MOSFET answer

[mc]

Answering my own question about a TO-220 MOSFET that switches on solidly
with 2.7 volts on the gate, the answer appears to be IRF3708.

 

[Joerg]

Answering my own question about a TO-220 MOSFET that switches on solidly
with 2.7 volts on the gate, the answer appears to be IRF3708.

The spec on those is for 2.8V and at 7.5A.

 

[Spehro Pefhany]

Answering my own question about a TO-220 MOSFET that switches on solidly
with 2.7 volts on the gate, the answer appears to be IRF3708.

This stuff is quite difficult to search. MOSFETs are one of those
components that's just exploding in options, and the parametric
searches often give you Rds(on) without telling you the Vgs for that
parameter. Many of the most attractive newer options (especially the
newer packages) turn out to be will-o'-the-wisp fata morgana parts
which are unavailable in small quantities.


Best regards,
Spehro Pefhany

 

[Joerg]

Hello Spehro,

This stuff is quite difficult to search. MOSFETs are one of those
components that's just exploding in options, and the parametric
searches often give you Rds(on) without telling you the Vgs for that
parameter. Many of the most attractive newer options (especially the
newer packages) turn out to be will-o'-the-wisp fata morgana parts
which are unavailable in small quantities.

It's also important to rely on guaranteed data and for Rdson at a
certain Vgs that would be the "max" column. I have seen many designs
where people just took the graph. Worked fine initially. Then another
batch arrived and it was like popcorn.

Quite a few devices that are marketed as logic level are too marginal
for my taste.

 

[mc]

It's also important to rely on guaranteed data and for Rdson at a certain

Vgs that would be the "max" column. I have seen many designs where people
just took the graph. Worked fine initially. Then another batch arrived and
it was like popcorn.

Quite a few devices that are marketed as logic level are too marginal for
my taste.

Exactly! That was my concern.

 

[mc]

IRF3708 is specified as Rds(on) max 29 milliohms with 2.8 V on the gate. I
think that will do, for driving from 3-volt microcontrollers and the like.

 

[Spehro Pefhany]

Hello Spehro,


It's also important to rely on guaranteed data and for Rdson at a
certain Vgs that would be the "max" column. I have seen many designs
where people just took the graph. Worked fine initially. Then another
batch arrived and it was like popcorn.

Of course. Fortunately there are more and more that are guaranteed to
have a certain Rds at 2.5V or even 1.8V. Unfortunately, that usually
goes along with a low breakdown voltage (I've seen as low as 8V, so
the D-S might break down before the gate). Still, the latter provides
a clue for searching, which will help until the people designing the
parametric searches get a clue and make it possible to search for what
you need.

Quite a few devices that are marketed as logic level are too marginal
for my taste.

4.5V ratings are usually safe enough with a 5V supply, assuming a
solid CMOS drive, but not with a 3.3V supply, or with BJT emitter
followers ( a resistor from (CMOS) input to output with the latter can
help with slow switching).


Best regards,
Spehro Pefhany

 

[Joerg]

mc wrote:

IRF3708 is specified as Rds(on) max 29 milliohms with 2.8 V on the gate. I
think that will do, for driving from 3-volt microcontrollers and the like.

Just keep in mind that you have to slosh around Cgd in excess of 2000pF
with devices like this. Assuming that the Rdson of the high device on a
modern uC port can be above 25ohms it can take a long time to drag the
FET gate through the last few hundred mV up there. It'll be at the flat
portion of the RC curve. IOW the area where you normally would want to
be out of the linear range of a FET but might not be in this case.

Some uCs don't like driving capacitive loads of that magnitude.