EMRFD Message Archive 3303

Message Date From Subject
3303 2009-07-12 15:37:07 gt25psi2002 Changing position between a mixer's pad and diplexer?
Dear Sirs,

1. R2Pro inserts a mixer's pad (into the original R2) to reduce 1/f
mixer's noise.

2. A diplexer has two jobs: (a) it absorbs undesired mixer output
signals so that they are not reflected back into the mixer and (b) it
transmits desired signals to the output.

3. For a superhet Rx, the IF is fixed thus a Bridge-Tee type diplexer
is more appropirate.


Question: What will happen - especially 1/f noise performance - if we
change the position between the mixer's pad and diplexer?


My logic is that:

1. The higher Q value of a diplexer will increase its performance (i.e.
narrower bandwidth and thus a better SNR and IMD).

2. It becomes more dificult to have a higher Q value of a diplexer due
to constraint on parts. And I plan to cascade two diplexers in series to
increase its Q value.

3. Rather than using other part to connect/decouple the two diplexers
together, I would use a PAD instead.

4. Thus, the scenario will be mixer->diplexer->pad->diplexer.


Best regards,
Amin
3305 2009-07-14 06:13:48 timshoppa Re: Changing position between a mixer's pad and diplexer?
3306 2009-07-14 06:40:02 gt25psi2002 Re: Changing position between a mixer's pad and diplexer?
Dear Tim,

Thanks very much.

> I think the 1/f noise reduction factor has to do with gain
distribution. The R2pro has a preamp before the mixer and this is what
actually reduces the proportion of 1/f noise, by raising the signal. The
extra gain from the preamp allows the introduction of the pad.


Yes, I remember the formula now but don't realize it before you
mentioned it.

Anyway, will the 1/f noise also be reduced if just inserting the pad
without the preamp before the mixer?

> The diplexer really wants to see 50 ohms at its input to do its job
right. The pad makes the impedance the diplexer sees on its input be 50
ohms, even if the mixer isn't really 50 ohms.

So, the pad must always be before the diplexer. This is the lesson I
learn today.

Best regards,
Amin
3307 2009-07-14 07:06:40 timshoppa Re: Changing position between a mixer's pad and diplexer?
3308 2009-07-14 12:59:40 ehydra Re: Changing position between a mixer's pad and diplexer?
My understanding is that a diplexer is nothing more than a filter with
controlled reflection coefficient. For example, an crystal filter works
great as frequency controlling bandpass, but bad in absorbing unused
frequency bands (= other power).

So the question is: How much absorption is required for a particular
mixer topology?? Seems, that a diode-ring mixer needs very good
termination, a gilbert-mixer does not.

A bridged-tee bandpass diplexer looses 50% of the wanted input power AND
requires proper output termination.


- Henry


--
ehydra.dyndns.info


gt25psi2002 schrieb:
> Dear Sirs,
>
> 1. R2Pro inserts a mixer's pad (into the original R2) to reduce 1/f
> mixer's noise.
>
> 2. A diplexer has two jobs: (a) it absorbs undesired mixer output
> signals so that they are not reflected back into the mixer and (b) it
> transmits desired signals to the output.
>
> 3. For a superhet Rx, the IF is fixed thus a Bridge-Tee type diplexer
> is more appropirate.
>
>
> Question: What will happen - especially 1/f noise performance - if we
> change the position between the mixer's pad and diplexer?
>
>
> My logic is that:
>
> 1. The higher Q value of a diplexer will increase its performance (i.e.
> narrower bandwidth and thus a better SNR and IMD).
>
> 2. It becomes more dificult to have a higher Q value of a diplexer due
> to constraint on parts. And I plan to cascade two diplexers in series to
> increase its Q value.
>
> 3. Rather than using other part to connect/decouple the two diplexers
> together, I would use a PAD instead.
>
> 4. Thus, the scenario will be mixer->diplexer->pad->diplexer.
>
>
> Best regards,
> Amin
3309 2009-07-14 20:12:05 gt25psi2002 Re: Changing position between a mixer's pad and diplexer?
Dear Henry,

> My understanding is that a diplexer is nothing more than a filter with
controlled reflection coefficient. For example, an crystal filter works
great as frequency controlling bandpass, but bad in absorbing unused
frequency bands (= other power).
>
> So the question is: How much absorption is required for a particular
mixer topology?? Seems, that a diode-ring mixer needs very good
termination, a gilbert-mixer does not.
>
> A bridged-tee bandpass diplexer looses 50% of the wanted input power
AND requires proper output termination.


Well, in my opinion, everything in the world there is pro and con such
as bandwidth vs insertion loss. In depends on what you are looking for
in a particular design and what measure to be taken to reduce the
opposite unwanted side effect.

Best regards,
Amin
3310 2009-07-14 23:19:13 Dave - WB6DHW Re: Changing position between a mixer's pad and diplexer?
One definition of a diplexer is a circuit that splits the RF into two
paths. My KW Diplexer has a lowpass and a highpass path. It is
intended to go after an amplifier. The low pass goes to the antenna and
the highpass(consisting of the unwanted harmonics and other undesirable
out of bad products) goes to a 50 ohm load. The diplexer is designed to
have 50 ohms at all frequencies.

Dave - WB6DHW
<http://wb6dhw.com>

gt25psi2002 wrote:
> Dear Henry,
>
>
>> My understanding is that a diplexer is nothing more than a filter with
>>
> controlled reflection coefficient. For example, an crystal filter works
> great as frequency controlling bandpass, but bad in absorbing unused
> frequency bands (= other power).
>
>> So the question is: How much absorption is required for a particular
>>
> mixer topology?? Seems, that a diode-ring mixer needs very good
> termination, a gilbert-mixer does not.
>
>> A bridged-tee bandpass diplexer looses 50% of the wanted input power
>>
> AND requires proper output termination.
>
>
> Well, in my opinion, everything in the world there is pro and con such
> as bandwidth vs insertion loss. In depends on what you are looking for
> in a particular design and what measure to be taken to reduce the
> opposite unwanted side effect.
>
> Best regards,
> Amin
>
>
>
>
>
>
3311 2009-07-15 03:05:40 leon Heller Re: Changing position between a mixer's pad and diplexer?
----- Original Message -----
3312 2009-07-15 07:57:28 ehydra Re: Changing position between a mixer's pad and diplexer?
You mean a duplexer?
That is just an diplexer with the antenna in-between. If this antenna is
not a pure resistive load, the diplexer will go crazy. OK, say it won't
work really good. A circulator will there work better. Depends on how
different the transmit/receive frequency pair is.

But I'm not an expert in this field.


- Henry


--
ehydra.dyndns.info


leon Heller schrieb:
> ----- Original Message -----
>
3313 2009-07-15 11:03:20 Chris Trask Re: Changing position between a mixer's pad and diplexer?
>
> 1. R2Pro inserts a mixer's pad (into the original R2) to reduce 1/f
> mixer's noise.
>
> 2. A diplexer has two jobs: (a) it absorbs undesired mixer output
> signals so that they are not reflected back into the mixer and (b) it
> transmits desired signals to the output.
>
> 3. For a superhet Rx, the IF is fixed thus a Bridge-Tee type diplexer
> is more appropirate.
>

It's actually a bit more complicated than that. Diode ring mixers are
most susceptible to improper terminations, and all three ports need to be
terminated over the entire band of frequencies (RF, LO, IF) that the mixer
will experience. This can be quite difficult under any circumstances. Most
designers simply place a 3dB pad at all three terminals. This helps improve
the overall IMD performance, but destroys the NF performance.

I have, at times, made use of a fairly simple circuit that requires a
power divider, a "Magic-T" and a piece of transmission line that is a half
wavelength long at the port centre frequency. One output of the splitter
goes to the transmission line and the other goes to one input of the Magic
T. The other end of the transmission line goes to the other port of the
Magic T. At the half wavelength frequency of the transmission line, the
signal appears at the difference port. The input impedance seen at the
input port of the splitter is constant.

This is convenient when working over a fairly narrow range of
frequencies which high enough to where the length of the transmission line
does not become a problem. There are also multiple resonances that have to
be dealt with where the transmission line is at odd integral half-wavelength
long.

For lower frequencies, the circuit is reduced to a 2-pole passive
all-pass network and a 1:2CT hybrid transformer. The all-pass network is
designed for twice the characteristic frequency and with a Q that gives the
desired bandwidth performance. The input of the circuit is connected to one
end of the all-pass network and one input of the hybrid. The other end of
the all-pass network goes to the other input of the hybrid. Signals at the
centre frequency of the all-pass network appear at the difference port and
the circuit input inpedance is constant. Neither NF nor IMD performance are
compromised, save for the insertion loss of the circuit.

These methods may seem a bit busy, but in situations where performance
needs to be optimized they become a necessity.


Chris

,----------------------. High Performance Mixers and
/ What's all this \ Amplifiers for RF Communications
/ extinct stuff, anyhow? /
\ _______,--------------' Chris Trask / N7ZWY
_
3314 2009-07-15 12:38:41 ehydra Re: Changing position between a mixer's pad and diplexer?
Interesting. The conventional bridged-tee bandpass diplexer introduces
losses of 6dB overall. Is it possible to see it as a circuit diagram?

I played around with just a new concept in LTspice:
http://ehydra.dyndns.info/ng/Mixer_2.asc (LTspice)
http://ehydra.dyndns.info/ng/Mixer_2.png (as picture)

But I don't know if it will work in practice. What do the experts here
think about it?
So, how can I calculate NF, IMD, best transistor type?

My preferred interest is using it for shorted antennas in the frequency
range of 10MHz to 160MHz. Power consumption must be reasonable because
it is for hand-held mobiles.


regards -
Henry



--
ehydra.dyndns.info


Chris Trask schrieb:
> It's actually a bit more complicated than that. Diode ring mixers are
> most susceptible to improper terminations, and all three ports need to be
> terminated over the entire band of frequencies (RF, LO, IF) that the mixer
> will experience. This can be quite difficult under any circumstances. Most
> designers simply place a 3dB pad at all three terminals. This helps improve
> the overall IMD performance, but destroys the NF performance.
>
> I have, at times, made use of a fairly simple circuit that requires a
> power divider, a "Magic-T" and a piece of transmission line that is a half
> wavelength long at the port centre frequency. One output of the splitter
> goes to the transmission line and the other goes to one input of the Magic
> T. The other end of the transmission line goes to the other port of the
> Magic T. At the half wavelength frequency of the transmission line, the
> signal appears at the difference port. The input impedance seen at the
> input port of the splitter is constant.
>
> This is convenient when working over a fairly narrow range of
> frequencies which high enough to where the length of the transmission line
> does not become a problem. There are also multiple resonances that have to
> be dealt with where the transmission line is at odd integral half-wavelength
> long.
>
> For lower frequencies, the circuit is reduced to a 2-pole passive
> all-pass network and a 1:2CT hybrid transformer. The all-pass network is
> designed for twice the characteristic frequency and with a Q that gives the
> desired bandwidth performance. The input of the circuit is connected to one
> end of the all-pass network and one input of the hybrid. The other end of
> the all-pass network goes to the other input of the hybrid. Signals at the
> centre frequency of the all-pass network appear at the difference port and
> the circuit input inpedance is constant. Neither NF nor IMD performance are
> compromised, save for the insertion loss of the circuit.
>
> These methods may seem a bit busy, but in situations where performance
> needs to be optimized they become a necessity.
>
>
> Chris
3315 2009-07-15 13:14:03 victorkoren Re: Changing position between a mixer's pad and diplexer?
I have two points to make in reference to your schematic and message.
First, The diplexer do not (ideally, when L and C are lossless) attenuate. You may be thinking that it attenuates 6dB if you include R8 as part of the diplexer. Consider using a post diplexer amplifier with inherent Z0 input impedance. Then inserting or removing the diplexer will give the same gain. It is exactly similar to how you terminate any passive LC filter which was designed to be loaded by Z0.
Second, a Diplexer is only needed when using a mixer that the RF signal can flow at both directions, from input to its output and from output to input by its own mixing operation (as in diode or active switched mixers). The problem in such mixers is that the signal is mixed again and again as it gets reflected at its output and its input, ans so more spurs are generated. The mixer in your schematic (a single balanced active mixer) does NOT need a diplexer at its output. The best termination of such a mixer (if its output is always at a constant frequency) is just a resonant circuit. This is the same with Gilbert cell mixers - The mixer output signal can not be reflected and mixed again as the mixer transistors are unidirectional (reflecti
3316 2009-07-15 13:30:43 Chris Trask Re: Changing position between a mixer's pad and diplexer?
>
> > I have, at times, made use of a fairly simple circuit that requires
a
> > power divider, a "Magic-T" and a piece of transmission line that is a
half
> > wavelength long at the port centre frequency. One output of the
splitter
> > goes to the transmission line and the other goes to one input of the
Magic
> > T. The other end of the transmission line goes to the other port of the
> > Magic T. At the half wavelength frequency of the transmission line, the
> > signal appears at the difference port. The input impedance seen at the
> > input port of the splitter is constant.
> >
> > This is convenient when working over a fairly narrow range of
> > frequencies which high enough to where the length of the transmission
line
> > does not become a problem. There are also multiple resonances that have
to
> > be dealt with where the transmission line is at odd integral
half-wavelength
> > long.
> >
> > For lower frequencies, the circuit is reduced to a 2-pole passive
> > all-pass network and a 1:2CT hybrid transformer. The all-pass network
is
> > designed for twice the characteristic frequency and with a Q that gives
the
> > desired bandwidth performance. The input of the circuit is connected to
one
> > end of the all-pass network and one input of the hybrid. The other end
of
> > the all-pass network goes to the other input of the hybrid. Signals at
the
> > centre frequency of the all-pass network appear at the difference port
and
> > the circuit input inpedance is constant. Neither NF nor IMD performance
are
> > compromised, save for the insertion loss of the circuit.
> >
> > These methods may seem a bit busy, but in situations where
performance
> > needs to be optimized they become a necessity.
>
>
> Interesting. The conventional bridged-tee bandpass diplexer introduces
> losses of 6dB overall. Is it possible to see it as a circuit diagram?
>
>

Here's a sketch for both the distributed and lumped element versions:

http://www.home.earthlink.net/~christrask/Mixer%20Diplexers.pdf


Chris

,----------------------. High Performance Mixers and
/ What's all this \ Amplifiers for RF Communications
/ extinct stuff, anyhow? /
\ _______,--------------' Chris Trask / N7ZWY
_
3317 2009-07-15 17:32:22 timshoppa Re: Changing position between a mixer's pad and diplexer?
3318 2009-07-16 00:03:04 gt25psi2002 Re: Changing position between a mixer's pad and diplexer?
Dear Chris,

> Here's a sketch for both the distributed and lumped element
versions:
> http://www.home.earthlink.net/~christrask/Mixer%20Diplexers.pdf


I'm very fascinated with the Figure 2 - Lumped Element Diplexer.

1. Please advise me where can I get the detailed info as well as
formula of the Lumped Element Diplexer especially the same type as the
drawing.

2. Is it LP/HP or BP/BS type?

Best regards,
Amin
3319 2009-07-16 00:18:46 gt25psi2002 Re: Changing position between a mixer's pad and diplexer?
Dear Tim,

> Kangaus.com has the R2pro schematics along with KK7B's excellent
circuit description on their website. Also very worthwhile reading are
KK7B's other articles in QST (available at arrl.org).
>
> EMRFD has other examples of lossless diplexers for terminating mixer
outputs.
>

I've read the R2Pro schematic and found out it is a LP/HP type
diplexer, not a bandpass type (Bridge-Tee type) that I'm looking for
(i.e. it will be used in a fixed IF of a superhet receiver).

Neither have I found any lossless diplexer (especially bandpass type) in
EMRFD. Please advise me what page or what figure number you're referring
to.

Best regards,
Amin
3320 2009-07-16 00:23:04 Vaclav Peroutka Re: [emrfd] Re: Changing position between a mixer's pad and diplexer
Dear Amin,

>
> I've read the R2Pro schematic and found out it is a LP/HP type
> diplexer, not a bandpass type (Bridge-Tee type) that I'm looking for
> (i.e. it will be used in a fixed IF of a superhet receiver).
>

Hopefully this page http://www.qrp.pops.net/dip_sup.asp will help you ?

73,
Vaclav
3321 2009-07-16 01:01:23 gt25psi2002 Re: Changing position between a mixer's pad and diplexer?
Dear Vaclav,

> Hopefully this page http://www.qrp.pops.net/dip_sup.asp will help you
?
>

Yes, that's the circuit I've been using so far. But I don't know how
many dB loss factor.

And I'm very interested in the lossless version of diplexer especially
the one proposed by Chris. Hope it is a bandpass type.

Best regards,
Amin
3322 2009-07-16 06:40:15 Chris Trask Re: Changing position between a mixer's pad and diplexer?
>
> > Here's a sketch for both the distributed and lumped element
> > versions:
> >
> > http://www.home.earthlink.net/~christrask/Mixer%20Diplexers.pdf
>
>
> I'm very fascinated with the Figure 2 - Lumped Element Diplexer.
>
> 1. Please advise me where can I get the detailed info as well as
> formula of the Lumped Element Diplexer especially the same type as the
> drawing.
>

It will take me a while to come up with all of that, and I should
probably do it as a formal document as there are a few details that need to
be fully covered.

>
> 2. Is it LP/HP or BP/BS type?
>

It's BP/BS. The BP function is the IF output terminal, and the BS is
the termination of the primary winding centre tap. The two-winding inductor
is variable, and the two capacitors are fixed, usually two capacitors in
parallel to get as close as possible (1-2%) to the design value with readily
available parts.

Chris

,----------------------. High Performance Mixers and
/ What's all this \ Amplifiers for RF Communications
/ extinct stuff, anyhow? /
\ _______,--------------' Chris Trask / N7ZWY
_
3323 2009-07-16 07:27:58 ehydra Re: Changing position between a mixer's pad and diplexer?
Hm. I don't think that any diplexer can be lossless. There must be a
port where unuseful power is absorbed. It can be part of a diplexer
circuit, or elsewhere realized, or a mixture. The mixer spreads the
energy away: In the right frequency band for the useful signal, and in
another band the not used part. The diplexer open the door to the right
direction based on actual frequency. After the door, the signals come
nether back in the ideal case.

I have no time to simulate at the moment. So all is speculation.

regards -
Henry



gt25psi2002 schrieb:
> Dear Vaclav,
>
>> Hopefully this page http://www.qrp.pops.net/dip_sup.asp will help you
> ?
>
> Yes, that's the circuit I've been using so far. But I don't know how
> many dB loss factor.
>
> And I'm very interested in the lossless version of diplexer especially
> the one proposed by Chris. Hope it is a bandpass type.
>
> Best regards,
> Amin
3324 2009-07-16 07:33:16 ehydra Re: Changing position between a mixer's pad and diplexer?
Sorry Tim -

I meant the Trask circuits.
As I don't have a EMRFD book, I cannot see the lossless diplexer there.

Most textbooks are just copied ideas of an aera before. As most work of
mankind. It is seldom that really new ideas come in
3325 2009-07-16 11:58:53 timshoppa Re: Changing position between a mixer's pad and diplexer?
3326 2009-07-16 12:11:54 timshoppa Re: Changing position between a mixer's pad and diplexer?
3329 2009-07-16 16:25:03 Chris Trask Re: Changing position between a mixer's pad and diplexer?
>
> > Here's a sketch for both the distributed and lumped element
> versions:
> > http://www.home.earthlink.net/~christrask/Mixer%20Diplexers.pdf
>
>
> I'm very fascinated with the Figure 2 - Lumped Element Diplexer.
>
> 1. Please advise me where can I get the detailed info as well as
> formula of the Lumped Element Diplexer especially the same type as the
> drawing.
>


I just finished a brief monograph on the diplexer I described earlier,
which can be found at:

http://www.home.earthlink.net/~christrask/Mixer%20Diplexers.pdf

Chris

,----------------------. High Performance Mixers and
/ What's all this \ Amplifiers for RF Communications
/ extinct stuff, anyhow? /
\ _______,--------------' Chris Trask / N7ZWY
_
3331 2009-07-17 07:28:48 gt25psi2002 Re: Changing position between a mixer's pad and diplexer?
Dear Chris,

Thanks very very much for the paper.

Best regards,
Amin

>
> I just finished a brief monograph on the diplexer I described
earlier,
> which can be found at:
>
> http://www.home.earthlink.net/~christrask/Mixer%20Diplexers.pdf
>
> Chris
>
> ,----------------------. High Performance Mixers and
> / What's all this \ Amplifiers for RF Communications
> / extinct stuff, anyhow? /
> \ _______,--------------' Chris Trask / N7ZWY
> _
3332 2009-07-17 07:40:32 Chris Trask Re: Changing position between a mixer's pad and diplexer?
>
>>
>> I just finished a brief monograph on the diplexer I described
>earlier,
>> which can be found at:
>>
>> http://www.home.earthlink.net/~christrask/Mixer%20Diplexers.pdf
>
>Thanks very very much for the paper.
>

There are a few typographical errors, which I will take care of later today.


Chris

,----------------------. High Performance Mixers and
/ What's all this \ Amplifiers for RF Communications
/ extinct stuff, anyhow? /
\ _______,--------------' Chris Trask / N7ZWY
_
3333 2009-07-18 14:43:56 ehydra Re: Changing position between a mixer's pad and diplexer?
Hello Tim -

I read the R2pro doc. Thank you for the reference.

But I don't think I can get good infos there. The format is a mess.
Maybe it is from a technical standpoint a high-performance circuit, the
documentati
3334 2009-07-19 06:24:44 timshoppa Re: Changing position between a mixer's pad and diplexer?
3335 2009-07-19 12:26:52 Dave Brainerd Re: Changing position between a mixer's pad and diplexer?
If you are an ARRL member, you can download any QST article except the
last couple of years from the ARRL site.
I am using the comm
3337 2009-07-20 04:22:38 ehydra Re: Changing position between a mixer's pad and diplexer?
Hi Dave -

I don't have access to ARRL.

Pasting the KK7B circuit in SPICE should make the function more readable.
Most of the ham circuits just draw to much DC power. My main interest is
for circuits having a good relationship between power consumption and
other parameters.

I think here is an interesting design with description. Unfortunately,
it is in german and I don't know of an english version.
Google-Translation should help. See page 2 and 7:
http://www.mydarc.de/dc4ku/HF_Eingangsteil.pdf

regards -
Henry

--
ehydra.dyndns.info


Dave Brainerd schrieb:
> If you are an ARRL member, you can download any QST article except the
> last couple of years from the ARRL site.
> I am using the common base amp("borrowed" from one of the KK7B or W7EL
> designs and modified) in my UHFSDR. Just learned how varying the
> various resisters interact and affect the circuit.
>
> Dave - WB6DHW
> <http://wb6dhw.com>
>