An improved 2n7002

2n7002k.

formatting link
Crss and Coss are about halved compared to the old-guard part, the rest of the specs eyeball out about the same. Rdson may be a tad lower, depending on Vgs & such. YMMV.

I may have to drive some switches at 15MHz, so low Crss is appealing.

Several mfrs make 'em. Gate's protected, too.

You guys would like the very nice OnSemi datasheet I have for the OnSemi version. Online I only the see the Fairchild version though -- not as detailed & helpful.

Cheers, James Arthur

Reply to
dagmargoodboat
Loading thread data ...

I've got some products where the classic 2N7002 is better than any more modern stuff we've tried. In particular, the Fairchild part will switch 50 volts in under 1 ns, wildly different from what the data sheet suggests.

The K version would be interesting to try for fast switching. I wonder how gate protection might affect switching speed.

In the Vishay data sheet above, the Output and Transfer curves on sheet 3 seem to disagree. The transfer curve also cruises well above the abs max current.

(The Fairchild 2N7002 transfer curve is going great guns at 2 amps!)

Data sheets these days!

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
jlarkin

Octopart has ON Semi version if you go down to 2N7002KT1G.

What we need is a fast p-channel equivalent.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

You shouldn't expect an 'improved' jedec registered part.

Variations between mfrs fabs of the same part number are a nuisance, if not an actual hazard - not an advantage.

You CAN find 'better' parts with different part numbers and that is what you should be doing, if you need parameters that the registered part cannot guarantee.

RL

Reply to
legg

Yes, that shows the typical capacitance numbers of interest. Thanks Win.

The OnSemi datasheet I prefer is OnSemi publication

2N7002K/D, rev. 16, October 2016 '2N7002K-D.pdf'. That version has typical and worst-case numbers I don't see elsewhere, not even on OnSemi's website, downloaded three years ago. But I don't remember where I got it.

Ciss = 24.5pF typ, 45pF max. Crss = 2.2pF typ, 5.0pF max. Coss = 4.2pF typ, 8.0pF max.

These capacitances are already close to their minimums around ~2V.

For example, the classic Fairchild 2n7002 datasheet I've got on my screen shows Crss = 20pF at 1V, and still 10pF at 10V. OnSemi's 2n7002k: Crss = 8pF/0V, 4pF/1V, and 2.5pF/10V.

Also interesting is this Nexperia version, for its 40K/W thermal resistance...

formatting link

As part of a much more complicated design with many moving parts I've inherited a less-than-optimal class-C r.f. amplifier, which prompted this bit of browsing...

Cheers, James Arthur

Reply to
dagmargoodboat

I wasn't expecting an improved part, I was reporting the existence of them. :-)

Cheers, James Arthur

Reply to
dagmargoodboat

And some fast p-GaN fets. I'm using them in totem poles, but the upper gate driver gets really messy.

formatting link

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
jlarkin

I think some legal genius decided that one can't trademark a part number. So LM1117 or MAX809 or whatever can be fabbed and sold by all sorts of unsavory characters. We often specify ONSEMI ONLY or some such for some of our stocked parts, and hope that they don't change the recipe.

We have made deals with distributors to set aside a reel of mosfets, send us 5 to evaluate, and sell us the same reel if we like them.

--
John Larkin         Highland Technology, Inc 

lunatic fringe electronics
 Click to see the full signature
Reply to
jlarkin

For a 2 or 3 terminal device you should probably stick to the worst case guaranteed performance, in multisourcing.

If you're pushing the spec beyond that, you have to generate your own internal part number, and qualify/restrict vendors.

With more pins and internal complexity, ALL part numbers are basically an internal qual excercise. I have to qualify at least three sources for something as basic as a UC3842, in a new product build. I'll often rely on past known-good triples as a starting point, but they all get fitted extensively enough to know if they're suited to the new app.

If purchasing wants to add a vendor after the papers are finalized, it's their tough luck - they've already got the three and there has to be SIGNIFIGANT cost savings to justify even thinking about it, never mind running the 'new' part through it's paces, after-the-fact. The best you can offer is a new internal number, including the new part, to be adopted in the 'next' product, where used. It can be vetted then, for that new use, then possibly be worked into the older builds, based on proven performance.

If whoever's money we're (I'm) spending says 'go do it', well, nuff said. If it's my own nickel, I'll usually find the time to beat up a new candidate to my own satisfaction.

The worst position you can be in, is to find out that a single-sourced part isn't doing what it claims as a basic minimum function - you need a completely different approach, that avoids the part entirely, in your back pocket, as an escape route. - and about two bushels of unscheduled man-hours in which to make the escape.

RL

Reply to
legg

All our BOMs use internal 7-digit part numbers. And every internal part has a list of qualified vendors and their part numbers. We have a stock bin for 2N7002 Fairchild Only. Another stock number is 2N7002 Fairchild or ON or Vishay.

We do often push specs. Like getting 1 ns out of a part whose data sheet says it switches in 25. Sometimes we exceed abs max if the payoff is big and we think we have margin. We can have a stock bin for a part, and another bin for the tested part, and a test procedure.

Engineering qualifies parts here; never purchasing.

There are tons of sole-source parts, and some turn out to be bad. It usually takes a board spin to fix that.

Reply to
John Larkin

Good advice!

--
 Thanks, 
    - Win
Reply to
Winfield Hill

Very pretty. Fast GaN PFETs might be possible!

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
 Click to see the full signature
Reply to
Phil Hobbs

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.