Those GaN parts are interesting, if you can solder them down. They are not avalanche rated, so may be easy to blow out. They'd make good mosfet gate drivers!
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
DEI/IXYS makes true Kelvin connected mosfets. Expensive.
-- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Phil Hobbs wrote in news: snipped-for-privacy@electrooptical.net:
25 ohm is the parallel of 50 ohm terminator at the gate and 50 ohm cable. You can not do otherwise for test.In general, this resistor limit the amount of charge going into the gate (joule limiter,) so if one gets faster response with bigger resistor, that is better device.
Mass.
--- news://freenews.netfront.net/ - complaints: snipped-for-privacy@netfront.net ---
Unfortunately their support is quite lousy. My experience with domestic mfgs is much better. It even was while I still lived in Europe.
But they don't seem to have much in the isolated TO220 Full Pak package. It's a hi-rel app where this is quite important.
-- Regards, Joerg http://www.analogconsultants.com/
Oh I know it'll do the job. But in the more regulated fields (med, aero, space and such) one should be able to vet the whole design against formal data and I can't do that if the turn-on rise time max is formally only listed at a whopping 330nsec. In modern switcher electronics that's enough time to grow cobb webs.
-- Regards, Joerg http://www.analogconsultants.com/
faster when driven hard. But even with 25ohms driving resistance which any sane engineer would never do it should be faster than this. I tend to driv e with zero additional gate resistance, pedal to the metal. Why is such unf avorable data published? I need them in a class D driver and want to nail t he dead time well before turning it on.
r a decade, faster than the data sheet numbers. Just drive the gate hard. I 've switched 50 volts, 1 amp, with 2N7002s, in under 1 ns, well over 20x fa ster than the data sheet suggests.
ng smaller fets can help there.
h is not often specified. Some are fractions of an ohm, some are multiple o hms. The kilovolt SiC fets look great until you see the Rg numbers.
re I'll need about 100 amps of gate drive. Somewhere in the 2 ns range, pea k gate current equals switched drain current!
ld be better, of course. Thing is, I need to keep the bottom FET in a class D amp away from where its body diode comes on and that's why I need an Rds on in the 50mOhm range.
ate charge. The EPC2010 (digikey it) don't even say anything about timings . But with 500pf cap and 5nC gate charge, i am sure you can switch it in l ess than 10ns.
t avalanche rated, so may be easy to blow out. They'd make good mosfet gate drivers!
I'll keep them well below 1/2 Voltage and 1/2 current. I am thinking about die-bonding a bunch of them in TO-220 THIN (T3) package as show in this bo ard
http://173.224.223.62/motor
BTW, the PCB test pattern in this board is double width, my mistake.
will 150V, 15A do?
-Lasse
There's nothing preventing anyone from using a beefier pulser and a voltage divider.
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
Unfortunately not in this case. But it shows that they can also provide meaningful data. I sometimes wonder who really reviews this stuff before a datasheet gets released. As a reviever 310nsec would have made my neck hair stand up.
-- Regards, Joerg http://www.analogconsultants.com/
Well, I'll use the IRFI4227 and brute-force the drive. It's the only device left with a SPICE model as well as a clean good datasheet. A
10-amp gate drive should be enough of a hammer.-- Regards, Joerg http://www.analogconsultants.com/
Exactly, why aren't they aware of this very negative aspect of their business strategy.
Mikek
My guess is the usual, key-account thinking. But it's short-sighted because the little guys of today would have been the key accounts 10 years from now. Once alienated it is very difficult to win back customers, especially engineers.
-- Regards, Joerg http://www.analogconsultants.com/
Well, don't have more than about 2 ohms of gate resistance. Less is more.
Fairchild tends to do that, for some reason.
(If your design team isn't smart enough to read the part where it clearly says "R_G = 25 ohm", you should consider getting a new design team. XD )
I would recommend against the part because it doesn't specify R_G, whereas many devices do, nowadays. Unless you like batch characterization, you don't know how fast it can / will go...
Tim
-- Seven Transistor Labs Electrical Engineering Consultation Website: http://seventransistorlabs.com
Why?
I am the design team. I let the big Labrador look at it as well and he didn't understand either why they spec with 25ohms. That's like spec'ing a sports car saying it'll accelerate from 0-60mph in 30 seconds. Underground: Mud.
The internal gate resistance is not 25ohms. I have used this FET before and if driven with gusto it switches almost 10 times as fast as the datasheet say it does.
It's designed-out already. Also because Fairchild did not care to answer in due course.
-- Regards, Joerg http://www.analogconsultants.com/
Joerg
You may not believe it, but there are still people out there looking for "nice and slow" FETs "because it will pass EMC" that way.
Imagine a small almost-forgotten engineering department at a big multinational company designing an obscure industrial electronics "thing". The "thing" needs a power supply, and the design methods of said department are from aeons past, when men were real men and MOSFETs did not break the sonic wall of one microsecond. Imagine that said department designs SMPS by the tried and true re-use method, and gets the red card from the EMC test lab every time they change their SMPS because of parts becoming obsolete. The Dark Voodoo approaches that were used by the Wise Engineers Of Old who had designed The Father Of All SMPS magically no longer work with modern fast transistors. Unfortunately no other design methods can or may exist in their world (department) except those listed in a data sheet or appnote, and careers depend on things passing EMC, so everybody is desperate to somehow make the miracle happen and tame that SMPS, whose refusal to be tamed is growing year by year and decade by decade, like the resistance in an Oppressed Province around 29.64N,91.14E.
Now imagine that it's not only one department working that way, but a highly secret number approaching the mathematical set theory definition of "almost all", namely "all except a finite subset", add to this an overly formal design review process (which you surely will know from experience) throw in some of the pressure that the big companies can exert over their suppliers and Voila -> A Special Datasheet!
Greetings from your former Heimat Dimitrij
P.S. Legal note for the humor impaired: This message is strictly confidential. It is intended for use only by the intended recipient. Unless you can make head or tail of this passage, you are not the intended recipient. Unauthorized use is strictly prohibited and may be unlawful. If you received this message in error, you must not make use of the information contained herein and you must delete it immediately.
I just read datasheets, I don't work there...
Alright, then it has at least R_g < 2.5 ohm or thereabouts. If that's not good enough, then disregard and look at the next.
A better question is why MOSFETs have times listed at all. There's no semiconductor-physics reason to rate them as some speed, it's all about R and C. (Well, maybe a carrier drift term, but I'd be surprised if that mattered outside of hypothetical 4000V, GHz band RF transistors -- assuming such could even be made.)
All that should matter is Qg(tot) and R_g. If they don't give those, it doesn't count.
Because... after all... in your particular application, what are the chances that you'll have exactly the same setup their test did? Let alone if you use oddball drive voltages, or networks (turn-on / turn-off diode?). Specifying a time is useless and redundant!
What is (or should be) far more irritating is datasheets which claim none at all. I read a TI NextFET datasheet which made the laughable claim "R_G = 0". Inquiry yielded no clarification, of course...
In contrast, BJTs and PN junction diodes have clear limits on switching speed; there's still a lot of play in those parameters, and spooky physics things like charge diffusion and step recovery, but in general and for conventional setups (e.g., not trying to 10x overdrive the base for a few nanoseconds, or whatever), you aren't getting much faster than what the datasheet says (the main exception being t_stg, which is almost always excessive, because the saturation hFE chosen is almost always excessively low for the device).
Tim
-- Seven Transistor Labs Electrical Engineering Consultation Website: http://seventransistorlabs.com
which
tend
cable.
Really? Using absolutely inappropriate test setups that do NOT reflect normal usage, e.g. using a 1 A gate driver chip, is supposed to inspire confidence in the manufacturer? The 1 A gate driver chip is much closer to realistic conditions.
Really?
?-)
1 amp is actually at the low end of the world of mosfet gate drivers. There are lots of 6 and 10 amp parts, and a few real beasts, like IXRFD630.
LM5112 is a cute part, useful for other things than gate driving, too.
-- John Larkin Highland Technology, Inc jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
Ok, but this one is many times faster than the datasheet says when driven by a generator with much smaller R-out than 25 ohms. So if people hang it onto a run-of-the-mills modern switcher chip, which most will do, they'd be in for a serious surprise if they expected something slow.
That's why the world has consultants 8-)
Still makes no sense because most engineers will hang this FET to a regular modern switcher chip and it'll be fast. They all are unless their inner design or the foundry process are totally screwed up.
Did you just move there? Which area? I lived in the foothills east of Cologne before moving to the US. But originally I am from Westphalia.
And do it before the goons come :-)
-- Regards, Joerg http://www.analogconsultants.com/
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.