Opinions please -- 200V, 70mA from 3.2V

Completely reasonable. There's no reason you shouldn't be able to, and it saves considerable nuisance.

(I've done both, and COTS makes perfect sense in most cases.)

Cheers, James Arthur

Feeling evil, I hooked a 1n5817 from M1's drain to the LT3757A's IntVcc, bootstrapping the LT3757A's power supply off said snubber. Works, supplies the LT3757A with ~+9V, boosting its p-p gate-drive accordingly. No drivers needed. (I don't trust the simulation though--it's acting weird all over).

You could do that, if you really, really can't stand burning a couple percent in a passive snubber.

Grins, James Arthur

I can't find who mentioned it, but I had my coupling constant wrong -- I actually did the math instead of just applying all the correct operations in a random order and came up with a coupling coefficient of 0.985 -- and my efficiency is now 92%. I'm still going to play with FET selection.

--

Tim Wescott 
Wescott Design Services 
http://www.wescottdesign.com

Guilty.

92% is getting decent. Getting rid of those runt pulses should be high on the efficiency-quest list.

Cheers, James Arthur

Transformers have a lot more adjustable variables than regular components, and finding an off-the-shelf transformer that does your job is rare.

Commercial off the shelf does make sense in some cases - mostly when you are doing something that everybody else is doing, which usually means that you should be buying in a commercial unit.

Doing it because you are scared silly of transformers is less than sensible.

--
Bill Sloman, Sydney

They went away with the (hopefully) more accurate coupling constant.

--

Tim Wescott 
Wescott Design Services 
http://www.wescottdesign.com

Then maybe Winfield's idea is the better one here?

Then it buys some more college time for the kids :-)

It looked a little like rollercoaster when I ran it. Just MHO, I would not want the FET transition time outside the LEB window of a switcher IC.

I don't know of any app notes but it's essentially just a PNP/NPN follower pair with the bases and emitters tied together. Zetex/Diodes offers good ones in SOT packages, cheap. The a resistor of a few ten ohms from bases to emitters so it can reach the DC levels. The emitter followers take over the "arm wrestling" part of the job in cases where a switcher IC is too weak. But the LT3757 normally does not need that.

--
Regards, Joerg 

http://www.analogconsultants.com/

I can't run that one without massaging because of missing symbols and stuff. But it's funny, a more digital oriented engineer friend calls SEPICs "septics" all the time.

BTW, Art of Electronic got a nice endorsement in the Dec-2015 issue of IEEE Spectrum.

--
Regards, Joerg 

http://www.analogconsultants.com/

"Memorial"? Has Vlad passed away? When I saw the reference to his circuit, I was thinking I hadn't seen a post from him in a while.

George H.

Yes, suddenly, the middle of last year. I tried Googling for an obit but did not find one -- he lived in the Kansas City area so you may find something if you dig.

--
www.wescottdesign.com

That is very sad. I believe Vladimir was quite young.

:-(

--
Regards, Joerg 

http://www.analogconsultants.com/

Weird. It's all self-contained--I'm using only built-in symbols:

SYMBOL Misc\\2N7002 SYMBOL bv SYMBOL cap SYMBOL diode SYMBOL ind2 SYMBOL nmos SYMBOL res SYMBOL schottky SYMBOL voltage

The only model that doesn't come with LTspice is the bigger FET, and that model is in the schematic. What does it complain about exactly?

Took them awhile!

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

On 2016-01-25 10:30, Phil Hobbs wrote:

The normal version doesn't have a 2N7002 symbol but that can be replaced. Then it complains about a missing RFU01SM4S model. Followed by a message crescendo at the end :-)

Circuit: * E:\Simulations\JSC\Simulations\Hill\Antiseptic_Converter.asc

Error on line 9 : d5 catch n005 rfu01sm4s Unable to find definition of model "rfu01sm4s" Error on line 13 : d2 catch n001 rfu01sm4s Unable to find definition of model "rfu01sm4s" Error on line 14 : d3 n001 +hv rfu01sm4s Unable to find definition of model "rfu01sm4s" Error on line 15 : d4 n005 +hv rfu01sm4s Unable to find definition of model "rfu01sm4s" Instance "m:m1:_u1:2": Length shorter than recommended for a level 3 MOSFET. Instance "m:m1:_u1:1": Length shorter than recommended for a level 3 MOSFET. Direct Newton iteration for .op point succeeded. .step ccoup=1e-008 Heightened Def Con from 4.91833e-011 to 4.04918e-009 Heightened Def Con from 4.47735e-009 to 1.04774e-008 Heightened Def Con from 1.32865e-008 to 1.72865e-008 Heightened Def Con from 2.00317e-008 to 2.30317e-008 Heightened Def Con from 2.60201e-008 to 2.90201e-008 Heightened Def Con from 2.90336e-008 to 3.20336e-008 Heightened Def Con from 3.50768e-008 to 4.10768e-008 Heightened Def Con from 4.47963e-008 to 4.77963e-008 Heightened Def Con from 5.18043e-008 to 5.68043e-008 Heightened Def Con from 5.85142e-008 to 6.15142e-008 Heightened Def Con from 6.54126e-008 to 6.84126e-008 Heightened Def Con from 7.33175e-008 to 7.73175e-008 Heightened Def Con from 7.88568e-008 to 8.18568e-008 Heightened Def Con from 8.31068e-008 to 8.61068e-008 Heightened Def Con from 9.85817e-008 to 1.02582e-007 Heightened Def Con from 1.05941e-007 to 1.09941e-007 Heightened Def Con from 1.16484e-007 to 1.19484e-007 Heightened Def Con from 1.2414e-007 to 1.2714e-007 Heightened Def Con from 1.30467e-007 to 1.34467e-007 Heightened Def Con from 1.40745e-007 to 1.43745e-007 Heightened Def Con from 1.4702e-007 to 1.5002e-007 Heightened Def Con from 1.52598e-007 to 1.55598e-007 Heightened Def Con from 1.56716e-007 to 1.59716e-007 Heightened Def Con from 1.67573e-007 to 1.71573e-007 Heightened Def Con from 1.75446e-007 to 1.78446e-007 Heightened Def Con from 1.79127e-007 to 1.82127e-007 Heightened Def Con from 1.87946e-007 to 1.90946e-007 Heightened Def Con from 1.95159e-007 to 1.98159e-007 Heightened Def Con from 1.98228e-007 to 2.01228e-007 Heightened Def Con from 2.04232e-007 to 2.07232e-007 Heightened Def Con from 2.11329e-007 to 2.14329e-007 Heightened Def Con from 2.14395e-007 to 2.17395e-007 Heightened Def Con from 2.21231e-007 to 2.23231e-007 Heightened Def Con from 8.07424e-007 to 8.10424e-007 Heightened Def Con from 8.14292e-007 to 8.17292e-007 Heightened Def Con from 8.22614e-007 to 8.25614e-007 Heightened Def Con from 8.33907e-007 to 8.36907e-007 Heightened Def Con from 8.41105e-007 to 8.44105e-007 Heightened Def Con from 8.46885e-007 to 8.49885e-007 Heightened Def Con from 8.52639e-007 to 8.55639e-007 Heightened Def Con from 8.57652e-007 to 8.60652e-007 Heightened Def Con from 8.64234e-007 to 8.68234e-007 Heightened Def Con from 8.7448e-007 to 8.7748e-007 Heightened Def Con from 8.80634e-007 to 8.83634e-007 Heightened Def Con from 8.91809e-007 to 8.94809e-007 Heightened Def Con from 9.28461e-007 to 9.31461e-007 Heightened Def Con from 9.46108e-007 to 9.50108e-007 Heightened Def Con from 9.81297e-007 to 9.84297e-007 Heightened Def Con from 1e-006 to 1e-006 Heightened Def Con from 1e-006 to 1.003e-006 Heightened Def Con from 1.02801e-006 to 1.03101e-006 Heightened Def Con from 1.03774e-006 to 1.04074e-006 Heightened Def Con from 1.04599e-006 to 1.04899e-006 Heightened Def Con from 1.05178e-006 to 1.05478e-006 Heightened Def Con from 1.05828e-006 to 1.06128e-006 Heightened Def Con from 1.06244e-006 to 1.06544e-006 Heightened Def Con from 1.07817e-006 to 1.08217e-006 Heightened Def Con from 1.08683e-006 to 1.08983e-006 Heightened Def Con from 1.0947e-006 to 1.0977e-006 Heightened Def Con from 1.09924e-006 to 1.1001e-006 Heightened Def Con from 1.10053e-006 to 1.10353e-006 Heightened Def Con from 1.10778e-006 to 1.11078e-006 Heightened Def Con from 1.11402e-006 to 1.11702e-006 Heightened Def Con from 1.12161e-006 to 1.12461e-006 Heightened Def Con from 1.12747e-006 to 1.13047e-006 Heightened Def Con from 1.13402e-006 to 1.13702e-006 Heightened Def Con from 1.13768e-006 to 1.14068e-006 Heightened Def Con from 1.14164e-006 to 1.14464e-006 Heightened Def Con from 5.10034e-006 to 5.10334e-006 Heightened Def Con from 5.13972e-006 to 5.14172e-006 Heightened Def Con from 5.1434e-006 to 5.1454e-006 Heightened Def Con from 5.14605e-006 to 5.14805e-006 Heightened Def Con from 5.47242e-006 to 5.47442e-006 Heightened Def Con from 7.043e-006 to 7.046e-006 Heightened Def Con from 7.16179e-006 to 7.16379e-006 Heightened Def Con from 7.17203e-006 to 7.17403e-006 Heightened Def Con from 7.31148e-006 to 7.31548e-006 Heightened Def Con from 7.33351e-006 to 7.33551e-006 Heightened Def Con from 7.4898e-006 to 7.4928e-006 Heightened Def Con from 7.50466e-006 to 7.50766e-006 Heightened Def Con from 7.55023e-006 to 7.55323e-006 Heightened Def Con from 7.60186e-006 to 7.60486e-006 Heightened Def Con from 7.92631e-006 to 7.93231e-006 Heightened Def Con from 8.04427e-006 to 8.04627e-006 Heightened Def Con from 1.35959e-005 to 1.35979e-005 Heightened Def Con from 1.94021e-005 to 1.94081e-005 Heightened Def Con from 1.94181e-005 to 1.94251e-005 Heightened Def Con from 1.9431e-005 to 1.9435e-005 Heightened Def Con from 1.94447e-005 to 1.94527e-005 Heightened Def Con from 1.9456e-005 to 1.946e-005 Heightened Def Con from 1.94632e-005 to 1.94702e-005 Heightened Def Con from 1.94771e-005 to 1.94881e-005 Heightened Def Con from 1.94989e-005 to 1.95069e-005 Heightened Def Con from 1.95127e-005 to 1.95187e-005 Heightened Def Con from 1.95547e-005 to 1.95577e-005 Heightened Def Con from 0.000326051 to 0.000326054 Heightened Def Con from 0.000328735 to 0.000328738 Heightened Def Con from 0.000328885 to 0.000328888 Heightened Def Con from 0.000331095 to 0.0003311 Heightened Def Con from 0.000333108 to 0.000333111 Heightened Def Con from 0.000333151 to 0.000333154 Heightened Def Con from 0.000334446 to 0.000334449 Heightened Def Con from 0.000335205 to 0.000335207 Heightened Def Con from 0.000335209 to 0.000335212 Heightened Def Con from 0.00033537 to 0.000335373 Heightened Def Con from 0.000335396 to 0.000335399 Heightened Def Con from 0.000335501 to 0.000335504 Heightened Def Con from 0.000335906 to 0.000335909 .step ccoup=3.16228e-008

[...]

--
Regards, Joerg 

http://www.analogconsultants.com/

Works fine for me (once I'd added a 2N7002 model).

Have you updated your LTspice lately? Those are in the bog-standard LT library on mine. I downloaded the posted one, and it runs fine here.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

Sigh, thanks for posting that Joerg.

George H.

Uh, unless I read it wrong, Winfield's idea was a two-stage boost.

At any rate, the circuit with various suggested mods, mostly from you and James Arthur, is now cooking along in simulation at a hair over 90% efficiency. I could maybe do better, but other parts of the circuit will be taking the ever-so-efficiently generated 200V and burning up parts of it in a linear pass regulator, so there's a point at which the pursuit of excellence becomes a bit absurd.

(BTW, the original "active" snubbing has been scrapped, and the snubber voltage is being used to power the gate drivers through a teeny SEPIC.)

--

Tim Wescott 
Wescott Design Services 
http://www.wescottdesign.com

I haven't updated in roughly a year now because almost every time I did in the past something became unglued. Why not use the regular NMOS symbol for things like the 2N7002?

--
Regards, Joerg 

http://www.analogconsultants.com/

Sorry, was a misunderstanding then. If one can deal with the losses in the first stage diode that idea should also work well.

90% is good, almost Royer territory. The nice thing about LTC parts is that the boards come out almost verbatim as simulated as long as your assumptions on the magnetics are correct. Plus core losses, of course, which are difficult to simulate. So in reality the efficiency is often a percentage point or two lower. Something to keep in mind for the thermal estimation.

One of the trademarks of switcher designers are square burn scars from SMT inductors.

Then you can possibly affix a "Green Power" sticker to the board :-)

--
Regards, Joerg 

http://www.analogconsultants.com/