivly take the branch of the circuit away.
From: Nico Coesel on 10 Oct 2008 13:14 Jamie Morken <jmorken(a)shaw.ca> wrote: >Hi, > >I am testing a flyback and trying to size the primary side RC snubber, >without the snubber my fet blows up, I started with a 100ohm 1206 >package resistor and a 47nF 1kV 1812 package capacitor, and now am using >a 100ohm resistor and 10nF 1kV 1206 package capacitor, this still smokes >a 1206 package 1/4 watt resistor. The flyback is low power, 10watts >maximum output, and I would like to dissipate as little power in the >snubber as possible for efficiency. Switching to a smaller snubber cap If this is your goal, you should attack the real problem. The snubber is necessary to dissipate the energy which is stored in the leakage induction in the transformer. If you can find (or have made) a transformer with a lower leakage induction, you'll dissipate less heat in the snubber or clamp circuit. >reduces the heat in the resistor, and also I noticed in an ltspice sim >that a smaller value R also decreases the heat in the resistor, so which >method is more efficient to snub voltages, dropping the capacitance >value or the resistance value for an RC snubber? You can also consider a zener diode to clamp the peak voltage to what the FET can handle. -- Programmeren in Almere? E-mail naar nico(a)nctdevpuntnl (punt=.)
From: Vladimir Vassilevsky on 10 Oct 2008 13:42 Nico Coesel wrote: > Jamie Morken <jmorken(a)shaw.ca> wrote: > > >>Hi, >> >>I am testing a flyback and trying to size the primary side RC snubber, >>without the snubber my fet blows up, I started with a 100ohm 1206 >>package resistor and a 47nF 1kV 1812 package capacitor, and now am using >>a 100ohm resistor and 10nF 1kV 1206 package capacitor, this still smokes >>a 1206 package 1/4 watt resistor. The flyback is low power, 10watts >>maximum output, and I would like to dissipate as little power in the >>snubber as possible for efficiency. Switching to a smaller snubber cap > > > If this is your goal, you should attack the real problem. The snubber > is necessary to dissipate the energy which is stored in the leakage > induction in the transformer. If you can find (or have made) a > transformer with a lower leakage induction, you'll dissipate less heat > in the snubber or clamp circuit. If the goal is the dissipation of the leakage inductance energy, then the solution could be just slowing down the turn off process of the FET. > You can also consider a zener diode to clamp the peak voltage to what > the FET can handle. Depending on the application, there could be the significant amount of power in the leakage inductance. Also, the zener should be able to handle peak current. For that matter, I prefer non-dissipative clamping which returns leakage power back to the supply rail. Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com
From: Eeyore on 10 Oct 2008 16:34 Nico Coesel wrote: > Jamie Morken <jmorken(a)shaw.ca> wrote: > > >Hi, > > > >I am testing a flyback and trying to size the primary side RC snubber, > >without the snubber my fet blows up, I started with a 100ohm 1206 > >package resistor and a 47nF 1kV 1812 package capacitor, and now am using > >a 100ohm resistor and 10nF 1kV 1206 package capacitor, this still smokes > >a 1206 package 1/4 watt resistor. The flyback is low power, 10watts > >maximum output, and I would like to dissipate as little power in the > >snubber as possible for efficiency. Switching to a smaller snubber cap > > If this is your goal, you should attack the real problem. The snubber > is necessary to dissipate the energy which is stored in the leakage > induction in the transformer. If you can find (or have made) a > transformer with a lower leakage induction, you'll dissipate less heat > in the snubber or clamp circuit. And knowing how to do that is an art/science in itself. Especially of it's running off the mains and you have to take safety margins into account. > >reduces the heat in the resistor, and also I noticed in an ltspice sim > >that a smaller value R also decreases the heat in the resistor, so which > >method is more efficient to snub voltages, dropping the capacitance > >value or the resistance value for an RC snubber? > > You can also consider a zener diode to clamp the peak voltage to what > the FET can handle. Crikey. You said almost exactly the same as me. But then again those are the facts. Graham
From: legg on 10 Oct 2008 17:59
On Thu, 09 Oct 2008 04:49:46 -0700, Jamie Morken <jmorken(a)shaw.ca> wrote: >Jamie Morken wrote: >> Hi, >> >> I am testing a flyback and trying to size the primary side RC snubber, >> without the snubber my fet blows up, I started with a 100ohm 1206 >> package resistor and a 47nF 1kV 1812 package capacitor, and now am using >> a 100ohm resistor and 10nF 1kV 1206 package capacitor, this still smokes >> a 1206 package 1/4 watt resistor. The flyback is low power, 10watts >> maximum output, and I would like to dissipate as little power in the >> snubber as possible for efficiency. Switching to a smaller snubber cap >> reduces the heat in the resistor, and also I noticed in an ltspice sim >> that a smaller value R also decreases the heat in the resistor, so which >> method is more efficient to snub voltages, dropping the capacitance >> value or the resistance value for an RC snubber? > > >Just a bit of an update, I removed the RC snubbers on the flybacks I am >testing, and they survived, so I think my original mosfet failures were >from an unrelated cause, I then put on a 100ohm/68pF RC snubber and >there is a 100mW extra draw on the bench supply, with a 2watt output >load. I'm not sure if this small of an RC snubber is very useful, but >it seems to be a good fit (ie. low power!) > Depending on what you are trying to snub, differing configurations of RC or RCD in current or voltage are applicable. It takes considerable abuse to pop fets from switching loss alone - they are not as SOA-sensitive as bipolars were (and still are). Power loss in the simplest RC snubber, with no ringing, is approximated as: C x V^2 x f / 2 for each specific voltage change in the waveform; all power within the waveform being cumulative when correlated in the same waveform. If all transformer leakage inductance energy was expected to be absorbed in a voltage clamp, the energy would be: Lleak x Ipk^2 x f / 2. Not all of the energy need necessarily be dissipative - as some is redirected to produce the dynamic waveform's edges. If a relatively slow Trr rectifier is used in the clamp, some of the energy may also be returned to the circuit during the clamp diode's reverse recovery, without being burned off in the clamp ( a 'damping' effect ). As suggested by Jim T, scoping the current and voltage waveforms is educational in identifying and reducing switch stress - and to determine effectiveness of any measures taken to alter the stress profile. Also keep an eyeball on input power consumption - this can be a useful relative indicator. Worse case is likely at the heaviest peak switch current turn-off condition, at some worst-case temperature and input voltage condition, though the addition of snubbing parts or a stray-capacitive-laden or rectifier-diode-lossy turn-on may contribute to turn-on losses in the profile. RL |