Espruino JavaScript
Heads up! Only through-hole (TO-220) power MOSFET that switches with 3.3v on gate is being discontinued! #1875
Replies: 2 comments
-
|
Posted at 2020-07-20 by @gfwilliams Thanks! That's frustrating - maybe it's time to stockpile ;) Very surprising though. Hopefully someone will start selling a breakout for a surface mount variant. I've been using the ST P36NF06L as my go-to. That can pull 10A @ 15v with 3v gate voltage. It's not amazing, but has been good enough for most of the stuff I end up doing. |
Beta Was this translation helpful? Give feedback.
-
|
Posted at 2020-07-24 by DrAzzy Well funny you should mention that ;-) I do sell beefy SMD MOSFETs on breakout boards. Including with driver chips that not only put a supplied 5~10V on the gate with a low voltage signal, they can dump a slug of current into it to slam the gate open and minimize switching losses, so you can PWM stuff at high current and insane frequencies without overheating the FETs. https://www.tindie.com/products/drazzy/4-channel-logic-level-mosfet-driver-100a-up-to100v/ Even if you're willing to use SMD parts with really nasty (as in, don't even think of using them without a reflow oven nasty, because you can't solder them realistically with an iron, nor with hot air if you have enough 2 oz copper to help pull the heat away), the there just aren't that many decent power MOSFETs that can hold off more than 20V and switch at 2.5 (the numbers you want to use for 3~3.3v logic level signals). The IRFR3708 (also discontinued - was just a different package) was one of the real standouts on that list, and you could solder it with an iron! At 20V, there are a few good options in DPAK though, thankfully, but if you need to hold off 30v because you're using 24V devices, that doesn't help you. All the good 30V FETs that turned on with 2.5 on the gate and handled tens of amps are discontinued! Re: STP36NF06L They look like pretty lousy FETs at 3v... (like, well, basically every other through-hole MOSFET). I reckon you'd see a volt across drain and source with only 2-3A flowing, at which point you almost might as well use a BJT! On an IRF3708, even with 10A, you'd see around 0.15v between drain and source. Quick back of envelope suggests that without heroic cooling measures, you'd at best be able to get maybe 4A through them, probably less, with a significant portion of the power being dissipated in the FET and wasted as heat.... 4A is within the capabilities of a rice-grain sized SOT-23 MOSFET, even with only 3V on the gate, top end FETs in SOT-23 are up to that, with up to 30Vds (I just so happen to sell some nice FETs in SOT-23 mounted on breakout boards too! https://www.tindie.com/products/drazzy/simple-sot-23-mosfet-six-pack-logic-level/ ). SOT-23 FETs are insanely good up to a few amps, especially for the price - and that's even if you're buying my assembled boards with my margins! Also, unlike the beefier packages, they almost always spec max current at Ta (ambient temperature) not Tc (case temperature), so the datasheet numbers are actually realistic without heroic cooling measures... Edit: Oh, and I couldn't help myself. I bought the 381 IRF3708's in TO-220 that Rochester Electronics had on close-out special (they are weird - the moment something gets marked last time buy, it's like it turns radioactive and they feel an urgent need to get rid of it - whereas I have a desperate desire (at least when there's nothing as good available otherwise, like now) to buy a stockpile). Assuming someone else doesn't start making baller low-voltage 30V+ FETs in nice beefy throughhole packages (in which case I would be left holding the anti-static bag), I'll wait 'til they're out of stock from all the usual suspects and try to flip em :-P |
Beta Was this translation helpful? Give feedback.
-
Posted at 2020-07-18 by DrAzzy
The IRF3708PBF has long been my go-to through-hole MOSFET - a beefy TO-220 FET, good to 30Vds, and can handle tens of amps with added cooling.... but the real magic was that it was rated for Vgs = 2.8V, meaning you could comfortably control it with a 3.3v device and count on Rds(on) of 29 milliohm (max, 15 typ) (and that's at 2.8v, so it would be even better). A perfect power MOSFET for 3.3v boards like Espruino.
Well, unfortunately, Infineon is shutting down the line they make them at, and discontinuing the product. Their recommended replacements don't switch at 3.3v, and there aren't any other beefy through-hole MOSFETs that switch at 3.3v... at all. In fact, there aren't even any easily-soldered (ex, DPAK) SMD power MOSFETs that switch at 3.3v and can handle more than 20v (though the MCC90N02, recently released, has specs to die for in DPAK, but 20Vds maximum), as the DPAK version is being discontinued too.
So, if you think you ever might want to switch a load too heavy for a SOT-23 FET (those cap out at ~6A, manufacturer-spec-numbers, which one should never expect to push to the max) with an Espruino, and don't want to commit to the expense and complexity of a gate driver, and don't have the means to solder weird parts like Power-TDSON-8, you should order some with a sense of urgency!
Yes, I am aware that $parkfun says that you can use an FQP30N06L with a 3.3v gate voltage. The performance curves in the datasheet say otherwise - at least for loads that couldn't be handled by the top-end, much smaller, SOT-23 fets...
Just thought I'd let y'all know the bad news over here, since I clearly remember talking about it here - I think it's even in the MOSFET guide in the docs, assuming it hasn't been edited.
Beta Was this translation helpful? Give feedback.
All reactions