Ultra Low noise N Channel Fet

[amdx]

 Saw this on audio Forum. That some may find an interest. I think it is
a recent release.

> https://www.ti.com/lit/ds/symlink/jfe150.pdf?ts=1627646886465&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FJFE150

                                               Mikek


--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

[Spehro Pefhany]

On Tue, 3 Aug 2021 08:10:53 -0500, amdx <am...@knology.net> wrote:

>  Saw this on audio Forum. That some may find an interest. I think it is
>a recent release.
>
>> https://www.ti.com/lit/ds/symlink/jfe150.pdf?ts=1627646886465&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FJFE150
>
>                                                Mikek

Interesting, but a *lot* of input capacitance.
--
Best regards,
Spehro Pefhany

[Jeroen Belleman]

amdx wrote:
> Saw this on audio Forum. That some may find an interest. I think it is
> a recent release.
>
>> https://www.ti.com/lit/ds/symlink/jfe150.pdf?ts=1627646886465&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FJFE150
>>
>
> Mikek

Still not as nice as the late lamented BF862, which had about the
same noise for *much* lower C_iss and C_rss.

Jeroen Belleman

[Jan Panteltje]

On a sunny day (Tue, 3 Aug 2021 08:10:53 -0500) it happened amdx
<am...@knology.net> wrote in <sebf8v$ih7$1...@dont-email.me>:

Does not look all that linear to me, first thing required for audio.
May work for very small signals, but when people want low noise they
usually have a high dynamic range.

?

[jla...@highlandsniptechnology.com]

Could be the giant chip and 6 ESD diodes. Input current is high too.
Best for low impedance sources.

Interesting part; Mouser has it. I didn't realize that TI still made
discretes.



--

John Larkin Highland Technology, Inc

The best designs are necessarily accidental.

[Phil Hobbs]

Yeah, weird--about the same noise as a CPH3910, but higher
transconductance and I_DSS. You'd do a lot, a lot better with a couple
of CPH3910s in parallel.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC / Hobbs ElectroOptics
Optics, Electro-optics, Photonics, Analog Electronics
Briarcliff Manor NY 10510

http://electrooptical.net
http://hobbs-eo.com

[Joe Gwinn]

On Tue, 03 Aug 2021 07:58:59 -0700, jla...@highlandsniptechnology.com
wrote:

>On Tue, 03 Aug 2021 09:18:35 -0400, Spehro Pefhany
><spef...@interlogDOTyou.knowwhat> wrote:
>
>>On Tue, 3 Aug 2021 08:10:53 -0500, amdx <am...@knology.net> wrote:
>>
>>>  Saw this on audio Forum. That some may find an interest. I think it is
>>>a recent release.
>>>
>>>> https://www.ti.com/lit/ds/symlink/jfe150.pdf?ts=1627646886465&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FJFE150
>>>
>>>                                                Mikek
>>
>>Interesting, but a *lot* of input capacitance.
>
>Could be the giant chip and 6 ESD diodes. Input current is high too.
>Best for low impedance sources.

Yes, like studio microphones. The giveaway is the bit about handling
48-volt phantom power.

Joe Gwinn

[Gerhard Hoffmann]

Am 03.08.21 um 15:10 schrieb amdx:

>  Saw this on audio Forum. That some may find an interest. I think it is
> a recent release.
>
>> https://www.ti.com/lit/ds/symlink/jfe150.pdf?ts=1627646886465&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FJFE150

Hope you saw the plea for a fast PNP. :-)
The TI guy seemed not _completely_ uninterested.
But his proposal for 24V SiGe tells he's probably not the process
specialist.

Gerhard

[jla...@highlandsniptechnology.com]

On Tue, 3 Aug 2021 18:35:17 +0200, Gerhard Hoffmann <dk...@arcor.de>
wrote:

Can SiGe do PNPs? And high voltage?

There must not be enough demand for fast PNPs to make them worth
fabbing.

I want P-channel phemts and GaN!

[Gerhard Hoffmann]

Am 03.08.21 um 19:24 schrieb jla...@highlandsniptechnology.com:

> On Tue, 3 Aug 2021 18:35:17 +0200, Gerhard Hoffmann <dk...@arcor.de>
> wrote:
>
>> Am 03.08.21 um 15:10 schrieb amdx:
>>>  Saw this on audio Forum. That some may find an interest. I think it is
>>> a recent release.
>>>
>>>> https://www.ti.com/lit/ds/symlink/jfe150.pdf?ts=1627646886465&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FJFE150
>>
>> Hope you saw the plea for a fast PNP. :-)
>> The TI guy seemed not _completely_ uninterested.
>> But his proposal for 24V SiGe tells he's probably not the process
>> specialist.
>>
>> Gerhard
>
> Can SiGe do PNPs? And high voltage?
>
> There must not be enough demand for fast PNPs to make them worth
> fabbing.
>
> I want P-channel phemts and GaN!

I only asked for a fast PNP. When they make yet another mediocre JFET,
then why not a fast PNP?

Can't get my LMX2594; next summer from DK, perhaps.
On Mouser, they had a few hundred until I clicked on "buy".
Newark had a few 100 at twice the price.
I won't try to catch the falling knive.
Just when my test boards arrived at Cologne airport.

JLCpcb was really fast, as usual.
BTW they have 2 4-layer stackups with thicker or thinner lines
for the same impedance.
My board is on process JLC27-something, with 6 and 12 mil microstrips
for the VNA or TDR, so the $5 are not completely wasted. 1/2 :-)
11.55 mil should be 50 Ohms.

Gerhard

[John Miles, KE5FX]

On Tuesday, August 3, 2021 at 10:57:16 AM UTC-7, Gerhard Hoffmann wrote:
> Can't get my LMX2594; next summer from DK, perhaps.

LCSC has them at a decent price, it looks like. About US $50/qty 1.

-- john, KE5FX

[John Larkin]

On Tue, 3 Aug 2021 19:57:08 +0200, Gerhard Hoffmann <dk...@arcor.de>

wrote:

>Am 03.08.21 um 19:24 schrieb jla...@highlandsniptechnology.com:
>> On Tue, 3 Aug 2021 18:35:17 +0200, Gerhard Hoffmann <dk...@arcor.de>
>> wrote:
>>
>>> Am 03.08.21 um 15:10 schrieb amdx:
>>>>  Saw this on audio Forum. That some may find an interest. I think it is
>>>> a recent release.
>>>>
>>>>> https://www.ti.com/lit/ds/symlink/jfe150.pdf?ts=1627646886465&ref_url=https%253A%252F%252Fwww.ti.com%252Fproduct%252FJFE150
>>>
>>> Hope you saw the plea for a fast PNP. :-)
>>> The TI guy seemed not _completely_ uninterested.
>>> But his proposal for 24V SiGe tells he's probably not the process
>>> specialist.
>>>
>>> Gerhard
>>
>> Can SiGe do PNPs? And high voltage?
>>
>> There must not be enough demand for fast PNPs to make them worth
>> fabbing.
>>
>> I want P-channel phemts and GaN!
>
>I only asked for a fast PNP. When they make yet another mediocre JFET,
>then why not a fast PNP?
>
>Can't get my LMX2594; next summer from DK, perhaps.
>On Mouser, they had a few hundred until I clicked on "buy".
>Newark had a few 100 at twice the price.
>I won't try to catch the falling knive.
>Just when my test boards arrived at Cologne airport.

Mouser expects 3100 tomorrow.

Wow, a 15 GHz synthesizer. That would have filled a couple of racks
and cost a few hundred K$ not so long ago.

I remember seeing an RF frequency measurement system that filled an
entire rack. The equivalent counter is now maybe $25.

[Gerhard Hoffmann]

Am 05.08.21 um 01:49 schrieb John Larkin:

> On Tue, 3 Aug 2021 19:57:08 +0200, Gerhard Hoffmann <dk...@arcor.de>
> wrote:

>> Can't get my LMX2594; next summer from DK, perhaps.
>> On Mouser, they had a few hundred until I clicked on "buy".
>> Newark had a few 100 at twice the price.
>> I won't try to catch the falling knive.
>> Just when my test boards arrived at Cologne airport.
>
> Mouser expects 3100 tomorrow.

Yesterday they said next March and heavy demand. I'll check that when I
wake up tomorrow.

Up to now I use ADF5356, that needs an internal frequency doubler that
is a royal pain since the fundamental gets through nearly as strong as
the wanted output.

<
https://ez.analog.com/rf/f/q-a/536893/adf5356-doubts-about-bleeding-current-programming
>

This here is a fundamental notch filter that works quite OK:

<
https://www.flickr.com/photos/137684711@N07/50404949846/in/album-72157662535945536/
>

lambda/4 cancels 3/4 lambda.


<
https://www.flickr.com/photos/137684711@N07/50404249018/in/album-72157662535945536/
>

quite OK for the simple structure.



BTW AD engineering zone was not very helpful. Seen by 380 people, no
answer.I set the bleeding current that the eval software choose, but
don't know why it did.

The TI synthesizer eval board seems to have less phase noise and does
the 15 GHz with the fundamental. They also have a 20 GHz chip with
frequency doubler and tracking filter, but I need only 10.

> Wow, a 15 GHz synthesizer. That would have filled a couple of racks
> and cost a few hundred K$ not so long ago.

Ah, a R&S SMP-20 is 3 or 4 height units 19".
Methinks I want one. But there's the price...

>
> I remember seeing an RF frequency measurement system that filled an
> entire rack. The equivalent counter is now maybe $25.

If that counter is a chip, I'd be most interested.


I've done a 13 GHz 1/8 prescaler for triggering my scope or to extend
the SR620 counter. Boards just flying in from JLCPCb for the newest
update. $2 or so for 20 pcs. incremental cost. :-)

cheers, Gerhard

[jla...@highlandsniptechnology.com]

On Thu, 5 Aug 2021 03:22:04 +0200, Gerhard Hoffmann <dk...@arcor.de>

wrote:

>Am 05.08.21 um 01:49 schrieb John Larkin:
>> On Tue, 3 Aug 2021 19:57:08 +0200, Gerhard Hoffmann <dk...@arcor.de>
>> wrote:
>
>>> Can't get my LMX2594; next summer from DK, perhaps.
>>> On Mouser, they had a few hundred until I clicked on "buy".
>>> Newark had a few 100 at twice the price.
>>> I won't try to catch the falling knive.
>>> Just when my test boards arrived at Cologne airport.
>>
>> Mouser expects 3100 tomorrow.
>
>Yesterday they said next March and heavy demand. I'll check that when I
>wake up tomorrow.
>
>Up to now I use ADF5356, that needs an internal frequency doubler that
>is a royal pain since the fundamental gets through nearly as strong as
>the wanted output.
>
><
>https://ez.analog.com/rf/f/q-a/536893/adf5356-doubts-about-bleeding-current-programming
> >
>
>This here is a fundamental notch filter that works quite OK:
>
><
>https://www.flickr.com/photos/137684711@N07/50404949846/in/album-72157662535945536/
> >
>
>lambda/4 cancels 3/4 lambda.

Fun. What's the substrate? I'm expecting a proto board tomorrow, 4
layers of Isola Tachyon 100G. Expensive.

https://www.dropbox.com/s/a37ctm6rymzf1fy/Z508A.jpg?raw=1

I expect that one or maybe two of those circuits might work.

>
>
><
>https://www.flickr.com/photos/137684711@N07/50404249018/in/album-72157662535945536/
> >
>
>quite OK for the simple structure.
>
>
>
>BTW AD engineering zone was not very helpful. Seen by 380 people, no
>answer.I set the bleeding current that the eval software choose, but
>don't know why it did.
>
>The TI synthesizer eval board seems to have less phase noise and does
>the 15 GHz with the fundamental. They also have a 20 GHz chip with
>frequency doubler and tracking filter, but I need only 10.
>
>> Wow, a 15 GHz synthesizer. That would have filled a couple of racks
>> and cost a few hundred K$ not so long ago.
>
>Ah, a R&S SMP-20 is 3 or 4 height units 19".
>Methinks I want one. But there's the price...
>
>
>>
>> I remember seeing an RF frequency measurement system that filled an
>> entire rack. The equivalent counter is now maybe $25.
>
>If that counter is a chip, I'd be most interested.

Entire boards from Amazon!

>
>
>I've done a 13 GHz 1/8 prescaler for triggering my scope or to extend
>the SR620 counter. Boards just flying in from JLCPCb for the newest
>update. $2 or so for 20 pcs. incremental cost. :-)
>
>cheers, Gerhard

[Chris Jones]

On 05/08/2021 11:22, Gerhard Hoffmann wrote:
> Am 05.08.21 um 01:49 schrieb John Larkin:
>> On Tue, 3 Aug 2021 19:57:08 +0200, Gerhard Hoffmann <dk...@arcor.de>
>> wrote:
>
>>> Can't get my LMX2594; next summer from DK, perhaps.
>>> On Mouser, they had a few hundred until I clicked on "buy".
>>> Newark had a few 100 at twice the price.
>>> I won't try to catch the falling knive.
>>> Just when my test boards arrived at Cologne airport.
>>
>> Mouser expects 3100 tomorrow.
>
> Yesterday they said next March and heavy demand. I'll check that when I
> wake up tomorrow.
>
> Up to now I use ADF5356, that needs an internal frequency doubler that
> is a royal pain since the fundamental gets through nearly as strong as
> the wanted output.
>
> <
> https://ez.analog.com/rf/f/q-a/536893/adf5356-doubts-about-bleeding-current-programming
>     >
>
> This here is a fundamental notch filter that works quite OK:
>
> <
> https://www.flickr.com/photos/137684711@N07/50404949846/in/album-72157662535945536/
>   >
>
> lambda/4 cancels 3/4 lambda.
>
>
> <
> https://www.flickr.com/photos/137684711@N07/50404249018/in/album-72157662535945536/
>     >
>
> quite OK for the simple structure.
>

Does anyone know how the ATEK950P6 works? Seems to good to be on-chip LC
filters.

[Gerhard Hoffmann]

Am 05.08.21 um 07:51 schrieb Chris Jones:

>
> Does anyone know how the ATEK950P6 works? Seems to good to be on-chip LC
> filters.
>

WOW.
Can only be integrated on-chip, given how flat the package is.

cheers, Gerhard

[Gerhard Hoffmann]

Am 05.08.21 um 05:29 schrieb jla...@highlandsniptechnology.com:

> On Thu, 5 Aug 2021 03:22:04 +0200, Gerhard Hoffmann <dk...@arcor.de>
> wrote:

>> This here is a fundamental notch filter that works quite OK:
>>
>> <
>> https://www.flickr.com/photos/137684711@N07/50404949846/in/album-72157662535945536/
>> >
>>
>> lambda/4 cancels 3/4 lambda.
>
> Fun. What's the substrate? I'm expecting a proto board tomorrow, 4
> layers of Isola Tachyon 100G. Expensive.

Rogers TMM-6
looks like a rubber eraser filled with ceramic. Quite unstable.
You can bend or even break it when fastening an SMA.

I think it works on FR-4, too. The lower Q just makes more over-all
attenuation, but the notch effect is just run time difference.

>>> I remember seeing an RF frequency measurement system that filled an
>>> entire rack. The equivalent counter is now maybe $25.
>>
>> If that counter is a chip, I'd be most interested.
>
> Entire boards from Amazon!

I'll look into this. I was about to augment a CPLD design of mine
with some CML Flipflops.

>> I've done a 13 GHz 1/8 prescaler for triggering my scope or to extend
>> the SR620 counter. Boards just flying in from JLCPCb for the newest
>> update. $2 or so for 20 pcs. incremental cost. :-)

Wow they are fast! It's on my table. Gerber uploaded 2021-07-27 evening.
Arrived from China August 5 9 am.

Two 2layer boards, 5 & 10 pcs.,
One 4layer board, 5 pcs. JLC7628 stacking

€29.50 all included.

They send the daily production for Europe in one large box to Cologne,
slaughter it there and resend the small boxes via inland mail.
They handle all the customs & inland VAT.

cheers, Gerhard

[jla...@highlandsniptechnology.com]

On Thu, 5 Aug 2021 11:19:52 +0200, Gerhard Hoffmann <dk...@arcor.de>

wrote:

>Am 05.08.21 um 05:29 schrieb jla...@highlandsniptechnology.com:
>> On Thu, 5 Aug 2021 03:22:04 +0200, Gerhard Hoffmann <dk...@arcor.de>
>> wrote:
>
>>> This here is a fundamental notch filter that works quite OK:
>>>
>>> <
>>> https://www.flickr.com/photos/137684711@N07/50404949846/in/album-72157662535945536/
>>> >
>>>
>>> lambda/4 cancels 3/4 lambda.
>>
>> Fun. What's the substrate? I'm expecting a proto board tomorrow, 4
>> layers of Isola Tachyon 100G. Expensive.
>
>Rogers TMM-6
>looks like a rubber eraser filled with ceramic. Quite unstable.
>You can bend or even break it when fastening an SMA.

Some of the Rogers teflon stuff is like copperclad leather. Mounting
spacers and screws get interesting.

>
>I think it works on FR-4, too. The lower Q just makes more over-all
>attenuation, but the notch effect is just run time difference.

In time domain, we try to keep all the signal traces short. At 100 ps
rise/fall, and fraction of an inch traces, FR4 (and cheap edge-launch
SMAs!) are good enough.

This cheap connector is quite good if you handle the pcb layers right.
We simulated that with ATLC.

https://www.dropbox.com/s/ag8vj0fj2htkjsm/SS_SMA_Edge.JPG?raw=1

My Tachyon board has test traces and various SMA footprints, to
compare to similar patterns that I have on FR4.

[jla...@highlandsniptechnology.com]

That's beautiful. What does it cost?

[Chris Jones]

I was wondering whether it could be some sort of FBAR acoustic thing
plus or on a silicon switch die. HP/Agilent/Avago/Broadcom/whatever had
some technology like that about 15 years ago for cellphone diplexers and
duplexers IIRC.

[Chris Jones]

No idea, and I'd never heard of the company either.

[Clifford Heath]

Has anyone figured out how to buy these, or how much they cost?

[Gerhard Hoffmann]

Am 12.08.21 um 02:35 schrieb Clifford Heath:

I did not really try to find out, although I find them definitely
interesting. The producer is in Istanbul, Turkey.

< https://www.atekmidas.com/about/ >


I got my LMX2594s btw. from Mouser, seems to have been a computer
hickup. But supply does dry out.

Fedex could work on their web software. It showed that my box was in
Memphis, TN and could be expected in 10 minutes here in smalltown, de.

cheers, Gerhard

[Jasen Betts]

On 2021-08-12, Gerhard Hoffmann <dk...@arcor.de> wrote:
>
> I got my LMX2594s btw. from Mouser, seems to have been a computer
> hickup. But supply does dry out.
>
> Fedex could work on their web software. It showed that my box was in
> Memphis, TN and could be expected in 10 minutes here in smalltown, de.

Yeah they don't do time zones.

--
Jasen.

[Gerhard Hoffmann]

Am 12.08.21 um 22:56 schrieb Jasen Betts:

Nono, they do; I can select my local time.
It simply took a day longer with wrong data on the web.
No problem.

Gerhard