Jelly bean N FETS

[Spehro Pefhany]

Now I wonder why they don't have SOA curves for a 20-V, 3.2-A MOSFET in
a SOT-23 package....probably because the curves wouldn't be
distinguishable from the Y axis...

Cheers,

Phil Hobbs

I just had one of those high-current SOT-23 suckers quit on a digital
photo frame... took out the switching power supply adapter too. It was
the switching element in LED backlight power supply. Wonder what the
most rugged SOT-23 MOSFET is that I can buy from Digikey.



Best regards,
Spehro Pefhany

 

[Martin Griffith]

http://en.wiktionary.org/wiki/Citations:embrighten

But it does sound rightpondian, I've never heard it in the US. And
Martin mentioned he'd already had some rioja

I wondered who would spoy that!
http://en.wikipedia.org/wiki/The_Simpsons#Influences_on_language


martin

 

[Martin Griffith]

Have you seen the the BSS138.

http://www.fairchildsemi.com/ds/BS/BSS138.pdf


Newark has the BSS138LT1G for 0.04 cents small quantity SOT-23.

Another possiblity may be the NTS4001.

10 cents from Newark sc-70

http://www.onsemi.com/pub_link/Collateral/NTS4001N-D.PDF

I have about fifty of each,easy drive requirments.

The 138 looks quite nice, thanks


martin

 

[Joerg]

I just had one of those high-current SOT-23 suckers quit on a digital
photo frame... took out the switching power supply adapter too. It was
the switching element in LED backlight power supply. Wonder what the
most rugged SOT-23 MOSFET is that I can buy from Digikey.

You'd have to put target numbers to the definition of "rugged". But mind
that something will then have to drive its largish Cgs.

 

[Joerg]

I wondered who would spoy that!
http://en.wikipedia.org/wiki/The_Simpsons#Influences_on_language

But they didn't say "enbrighten"

I remember that Perry Rhodan (Sci-Fi books) had a serious influence on
language in Germany when I was a kid.

 

[Spehro Pefhany]

You'd have to put target numbers to the definition of "rugged". But mind
that something will then have to drive its largish Cgs.

Good point-- the Taiwan-designed PWM chip (BIT3251) is really rather
wimpy-- 50 ohms output impedance (typical, no limits given), probably
running at 5V, and it operates at a fairly high frequency (330kHz). No
idea what the dead part is/was other than it's obviously an n-channel
SOT-23 MOSFET.

Suggestions? If 20V is enough, I like the looks of the Zetex
ZXMN2F30FHTA 4.9A, 0.045 ohms at 4.5V and only 4.8nC total gate
charge.

Best regards,
Spehro Pefhany

 

[Hammy]

The 138 looks quite nice, thanks


martin

I was curious on just what a tiny NTS4001 could take so I did some
test. Here is the test setup I did on the breadboard.

http://i33.tinypic.com/idvssp.png

There is about three inches of 22AWG between the TLC555 output and the
NTS4001 gate.

I ran Vcc between 2.5VDC to 6VDC, Vcc equals Vgs. I have no way to
adjust the ambient temp accurately but its about 25 degrees C. The
NTS4001 is mounted on one of my homemade proto adapters. L=0.75 cm X
0.25cm. The drain tab is on a 0.4cm x 0.4cm 1oz copper pour.

This is a non magnified view (scope capture) showing CH1= Vgs and
CH2=VDS. The pulse length is 112mS.

Vcc = 6VDC, IDS = 0.4A, VDS= 272mV. This was consistent even over 10
one shots at about ten second intervals.

http://i34.tinypic.com/r1n828.jpg

The rest of the scope captures are higher resolution zooms of VDS.
Magnified view 6VDC.

http://i36.tinypic.com/r8t4ph.jpg

For Vcc = 3VDC,VDS= 0.228mV,Ids= 0.18A

http://i36.tinypic.com/r8t4ph.jpg

For Vcc= 2.5Vdc , Vds= 292mV,Ids=0.147A

http://i34.tinypic.com/vy4cp1.jpg

Not to shabby for a FET I paid 0.05 cents for (During one of there
sales). Newark raised the price a bit to ten cents but still worth the
money. As with most single and dual mosfets and bipolars I find Newark
consistently is cheaper then Digikey most times significantly.

 

[Hammy]

I was curious on just what a tiny NTS4001 could take so I did some
test. Here is the test setup I did on the breadboard.

http://i33.tinypic.com/idvssp.png

There is about three inches of 22AWG between the TLC555 output and the
NTS4001 gate.

I ran Vcc between 2.5VDC to 6VDC, Vcc equals Vgs. I have no way to
adjust the ambient temp accurately but its about 25 degrees C. The
NTS4001 is mounted on one of my homemade proto adapters. L=0.75 cm X
0.25cm. The drain tab is on a 0.4cm x 0.4cm 1oz copper pour.

Whoops made a boo-boo the dimensions are L=0.75cm x W=0.6cm

 

[Tim Williams]

I wondered who would spoy that!

It's a perfectly cromulent word.

Tim

 

[Martin Griffith]

It's a perfectly cromulent word.

Tim

exxxcellent


martin

 

[Martin Griffith]

Whoops made a boo-boo the dimensions are L=0.75cm x W=0.6cm

Thanks mucho for that, I don't really use Digikey, since I'm based in
spain, only use Digikey for some eurounobtainium stuff like panasonic
jog wheels


martin

 

[Hammy]

Thanks mucho for that, I don't really use Digikey, since I'm based in
spain, only use Digikey for some eurounobtainium stuff like panasonic
jog wheels


martin

No problem

 

[Joerg]

Good point-- the Taiwan-designed PWM chip (BIT3251) is really rather
wimpy-- 50 ohms output impedance (typical, no limits given), probably
running at 5V, and it operates at a fairly high frequency (330kHz). No
idea what the dead part is/was other than it's obviously an n-channel
SOT-23 MOSFET.

Suggestions? If 20V is enough, I like the looks of the Zetex
ZXMN2F30FHTA 4.9A, 0.045 ohms at 4.5V and only 4.8nC total gate
charge.

I usually go by whatever Digikey has, preferably a five-digit qty of
stocked parts. Got to throw it on Spice to see what the switching losses
will be with a 50ohm driver. At 330kHz probably not a pretty picture.

 

[Joerg]

If you are going to use an external current limiting resistor, what the hell
difference does it make if your fet has a few ohms of Rds unless you can't
heat sink a watt and a half or so for a few milliseconds (and you really
didn't specify duty cycle)? There really are only a couple of choices ...
heat sink the fet or heat sink the cl resistor. Me? I'd look at a SMD
variant of the 2N7000 and see what the peak pulse current/power capability
is.

Looks like he wants to do short pulses of defined brightness. Then the
Rdson does matter. At low duty cycle it won't blow up.