varicaps?

On Wed, 1 Apr 2020 15:38:26 -0700 (PDT), whit3rd <whit3rd@gmail.com>
wrote:

On Wednesday, April 1, 2020 at 2:01:15 PM UTC-7, John Larkin wrote:

I need a small range, maybe a change of 0.2 pF over 5 volts bias
swing. The c-b junction of some small transistor might work.

A power Zener will have larger area and a good range of reverse-bias with low leakage.
Realistically, though, probably ANY diode other than a RF type will have
significant capacitance and likely can do this.

Because Rbb can be high in a 'small transistor', I'd try a cheap zener for
tuning a high-Q oscillator.

Zeners tend to have a LOT of capacitance. I only want a tiny bit.

I am finding varicaps available. I just got concerned when some of our
stock parts went EOL, and some distributors had no stock on others,
and coincidentally the Skyworks web site was down and I couldn't see
any of their data sheets.

I was partial to a dual, but it should work with a single, which are
more common. Nice thread on varicaps, a change from you-know-what.




--

John Larkin Highland Technology, Inc

Science teaches us to doubt.

Claude Bernard

 

[Phil Hobbs]

On 2020-04-02 10:51, Piglet wrote:

On 02/04/2020 15:45, Phil Hobbs wrote:
On 2020-04-02 10:41, Piglet wrote:
On 02/04/2020 15:28, George Herold wrote:
On Thursday, April 2, 2020 at 10:12:28 AM UTC-4, piglet wrote:
On 02/04/2020 14:29, jlarkin@highlandsniptechnology.com wrote:
On Thu, 2 Apr 2020 10:24:49 +0100, piglet <erichpwagner@hotmail.com
wrote:

On 01/04/2020 11:34 pm, John Larkin wrote:
On Wed, 1 Apr 2020 16:56:48 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-04-01 16:20, sea moss wrote:

I wonder if it's practical to use a MOSFET's Coss as a
varicap.  I haven't
seen it done.

Standard technique in ASIC design

-- Kevin Aylward
http://www.anasoft.co.uk - SuperSpice
http://www.kevinaylward.co.uk/ee/index.html

Cool.  I should have googled it first, there's a few examples
out there.

There are pots of low-pF ones on Digikey.  The 1000-pFish ones
(MVAM108
etc.) are long gone, and the tens-to-hundreds range is
declining.  My
fave MV209 is a distant memory, but BB201s are still in stock.

Cheers

Phil Hobbs

According to my Boonton, a BFT25 c-b junction only goes from
about 0.7
pF to around 0.65 from 0 to 6 volts. There's probably a bit of
fixture
capacitance too. Not a very good varicap.

I want a delta-c around 0.15 pF roughly, maybe 0.2. My dac+opamp
could
go to 20 volts.


You're going to need a bigger transistor. Try a BC847/2N3904 ?

piglet

I'm guessing that a varicap is doped to be, well, a varicap. They
have
delta-c catios like 6:1 over some voltage range. Looks like
transistor
junctions have much flatter c:v curves. So the tempco contribution
will be a lot higher if I use a transistor.

I've ordered some singles, Skyworks parts in the horrible SC79
package, and my 4-layer proto board should be here soon. If the board
house doesn't shut down.




Yes I think you're right about the flatter c:v curves. You only want
0.2pF swing so will have quite a lot of padding down.

I hooked up an SMD BC848C to the Boonton and got these:
Ccb (E open*)
0V 4.0pF
0.5V 3.2pF
1V 2.8pF
2V 2.36pF
3V 2.05pF
4V 1.85pF
5V 1.71pF
6V 1.60pF
7V 1.50pF
8V 1.42pF

Cceb (E tied to B)
0V 4.29pF
0.5V 3.54pF
1V 3.16pF
2V 2.69pF
3V 2.40pF
4V 2.20pF
5V 2.05pF
6V 1.92pF
7V 1.83pF
8V 1.75pF


A thru-hole version was Ccb (e open):
0V 5.35pF
0.5V 4.55pF
1V 4.12pF
2V 3.56pF
3V 3.19pF
4V 2.95pF
5V 2.72pF
6V 2.52pF
7V 2.33pF
8V 2.16pF

A thru hole RF part MPSH10 Ccb (e open):
0V 1.23pF
0.5V 1.09pF
1V 1.02pF
2V 0.94pF
3V 0.89pF
4V 0.86pF
5V 0.84pF
6V 0.82pF
7V 0.8pF
8V 0.79pF


A 4.3V 400mW zener was:
0V 211pF
0.5V 183pF
1V 168pF
2V 149pF

A junk box T1 red LED:
0V 28.4pF
0.5V 25.9pF
1V 24.4pF
2V 22pF
3V 20.4pF
4V 19.1pF
5V 18pF
6V 17.3pF
7V 16.6pF
8V 16pF

* I tried the NPN bjts Ccb-e and Cb-ce i.e. involving the base-emitter
junction - as expected the capacitance was higher at 12 to 16pF at
zero
bias falling to 6-8pF at 6V (didnt go higher to avoid breakdown) but I
noticed that the low bias capacitance values looked jittery and only
settled down above about 2V.

My junkbox has some wire ended varicaps somewhere so next coffee
break I
will try those too.

piglet

Fun,  If I can riff on the physics.  The variable C is caused by
changes in the depletion width.  So what sort of devices would have
a large variation?
PIN photodiodes?
maybe c-b junc of high voltage transistors?
What else?

George H.


Yes, one can get the depletion zone narrower (higher C) by approaching
forward bias but that quickly stops getting useful above 0.6 - 0.7V.

Heh - that reminds me try that Red LED in the pre-forward bias
direction. Thanks.

piglet


Besides C-ratio, the other important FOM for a varactor is Q.
Highly-doped diodes are the ticket for that.

Cheers

Phil Hobbs


Aren't emitters highly doped? Does that mean I could explore B-E
junctions as poor-man's varicaps?

piglet

That was commonly done BITD. Unfortunately Rbb' is generally much
higher than the emitter or collector resistances.

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

 

On Thu, 2 Apr 2020 15:41:48 +0100, Piglet <erichpwagner@hotmail.com>
wrote:

On 02/04/2020 15:28, George Herold wrote:
On Thursday, April 2, 2020 at 10:12:28 AM UTC-4, piglet wrote:
On 02/04/2020 14:29, jlarkin@highlandsniptechnology.com wrote:
On Thu, 2 Apr 2020 10:24:49 +0100, piglet <erichpwagner@hotmail.com
wrote:

On 01/04/2020 11:34 pm, John Larkin wrote:
On Wed, 1 Apr 2020 16:56:48 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-04-01 16:20, sea moss wrote:

I wonder if it's practical to use a MOSFET's Coss as a varicap. I haven't
seen it done.

Standard technique in ASIC design

-- Kevin Aylward
http://www.anasoft.co.uk - SuperSpice
http://www.kevinaylward.co.uk/ee/index.html

Cool. I should have googled it first, there's a few examples out there.

There are pots of low-pF ones on Digikey. The 1000-pFish ones (MVAM108
etc.) are long gone, and the tens-to-hundreds range is declining. My
fave MV209 is a distant memory, but BB201s are still in stock.

Cheers

Phil Hobbs

According to my Boonton, a BFT25 c-b junction only goes from about 0.7
pF to around 0.65 from 0 to 6 volts. There's probably a bit of fixture
capacitance too. Not a very good varicap.

I want a delta-c around 0.15 pF roughly, maybe 0.2. My dac+opamp could
go to 20 volts.


You're going to need a bigger transistor. Try a BC847/2N3904 ?

piglet

I'm guessing that a varicap is doped to be, well, a varicap. They have
delta-c catios like 6:1 over some voltage range. Looks like transistor
junctions have much flatter c:v curves. So the tempco contribution
will be a lot higher if I use a transistor.

I've ordered some singles, Skyworks parts in the horrible SC79
package, and my 4-layer proto board should be here soon. If the board
house doesn't shut down.




Yes I think you're right about the flatter c:v curves. You only want
0.2pF swing so will have quite a lot of padding down.

I hooked up an SMD BC848C to the Boonton and got these:
Ccb (E open*)
0V 4.0pF
0.5V 3.2pF
1V 2.8pF
2V 2.36pF
3V 2.05pF
4V 1.85pF
5V 1.71pF
6V 1.60pF
7V 1.50pF
8V 1.42pF

Cceb (E tied to B)
0V 4.29pF
0.5V 3.54pF
1V 3.16pF
2V 2.69pF
3V 2.40pF
4V 2.20pF
5V 2.05pF
6V 1.92pF
7V 1.83pF
8V 1.75pF


A thru-hole version was Ccb (e open):
0V 5.35pF
0.5V 4.55pF
1V 4.12pF
2V 3.56pF
3V 3.19pF
4V 2.95pF
5V 2.72pF
6V 2.52pF
7V 2.33pF
8V 2.16pF

A thru hole RF part MPSH10 Ccb (e open):
0V 1.23pF
0.5V 1.09pF
1V 1.02pF
2V 0.94pF
3V 0.89pF
4V 0.86pF
5V 0.84pF
6V 0.82pF
7V 0.8pF
8V 0.79pF


A 4.3V 400mW zener was:
0V 211pF
0.5V 183pF
1V 168pF
2V 149pF

A junk box T1 red LED:
0V 28.4pF
0.5V 25.9pF
1V 24.4pF
2V 22pF
3V 20.4pF
4V 19.1pF
5V 18pF
6V 17.3pF
7V 16.6pF
8V 16pF

* I tried the NPN bjts Ccb-e and Cb-ce i.e. involving the base-emitter
junction - as expected the capacitance was higher at 12 to 16pF at zero
bias falling to 6-8pF at 6V (didnt go higher to avoid breakdown) but I
noticed that the low bias capacitance values looked jittery and only
settled down above about 2V.

My junkbox has some wire ended varicaps somewhere so next coffee break I
will try those too.

piglet

Fun, If I can riff on the physics. The variable C is caused by
changes in the depletion width. So what sort of devices would have
a large variation?
PIN photodiodes?
maybe c-b junc of high voltage transistors?
What else?

George H.


Yes, one can get the depletion zone narrower (higher C) by approaching
forward bias but that quickly stops getting useful above 0.6 - 0.7V.

Heh - that reminds me try that Red LED in the pre-forward bias
direction. Thanks.

piglet

I expect that negative base bias would modulate Cce, which makes a
3-terminal varicap. Somebody sells a ceramic-based 3-terminal varicap.



--

John Larkin Highland Technology, Inc

Science teaches us to doubt.

Claude Bernard

 

[Piglet]

On 02/04/2020 16:11, Phil Hobbs wrote:

On 2020-04-02 10:51, Piglet wrote:
On 02/04/2020 15:45, Phil Hobbs wrote:
On 2020-04-02 10:41, Piglet wrote:
On 02/04/2020 15:28, George Herold wrote:
On Thursday, April 2, 2020 at 10:12:28 AM UTC-4, piglet wrote:
On 02/04/2020 14:29, jlarkin@highlandsniptechnology.com wrote:
On Thu, 2 Apr 2020 10:24:49 +0100, piglet <erichpwagner@hotmail.com
wrote:

On 01/04/2020 11:34 pm, John Larkin wrote:
On Wed, 1 Apr 2020 16:56:48 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-04-01 16:20, sea moss wrote:

I wonder if it's practical to use a MOSFET's Coss as a
varicap. I haven't
seen it done.

Standard technique in ASIC design

-- Kevin Aylward
http://www.anasoft.co.uk - SuperSpice
http://www.kevinaylward.co.uk/ee/index.html

Cool. I should have googled it first, there's a few examples
out there.

There are pots of low-pF ones on Digikey. The 1000-pFish ones
(MVAM108
etc.) are long gone, and the tens-to-hundreds range is
declining. My
fave MV209 is a distant memory, but BB201s are still in stock.

Cheers

Phil Hobbs

According to my Boonton, a BFT25 c-b junction only goes from
about 0.7
pF to around 0.65 from 0 to 6 volts. There's probably a bit of
fixture
capacitance too. Not a very good varicap.

I want a delta-c around 0.15 pF roughly, maybe 0.2. My dac+opamp
could
go to 20 volts.


You're going to need a bigger transistor. Try a BC847/2N3904 ?

piglet

I'm guessing that a varicap is doped to be, well, a varicap. They
have
delta-c catios like 6:1 over some voltage range. Looks like
transistor
junctions have much flatter c:v curves. So the tempco contribution
will be a lot higher if I use a transistor.

I've ordered some singles, Skyworks parts in the horrible SC79
package, and my 4-layer proto board should be here soon. If the
board
house doesn't shut down.




Yes I think you're right about the flatter c:v curves. You only want
0.2pF swing so will have quite a lot of padding down.

I hooked up an SMD BC848C to the Boonton and got these:
Ccb (E open*)
0V 4.0pF
0.5V 3.2pF
1V 2.8pF
2V 2.36pF
3V 2.05pF
4V 1.85pF
5V 1.71pF
6V 1.60pF
7V 1.50pF
8V 1.42pF

Cceb (E tied to B)
0V 4.29pF
0.5V 3.54pF
1V 3.16pF
2V 2.69pF
3V 2.40pF
4V 2.20pF
5V 2.05pF
6V 1.92pF
7V 1.83pF
8V 1.75pF


A thru-hole version was Ccb (e open):
0V 5.35pF
0.5V 4.55pF
1V 4.12pF
2V 3.56pF
3V 3.19pF
4V 2.95pF
5V 2.72pF
6V 2.52pF
7V 2.33pF
8V 2.16pF

A thru hole RF part MPSH10 Ccb (e open):
0V 1.23pF
0.5V 1.09pF
1V 1.02pF
2V 0.94pF
3V 0.89pF
4V 0.86pF
5V 0.84pF
6V 0.82pF
7V 0.8pF
8V 0.79pF


A 4.3V 400mW zener was:
0V 211pF
0.5V 183pF
1V 168pF
2V 149pF

A junk box T1 red LED:
0V 28.4pF
0.5V 25.9pF
1V 24.4pF
2V 22pF
3V 20.4pF
4V 19.1pF
5V 18pF
6V 17.3pF
7V 16.6pF
8V 16pF

* I tried the NPN bjts Ccb-e and Cb-ce i.e. involving the
base-emitter
junction - as expected the capacitance was higher at 12 to 16pF at
zero
bias falling to 6-8pF at 6V (didnt go higher to avoid breakdown)
but I
noticed that the low bias capacitance values looked jittery and only
settled down above about 2V.

My junkbox has some wire ended varicaps somewhere so next coffee
break I
will try those too.

piglet

Fun, If I can riff on the physics. The variable C is caused by
changes in the depletion width. So what sort of devices would have
a large variation?
PIN photodiodes?
maybe c-b junc of high voltage transistors?
What else?

George H.


Yes, one can get the depletion zone narrower (higher C) by approaching
forward bias but that quickly stops getting useful above 0.6 - 0.7V.

Heh - that reminds me try that Red LED in the pre-forward bias
direction. Thanks.

piglet


Besides C-ratio, the other important FOM for a varactor is Q.
Highly-doped diodes are the ticket for that.

Cheers

Phil Hobbs
n

Aren't emitters highly doped? Does that mean I could explore B-E
junctions as poor-man's varicaps?

piglet


That was commonly done BITD. Unfortunately Rbb' is generally much
higher than the emitter or collector resistances.

Cheers

Phil Hobbs

Ah yes of course, we had fun last year disecting the AD310 parametric
input op-amp which used back to back transdiode (CB-E) bjts as varactors.

piglet

 

[John Larkin]

On Thu, 02 Apr 2020 09:32:36 +0300, upsidedown@downunder.com wrote:

On Wed, 1 Apr 2020 16:30:55 -0700 (PDT), Phil Allison
pallison49@gmail.com> wrote:

Joerg wrote:

-------------

Up to about 30pf they'll likely remain. Skyworks and such. Those with
large capacitances lost their market.



** AM radio has disappeared ??

In many countries in Europe the domestic MW/SW stations have been
closed for more than a decade ago and also foreign services have also
been shut down a few years ago.

Domestic services using ground wave (and NVIS) was useful when the
distance between transmitter sites was large. These days when
cellular phones need quite close distances. there are also a large
number of sites for low or medium power FM/DAB transmitters and more
and more radio programs are received directly through the cellular
network.


News to me.


... Phil

50 kilowatts of radiated RF must be expensive. 5G will probably take
over everything at a huge energy saving.



--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[Gerhard Hoffmann]

Am 02.04.20 um 18:42 schrieb jlarkin@highlandsniptechnology.com:

I expect that negative base bias would modulate Cce, which makes a
3-terminal varicap. Somebody sells a ceramic-based 3-terminal varicap.

<
https://www.mouser.de/ProductDetail/STMicroelectronics/STPTIC-27L2C5?qs=%2Fha2pyFaduiW3aaIsUg5X3hLBHWOHwb4keLGrqq6Qt%2FsG81BTg1ojY8XtaXGdNNv
>

https://www.mouser.de/datasheet/2/389/dm00245248-1798943.pdf


This one? I have 10 pcs. here, but never found the time to play with them.

cheers, Gerhard

(Oh, yes, I use the SR620!)

 

[George Herold]

On Thursday, April 2, 2020 at 12:02:46 PM UTC-4, piglet wrote:

On 02/04/2020 16:11, Phil Hobbs wrote:
On 2020-04-02 10:51, Piglet wrote:
On 02/04/2020 15:45, Phil Hobbs wrote:
On 2020-04-02 10:41, Piglet wrote:
On 02/04/2020 15:28, George Herold wrote:
On Thursday, April 2, 2020 at 10:12:28 AM UTC-4, piglet wrote:
On 02/04/2020 14:29, jlarkin@highlandsniptechnology.com wrote:
On Thu, 2 Apr 2020 10:24:49 +0100, piglet <erichpwagner@hotmail.com
wrote:

On 01/04/2020 11:34 pm, John Larkin wrote:
On Wed, 1 Apr 2020 16:56:48 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-04-01 16:20, sea moss wrote:

I wonder if it's practical to use a MOSFET's Coss as a
varicap. I haven't
seen it done.

Standard technique in ASIC design

-- Kevin Aylward
http://www.anasoft.co.uk - SuperSpice
http://www.kevinaylward.co.uk/ee/index.html

Cool. I should have googled it first, there's a few examples
out there.

There are pots of low-pF ones on Digikey. The 1000-pFish ones
(MVAM108
etc.) are long gone, and the tens-to-hundreds range is
declining. My
fave MV209 is a distant memory, but BB201s are still in stock.

Cheers

Phil Hobbs

According to my Boonton, a BFT25 c-b junction only goes from
about 0.7
pF to around 0.65 from 0 to 6 volts. There's probably a bit of
fixture
capacitance too. Not a very good varicap.

I want a delta-c around 0.15 pF roughly, maybe 0.2. My dac+opamp
could
go to 20 volts.


You're going to need a bigger transistor. Try a BC847/2N3904 ?

piglet

I'm guessing that a varicap is doped to be, well, a varicap. They
have
delta-c catios like 6:1 over some voltage range. Looks like
transistor
junctions have much flatter c:v curves. So the tempco contribution
will be a lot higher if I use a transistor.

I've ordered some singles, Skyworks parts in the horrible SC79
package, and my 4-layer proto board should be here soon. If the
board
house doesn't shut down.




Yes I think you're right about the flatter c:v curves. You only want
0.2pF swing so will have quite a lot of padding down.

I hooked up an SMD BC848C to the Boonton and got these:
Ccb (E open*)
0V 4.0pF
0.5V 3.2pF
1V 2.8pF
2V 2.36pF
3V 2.05pF
4V 1.85pF
5V 1.71pF
6V 1.60pF
7V 1.50pF
8V 1.42pF

Cceb (E tied to B)
0V 4.29pF
0.5V 3.54pF
1V 3.16pF
2V 2.69pF
3V 2.40pF
4V 2.20pF
5V 2.05pF
6V 1.92pF
7V 1.83pF
8V 1.75pF


A thru-hole version was Ccb (e open):
0V 5.35pF
0.5V 4.55pF
1V 4.12pF
2V 3.56pF
3V 3.19pF
4V 2.95pF
5V 2.72pF
6V 2.52pF
7V 2.33pF
8V 2.16pF

A thru hole RF part MPSH10 Ccb (e open):
0V 1.23pF
0.5V 1.09pF
1V 1.02pF
2V 0.94pF
3V 0.89pF
4V 0.86pF
5V 0.84pF
6V 0.82pF
7V 0.8pF
8V 0.79pF


A 4.3V 400mW zener was:
0V 211pF
0.5V 183pF
1V 168pF
2V 149pF

A junk box T1 red LED:
0V 28.4pF
0.5V 25.9pF
1V 24.4pF
2V 22pF
3V 20.4pF
4V 19.1pF
5V 18pF
6V 17.3pF
7V 16.6pF
8V 16pF

* I tried the NPN bjts Ccb-e and Cb-ce i.e. involving the
base-emitter
junction - as expected the capacitance was higher at 12 to 16pF at
zero
bias falling to 6-8pF at 6V (didnt go higher to avoid breakdown)
but I
noticed that the low bias capacitance values looked jittery and only
settled down above about 2V.

My junkbox has some wire ended varicaps somewhere so next coffee
break I
will try those too.

piglet

Fun, If I can riff on the physics. The variable C is caused by
changes in the depletion width. So what sort of devices would have
a large variation?
PIN photodiodes?
maybe c-b junc of high voltage transistors?
What else?

George H.


Yes, one can get the depletion zone narrower (higher C) by approaching
forward bias but that quickly stops getting useful above 0.6 - 0.7V.

Heh - that reminds me try that Red LED in the pre-forward bias
direction. Thanks.

piglet


Besides C-ratio, the other important FOM for a varactor is Q.
Highly-doped diodes are the ticket for that.

Cheers

Phil Hobbs
n

Aren't emitters highly doped? Does that mean I could explore B-E
junctions as poor-man's varicaps?

piglet


That was commonly done BITD. Unfortunately Rbb' is generally much
higher than the emitter or collector resistances.

Cheers

Phil Hobbs


Ah yes of course, we had fun last year disecting the AD310 parametric
input op-amp which used back to back transdiode (CB-E) bjts as varactors.

piglet

Ahh, parametric amps... be fun to build one.
Is there some easy way to make a (demonstration) para-amp from
the really crappy ceramic caps?

George H.

 

[whit3rd]

On Thursday, April 2, 2020 at 8:28:16 AM UTC-7, jla...@highlandsniptechnology.com wrote:

I could use a high-capacitance part and pad it down with a series NPO,
but that's messy. I want a high C ratio part, like a real varicap, to
avoid the high baseline value of a low-slope diode.

But you need an isolation capacitor anyway, so the series NPO
is required. The series pair ( C2 *C2/(C1 + C2) ) is going to have
a low-slope characteristic regardless. Why worry, if you only need
a fraction of a pF as your delta? A main issue will be stray/PCB capacitance.

 

[piglet]

On 02/04/2020 6:05 pm, George Herold wrote:

On Thursday, April 2, 2020 at 12:02:46 PM UTC-4, piglet wrote:

Ah yes of course, we had fun last year disecting the AD310 parametric
input op-amp which used back to back transdiode (CB-E) bjts as varactors.

piglet

Ahh, parametric amps... be fun to build one.
Is there some easy way to make a (demonstration) para-amp from
the really crappy ceramic caps?

George H.

Varicap diode paramps (like the AD310 we enjoyed last year) have pF
scale capacitance and were driven in 00s kHz or lo-MHz but since the
ceramic caps with large voltage coefficient are uF range I like to think
one could get away with power line frequencies. Which opens up the
delightful steampunk idea I sketched here...

<https://www.dropbox.com/s/jqubvy92xuc2788/mains_paramp_idea.pdf?raw=1>

Of course if the input varactor caps aren't closely matched the offset
voltage could be of the order of volts but still the DC input bias
current is near zero! And 60Hz pump frequency rather limits GBW and
output slew rate!

piglet

 

[John Larkin]

On Thu, 2 Apr 2020 10:27:30 -0700 (PDT), whit3rd <whit3rd@gmail.com>
wrote:

On Thursday, April 2, 2020 at 8:28:16 AM UTC-7, jla...@highlandsniptechnology.com wrote:

I could use a high-capacitance part and pad it down with a series NPO,
but that's messy. I want a high C ratio part, like a real varicap, to
avoid the high baseline value of a low-slope diode.

But you need an isolation capacitor anyway, so the series NPO
is required. The series pair ( C2 *C2/(C1 + C2) ) is going to have
a low-slope characteristic regardless. Why worry, if you only need
a fraction of a pF as your delta? A main issue will be stray/PCB capacitance.

I currently have a dual common-anode varicap on the test layout. One
end is ground, one is the LC, and I'm applying the bias to the common
anode through a resistor. That makes for a fast control loop.

Our loop bandwidth will be around 250 KHz, and the frequency about 150
MHz, so an RC between the DAC and a single-ended varicap won't be a
problem.

I'm going to experiment with an active layer 2 guard to reduce the
effective PCB capacitance, specifically the horrible TC of FR4.



--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

AD310K
http://ixbt.photo/photo/358797/30781Lv6MstoAez/aAL2LDV3Bs/956874.png

 

[piglet]

On 02/04/2020 8:52 pm, Jeroen Belleman wrote:

On 2020-04-02 20:46, piglet wrote:
On 02/04/2020 6:05 pm, George Herold wrote:
On Thursday, April 2, 2020 at 12:02:46 PM UTC-4, piglet wrote:

Ah yes of course, we had fun last year disecting the AD310 parametric
input op-amp which used back to back transdiode (CB-E) bjts as
varactors.

piglet

Ahh, parametric amps...  be fun to build one.
Is there some easy way to make a (demonstration) para-amp from
the really crappy ceramic caps?

George H.


Varicap diode paramps (like the AD310 we enjoyed last year) have pF
scale capacitance and were driven in 00s kHz or lo-MHz but since the
ceramic caps with large voltage coefficient are uF range I like to think
one could get away with power line frequencies. Which opens up the
delightful steampunk idea I sketched here...

https://www.dropbox.com/s/jqubvy92xuc2788/mains_paramp_idea.pdf?raw=1


I see two problems: The input signal makes the capacitors
change in the same direction, so do not unbalance the bridge.
The diodes in the demodulator should be a ring quad, of course,
not a rectifier bridge.

This needs a bit more work.

Jeroen Belleman

Thanks Jeroen, yes I thought I got the ring modulator wrong. Might have
to rethink the front end! Idiot first pass solution might be to have one
capacitor fixed (loses all cancellation properties of identical parts
though).

piglet

 

[Jeroen Belleman]

On 2020-04-02 20:46, piglet wrote:

On 02/04/2020 6:05 pm, George Herold wrote:
On Thursday, April 2, 2020 at 12:02:46 PM UTC-4, piglet wrote:

Ah yes of course, we had fun last year disecting the AD310 parametric
input op-amp which used back to back transdiode (CB-E) bjts as
varactors.

piglet

Ahh, parametric amps... be fun to build one.
Is there some easy way to make a (demonstration) para-amp from
the really crappy ceramic caps?

George H.


Varicap diode paramps (like the AD310 we enjoyed last year) have pF
scale capacitance and were driven in 00s kHz or lo-MHz but since the
ceramic caps with large voltage coefficient are uF range I like to think
one could get away with power line frequencies. Which opens up the
delightful steampunk idea I sketched here...

https://www.dropbox.com/s/jqubvy92xuc2788/mains_paramp_idea.pdf?raw=1

I see two problems: The input signal makes the capacitors
change in the same direction, so do not unbalance the bridge.
The diodes in the demodulator should be a ring quad, of course,
not a rectifier bridge.

This needs a bit more work.

Jeroen Belleman

 

[DemonicTubes]

On Wednesday, April 1, 2020 at 3:01:15 PM UTC-6, John Larkin wrote:

On Wed, 1 Apr 2020 12:39:39 -0700 (PDT), sea moss
danluster81@gmail.com> wrote:

On Wednesday, April 1, 2020 at 12:02:21 PM UTC-7, John Larkin wrote:
They seem to be getting obsolete. Maybe I can cheat and use a regular
pn diode.

I wonder if it's practical to use a MOSFET's Coss as a varicap. I haven't seen it done.

I need a small range, maybe a change of 0.2 pF over 5 volts bias
swing. The c-b junction of some small transistor might work. I think
varicaps have special doping profiles to tune the c-v curve, but I'll
be working at fairly high voltages, 10V maybe, and that looks to be
out of the hyperabrupt zone. I'd like a constant voltage-frequency
curve, but that ain't going to happen.

I'm tuning a 150 MHz LC oscillator a few thousand PPM, part of a
bizarre digital PLL.

Looks like I can get some low-capacitance Skyworks singles in SC79.



--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

Would a 1SV280 work, or am I misunderstanding your needs?

1.5 - 2.0 pF @ 10 V
3.8 - 4.7 pF @ 2 V

 

[John Larkin]

On Thu, 2 Apr 2020 14:58:33 -0700 (PDT), DemonicTubes
<tlackie@gmail.com> wrote:

On Wednesday, April 1, 2020 at 3:01:15 PM UTC-6, John Larkin wrote:
On Wed, 1 Apr 2020 12:39:39 -0700 (PDT), sea moss
danluster81@gmail.com> wrote:

On Wednesday, April 1, 2020 at 12:02:21 PM UTC-7, John Larkin wrote:
They seem to be getting obsolete. Maybe I can cheat and use a regular
pn diode.

I wonder if it's practical to use a MOSFET's Coss as a varicap. I haven't seen it done.

I need a small range, maybe a change of 0.2 pF over 5 volts bias
swing. The c-b junction of some small transistor might work. I think
varicaps have special doping profiles to tune the c-v curve, but I'll
be working at fairly high voltages, 10V maybe, and that looks to be
out of the hyperabrupt zone. I'd like a constant voltage-frequency
curve, but that ain't going to happen.

I'm tuning a 150 MHz LC oscillator a few thousand PPM, part of a
bizarre digital PLL.

Looks like I can get some low-capacitance Skyworks singles in SC79.



--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

Would a 1SV280 work, or am I misunderstanding your needs?

1.5 - 2.0 pF @ 10 V
3.8 - 4.7 pF @ 2 V

That changes about half a pF from 8 to 15 volts. Not bad, but a bit
more than I'd like. Skyworks SMV1248 might be better.

Why do people make 15 volt varicaps and only give cv data up to 8
volts?


--

John Larkin Highland Technology, Inc
picosecond timing precision measurement

jlarkin att highlandtechnology dott com
http://www.highlandtechnology.com

 

[Clifford Heath]

On 3/4/20 1:28 am, George Herold wrote:

On Thursday, April 2, 2020 at 10:12:28 AM UTC-4, piglet wrote:
On 02/04/2020 14:29, jlarkin@highlandsniptechnology.com wrote:
On Thu, 2 Apr 2020 10:24:49 +0100, piglet <erichpwagner@hotmail.com
wrote:

On 01/04/2020 11:34 pm, John Larkin wrote:
On Wed, 1 Apr 2020 16:56:48 -0400, Phil Hobbs
pcdhSpamMeSenseless@electrooptical.net> wrote:

On 2020-04-01 16:20, sea moss wrote:

I wonder if it's practical to use a MOSFET's Coss as a varicap. I haven't
seen it done.

Standard technique in ASIC design

-- Kevin Aylward
http://www.anasoft.co.uk - SuperSpice
http://www.kevinaylward.co.uk/ee/index.html

Cool. I should have googled it first, there's a few examples out there.

There are pots of low-pF ones on Digikey. The 1000-pFish ones (MVAM108
etc.) are long gone, and the tens-to-hundreds range is declining. My
fave MV209 is a distant memory, but BB201s are still in stock.

Cheers

Phil Hobbs

According to my Boonton, a BFT25 c-b junction only goes from about 0.7
pF to around 0.65 from 0 to 6 volts. There's probably a bit of fixture
capacitance too. Not a very good varicap.

I want a delta-c around 0.15 pF roughly, maybe 0.2. My dac+opamp could
go to 20 volts.


You're going to need a bigger transistor. Try a BC847/2N3904 ?

piglet

I'm guessing that a varicap is doped to be, well, a varicap. They have
delta-c catios like 6:1 over some voltage range. Looks like transistor
junctions have much flatter c:v curves. So the tempco contribution
will be a lot higher if I use a transistor.

I've ordered some singles, Skyworks parts in the horrible SC79
package, and my 4-layer proto board should be here soon. If the board
house doesn't shut down.




Yes I think you're right about the flatter c:v curves. You only want
0.2pF swing so will have quite a lot of padding down.

I hooked up an SMD BC848C to the Boonton and got these:
Ccb (E open*)
0V 4.0pF
0.5V 3.2pF
1V 2.8pF
2V 2.36pF
3V 2.05pF
4V 1.85pF
5V 1.71pF
6V 1.60pF
7V 1.50pF
8V 1.42pF

Cceb (E tied to B)
0V 4.29pF
0.5V 3.54pF
1V 3.16pF
2V 2.69pF
3V 2.40pF
4V 2.20pF
5V 2.05pF
6V 1.92pF
7V 1.83pF
8V 1.75pF


A thru-hole version was Ccb (e open):
0V 5.35pF
0.5V 4.55pF
1V 4.12pF
2V 3.56pF
3V 3.19pF
4V 2.95pF
5V 2.72pF
6V 2.52pF
7V 2.33pF
8V 2.16pF

A thru hole RF part MPSH10 Ccb (e open):
0V 1.23pF
0.5V 1.09pF
1V 1.02pF
2V 0.94pF
3V 0.89pF
4V 0.86pF
5V 0.84pF
6V 0.82pF
7V 0.8pF
8V 0.79pF


A 4.3V 400mW zener was:
0V 211pF
0.5V 183pF
1V 168pF
2V 149pF

A junk box T1 red LED:
0V 28.4pF
0.5V 25.9pF
1V 24.4pF
2V 22pF
3V 20.4pF
4V 19.1pF
5V 18pF
6V 17.3pF
7V 16.6pF
8V 16pF

* I tried the NPN bjts Ccb-e and Cb-ce i.e. involving the base-emitter
junction - as expected the capacitance was higher at 12 to 16pF at zero
bias falling to 6-8pF at 6V (didnt go higher to avoid breakdown) but I
noticed that the low bias capacitance values looked jittery and only
settled down above about 2V.

My junkbox has some wire ended varicaps somewhere so next coffee break I
will try those too.

piglet

Fun, If I can riff on the physics. The variable C is caused by
changes in the depletion width. So what sort of devices would have
a large variation?
PIN photodiodes?
maybe c-b junc of high voltage transistors?
What else?

High voltage rectifier diodes

 

[Clifford Heath]

On 2/4/20 11:12 pm, Jeroen Belleman wrote:

On 2020-04-02 13:11, Chris Jones wrote:
On 02/04/2020 19:27, Phil Allison wrote:
  Chris Jones wrote:
-------------------



** AM radio has disappeared ??

     News to me.


...   Phil



If someone were designing an AM radio these days, quite likely they
would use a chip for the radio receiver circuitry,


** Pure speculation - not fact.

Well perhaps we should speculate instead that the chip companies make
the integrated radio ICs as some kind of decorative ornament, rather
than something that is sold in quantity for use in radio receivers. No,
we shouldn't speculate that, it is surely fact!

Times have changed for sure. Look at the Silicon Labs Si4735
for example. Just one chip for LW, SW, AM and FM bands. Oh yes,
it does RDS too. No filter coils, no varicaps. A radio with one
of these looks nothing like what radios used to look like inside.
The only giveaway they still retained is a ferrite antenna rod.

The manufacturer has a software patch for the DSP that adds support for
SSB, as well.

CH

 

[Joerg]

On 2020-04-01 17:29, Rick C wrote:

On Wednesday, April 1, 2020 at 7:31:01 PM UTC-4, Phil Allison wrote:
Joerg wrote:

-------------

Up to about 30pf they'll likely remain. Skyworks and such. Those
with large capacitances lost their market.



** AM radio has disappeared ??

News to me.

My car doesn't even receive AM, only FM, Sirius XM and Slacker
Internet radio.

I would not want that radio in my car. I listen to AM most of the time
when driving, and not just then.

.. Oh, and bluetooth from your phone. But no AM radio.
I think they could do that 100% in software if they wanted to. It's
only 1600 kHz max frequency.

Well, they don't go that far but everything is PLL or DDS nowadays.
There simply isn't a need any longer for large capacitance varicaps to
run the local oscillator. As for the preselector (adjustable filter in
front of the mixer), that has fallen victim to the red pencil of the
bean counter, which has cost performance when in the vicinity of other
strong signals.

--
Regards, Joerg

http://www.analogconsultants.com/

 

[Phil Allison]

upsid...@downunder.com wrote:

-------------------------------

** So misses out on 4000 or so stations in the US alone.

AM broadcast will never disappear, it's just too damn useful.

Here in Europe it has disappeared without a trace.


** Really??


https://en.wikipedia.org/wiki/List_of_European_medium_wave_transmitters

Some of which are quite low power and some are operated only once a
month.

** He says, clutching at straws.

** AM has major technical advantages in relation to range and simplicity/ reliability of receiving equipment.


Why do you need a large range,

** To get to lots of people - fuckhead.

when there are a huge number of
transmitter sites due to the cellular networks. ?

** One transmitter covers a massive area for a tiny cost.

Are you being tedious or the sake of it again?

Here in Australia we have thousands of AM transmitters, private and Govt owned and rely on them at times of emergency to get crucial info out to the public.

Example the recent, severe bush fires.

Do they have their own emergency generators capable of running for at
least a weak ?

** During the recent fires, crucial parts of the cellular network went out and folk had no way to charge their smart phones.

There was no such possibility with broadcast AM radio.

Why are you being such an ass?


..... Phil

 

[Joerg]

On 2020-04-02 00:01, Rick C wrote:

On Thursday, April 2, 2020 at 2:23:36 AM UTC-4, Chris Jones wrote:
On 02/04/2020 10:30, Phil Allison wrote:
Joerg wrote:

-------------

Up to about 30pf they'll likely remain. Skyworks and such.
Those with large capacitances lost their market.



** AM radio has disappeared ??

News to me.


... Phil



If someone were designing an AM radio these days, quite likely
they would use a chip for the radio receiver circuitry, and those
quite likely would have a PLL with the VCO integrated on the chip,
e.g. TDA7786. So, the market for discrete varactors suited to the
AM band may have gone away regardless of whether AM radios stay
popular.

Maybe for the base level Honda Civic and the Yugo. These days you
get an XM radio if you want it or not and the same electronics
receives AM/FM without an extra chip. That is likely why AM was left
out of the Tesla radio. It's not worth the bother of adding the
passives required to filter the band.

Maybe because Tesla drivers aren't supposed to be listening to Rush
Limbaugh?

--
Regards, Joerg

http://www.analogconsultants.com/