Low audio noise JFETs?

The Wikipedia entry matches my usage of the word pretty much exactly. For example, stating a "repeatability" specification in large print while the actual accuracy specification is elsewhere in small print and you have to add several error terms to *that* number to get the actual worst-case error.

Best regards, Spehro Pefhany

It's just a one-pager what I found but yes, the IF9031 and others do look nice. Might have to call them up to see how they are sold. The distributors I looked at don't carry that manufacturer.

Not allowed to tell ;-)

Fred Bartoli a écrit :

Forgot the Toshiba 2SK3320 (nice low noise dual in a small SMT package).

Aha, thanks. Nowadays it's more like that they claim low noise and then there isn't any noise spec at all in the datasheet.

Is your favourite BF862 noisy at audio freqs ?

Graham

Yes, Phil had already mentioned that one iand it's an excellent JFET. Toshiba doesn't seem to be a big believer in SMT though. But we can probably cram one in there, the others don't have to be that good.

It ain't my favorite because it's single-sourced. Usually RF FETs aren't that great at audio.

with some of the modern day rap music I hear blasting out of these cars going by while they attempt to knock out my windows with that distorted base, I would thing that noise factor in the audio spectrum would be attractive!

Yes, but the noise all has to be generated by an approved rap musician or group. Perfect, oxygen-free fidelity is required to reproduce it adequately.

datasheet.

...

Without something to support the supposition, yes, I think it is. Many parts designed specifically for RF work have rather terrible audio noise figures. The 1/f noise may well start at 10kHz or even above. The equipment we build that covers to 100MHz and above often has rather poor low-audio-range noise performance, but then very few of our customers are looking with this equipment below 1MHz even.

Cheers, Tom

People like NXP's BF862, but you don't like NXP (by the way I've seen their bipolar discretes fab. Second biggest discretes fab in the world, after Rohm, and continuously expanding). Specced 0.8nV/rtHz @ 100kHz (a lot of good that'll do you).

As a dual I've got NXPs PMBFJ620 in my drawer but haven't tried it out yet. Datasheet says 6nV/rtHz @ 100Hz.

Ciss of the 862 is roughly 2x that of the 620, so I'd expect the 862 to be less noisy than the 620 (but that's not even a hand-waving deduction any more, it's more like arms-flailing).

Then there are of course things like Interfet's IF3601 (single) and IF3602 (dual, 0.3nV/rtHz @ 100Hz), but it's my hunch that you're not keen on spending big $$$ on something in a TO-99 can.

robert

Speaking of low-noise FETs, Fred, have you ever gotten around to measuring the '147s and '620s I sent you in... let's seeee... March?

Take care,

--Daniel

Reply in private.

datasheet.

If you want the best of both worlds you'll have to design in two diplexers and two amp strips.

I attribute it to the California and Texas climate. Compare and .

Martin.

Daniel, I answered you but the email bounced back to me twice (Connection timed out)

Can you provide me with a working address?

t)

And the answer was...?

That's funny. Anyway, I just sent you a message from my regular account which so far has given me no trouble. Just reply to that one.

--Daniel

I hate the magazine ads for ICs that say "low cost" or "high speed" or "low power" and have no numbers.

John

I really hate it when they tout a fancy new chip, you scoot over to Digikey and then there ain't none. Whereas Costco, Walmart and others stuff their shelves before placing any ads. At least as far as I am concerned semi mfgs have blown countless Dollars on ads where I looked once and never returned. Same for fantasy pricing. Remember when TI mentioned under MSP430 "starting at 49c" whereas in reality the lowliest version cost almost a buck in qties?