EMRFD Message Archive 3899

Message Date From Subject
3899 2009-12-22 11:35:39 Vern VanZ Crystal filter matching transformers
Hi all,

I'm new to this forum and am working on my first solid-state (super?)hetrodyne receiver. The 4 pole crystal filter I've constructed is centered at 9 MHz (9,000,270 Hz) with 200 Ohm impedance at either end. I wound impedance matching transformers for both the input and output for 50 Ohm match with windings thus:

(using T44-2 cores)
Input primary: 12 turns, secondary: 25 turns #26 wire.
Output Prim: 25 turns, sec: 12 t same wire gauge.

The bottom of each winding is connected to common circuit ground.

I later connected a DDS signal source at -10dbm to the input and measured the output into 50 Ohms using the PIC power meter. The output measures 5 db lower than the input. Using a RLB, I measured a minimum VSWR of 1.97:1 at the center frequency...

As I connected the output of the filter to a 50 Ohm 'Dummy' load, the input appeared to change. Also the measurements of the transformers of course changed if loaded vs unloaded...

My question is, how critical is this, and is there a fool-proof method of getting the windings correct for each end?

Thanks - 73,
N7GTB
-Vern
3900 2009-12-22 11:50:53 Chris Trask Re: Crystal filter matching transformers
>
> I'm new to this forum and am working on my first solid-state
(super?)hetrodyne
> receiver. The 4 pole crystal filter I've constructed is centered at 9 MHz
> (9,000,270 Hz) with 200 Ohm impedance at either end. I wound impedance
matching
> transformers for both the input and output for 50 Ohm match with windings
thus:
>

<>

>
> My question is, how critical is this,
>

For any high-Q filter, such as when using crystals, the "internal
impedance" of the filter has everything to do with the passband ripple,
bandwidth, skirt selectivity, etc., and the transformer ratios determine
this impedance with respect to the source and load.

>
> and is there a fool-proof method of getting the windings correct for each
end?
>

Not 100% fool-proof, but if you can make transformers using bifilar and
trifilar combinations of twisted wire you'll get better and more consistent
results. In your case, a bifilar winding going all the way around the
toroid, then tying a pair of opposite ends together for the tap, will work
fine. Also, binocular cores are a better choice for small-signal designs.


Chris

,----------------------. High Performance Mixers and
/ What's all this \ Amplifiers for RF Communications
/ extinct stuff, anyhow? /
\ _______,--------------' Chris Trask / N7ZWY
_
3901 2009-12-22 11:53:26 Roelof Bakker Re: Crystal filter matching transformers
Hi Vern,

The problem is probably the type of core you used.
The simplest approach is to use a ferrite core, e.g. a FT-37-43.
Instead of a transformer with two differnt windings, you can also use a
autotransformer.
Wind ten bifilar turns on the core, connect these in series and use the
centre tap as the 50 ohm input and the top of the winding for the filter.

Best regards,
Roelof Bakker, pa0rdt
3903 2009-12-22 15:28:10 Vern VanZ Re: Crystal filter matching transformers
Chris, Roelof,

Thanks for the help. I'll try different (i.e.- ferrite) cores as well as winding scheme...!

73,
N7GTB
-Vern
Kennewick, WA



----- Original Message ----
3921 2009-12-24 09:35:14 ve7ca2 Re: Crystal filter matching transformers
Hi Vern. I have made a ton of matching transformers for xtal filters and by far the easiest to make is a L-Low Pass

For 9 MHz, connect a 1.5 uF inductor in series between the 50 ohm and 200 ohm ports and insert a 150 pf Cap to ground
3922 2009-12-24 16:32:31 Vern VanZ Re: Crystal filter matching transformers
Hi Markus,

Thanks for the tip. I've already wound two 4:1 (TLTs ?) for the input/output of the filter and they seem to be working out quite nicely. This is a learning exercise so I'll put your suggestion in my 'design notes', then build them and do some measurements just to learn what the differences are.

Should the caps be silver mica or will ceramic do...?

73,
N7GTB
-Vern



----- Original Message ----
3924 2009-12-25 14:24:33 ve7ca2 Re: Crystal filter matching transformers
HI VERN:

I have used both ceramic and silver mica at freq.'s up to 9 MHz. The Q is very low as previosly mentioned so even some drift caused by temperature variation shouldn't be a problem.

My practice is to use SM caps if I have them in my parts box.

I made several 12 MHZ xtal filters for a 6 meter rig that I am building and measured the loss, SWR bandwidth by connecting two identical L_LP matching ccts together back to back. I fed one side with a signal generator and 6 db pad between SG and one input and looked at the output with a VNA. The loss with ceramic caps was so low I could hardly see it and holding the ceramic cap's with my fingers to warm them up didn't result in the L-LP matching ccts to drift out of range.

Good luck. I have found that most everything I have learned in this hobby has been from others. Great way to learn.

73 MARKUS VE7CA

3928 2009-12-28 08:28:20 steve white Crystal filter impedance??
Hi All



I have been doing some work designing my own 6 pole filters using 11.998 Mhz
crystals and using the various design programs that are available. However,
I still am a bit confused about their input and output impedances.



Specifically, I do not understand how to design a filter for a specific
impedance. Or how to measure the impedance of the filter once it is
designed. It would be nice to be able to measure the actual impedance to
confirm the design. Is their a method to measure their impedance, if so can
some one enlighten me how to this or how to best approximate it.



I have an old receiver that uses a SA612 as the mixer which has a 1500 ohm
output impedance and that is interfaced into a MC1350 IF amp which has an
input impedance of about 330 ohms or so. The filter goes between the mixer
and IF amp of course. The filter is 500 hz wide.



Steve NU0P



_____

3929 2009-12-28 13:08:20 R Wall Re: Crystal filter impedance??
Hi Steve,

The way I test my homebrew crystal filters is with my N2PK VNA and myVNA or VNWA software. The N2PK VNA measures the s-parameters for the filter at 50 ohm. The software (myVNA or VNWA) has a matching tool that can check what the filter's impedance is and give you the component values for the selected matching circuit (can select different matching circuits). myVNA will also measure your crystal parameters and list them in a log file. That can be loaded into Microsoft Excel or OpenOffice Cal where you can sort them for your filter.

You can build your own N2PK VNA (by Paul N2PK), the software is free thanks to Dave G8KBB (myVNA) and Tom DG8SAQ (VNWA).

n2pk.com

Building a N2PK VNA may be a bit much just to check one filter, but then you'll have it forever. I now wouldn't be without my N2PK VNA as you can also use it to test components etc.

Regards,

Roderick Wall, vk3yc.

----- Original Message -----
3930 2009-12-28 14:43:51 john lawson Re: Crystal filter impedance??
Hi Steve...you mention in your post "Specifically, I do not understand how to design a filter for a specific impedance".....Wes' XLAD program let's you know the minimum termanation resistance  that is accepatable based on the specific information you have entered previously. This is step 4 "Match ends" of the design of the filter. Too small of a value entered.... the program will refuse to compute further values and you receive an error message, "Resistance too small, try another value".  You can then play with the termnation resistance value entered upward in the step 4 until it's accepted by the program and use that value or a higher one.....Once the filter is "built" you can then design a L network to transform the impedances to match the filter to the input and output circuits.....I have built filters that have matched from 50 ohms up to 3.3K with the L network and the filters have turned out nicely as sweepings of the filters have
indicated. 
 
Others may provide additional  information on your remaining concerns as well as this one. Using sweeping gear is about all I am familiar with to find out the resulting filters shape, etc.  Best of Luck, John K5IRK
 
3931 2009-12-30 05:51:19 steve white Re: Crystal filter impedance??
Roderick



Thanks for your reply. The VNA is one tool that I do not have but have been
wanting for a while now. I was aware of its use in filter work but I was
not aware of it total capabilities in this regard. I have contacted the VNA
folks and have the costs now to build one identified. So think I am going
to do that.



Did you have any problems in getting your VNA up and running or just the
normal sorts of problems ??



Steve NU0P



_____

3932 2009-12-30 06:06:39 steve white Re: Crystal filter impedance??
Hi John



Thanks for the reply. Well you have answered one of my questions. I had
tried using XLAD and was a bit frustrated with it and so went and used the
DISHAL program to build up a filter. I have also used the AADE filter
design program as well. I had always tried to understand how to design the
filter with two different impedances using these programs. Say 300 ohm
output and 1500 ohm input but could not determine how to do that that with
the software. Apparently I was not thinking correctly about this. You have
to design the filter with the same input and output impedance and design the
matching circuits. I will try this.



I have a sweep generator and have used it with some success to look at
passbands of IF transformers and the like and I assume I can do the same
with crystal filter? I also have a spectrum analyzer (Tek 7L12) but sadly
no tracking generator for it and I imagine that would be the best tool to
use to look at the filter passband shape.



I will try several things and see how it turns out if I design it for 300
ohm I/O impedance and then match the one end from 1500 to 300 using a
transformer or LC.



Thanks for your thoughts on this.



Steve NU0P



_____

3933 2009-12-30 07:05:15 john lawson Re: Crystal filter impedance??
Steve good morning. Just a short note.....I'm assuming you've read the instructions that comes with Wes' programs. However, if you have not, I encourage you to look at them as they are easy to follow. XLAD

3934 2009-12-30 07:14:38 john lawson Re: Crystal filter impedance??
Steve good morning. Just a short note.....I'm assuming you've read the instructions that comes with Wes' programs. I use them exclusively, not to say others are not good, it's just my programs of preference. 
 
However, if you have not, I encourage you to look at the instructions closely as they are easy to follow and then do some experimentation to get a comfortable feel for them. XLAD will design the filter, Ladbuild you can view the filter schmatically, note that the RL in the left hand beginning of the schematic is the output of the filter and the far right hand end of the schematic is the input of the filter.  GPLA you can view the filter's expected filter response based on the values selected in XLAD. I'm not sure whether or not your sweeper will sweep the narrower cystal filters or not to get the desired effect. You will just have to experiment to find out.  Best of Luck....73, John K5IRK


3935 2009-12-30 12:28:21 R Wall Re: Crystal filter impedance??
Hi Steve,

The only problem I had was not putting some bypass capacitors the correct way around (polarity). If you have any problems building your VNA, you only need to go to the Yahoo N2PK VNA group and you'll get all the help you need. I only wish I had my Lab in a box 20 years ago. Good luck if you decide to build your Lab in a box N2PK VNA.

Roderick vk3yc.

----- Original Message -----