EMRFD Message Archive 8047

Message Date From Subject
8047 2013-01-06 06:28:10 101 Bandpass Diplexer
Hello from a beginner,

1). Refer to Fig. 2.107 & Page 2:41 says "This circuit is called a diplexer and has characteristic that the input impedance is 50R at all frequency".

2). What I've noticed is that the circuit comprises of a series-tuned circuit and a parallel-tuned circuit resonant at the same frequency. While the series-tuned circuit has a minimum impedance (nearly zero) but parallel-tuned circuit has a maximum impedance at resonant frequency.

3). My question is "how does the diplexer have input impedance of 50R as stated in #1?"

Regards.
8048 2013-01-06 07:56:03 dixonglennb Re: Bandpass Diplexer
> 3). My question is "how does the diplexer have input impedance of 50R as stated in #1?"

1. At the resonant frequency the parallel LC is high impedance and the series LC is low impedance. The load on the series LC output (labeled 'antenna' and assumed to be 50 Ohms) is seen by the input.

2. At frequencies significantly different from resonance, the parallel LC is low impedance and the series LC is high impedance. The input sees the 50 ohm resistor attached to the parallel LC.

3. The LC circuit Q's are set so that as the frequency transitions from out-of-band to in-band, the combination of the two LC circuit's impedances and loads are fifty ohms within practical limits.

4. The in-band circuit energy is then the only energy that gets transferred to the antenna. All out-of-band energy is dissipated in the parallel LC's 50 Ohm load resistor.

73,

Glenn AC7ZN
8050 2013-01-06 14:24:21 101 Re: Bandpass Diplexer
Dear Sir,

Thanks for the answer.

1). Does it mean that we can use this diplexer circuit to terminate a mixer immediately without inserti
8051 2013-01-06 15:00:34 William Carver Re: Bandpass Diplexer
The Q of the series and shunt tuned circuits have to be the same.
Typically they're Q=1, with XL = XC = R at the frequency you want to
pass on.

When you use higher Q then the two coils and capacitors go in different
directions: one gets bigger by Q, the other smaller by Q. So for high Q
you end up with radically different inductor sizes, making it a bit more
difficult to achieve constant driving point impedance over many decades
of frequency due to self-resonances in the passive components.

W7AAZ



On Sun, 2013-01-06 at 22:24 +0000, 101 wrote:
>
> Dear Sir,
>
> Thanks for the answer.
>
> 1). Does it mean that we can use this diplexer circuit to terminate a
> mixer immediately without inserti
8055 2013-01-06 17:02:40 101 Re: Bandpass Diplexer
Dear Sir,

That's what I'm still not clear.

If I'm not wrong, 5-10 years ago I used to read articles by Ulrich Rohde mentioning of different Q value of series and shunt tuned circuits, furthermore the Q value is very high in series tuned circuit.

8058 2013-01-06 18:59:18 William Carver Re: Bandpass Diplexer
If you want 50 ohms, theoretically "DC to daylight" the Q's MUST be the
same. But there is no law that you MUST have perfect match: if you are
willing to allow some reactance and mismatch then you can relax and make
an exception.

Herr Rohde has said that diode mixers can be very sensitive to a good
wideband impedance match. If so, and high intercept is your goal, you
would probably start to provide a perfect 50 ohm load to it. Then, if
necessary to obtain other desired characteristics, relax that perfect
match while watching to see that the changed circuit does not seriously
damage the mixer intercept: a judgement call by the designer.

And interesting point: in the typical ham diplexer you use a 50 ohm
resistor from the generator to the parallel tuned circuit, and a second
50 ohm resistor from the parallel tuned circuit to the load. That
provides a 50+j0 load to the generator for all frequencies. It also
produces a 50+j0 source impedance at all frequencies to the load. It is
matched in both directions.

If your load does not care about the impedance it sees off-frequency,
then the second resistor can be removed from the circuit. It becomes two
LC and one 50 ohm resistor. The generator side will still see a perfect
50 ohm termination even with that missing resistor.

If you have some SPICE software, draw it with L=0.8842uH, C=353.7pF,
both R=50, with 50 ohm load and generator impedance to start with. You
can look at impedance (input voltage divided by input current) versus
frequency and convenice yourself this connector of elements works. Then
change and L/C, still resonant at 9 MHz, and see the still-perfect match
at 9 MHz, but degraded match on either side.

Bill - W7AAZ




8060 2013-01-06 20:38:12 Kerry Re: Bandpass Diplexer
I did some tests on a 5 MHz bridged-T diplexer.

Frequency response; there is a useful, if unspectacular, filtering effect;

http://i50.tinypic.com/97nxvo.jpg

and VSWR;

http://i48.tinypic.com/amay6o.jpg

So far as the maths is concerned, I cheat and use this calculator;

http://www.changpuak.ch/electronics/calc_16a.php

Here are the design values;

http://i50.tinypic.com/2akba60.jpg


Kerry VK2TIL.
8061 2013-01-06 21:53:19 William Carver Re: Bandpass Diplexer
That's a Q=10 version. It's selectivity is a heck of a lot more
spectacular than the Q=1 diplexer that's usually seen!

The problem with the higher Q network is getting parallel L and C, which
have only 5 ohms impedance at 5 MHz, to have high Q...milliohms of
equivalent series resistance. If you could pull that off the SWR would
be 1:1.

It's always nice to have the increased selectivity for the following
circuits.

W7AAZ



On Mon, 2013-01-07 at 04:38 +0000, Kerry wrote:
>
> I did some tests on a 5 MHz bridged-T diplexer.
>
> Frequency response; there is a useful, if unspectacular, filtering
> effect;
>
> http://i50.tinypic.com/97nxvo.jpg
>
> and VSWR;
>
> http://i48.tinypic.com/amay6o.jpg
>
> So far as the maths is concerned, I cheat and use this calculator;
>
> http://www.changpuak.ch/electronics/calc_16a.php
>
> Here are the design values;
>
> http://i50.tinypic.com/2akba60.jpg
>
> Kerry VK2TIL.
>
>
>
>
>
8062 2013-01-06 22:18:31 Kerry Re: Bandpass Diplexer
Yes; I thought the extra selectivity was worthwhile.

I only used very ordinary ceramic capacitors; I don't think that their Q is very high. (Might measure some one day).

I didn't think the job warranted the use of my small "stash" of porcelains! :)

http://i47.tinypic.com/2vm605f.jpg

Q of the inductors is probably about as good as can be obtained without resort to much larger cores.

Considering the limitations of the components I was pleased with the test results; I thought that a "real-world" example might interest the OP and others.

Re the part of the original question regarding attenuation; there is no attenuation on either the mixer side or the amplifier side of this diplexer.

Kerry.
8063 2013-01-07 00:26:50 William Carver Re: Bandpass Diplexer
That's a really nice job, very well done. Imagine someone calling that
"ugly construction": it's anything but ugly.

Its probably capacitor ESR that's raising the SWR as you go way above 5
MHz, but I'd also be saving the porcelain caps for something REALLY
serious. I have hoarded a lot of dipped silvered mica capacitors, I'd be
pawing through them with the LC Meter, trying to find some combination
of three of them that added up to 6366pF.

And on the low side I'd guess it's coil Q that pops the SWR up. I wonder
what would happen if you added a second winding just like the one on
there, and put them in parallel?

But you have to draw the line somewhere or nothing would get finished!

W7AAZ



On Mon, 2013-01-07 at 06:18 +0000, Kerry wrote:
>
> Yes; I thought the extra selectivity was worthwhile.
>
> I only used very ordinary ceramic capacitors; I don't think that their
> Q is very high. (Might measure some one day).
>
> I didn't think the job warranted the use of my small "stash" of
> porcelains! :)
>
> http://i47.tinypic.com/2vm605f.jpg
>
> Q of the inductors is probably about as good as can be obtained
> without resort to much larger cores.
>
> Considering the limitations of the components I was pleased with the
> test results; I thought that a "real-world" example might interest the
> OP and others.
>
> Re the part of the original question regarding attenuation; there is
> no attenuation on either the mixer side or the amplifier side of this
> diplexer.
>
> Kerry.
>
>
>
>
>
8064 2013-01-07 01:14:37 Kerry Re: Bandpass Diplexer
... you have to draw the line somewhere or nothing would get finished!

Yes; I'm happy with the performance of this module so I will leave well-enough alone.

I did some testing recently of this rather old but still unfinished receiver project; this is not just the diplexer, it's the entire mixer/diplexer/post-amp module as in the previous photo;

http://i49.tinypic.com/2pplj4g.jpg

http://i47.tinypic.com/es5345.jpg

Conversion gain of the complete module is about +7dB.

Kerry.
8086 2013-01-09 17:00:04 101 Re: Bandpass Diplexer - W7AAZ
Dear Sir,

Are you, William Carver, the same pers
8231 2013-02-05 22:38:44 Kirk Kleinschmidt Are These Relays Suitable for Front-End and IF Small-Signal Switchin
Hi, gang:

I have been searching for small-signal relays suitable for switching RX front-end and IF filters that were affordable and can reasonably be expected to work well (without degrading over time, needing supplementary dc, etc).
Link for Aromat RG1E-12V relays:

http://pdf1.alldatasheet.com/datasheet-pdf/view/129365/NAIS/RG1-12V.html

These were $.59 each from MPJ&A, are NOS, designed for small-signal switching up to 900 MHz and seem to be rated for 50-75 ohms.

I'm a bit concerned that the max rated switching power (24 V @ 1 A) is a bit low, but these will only be used in small-signals paths (not TX output filters, etc).

 

I bought 50...so I'm hoping they're okay :)

Any input about suitability will be appreciated.

73,

--Kirk, NT0Z


My book, "Stealth Amateur Radio," is now available from
www.stealthamateur.com and on the Amazon Kindle (soon)



>

[Non-text portions of this message have been removed]
8232 2013-02-06 07:14:37 Roelof Bakker Re: Are These Relays Suitable for Front-End and IF Small-Signal Swit
Hello Kirk,

These relays are very well suitable for the task at hand.

To maintain good contact over time when used with small signal levels, I
have seen a technique called contact wetting.
This is done by flowing a small DC current through the contacts.
The current should be in the order of about 5 mA, if I remember correct.

This requires a resistor from V+ to one side the relay switch and a second
one to ground from the other, plus DC blocking capacitors. But I am sure
there are people lurking on this list, who can tell more on the subject.

73,
Roelof Bakker, pa0rdt
Middelburg, Netherlands
8233 2013-02-06 09:42:31 Markus Hansen Re: Are These Relays Suitable for Front-End and IF Small-Signal Swit
Check this link out: http://www.w0qe.com/Technical_Topics/small_signal_relays_at_rf.html

I found the test very helpful in choosing relays for upgrading the BPF board in my home brew transceiver. I decided on the Omron G6S-2-Y-DC 12. I purchased mine from http://www.onlinecomponents.com for $1.98 in qualities of 50 or more.

I purchased the DPDT version and solder the contacts in parallel so have I have double the contact surface when activated. I also run approx 4 ma's through the contacts when activated to keep the contacts clean.

They are also amazingly fast. I tested the close and open time's and they average around 1.5 msec. With the high contact rate they can also be used for TR switching. I have tested one in a TR circuit at 100 Watts into a 50 ohm load and switched continually for 24 hours at a rate of 1 open and closure per second and there was no evidence of any harm to the relay, or my 100 amp. HI.


Markus VE7CA
WEB: ve7ca.net




[Non-text portions of this message have been removed]
8235 2013-02-06 13:32:41 Jerry Haigwood Re: Are These Relays Suitable for Front-End and IF Small-Signal Swit
Hi Kirk,

As someone else pointed out, Omron seems to be the choice for RF
relays. I know that Elecraft uses them as well as other kit producers.
Here is the Mouser part number for a popular one.

653-G6A-234P-DC12 This one is a 2P2T.

The G6 number appears to be the ones specified for RF use.

Jerry W5JH

"building something without experimenting is just solder practice"