Hi,

I am wanting to match the output of a SA602 Mixer to a 45MHz, 4-pole
crystal filter and then match the output of the filter to the input
of a SA605 IF receiver.

The filter datasheet specifies the termination impedance of 800 Ohms and
3pF.

I have not used crystal filters before and would appreciate some advice
for typology of a suitable match.

The SA605 has a input impedance of 4500 Ohms par 3.5pF. Apparently they
have a "sweet spot" if matched to 800R.
The SA602 has output impedance of 1500 Ohm (not sure of capacitance).

From what I can see, if I placed a 1k resistor across the SA605 input,
this would give me close to 800R//3pF but not max. transfer of signal.

Some of the confusion I have is what to do with the 3pF termination
requirement. I am not sure if this means that I could match the
resistive components using an "L" or "Tapped C match" and then add
another 3pF to each side of the filter ???


Any help much appreciated

Regards

David

David wrote:
Hi,

I am wanting to match the output of a SA602 Mixer to a 45MHz, 4-pole
crystal filter and then match the output of the filter to the input
of a SA605 IF receiver.

The filter datasheet specifies the termination impedance of 800 Ohms and
3pF.

I have not used crystal filters before and would appreciate some advice
for typology of a suitable match.

The SA605 has a input impedance of 4500 Ohms par 3.5pF. Apparently they
have a "sweet spot" if matched to 800R.
The SA602 has output impedance of 1500 Ohm (not sure of capacitance).

From what I can see, if I placed a 1k resistor across the SA605 input,
this would give me close to 800R//3pF but not max. transfer of signal.

Some of the confusion I have is what to do with the 3pF termination
requirement. I am not sure if this means that I could match the
resistive components using an "L" or "Tapped C match" and then add
another 3pF to each side of the filter ???


Any help much appreciated

Regards

David


Andy writes:

You've generally got it right....
A xtal filter wants to see a particular input SOURCE resistance and
a particular LOAD resistance to develop the passband it was designed
for....

If your source has an output impedance of 4500 ohms, you would
want to transform it down to the 800 ohms specified. This can be
done in a transformer, but I usually do it with a split capacitor tank
circuit..... making the inductor the tuneable part.

If the load wants to "see" a capacitance, I usually use a 2-20pf
trimmer , together with whatever else impedance transformation
is available. Then I can tune the variable cap for the desired
passband characterists... Note that both the input and the output
are tuned in my technique....

If you don't have a sweep generator, you can just sweep the
gen you have back and forth to see if there are any ripples in
the passband.... Optimum is generally with the minimum ripples...
So, first tune the input and output for max output... Then rock
the generator to check for ripples... Often a very small adjustment
in the tuning will minimize the ripples at that point...

It ain't rocket surgery, but if you try to use fixed components
it ain't gonna happen, since you don't know your circuit strays....

Andy W4OAH

Andy,

Thanks for the information. So my matching circuit may look like....

Mixer Out(1k5) - Series coupling cap(1n) - Var L to ground - C-Tap
across L - Crystal Filter - C-Tap - var L across C-Tap - Coupling
C(1n) - SA605 Input (4500//3.5 pF)

C-Tap caps to ground are also tuneable.

I suppose the Q of the match would not need to be particularly high as
the selectivity is provided by the Crystal filter.

Regards

David

Andy wrote:
David wrote:
Hi,

I am wanting to match the output of a SA602 Mixer to a 45MHz, 4-pole
crystal filter and then match the output of the filter to the input
of a SA605 IF receiver.

The filter datasheet specifies the termination impedance of 800 Ohms and
3pF.

I have not used crystal filters before and would appreciate some advice
for typology of a suitable match.

The SA605 has a input impedance of 4500 Ohms par 3.5pF. Apparently they
have a "sweet spot" if matched to 800R.
The SA602 has output impedance of 1500 Ohm (not sure of capacitance).

From what I can see, if I placed a 1k resistor across the SA605 input,
this would give me close to 800R//3pF but not max. transfer of signal.

Some of the confusion I have is what to do with the 3pF termination
requirement. I am not sure if this means that I could match the
resistive components using an "L" or "Tapped C match" and then add
another 3pF to each side of the filter ???


Any help much appreciated

Regards

David


Andy writes:

You've generally got it right....
A xtal filter wants to see a particular input SOURCE resistance and
a particular LOAD resistance to develop the passband it was designed
for....

If your source has an output impedance of 4500 ohms, you would
want to transform it down to the 800 ohms specified. This can be
done in a transformer, but I usually do it with a split capacitor tank
circuit..... making the inductor the tuneable part.

If the load wants to "see" a capacitance, I usually use a 2-20pf
trimmer , together with whatever else impedance transformation
is available. Then I can tune the variable cap for the desired
passband characterists... Note that both the input and the output
are tuned in my technique....

If you don't have a sweep generator, you can just sweep the
gen you have back and forth to see if there are any ripples in
the passband.... Optimum is generally with the minimum ripples...
So, first tune the input and output for max output... Then rock
the generator to check for ripples... Often a very small adjustment
in the tuning will minimize the ripples at that point...

It ain't rocket surgery, but if you try to use fixed components
it ain't gonna happen, since you don't know your circuit strays....

Andy W4OAH

David wrote:

(top posting fixed)

Andy wrote:

David wrote:

Hi,

I am wanting to match the output of a SA602 Mixer to a 45MHz, 4-pole
crystal filter and then match the output of the filter to the input
of a SA605 IF receiver.

The filter datasheet specifies the termination impedance of 800 Ohms and
3pF.

I have not used crystal filters before and would appreciate some advice
for typology of a suitable match.

The SA605 has a input impedance of 4500 Ohms par 3.5pF. Apparently they
have a "sweet spot" if matched to 800R.
The SA602 has output impedance of 1500 Ohm (not sure of capacitance).

From what I can see, if I placed a 1k resistor across the SA605 input,
this would give me close to 800R//3pF but not max. transfer of signal.

Some of the confusion I have is what to do with the 3pF termination
requirement. I am not sure if this means that I could match the
resistive components using an "L" or "Tapped C match" and then add
another 3pF to each side of the filter ???


Any help much appreciated

Regards

David



Andy writes:

You've generally got it right....
A xtal filter wants to see a particular input SOURCE resistance and
a particular LOAD resistance to develop the passband it was designed
for....

If your source has an output impedance of 4500 ohms, you would
want to transform it down to the 800 ohms specified. This can be
done in a transformer, but I usually do it with a split capacitor tank
circuit..... making the inductor the tuneable part.

If the load wants to "see" a capacitance, I usually use a 2-20pf
trimmer , together with whatever else impedance transformation
is available. Then I can tune the variable cap for the desired
passband characterists... Note that both the input and the output
are tuned in my technique....

If you don't have a sweep generator, you can just sweep the
gen you have back and forth to see if there are any ripples in
the passband.... Optimum is generally with the minimum ripples...
So, first tune the input and output for max output... Then rock
the generator to check for ripples... Often a very small adjustment
in the tuning will minimize the ripples at that point...

It ain't rocket surgery, but if you try to use fixed components
it ain't gonna happen, since you don't know your circuit strays....

Andy W4OAH

Andy,

Thanks for the information. So my matching circuit may look like....

Mixer Out(1k5) - Series coupling cap(1n) - Var L to ground - C-Tap
across L - Crystal Filter - C-Tap - var L across C-Tap - Coupling
C(1n) - SA605 Input (4500//3.5 pF)

C-Tap caps to ground are also tunable.

I suppose the Q of the match would not need to be particularly high as
the selectivity is provided by the Crystal filter.

Regards

David

Something like this:
Filter
.-------.
| | To '605
.---o----------o-------| |-----
'602 | | | | |
.-------. | --- | '-------'
| |---' --- | |
| | | C| |
| | o-----. C| ===
| | | | C| GND
| |---. --- === |
'-------' | --- GND |
| | |
'---o----------'
(created by AACircuit v1.28.6 beta 04/19/05 www.tech-chat.de)

with an equivalent network going to the '605. You need enough Q to make
the circuit appear balanced -- you're depending on the Q of the tank to
overwhelm the inherent imbalance of the filter impedance. You _should_
be able to do this just by adjusting the coil, but I haven't tried it
with crystal filters, so what do I know?

--

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

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

"Applied Control Theory for Embedded Systems" came out in April.
See details at http://www.wescottdesign.com/actfes/actfes.html

David wrote:
Andy,

Thanks for the information. So my matching circuit may look like....

Mixer Out(1k5) - Series coupling cap(1n) - Var L to ground - C-Tap
across L - Crystal Filter - C-Tap - var L across C-Tap - Coupling
C(1n) - SA605 Input (4500//3.5 pF)

C-Tap caps to ground are also tuneable.

I suppose the Q of the match would not need to be particularly high as
the selectivity is provided by the Crystal filter.

Regards

David


Andy writes:

There is a formula for the required Q to perform the match.

Qreq'd = Square Root of ( (Rhigh/Rlow ) - 1 )


1500
or Sqrt of ( ------------ - 1 )
800

It's been a while so, if I got this wrong, I am sure some of the
local RF engineers will correct me...

Anyway,,, this is the minimum Q required in the tank to allow the
match
to be accomplished....

This comes from the equvalency formulae for series to parallel
comversion:


2 2
2
Rp = Rs ( Q + 1 ) Xp = Xs ( Q + 1) / Q

for coverting parallel resistance or reactance into equivalent
series resistance or reactance.....

Sure you can use Z and Y, but the above are much much easier.....


I can't follow your diagram very well, but I think you have it....
.... I am assuming SINGLE ENDED OUTPUT. For balanced
output, it is different...

Andy W4OAH

On 5 Jun 2006 17:53:05 -0700, Andy wrote:
-snip-
This comes from the equvalency formulae for series to parallel
comversion:

2 2
2
Rp = Rs ( Q + 1 ) Xp = Xs ( Q + 1) / Q


Before you try posting something like that again, you should make sure
you are using a fixed-pitch font.

It probably is indecipherable to anyone except someone using the exact
same proportional spaced font _and_ font size as you when you banged so
many times on the space bar.

Jonesy

Andy,

Thanks. I am using single-ended output from the SA602 as well as for the
SA605 input. I decouple the unused output/input with 100n Cap in each case.

How do you know the match is working correctly ? My concern is that when
I build the circuit and sweep it with a sig gen and tune it for peak,
how do I know the match to is correct in the first place.

Regards

David


Andy wrote:
David wrote:
Andy,

Thanks for the information. So my matching circuit may look like....

Mixer Out(1k5) - Series coupling cap(1n) - Var L to ground - C-Tap
across L - Crystal Filter - C-Tap - var L across C-Tap - Coupling
C(1n) - SA605 Input (4500//3.5 pF)

C-Tap caps to ground are also tuneable.

I suppose the Q of the match would not need to be particularly high as
the selectivity is provided by the Crystal filter.

Regards

David


Andy writes:

There is a formula for the required Q to perform the match.

Qreq'd = Square Root of ( (Rhigh/Rlow ) - 1 )


1500
or Sqrt of ( ------------ - 1 )
800

It's been a while so, if I got this wrong, I am sure some of the
local RF engineers will correct me...

Anyway,,, this is the minimum Q required in the tank to allow the
match
to be accomplished....

This comes from the equvalency formulae for series to parallel
comversion:


2 2
2
Rp = Rs ( Q + 1 ) Xp = Xs ( Q + 1) / Q

for coverting parallel resistance or reactance into equivalent
series resistance or reactance.....

Sure you can use Z and Y, but the above are much much easier.....


I can't follow your diagram very well, but I think you have it....
... I am assuming SINGLE ENDED OUTPUT. For balanced
output, it is different...

Andy W4OAH

David wrote:
Andy,

Thanks. I am using single-ended output from the SA602 as well as for the
SA605 input. I decouple the unused output/input with 100n Cap in each case.

How do you know the match is working correctly ? My concern is that when
I build the circuit and sweep it with a sig gen and tune it for peak,
how do I know the match to is correct in the first place.

Regards

David


Andy replies:
Personally, I would use PSPICE to simulate the circuit to see if it
peaked
at the right point. Then I would measure the voltage at the top of the

tuned tank. Then I would replace the tank and load with a resistor
that
is equal to the source resistor and see that I have the same voltage...

Alternatively, the voltage on the output side of the Thevenin sourse
resistance should be half the voltage of the generator, if the load
is truly transformed up to the Thevenin source......

Little checks like that. Perhaps you can think of others..

I am sorry if your reader gave a problem with the formatting as I
was doing the best I could with what I had..... The "Qs" in the
last formulae are all " Q squared"..

You should look these formulae up in a book , like the ARRL handbook
or some other, to understand how they are used, and practice a little
with them... I wrote a BASIC program to do Series-Parallel conversion
using them, which makes it quicker (don't know where it is, tho).

Also, I would reccommend using PSPICE to verify all these small
circuits
to see how it works and verify that you are doing it right....
It's much easier that building and measuring, and you can try stuff out
and
get a better insight.....

Andy W4OAH

Andy write:

I found an App note on the web that will
show you how to use it.


http://www.ece.utah.edu/~cfurse/micr...E/L15/L15.html

You can probly find more by doing a google search on :

" Rp Impedance Matching "

Andy,

Thanks for the link.

Regards

David

Andy wrote:
Andy write:

I found an App note on the web that will
show you how to use it.


http://www.ece.utah.edu/~cfurse/micr...E/L15/L15.html

You can probly find more by doing a google search on :

" Rp Impedance Matching "