oscillator design with a long xtal lead

[E Minor 7]

Hello,

I am digging into a project requires a 10MHz crystal drives in vaccum
whereas the oscillator circuit oscillates at a 1 meter distance.


crystals terminals are connected to the signal and shield of a 1 meter
coaxial, rg316.


However my current colplitts design doesn't work now


1 meter rg316 coaxial
_______ . .... _______[op amp] __
| | |
| = (30pF) |
xtal |-------------------- | F.B.
| = (30pF)
|_______ ...... _____|


op amp is a HFA1130, non inverting config with gain about 2X


xtal intended to use as an inductor and oscillates with the 2x30pF
caps. and the feedback inserts in the mid way of the caps.


The design works OK with zero lenght, but doesn't work with a long
lead.


Does the phase delay / cap effect of the coaxial affecting the whole
matter. ?


Thanks for great help


EMinorSeven

 

[Andrew Holme]

Hello,

I am digging into a project requires a 10MHz crystal drives in vaccum
whereas the oscillator circuit oscillates at a 1 meter distance.


crystals terminals are connected to the signal and shield of a 1 meter
coaxial, rg316.


However my current colplitts design doesn't work now


1 meter rg316 coaxial
_______ . .... _______[op amp] __
| | |
| = (30pF) |
xtal |-------------------- | F.B.
| = (30pF)
|_______ ...... _____|


op amp is a HFA1130, non inverting config with gain about 2X


xtal intended to use as an inductor and oscillates with the 2x30pF
caps. and the feedback inserts in the mid way of the caps.


The design works OK with zero lenght, but doesn't work with a long
lead.


Does the phase delay / cap effect of the coaxial affecting the whole
matter. ?

You might try cutting the coax to a quarter wavelength. Don't forget
velocity factor.

Why can't you put the whole oscillator in the vacum?

 

[Rene Tschaggelar]

Hello,

I am digging into a project requires a 10MHz crystal drives in vaccum
whereas the oscillator circuit oscillates at a 1 meter distance.
[snip]
The design works OK with zero lenght, but doesn't work with a long
lead.

Does the phase delay / cap effect of the coaxial affecting the whole
matter. ?

Apparently yes. What is wrong with having the
oscillator in tha vacuum too ?

Rene

 

[Phil Hobbs]

Hello,

I am digging into a project requires a 10MHz crystal drives in vaccum
whereas the oscillator circuit oscillates at a 1 meter distance.
[snip]
The design works OK with zero lenght, but doesn't work with a long
lead.

Does the phase delay / cap effect of the coaxial affecting the whole
matter. ?

Apparently yes. What is wrong with having the
oscillator in tha vacuum too ?

Rene

Normally the problem is outgassing. With attention to the cooling, you
can make the oscillator work in the vacuum--but often you can't make the
vacuum work with the oscillator.

This sort of oscillator is used in quartz film thickness gauges, for
example. Tweaking the cable length is a good idea, but a half
wavelength will work better than a quarter--it will reproduce the
crystal impedance, whereas a quarter wave produces Z0**2/ZL. A quarter
wave will make the crystal look inductive except between its series and
parallel resonances, which will cause the oscillator to not work.

Cheers,

Phil Hobbs

 

[Rene Tschaggelar]

I am digging into a project requires a 10MHz crystal drives in vaccum
whereas the oscillator circuit oscillates at a 1 meter distance.
[snip]
The design works OK with zero lenght, but doesn't work with a long
lead.

Does the phase delay / cap effect of the coaxial affecting the whole
matter. ?

Apparently yes. What is wrong with having the
oscillator in tha vacuum too ?

Normally the problem is outgassing. With attention to the cooling, you
can make the oscillator work in the vacuum--but often you can't make the
vacuum work with the oscillator.

This sort of oscillator is used in quartz film thickness gauges, for
example. Tweaking the cable length is a good idea, but a half
wavelength will work better than a quarter--it will reproduce the
crystal impedance, whereas a quarter wave produces Z0**2/ZL. A quarter
wave will make the crystal look inductive except between its series and
parallel resonances, which will cause the oscillator to not work.

A coil, a cap and a SOT23 together with the XTAL
should be sufficient. No ? It the SOT23 producing
gasses ?

Rene

 

[colin]

Hello,

I am digging into a project requires a 10MHz crystal drives in vaccum
whereas the oscillator circuit oscillates at a 1 meter distance.


crystals terminals are connected to the signal and shield of a 1 meter
coaxial, rg316.


However my current colplitts design doesn't work now


1 meter rg316 coaxial
_______ . .... _______[op amp] __
| | |
| = (30pF) |
xtal |-------------------- | F.B.
| = (30pF)
|_______ ...... _____|


op amp is a HFA1130, non inverting config with gain about 2X


xtal intended to use as an inductor and oscillates with the 2x30pF
caps. and the feedback inserts in the mid way of the caps.


The design works OK with zero lenght, but doesn't work with a long
lead.


Does the phase delay / cap effect of the coaxial affecting the whole
matter. ?


Thanks for great help


EMinorSeven

The phase delay is about 30 degrees wich if it was just afecting the loop
delay it might just still work, however the impedance seen looking into the
op amp end of the coax will look nothing like the crystal impedance your
circuit needs, it will add a significant amnount of capacitance or
inductance depending on the frequency, it might be as well to model it in a
simulator, I assume your trying to operate the crystal in its inductive
region, however if you can arange to use the crystal in its series resonant
mode you might be a better off, especialy if your crystal has an esr of
about 50ohms, the cable would then not have any effect on the circuit at all
at the frequency of operation, you could then use a butler circuit.

Another thing is to try diferent lenghts of coax to see how long it can be
before it stops working.

Colin =^.^=

 

[Tim Wescott]

Hello,

I am digging into a project requires a 10MHz crystal drives in vaccum
whereas the oscillator circuit oscillates at a 1 meter distance.


crystals terminals are connected to the signal and shield of a 1 meter
coaxial, rg316.


However my current colplitts design doesn't work now


1 meter rg316 coaxial
_______ . .... _______[op amp] __
| | |
| = (30pF) |
xtal |-------------------- | F.B.
| = (30pF)
|_______ ...... _____|


op amp is a HFA1130, non inverting config with gain about 2X


xtal intended to use as an inductor and oscillates with the 2x30pF
caps. and the feedback inserts in the mid way of the caps.


The design works OK with zero lenght, but doesn't work with a long
lead.


Does the phase delay / cap effect of the coaxial affecting the whole
matter. ?


Thanks for great help


EMinorSeven

The easiest way to answer this in a short post is to refer you to
"Oscillator Design & Computer Simulation" by Randall W. Rhea. It's
currently out of stock at my favorite bookstore, but you can give a try
at ordering it anyway:
http://www.powells.com/partner/30696/biblio/1884932304. Amazon may have
something, too.

The book gives a systematic way of designing oscillators with linearized
circuit models.

Or you can look at the impedance present at the end of the cable at your
desired frequency -- with that cable length it will look like the
crystal in parallel with a honking big capacitor. You may be able to
fudge it just by putting an equivalent inductive reactance in parallel
with the cable -- or you may just get weird mode changes. Rhea's book
will tell you how to verify operation one way or another.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

 

[Joerg]

Hello Rene,

A coil, a cap and a SOT23 together with the XTAL
should be sufficient. No ? It the SOT23 producing
gasses ?

Only if it had chili con carne for lunch ;-)

You can encapsulate the oscillator and connect the crystal via
glass-insulated terminals if needed. Encapsulation is quite common for
medical implants where for obvious reasons outgassing would not be cool.

Of course if the medium to be measured is too hot that's another story.

Regards, Joerg

 

[Tim Wescott]

E Minor 7 wrote:


I am digging into a project requires a 10MHz crystal drives in vaccum
whereas the oscillator circuit oscillates at a 1 meter distance.
[snip]
The design works OK with zero lenght, but doesn't work with a long
lead.

Does the phase delay / cap effect of the coaxial affecting the whole
matter. ?


Apparently yes. What is wrong with having the
oscillator in tha vacuum too ?

Normally the problem is outgassing. With attention to the cooling,
you can make the oscillator work in the vacuum--but often you can't
make the vacuum work with the oscillator.

This sort of oscillator is used in quartz film thickness gauges, for
example. Tweaking the cable length is a good idea, but a half
wavelength will work better than a quarter--it will reproduce the
crystal impedance, whereas a quarter wave produces Z0**2/ZL. A
quarter wave will make the crystal look inductive except between its
series and parallel resonances, which will cause the oscillator to not
work.

A coil, a cap and a SOT23 together with the XTAL
should be sufficient. No ? It the SOT23 producing
gasses ?

Rene

In a high vacuum _everything_ produces gasses. You can't stop it, you
can only try to keep out things that persistently outgas. The amount of
outgassing you have to be concerned with varies with the level of vacuum
you're trying to attain, but things can get pretty nuts.

--

Tim Wescott, who's only watched high vacuum efforts from afar.
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

 

[Tim Wescott]

Hello Rene,


Only if it had chili con carne for lunch ;-)

You can encapsulate the oscillator and connect the crystal via
glass-insulated terminals if needed. Encapsulation is quite common for
medical implants where for obvious reasons outgassing would not be cool.

Of course if the medium to be measured is too hot that's another story.

Regards, Joerg

I suspect that the level of outgassing that would be absolutely
unacceptable for a high-vacuum application would be orders of magnitude
below the radar screen in many medical applications -- particularly if
the gas involved was biologically benign, like oxygen, water or hydrogen.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

 

[Joerg]

Hello Tim,

I suspect that the level of outgassing that would be absolutely
unacceptable for a high-vacuum application would be orders of magnitude
below the radar screen in many medical applications -- particularly if
the gas involved was biologically benign, like oxygen, water or hydrogen.

True. But then I wonder how he is going to get a cable or connector in
there that won't be outgassing.

Regards, Joerg

 

[Michael A. Terrell]

Hello Tim,


True. But then I wonder how he is going to get a cable or connector in
there that won't be outgassing.

Regards, Joerg

How about magnetic coupling through a glass window?


--
Service to my country? Been there, Done that, and I've got my DD214 to
prove it.
Member of DAV #85.

Michael A. Terrell
Central Florida

 

[Tim Wescott]

Hello Tim,



True. But then I wonder how he is going to get a cable or connector in
there that won't be outgassing.

When they put IR imaging arrays into dewars they use wire pins going
through glass beads set in a metal disk. I _think_ they fuse the whole
thing together, but they may just seal them by clamping (or it may be
that different companies do different things).

Metal and glass are pretty good for outgassing -- you only need to keep
them hot with the vacuum pump going for a day or five. Some organic
materials can be very bad indeed.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

 

[Phil Hobbs]

E Minor 7 wrote:


I am digging into a project requires a 10MHz crystal drives in vaccum
whereas the oscillator circuit oscillates at a 1 meter distance.
[snip]
The design works OK with zero lenght, but doesn't work with a long
lead.

Does the phase delay / cap effect of the coaxial affecting the whole
matter. ?


Apparently yes. What is wrong with having the
oscillator in tha vacuum too ?

Normally the problem is outgassing. With attention to the cooling,
you can make the oscillator work in the vacuum--but often you can't
make the vacuum work with the oscillator.

This sort of oscillator is used in quartz film thickness gauges, for
example. Tweaking the cable length is a good idea, but a half
wavelength will work better than a quarter--it will reproduce the
crystal impedance, whereas a quarter wave produces Z0**2/ZL. A
quarter wave will make the crystal look inductive except between its
series and parallel resonances, which will cause the oscillator to not
work.

A coil, a cap and a SOT23 together with the XTAL
should be sufficient. No ? It the SOT23 producing
gasses ?

Rene

High and ultrahigh vacuum is a world to itself, and the rules are really
different. I don't know what sort of vacuum the OP is contemplating,
but if it's below 10**-6 torr, the cable idea is the right one. You can
get low-outgassing cable, usually polyimide-insulated, with no jacket.
Cables like that (and water-cooled holders for quartz oscillator
crystals) are common throwaway vacuum system parts.

The circuit board is going to be the big killer. Also lots of vacuum
systems need to be baked out at > 200C for hours and hours, which would
get flux and various FR-4 exhalations all over *everything*. You could
probably do this, e.g. on a sintered alumina substrate, but it wouldn't
be a slam dunk, and you'd have to make it cleanable--really cleanable.

Cheers,

Phil Hobbs

 

[Joerg]

Hello Tim,

When they put IR imaging arrays into dewars they use wire pins going
through glass beads set in a metal disk. I _think_ they fuse the whole
thing together, but they may just seal them by clamping (or it may be
that different companies do different things).

If they can get whole IR iamging sensors in there it shouldn't be a big
deal to get a wee oscillator in there. It could be encapsulated in
glass. I have some very low frequency crystals that are housed in glass
and there is at least 1/4 of a cubic inch of space in there.

The simplest oscillator to run with a crystal would be a tiny unbuffered
logic single-inverter and a resistor. Pretty much the size of a few
grains of salt.

Regards, Joerg

 

[Joerg]

Hello Michael,

How about magnetic coupling through a glass window?

That is no problem, usually. You could feed power on one frequency and
the 10MHz comes back. But one has to mind the folks from the FCC.
13.56MHz would be an ISM frequency but 10MHz isn't. Actually 10MHz is
one of the frequencies where NIST transmits the time standard and they
won't be enthused if someone messes with that.

I always try to remain 100% ISM with magnetic coupling. Makes me sleep
better at night ;-)

Regards, Joerg

 

[Tim Wescott]

Hello Tim,


If they can get whole IR iamging sensors in there it shouldn't be a big
deal to get a wee oscillator in there. It could be encapsulated in
glass. I have some very low frequency crystals that are housed in glass
and there is at least 1/4 of a cubic inch of space in there.

The simplest oscillator to run with a crystal would be a tiny unbuffered
logic single-inverter and a resistor. Pretty much the size of a few
grains of salt.

Regards, Joerg

True, and for a few more grains of salt you could make it so you could
supply DC power on the same coax that delivers 10MHz.

--

Tim Wescott
Wescott Design Services
http://www.wescottdesign.com

Posting from Google? See http://cfaj.freeshell.org/google/

 

[Joerg]

Hello Phil,

High and ultrahigh vacuum is a world to itself, and the rules are really
different. I don't know what sort of vacuum the OP is contemplating,
but if it's below 10**-6 torr, the cable idea is the right one. You can
get low-outgassing cable, usually polyimide-insulated, with no jacket.
Cables like that (and water-cooled holders for quartz oscillator
crystals) are common throwaway vacuum system parts.

The circuit board is going to be the big killer. Also lots of vacuum
systems need to be baked out at > 200C for hours and hours, which would
get flux and various FR-4 exhalations all over *everything*. You could
probably do this, e.g. on a sintered alumina substrate, but it wouldn't
be a slam dunk, and you'd have to make it cleanable--really cleanable.

Ok then, here is another idea: Create a low end impedance analyzer on
the outside and instead of oscillating monitor the crystal parameters.
With clever software it should be possible to compensate for the cable
characteristics and their changes with temperature.

Regards, Joerg

 

[Ben Jackson]

I am digging into a project requires a 10MHz crystal drives in vaccum
whereas the oscillator circuit oscillates at a 1 meter distance.

Why can't you move the oscillator out to the crystal and drive the
coax from outside?

 

[Rich Grise]

When they put IR imaging arrays into dewars they use wire pins going
through glass beads set in a metal disk. I _think_ they fuse the whole
thing together, but they may just seal them by clamping (or it may be
that different companies do different things).

Metal and glass are pretty good for outgassing -- you only need to keep
them hot with the vacuum pump going for a day or five. Some organic
materials can be very bad indeed.

I once had a gig where I worked around Ultra-High-Vacuum systems, and one
day one of my coworkers came into the lab just livid - "These feedthroughs
- they're _brass_!!" It seems zinc outgasses terribly and contaminates the
whole system. But, that's "ultra-high" vacuum - the kind where you can
manipulate individual atoms and do molecular-beam epitaxy and stuff, so
needs to be excruciatingly clean.

Cheers!
Rich