Any T092 package MOSFET to replace IRLML6402 ?

On a lot of projects I often use an IRLML6402 P Channel MOSFET to switch VCC to devices, so as to turn them on and off.

The IRLML6402 has an RDS(on) of 0.08 ohm typical, a gate threshold of 0.55v and a drain current of 2A or so. It works well in 3.3V powered projects. Its a surface mount device.

Does anyone know of a TO92 packaged MOSFET that would do as a replacement, it only needs to handle 250mA or so.

A parameter search at Farnell\Element14 was not successful.

Modern MOSFETs are unlikely to be available in TO92, that’s an antique package, everything is surface mount or a large heatsinkable package like TO220, TO247. Modern electronics production is optimized for surface mount so new components are surface mount.

Packages like SOIC8 and DPAK and D2PAK are the easiest SMT MOSFET packages for hand-soldering.

Yeah. MarkT is right! I did a search on Mouser and found only 31 TO92 P-Channel MOSFETs, and none of them even come close to the IRLML6402. And, none of them quite reach the -250mA @ VG -3.3V characteristic, either.

I found this IPAK that seems do the trick: SFT1342-W

It should work quite well at -3.3V for -250mA on the Drain. Unless something like PWM [or other fast switching] is needed – in which case you might need a Gate Driver [since this part has greater Gate Capacitance than the IRLML6402].

ReverseEMF:
Yeah. MarkT is right! I did a search on Mouser and found only 31 TO92 P-Channel MOSFETs, and none of them even come close to the IRLML6402. And, none of them quite reach the -250mA @ VG -3.3V characteristic, either.

I found this IPAK that seems do the trick: SFT1342-W

It should work quite well at -3.3V for -250mA on the Drain. Unless something like PWM [or other fast switching] is needed – in which case you might need a Gate Driver [since this part has greater Gate Capacitance than the IRLML6402].

Nice package, max for gate threshold is 2.8v, a bit close maybe. But Farnell have them so may try one.

MarkT:
Modern MOSFETs are unlikely to be available in TO92, that's an antique package.

Sure its antique, but its for a project that uses a T092 regulator (MCP1702) a TO92 MOSFET (2N7000) a TO92 temperture sensor (DS18B20) and a 28pin DIP micro (ATMega328P). The project is meant to be easy to easy to solder.

I have used a 8pin SOIC in the past thats the possibilty I guess, know of any low cost options ?

There are lots of PNP transistors (e.g. PN2907 - max 600 mA collector current) that do fit the bill for current rating and package type. They're cheap, too.

It's very hard to find MOSFETs that open properly at 3.3V in THT package. The only ones that I know that open fully at 3.3V are SMD types, SOT-23 even. Not easy at all to hand solder.

I’m planning to use a Microchip LP0701 to switch power from an 18650 to a Wemos D1 Mini. It comes in a TO-92, but it has a couple ohms RDS(on). I’m just not sure how well it will handle the high-current WiFi activity, but will be testing it when the Minis arrive.

ShermanP:
I'm planning to use a Microchip LP0701 to switch power from an 18650 to a Wemos D1 Mini. It comes in a TO-92, but it has a couple ohms RDS(on). I'm just not sure how well it will handle the high-current WiFi activity, but will be testing it when the Minis arrive.

2 ohm is too high an on resistance really, at 250mA you would be loosing 0.5V.

I know that an alternative is an SOIC8 SI4459ADY which I have used in the past, that has an on resistance of 0.0039ohm. They are not cheap from Farnell\Element14 at £1.80 each, although Aliexpress has someone advertising them at 22p, but whether they are genuine parts who knows.

I dont myself have a problem with soldering the IRLML6402 SOT-23 devices, but I wanted to create a board that has easy to solder parts for those not familiar with SMT stuff.

srnet:
2 ohm is too high an on resistance really, at 250mA you would be loosing 0.5V.

Just a suggestion; MOSFETs work very well in parallel. If it makes economical sense to use two, or even three, LP0701 MOSFETs in parallel, that might be your solution.

srnet:
although Aliexpress has someone advertising them at 22p, but whether they are genuine parts who knows.

Fair chance it's a mainland-made clone. I don't shop normally at Aliexpress, they're too expensive. Usually go to Taobao. There you can often find the same part (by number/outlook) at vastly different prices, where the cheap ones are often marked "domestic" and the more expensive ones "imported".
I've used some mainland clones of various parts, generally works for me but I don't know how well they match the original's specifications.
By the way, I see the SI4459ADY for USD 1.45 per piece on Digikey, same price on RS Components, that's about GBP 1.13.

srnet:
Nice package, max for gate threshold is 2.8v, a bit close maybe. But Farnell have them so may try one.

2.6V threshold, not 2.8. That's the off voltage - its definitely not going to be usable at 3.3V if it might
be totally off at 2.6V.

That mosfet, SFT1342, has its on-resistance rated down to 4.0V gate-source. Not below that, so good
for 5V logic, not for 3.3V logic.

You are looking at "typical" device graphs, but that's specious alas - actual devices have a large spread of threshold voltages (over a volt or more), so the only figure on the datasheet you can rely on is the Rds(on) at a particular Vgs.

Of course you can select a device based on measured performance, but you still have the problem that MOSFET threshold voltages drift with age and applied voltage.

Just trust the Rds(on) specification, its a guarantee from the manufacturer, not a "typical device" measurement.

DIY TO92

Maybe make 100 or so small PCBs to solder the SMD onto.

However, you are probably making new designs so just use the footprint of the SMD :wink:

OR

Solder wires to the SMD pads, cover the SMD with UV glue, expose . . . voilà.

Some larger SMT packages can be soldered onto stripboard easily.
SOIC8 MOSFETs can have the gate pin carefully bent up (not too far, its fragile), then the other pins
soldered down to stripboard and flying wire wrapped round gate pin and soldered.

For pities sake that's the Vds = -10V graph, completely irrelevant.

MarkT:
For heavens sake that's the Vds = -10V graph, irrelevant to switching.

MarkT:
For pities sake that's the Vds = -10V graph, completely irrelevant.

:stuck_out_tongue: :-[ :stuck_out_tongue: :-[ :stuck_out_tongue: :-[ :stuck_out_tongue: :-[ :stuck_out_tongue: :-[ :stuck_out_tongue: :-[ :stuck_out_tongue: :-[ :stuck_out_tongue: :-[ :stuck_out_tongue: :-[ :stuck_out_tongue: :-[ :stuck_out_tongue: :-[ :stuck_out_tongue: :-[ :stuck_out_tongue: :-[

Well, then -- how about this:

And, yes, once again, I know this is the Typical case, but even if the VGS lines were moved down, such that 2.6V was right on the ID = 0 line, the -3.3V line would still be well above 250mA! And the channel voltage, at VGS = -3.3V, would be well under -0.5V -- more like -0.2V or below!

And, yes, I agree, for a production run this is risky thinking -- any characteristic not clearly specified by the Datasheet is outside the manufacturer support system. But for a one-off home project? This is hobbyist manna, man!

ShermanP:
I'm planning to use a Microchip LP0701 to switch power from an 18650 to a Wemos D1 Mini. It comes in a TO-92, but it has a couple ohms RDS(on). I'm just not sure how well it will handle the high-current WiFi activity, but will be testing it when the Minis arrive.

I just wanted to confirm that the LP0701 is NOT good for switching power to a Wemos D1 Mini. I rigged up my battery-powered scope to measure the voltage drop from source to drain when the mosfet is on, and got spikes up to about 0.4V during wifi transmission. Using a charged 18650, the transmission still worked, but I think the battery would have to be recharged more frequently. Just too much RDSon.

I ended up using an SSM3j328r (SOT23F), and the spikes disappeared. I never found anything in a TO92 to match that performance. But that "F" in SOT23F is for "flat", which is very short stubs that stick out along the bottom. Not much area for the solder to grip. I think the IRLML6401 has normal SOT23 legs. That's the 12V version of the 6402, with moderately improved gate threshold and RDS numbers vs the 6402, and pretty much equivalent to the SSM part.

I decided to hedge my bets a little on this, I have done a PCB that has the option of using either a SI4459ADY or a IRLML6401 based power switch.

The SI4459ADY is the easier to solder option, I have some coming from the far East at 13p each, we shall see how good they are.

srnet:
I decided to hedge my bets a little on this, I have done a PCB that has the option of using either a SI4459ADY or a IRLML6401 based power switch.

The SI4459ADY is the easier to solder option, I have some coming from the far East at 13p each, we shall see how good they are.

The SI4459ADY has a bit higher gate threshhold than I'd like, particularly for supply voltages down to about 3.5V. But soldering would certainly be easier. Except for the gate, you can even bridge the pins with solder. But I do wonder about ordering this kind of thing from the far East at 13p, particularly when Digikey sells them for $1.45. In the US I tend to resort to Digikey when it matters if it's genuine. But in this case, I guess you can test and measure whether the RDS is right.