"cheap" JFET???????

I see that the lowest price for the 2N5114/5115 is in the $11.00 region. WHY so friggin much? Any lower priced JFETs,say $1.00 region?

Yeah, it's amazing. Noise is below 1 nv/rthz, and the Idss max/min spec is only 2.5:1.

And it has a **12 page** data sheet.

Metal-can parts? Let me guess, from Central Semi, too? Who still buys those anymore!

PN439x and 2N548x are going discontinued, too, but MMBT versions (SOT-23) are still available. Besides the more popular BFxxx ones, too.

Tim

-- Seven Transistor Labs Electrical Engineering Consultation Website:

We should have had a through discussion of JFETs in AoE 2nd edition, 25 years ago, but there wasn't time. For the 3rd edition, we took the extra time. Hah, over two man years, buried in Chapters 3, 3x and 8. While y'all were waiting! Tables 3.1, 3.7 (pg 217) and 8.2 (pg 516) of over 80 types, most of which are still available, if you know where (Mouser, B&D, or direct from InterFET, Linear Integrated Systems).

The tables include detailed bench measurements of Idss, Vgs, gm and gos. The latter, via our new Id-independent parameter, Gmax, where gos=gm/Gmax.

You get noise measurements vs freq, for 40 types, and over 20 types are plotted, e.g., Figure 8.52.

Yes, these are usually single measurements of one part, but hey that's better than you get from a datasheet's min/max values.

So, yes, discrete JFETs are still alive and well. Although often you'll prefer a JEFT or CMOS op-amp instead, or maybe a depletion-mode MOSFET.

Am 13.04.2015 um 07:42 schrieb Robert Baer:

JFET manufacturers seldom provide more than 1 page, most less than 1 page, and even that is, ehm, euphemistic.

regards, Gerhard

Thanks.

Thanks; but i do not count those last 3 pages..

DMOS FETS "work" differently Vgs versus Id so they are not interchangeable for a given circuit (see attached GIF).

Attachments don't work on usenet. :-( Actually, depletion-mode MOSFETs aren't all that dissimilar to JFETs, which are also depletion mode. See AoE-III, Figure 3.9 on page 135, Id vs Vgs, where two MOSFET types are nested with a bunch of JFETs. Note, DMOS doesn't mean depletion mode. It's the enchancement-mode MOSFETs that are another beast.

But we'd never suggest direct replacements anyway, as the plots in that figure illustrate, Vgs can be all over the map, and designs need to be specific to the part, as much as is practical anyway.

• Actually, i have never seen a problem with them (years of TerraNews and a week of theCubeNet) - both posting and "reading" them.

• Sorry; i thought that the "D" meant "depletion"; i stand corrected.

• Tell me about it; finding an inexpensive replacement to the 2N5115 is a bitch; it seems that the Fairchild J175 might do.

Whoa, this is news! Usenet server-to-server communications retaining attachments? I use NewsGuy, and Google (DejaNews) on occasion.

Yep, looking at my table, it doesn't look so bad.

I saw it on Agent news. First time I've seen a binary on a text group, though.

Here, a bit of insight from National Semiconductor "Discrete Semiconductor Products" (1989 edition): both 2N5115 and J175 are P-channel jFETs made with Process 88. So, except for packaging and test limits, they were the same die (if you got them from NS).

Others from Process 88 are J174, J176, J177, J270, J271, MMBFJ174, MMBFJ175, MMBFJ176, MMBFJ177,

The 2N5115 (and a few others) in metal cans (TO-18) are unobtainium, I suspect.

I've always thought they didn't. At least not with most groups, with my usenet server news.albasani.net and with my newsreader Outlook Express.

Well, Mr.Baer has proven me wrong on all counts! His image is clearly displayed here.

Thanks for the reference; maybe Vishay, Interfet, and Fairchild use the same process: MIL 2nd source requirements OR bought rights to make. Yep, all metal can versions are unobtanium or very $$$ whereas the plastic ones are decently priced.

BTW, the entire NS small-signal line went to Fairchild.

Although many JFET types are second sourced, and share the same datasheet specs, there are still in fact substantial differences between them. E.g., see AoE-III, Figure 3.52, comparing Fairchild and ON Semi 2n5457 in Id vs Vgs plots. See also the similar plots in Figure 3.55, comparing seven examples: apparently all over the map. However, then see Figure 3.56, where transconductace vs Id curves for these same seven different 2n5486 parts are on top of each other. So, as we know, setting the bias current for a JFET is challenging, due to Vgs variablity (a BJT current source is a good answer), but at least the gm predictability for a given type is not bad at all, even across manufacturers and date codes, e.g., see footnote 62.

Attachments do work, as a technical matter.

However, each newsgroup decides if they will accept attachments, or strip them, and configures accordingly.

Most text groups disallow attachments, to cut down on server bloat and malware problems.

Joe Gwinn

It's the people who run the news servers who decide. For example, the one I am using has a strict policy of no binaries. Therefore, attachments do not propagate and thus everyone on that server cannot see them. I've asked them whether they could allow at least small attachments under 100kB but they said no.