Comparing Mosfet Drivers

Well 710pF is wimpy for such big drivers, they'll almost be idling. With higher capacitance it's the Rdson of the output devices that matters. I was faced with the same choice last year and (IIRC) Fairchild didn't spec Rdson so I went with the 4421. Plus that has second source, I ended up using the Micrel version because it had more favorable pricing. Also, it comes in a 12A version if you really, really need some muscle.

It goes without saying that these things need excellent bypassing and supply planes.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Sounds good.

I'm just shooting for the least mosfet power switching loss while still using a driver in SOT23 package. f=900Khz

D from BC myrealaddress(at)comic(dot)com BC, Canada Posted to usenet sci.electronics.design

d.pdf

N3100C.pdf

Just thinking out loud here, the driver needs to drive it's own capacitance too. At some point, too light a load means the speed of the device itself is the limiting factor.

I never built any discrete switchers, but in chips, I always worry about a pin falling too fast. You get coupling into other pins and indirectly other parts of the chip. There was a famous screwed up chip one place where I worked that the switching pumped up the bandgap, on little step at a time until the power supply failed. Now can you switch too fast for a discrete design?

Ok, the 4421 doesn't come in SOT23. If it has to be you'll be kind of stuck with the FAN3100. The 4421 comes in DFN, also quite small but nasty to hand-solder.

900kHz with zippy transitions into 710pF is no big deal for either.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

driver

Not really but it can mess up the EMC.

--
Regards, Joerg

http://www.analogconsultants.com/

"gmail" domain blocked because of excessive spam.
Use another domain or send PM.

Thanks for the confirmation..

ha..funny that you should mention soldering. On my previous project I had to solder the 4421 DFN and that's why I'm trying to switch to something in a SOT23 to make it a little easier..

D from BC myrealaddress(at)comic(dot)com BC, Canada Posted to usenet sci.electronics.design

"D from BC" a écrit dans le message de news: snipped-for-privacy@4ax.com...

Strange things sometimes rules components selection...

I do DFN by hand all the time. It's easy. As usual the magic thing is flux. The only package I don't solder manually is BGA. Not yet :-)

Next board DFNs, 0201, 5 balls uBGA - 17 components + connections => 7mm diameter... YES!

-- Thanks, Fred.

LM5112 is cute. It allows different logic V- and output V-, which can be useful.

John

The Micrel part is MOS output, ie resistive. This determines peak current at any voltage.

The Fairchild part is bipolar output, so able to sink larger currents at lower voltages (ie at the 'off' threshold). This allows better external tailoring (diodes and resistors) of on and off drive speed. The external resistor can also serve to reduce drive IC dissipation.

RL

A little semiconductor knowledge is a dangerous thing ;-)

...Jim Thompson

--
| James E.Thompson, P.E.                           |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC\'s and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

Barack Obama, King of the Democratic People\'s Republic of America

The LM5112 has differential inputs like the FAN3100.

LM5112 has higher peak current specs than the FAN3100 giving me the impression that it has a lower output resistance.

page. 5 The LM5112 can charge 1nF a few nS quicker than the FAN3100.

I'm just a bit disappointed with the title 'Tiny 7A MOSFET Gate Driver' Wow! I'm assuming that 7A driver have less output resistance than a

2.5A driver.. Perhaps more than 1/2 the resistance for more than double the current for less than 1/2 the charge up time. Not! Not for 1nF anyways..

D from BC myrealaddress(at)comic(dot)com BC, Canada Posted to usenet sci.electronics.design

When driving mosfets, driver "resistance" doesn't mean much. What matters is how much current it can push and pull into the gate during the "Miller slew", which is the typically 5 or so volt plateau of gate voltage.

John

For push/pull stages, it also matters how fast the gate voltage will reach the "active" area on turn on or turn off. The optimization is quite tricky, as the driver outputs have rather nonlinear V/I characteristics, and the parasytics comes into play.

Vladimir Vassilevsky DSP and Mixed Signal Design Consultant

Just a heads up on the Fairchild part. I like their parts, but have found some of their high-side drivers (FAN7380 series) to fail when the switch node pulled the output slightly below the high side reference ground. We had to use the IR2110 instead, had no such problem. Could be the same output stage, but since we are talking low side driver, then you have full control of the ground.

Regards

Klaus

This isn't unique to Fairchild. The FC part probably isn't as robust as some others but it's also half the price or less. I think if you spring a nickel for a clamping diode and place your gate resistor in the Vs pin you might get better reliability.

See slup169 pg 25

Yes, that is the good old Teledyne TSC4421.

You probably won=92t find any differences in the application.

even LGA?

That's good info Klaus. Thanks.

James Arthur

420d.pdf

FFAN3100C.pdf

Yeah, I figured the spacing in a discrete design is so big it would be hard to have coupling. But I also thought an unguarded PCB trace (long and parallel) could be an issue.