Discussion:
mixer: DBM or dual gate mosfet?
(too old to reply)
a***@gmail.com
2006-03-01 09:03:21 UTC
Permalink
hello,

I'm building my first reciver. I can't choose what kind of mixers
should I use. I have read that diode ring mixers are far superior
compared to dual gate mosfet mixers.

Is this true for both - first (RF / VFO) and second (IF / BFO) stages?
Or is there any real difference at all?

thanks
Eamon Skelton
2006-03-01 10:52:41 UTC
Permalink
Post by a***@gmail.com
hello,
I'm building my first reciver. I can't choose what kind of mixers should I
use. I have read that diode ring mixers are far superior compared to dual
gate mosfet mixers.
It depends on what you mean by superior. The mosfet mixer has gain
and usually has a lower noise figure. The diode mixer will have superior
strong signal handling (higher IP3), but will have about 7dB loss. The
diode mixer needs more local oscillator power. They both make excellent
mixers if they are applied properly.

To get the most bang for your buck, it is hard to beat a ring of
1N914 or 1N4148 diodes at a few cent each. The ferrite 'cups' from
scrap Toko IF transformers can be used as cores for the trifilar
wound transformers.

There are a few other options you should consider. High level IC mixers
like the AD831 are worth considering. Switching mixers using MOSFETs
are capable of very high performance. Search for info about the N6NWP
front-end from QST Feb 93 or the H-mode mixer used in several recent
homebrew designs.
http://uk.groups.yahoo.com/group/picastar/
http://xoomer.virgilio.it/sergiocartoceti/article_7.htm
http://www.warc.org.uk/cdg2000/introduction.htm

If you build the receive mixer as a separate module,
you can try them all and pick the one that works best for you.

73, Ed. EI9GQ.
--
Linux 2.6.15
Remove 'X' to reply by e-mail.
Yes, my username really is: nospam
xpyttl
2006-03-01 12:03:01 UTC
Permalink
Post by a***@gmail.com
hello,
I'm building my first reciver. I can't choose what kind of mixers
should I use. I have read that diode ring mixers are far superior
compared to dual gate mosfet mixers.
"Superior" is something of a loaded word. Whether a particular parts is
superior or not depends on your design intent.

Probably the most popular mixer for simple HF receivers is the NE/SA
602/612. This is an active mixer. It has amazing amounts of gain, such
that an RF stage is almost never needed. It is extremely simple to deploy,
and it requires almost no power. Thus, in portable/battery powered circuits
it is almost always the mixer of choice. It has an absolutely horrid TOIP.
There are other, mostly older, even poorer, active mixers, but the 602 is a
very versatile part, so it seems to show up everywhere.

At the other extreme are diode ring mixers. These can have stellar TOIPs,
but take a lot of oscillator power. Further, they need lots of RF as well,
so some sort of RF stage is needed. All this adds up to a need for plenty
of power. The best diode ring mixers will use matched, Schottky diodes, but
good old 1N4148's do work, and quite well. Many designs use packaged diode
ring mixers such as those from Minicircuits.

The dual gate MOSFET falls kind of in the middle. It doesn't have the
horrible TOIP problems of an active mixer, but it's not as power hungry as a
diode ring. The MOSFET seems to have fallen out of favor lately, in spite
of being a "balanced" sort of solution. I suspect most designers are either
going for power consumption or performance, and really, quite good
performance can be had with the active mixers with careful design.
Post by a***@gmail.com
Is this true for both - first (RF / VFO) and second (IF / BFO) stages?
Or is there any real difference at all?
Careful design can manage what the second mixer sees more easily than than
the first. This can make the dynamic range problems of an active mixer less
of an issue. For that reason, balanced designs that tend toward management
of power consumption will sometimes use a diode ring for the first IF and an
active mixer for the second. But a superhet bent all out on power
conservation will almost always use a pair of 602's. Designers who want to
avoid ICs for whatever reason will use a pair of diode rings.

I shouldn't sound so down on the 602. A WELL-DESIGNED 602 receiver can
easily match the performance of the $1000 class rice box rigs. It cannot,
however, come close to the performance of an equally well-designed diode
ring rig. But the diode ring rig will probably consume three times the
power, meaning three times the heat to deal with and the associated
oscillator compensation issues.

So you picks your poison.

..
Paul Keinanen
2006-03-02 20:37:10 UTC
Permalink
On Wed, 1 Mar 2006 07:03:01 -0500, "xpyttl"
Post by xpyttl
Probably the most popular mixer for simple HF receivers is the NE/SA
602/612. This is an active mixer. It has amazing amounts of gain, such
that an RF stage is almost never needed. It is extremely simple to deploy,
and it requires almost no power. Thus, in portable/battery powered circuits
it is almost always the mixer of choice. It has an absolutely horrid TOIP.
There are other, mostly older, even poorer, active mixers, but the 602 is a
very versatile part, so it seems to show up everywhere.
If you really intend to use mixers with such horrible IP3 figures, I
would suggest using a very selective front end ahead of it. For a
single band CW receiver some fixed tuned stages might suffice, but
otherwise some tunable input filters should be used.

In Europe, there are several high power broadcasters starting at 7100
kHz, which would easily overload the 40 m receiver. Assuming loaded Q
of 100 and the front end tuned to 7000 kHz, the -3 dB bandwidth would
be +/-35 kHz from the centre frequency with some usable attenuation at
7100 kHz.

Using fixed tuned octave wide front end filters with the 602 is just
asking for trouble.

Paul OH3LWR
xpyttl
2006-03-03 02:11:16 UTC
Permalink
Post by Paul Keinanen
In Europe, there are several high power broadcasters starting at 7100
kHz, which would easily overload the 40 m receiver. Assuming loaded Q
OH2BT's comments about how much better things have gotten in Europe really
made me say hmmmm.... I only recently heard actual measurements, rather
than whining, and things are pretty horrible today - they must have been
intolerable decades ago.

Tight front ends and careful control of levels obviously are important with
any mixer, but especially something with the gain of a 602. Nevertheless, I
doubt there are many cases where a 602 would be even useable in Europe, let
alone "good".

..
Paul Keinanen
2006-03-03 07:29:30 UTC
Permalink
On Thu, 2 Mar 2006 21:11:16 -0500, "xpyttl"
Post by xpyttl
Post by Paul Keinanen
In Europe, there are several high power broadcasters starting at 7100
kHz, which would easily overload the 40 m receiver. Assuming loaded Q
OH2BT's comments about how much better things have gotten in Europe really
made me say hmmmm.... I only recently heard actual measurements, rather
than whining, and things are pretty horrible today - they must have been
intolerable decades ago.
We had problems keeping Radio Moscow out of _audio_ equipment :-).

Guitar amplifiers were quite problematic with long cables and a top
capacitance (the guitar and the player) at the end, bringing quite
large RF voltages into the audio stages, causing rectification in
unfiltered input stages.

Paul OH3LWR
A***@nouce.bellatlantic.net
2006-03-01 12:58:06 UTC
Permalink
Post by a***@gmail.com
hello,
I'm building my first reciver. I can't choose what kind of mixers
should I use. I have read that diode ring mixers are far superior
compared to dual gate mosfet mixers.
DBMs are very good if properly applied but if improperly used they
will disappoint the user. Their advatages is they are rugged, hard to
overload and easy to build. Disadvantage, no gain rfamp may be
needed, must have at least 5mw of LO power(some need more), are
designed for around 50 ohm impedences at all ports.

Dual gate MOSFET, popular many years ago, and generaly easy to
apply with moderate overload resistance. The common reason for not
being used as much is simply availbility. Common Jfets (u310, MPF102)
inpairs can be used in a cascode compound connection with nearly equal
performance as the MOSFETs without cost and ESD considerations.
Advantage is good gain, low noise and low power.
Post by a***@gmail.com
Is this true for both - first (RF / VFO) and second (IF / BFO) stages?
Or is there any real difference at all?
Depends on reciever design. I've seen Mosfets(or two Jfets) used for
first mixer and DBM used for product detector. The idea there is a
DBM and the end of an IF will see large signals and the overload
resistance is valuable there.

Like others have said superior varies depending on goals. One
parmeter in this case of DBMs is the oscillator power needs and often
the need for more gain stages. If you building a radio that runs on
batteries then using more power may be bad.

Alllison
xpyttl
2006-03-01 14:04:13 UTC
Permalink
Post by A***@nouce.bellatlantic.net
If you building a radio that runs on
batteries then using more power may be bad.
Also keep in mind that more power=more heat

If you are building a simple analog VFO, temperature compensating the VFO
can be the most tedious part of designing a receiver. Depending on how
tight your box is, the difference in heat could be an issue. Keeping the
oscillator stable while delivering more power also means more buffer stages
between the VFO and the mixer.

If you are designing with a DDS, of course, all this is pretty much moot.
With a typical DDS chip and a packaged clock oscillator at some high
frequency, the oscillator will draw so much current and generate so much
heat that what the mixer requires is invisible.

..
A***@nouce.bellatlantic.net
2006-03-01 16:22:19 UTC
Permalink
On Wed, 1 Mar 2006 09:04:13 -0500, "xpyttl"
Post by xpyttl
Post by A***@nouce.bellatlantic.net
If you building a radio that runs on
batteries then using more power may be bad.
Also keep in mind that more power=more heat
If you are building a simple analog VFO, temperature compensating the VFO
can be the most tedious part of designing a receiver. Depending on how
tight your box is, the difference in heat could be an issue. Keeping the
oscillator stable while delivering more power also means more buffer stages
between the VFO and the mixer.
Actually even without the heat issue you still have to compensate it
or ambient variation will drive you nuts. Granted a few transistors
delivering 5-10mW of power is not a great heat generator when you add
all the surrouding possible sources.
Post by xpyttl
If you are designing with a DDS, of course, all this is pretty much moot.
With a typical DDS chip and a packaged clock oscillator at some high
frequency, the oscillator will draw so much current and generate so much
heat that what the mixer requires is invisible.
Since buffereing the VFO is a good idea anyway the buffer and later
statges can supply the 5 or more milliwatts needed for level 7 rings.
Since those stages can be "remote" the small heat generated is not a
big issue. However between a VFO, buffer and a buffer to deliver
power you can be hitting 30-50mA and on batteries thats a bigger
issue.

If you using DDS, likely power is not an issue and the combined DDS
and control plus display could be surprizingly high or at least has to
be managed.

However you approach the problem a little though to the overall
effects are important. After all what usually seperates a great
reciever from a passable one is attention to the little details.


Allison
Ken Scharf
2006-03-02 03:31:17 UTC
Permalink
Post by A***@nouce.bellatlantic.net
On Wed, 1 Mar 2006 09:04:13 -0500, "xpyttl"
Post by xpyttl
Post by A***@nouce.bellatlantic.net
If you building a radio that runs on
batteries then using more power may be bad.
Also keep in mind that more power=more heat
If you are building a simple analog VFO, temperature compensating the VFO
can be the most tedious part of designing a receiver. Depending on how
tight your box is, the difference in heat could be an issue. Keeping the
oscillator stable while delivering more power also means more buffer stages
between the VFO and the mixer.
Actually even without the heat issue you still have to compensate it
or ambient variation will drive you nuts. Granted a few transistors
delivering 5-10mW of power is not a great heat generator when you add
all the surrouding possible sources.
Post by xpyttl
If you are designing with a DDS, of course, all this is pretty much moot.
With a typical DDS chip and a packaged clock oscillator at some high
frequency, the oscillator will draw so much current and generate so much
heat that what the mixer requires is invisible.
Since buffereing the VFO is a good idea anyway the buffer and later
statges can supply the 5 or more milliwatts needed for level 7 rings.
Since those stages can be "remote" the small heat generated is not a
big issue. However between a VFO, buffer and a buffer to deliver
power you can be hitting 30-50mA and on batteries thats a bigger
issue.
If you using DDS, likely power is not an issue and the combined DDS
and control plus display could be surprizingly high or at least has to
be managed.
However you approach the problem a little though to the overall
effects are important. After all what usually seperates a great
reciever from a passable one is attention to the little details.
Allison
Some of the modern DDS chips require little power. Analog Devices
has some DDS chips that draw less than 50ma at 5v, I think there is
one that takes but 15ma. True a vfo will draw even less, but we
are not talking about gobs of power in any case.

DDS vfo's have very low phase noise, and the ones that can be clocked
at 100mhz or higher can deliver quite low spurs. The AD9954 series
have a 14 bit DAC and can make a very good HFO for a single conversion
receiver with no PLL loop filter needed to clean up the output.
Chris Jones
2006-03-01 22:33:27 UTC
Permalink
Post by a***@gmail.com
hello,
I'm building my first reciver. I can't choose what kind of mixers
should I use. I have read that diode ring mixers are far superior
compared to dual gate mosfet mixers.
Is this true for both - first (RF / VFO) and second (IF / BFO) stages?
Or is there any real difference at all?
thanks
I have recently bought the books from the RSGB which contain all of the
Technical Topics columns from RadCom for the last couple of decades or so.
It seems like they really like making mixers from FST3125 Bus Switch ICs,
and up to perhaps 50MHz these are supposed to be much better than the
average diode ring mixer. They call the configuration "H-mode" and the guy
who I believe is supposed to have come up with the idea is called Colin
Horrabin. Here is an article randomly selected from a google search:
http://
xoomer.virgilio.it/sergiocartoceti/pdf%20files/IK4AUY_%20qex_07-2004.pdf
I don't like the way they generate the LO signals with XOR gates but apart
from that it is interesting.

Chris
Chris Jones
2006-03-02 21:48:16 UTC
Permalink
Post by Chris Jones
Post by a***@gmail.com
hello,
I'm building my first reciver. I can't choose what kind of mixers
should I use. I have read that diode ring mixers are far superior
compared to dual gate mosfet mixers.
Is this true for both - first (RF / VFO) and second (IF / BFO) stages?
Or is there any real difference at all?
thanks
I have recently bought the books from the RSGB which contain all of the
Technical Topics columns from RadCom for the last couple of decades or so.
It seems like they really like making mixers from FST3125 Bus Switch ICs,
and up to perhaps 50MHz these are supposed to be much better than the
average diode ring mixer. They call the configuration "H-mode" and the
guy who I believe is supposed to have come up with the idea is called
Colin
http://
xoomer.virgilio.it/sergiocartoceti/pdf%20files/IK4AUY_%20qex_07-2004.pdf
I don't like the way they generate the LO signals with XOR gates but apart
from that it is interesting.
Chris
One thing I forgot: I7SWX I believe is responsible for many of the H-mode
mixer circuits in the the Technical Topics column.
http://www.qsl.net/i7swx/index.htm

Chris
Saandy , 4Z5KS
2006-03-02 08:33:22 UTC
Permalink
...a long explanation, but a needed one.
One of the most important characteristics of a mixer is its ability to
handle large input signals without overloading. if overloaded by an
unwanted signal, chances are that it will not be able to handle the
weak signal you're interested in. the physical mechanism involved is
not important, except for the fact that increasing current through an
active mixer (re FET or transistor, or the injection power in a DBM
tend to alleviate the problem in some measure. the nois figure of both
is more or less equal and adequate in the HF range. the main advantage
of the dual gate MOSFET is that it needs much less power from the local
oscillator and can save you an amplifying stage. Not really crushing,
but sometimes neeedful.
If you're talking about a simple first receive I'd go with the MOSFET
mixer. there are very good examples in the hand book. iI used them for
years with success and still using them in one form or another.
the problem is much less severe at the second mixer, because the IF
filters tend to defend it from large unwanted signals and the amplituse
variations are also smaller due to AGC action.
Soooo , KISS (keep it simple, stupid!!!), a very good adage. goood luck
with your first.
Saandy 4Z5KS
Risto Tiilikainen
2006-03-02 14:20:07 UTC
Permalink
Post by a***@gmail.com
hello,
I'm building my first reciver. I can't choose what kind of mixers
should I use. I have read that diode ring mixers are far superior
compared to dual gate mosfet mixers.
Is this true for both - first (RF / VFO) and second (IF / BFO) stages?
Or is there any real difference at all?
thanks
Hi

If you think the situation in short waves today:
Russian and their previous satellite country jammers are quiet
Local broadcast is nearly completely in FM
Propaganda is no more effective to transmit in short waves
Commercial data is practically in Internet. and in satellites
Marine communication is in satellites.

All this means less cross modulation products in first mixer than some
sixteen years ago
Atmospheric noise in sw is much higher than the noise of modern front
and mixer stage
Advantage of diode mixer is marginal
IGFET mixer is simple and advanced solution for DIY project.
IGFET mixer doesn't need any front amplifier stage.
A selective band filter in front of mixer is superior to broadband
transformers
I am using loosely coupled 3 stage band filter tuned by variable triplet
air capacitor
3 coils for low end of sw and 3 coils for upper end of sw.
Coils are DC selected by small reed relays
KISS

If you are constructing premixer then I recommend DBM to keep birdies
in low levell
For IF/BFO my recommendation is also DBM or "semi DBM"
For both of those DBM solutions I recommend you to Google a nice
advanced component MC1496.
In some Motorola handbooks and ARRL handbooks are examples for MC1496 as
DBM, product detector and balanced modulator.
It is mostly used in single ended circuits in RF meaning and balanced
for DC

73, Risto OH2BT
Dave Platt
2006-03-02 19:33:23 UTC
Permalink
Post by Risto Tiilikainen
Russian and their previous satellite country jammers are quiet
Local broadcast is nearly completely in FM
Propaganda is no more effective to transmit in short waves
Commercial data is practically in Internet. and in satellites
Marine communication is in satellites.
All this means less cross modulation products in first mixer than some
sixteen years ago
Atmospheric noise in sw is much higher than the noise of modern front
and mixer stage
Advantage of diode mixer is marginal
There's a good discussion about the advantages and disadvantages of
various mixer types, for different applications, in "Experimental
Methods in Radio Frequency Design", a book I strongly recommend.

As others have pointed out, there's no one right solution. Even for
use in what seems like a simple, constant application (e.g. a CW
receiver for the 40-meter band), the choice of which is more
appropriate can swing one way or the other based on how you intend to
use the receiver.

As one example given in EMiRFD, if you're looking for a simple
receiver which is intended for QRP operation on backpacking trips,
then the low power consumption, and the mixer gain of an active mixer
such as a Gilbert cell (e.g. SA602 and similar) can make this the
ideal. Out in the woods, the RF levels will be low, and the
relatively low IP3 of these sorts of mixers isn't likely to be a
problem.

On the other hand, if you're planning to build a receiver which may
have to operate in a strong-adjacent-signal environment (e.g. for
Field Day or other contesting), then you may want to favor a
diode-ring double-balanced mixer operating at a high LO-injection
level, and the devil take the power consumption :-)

The FET-switch mixers seem to be a really nice alternative, and
although they've gotten relatively little visibility in amateur-radio
applications they've become very popular in commercial use (e.g.
cell-phone handsets). I haven't yet had a chance to play with these
myself but they look like fun!
--
Dave Platt <***@radagast.org> AE6EO
Hosting the Jade Warrior home page: http://www.radagast.org/jade-warrior
I do _not_ wish to receive unsolicited commercial email, and I will
boycott any company which has the gall to send me such ads!
A***@nouce.bellatlantic.net
2006-03-02 22:06:10 UTC
Permalink
Post by Dave Platt
Post by Risto Tiilikainen
Russian and their previous satellite country jammers are quiet
Local broadcast is nearly completely in FM
Propaganda is no more effective to transmit in short waves
Commercial data is practically in Internet. and in satellites
Marine communication is in satellites.
All this means less cross modulation products in first mixer than some
sixteen years ago
Atmospheric noise in sw is much higher than the noise of modern front
and mixer stage
Advantage of diode mixer is marginal
There's a good discussion about the advantages and disadvantages of
various mixer types, for different applications, in "Experimental
Methods in Radio Frequency Design", a book I strongly recommend.
As others have pointed out, there's no one right solution. Even for
use in what seems like a simple, constant application (e.g. a CW
receiver for the 40-meter band), the choice of which is more
appropriate can swing one way or the other based on how you intend to
use the receiver.
As one example given in EMiRFD, if you're looking for a simple
receiver which is intended for QRP operation on backpacking trips,
then the low power consumption, and the mixer gain of an active mixer
such as a Gilbert cell (e.g. SA602 and similar) can make this the
ideal. Out in the woods, the RF levels will be low, and the
relatively low IP3 of these sorts of mixers isn't likely to be a
problem.
The lowly SA/NE602 isn't so bad considering the power it uses and the
15-17db of gain it offers. Like any power power mixer care in use is
important.
Post by Dave Platt
On the other hand, if you're planning to build a receiver which may
have to operate in a strong-adjacent-signal environment (e.g. for
Field Day or other contesting), then you may want to favor a
diode-ring double-balanced mixer operating at a high LO-injection
level, and the devil take the power consumption :-)
Even then with care in generating the power it's not that bad.
Post by Dave Platt
The FET-switch mixers seem to be a really nice alternative, and
although they've gotten relatively little visibility in amateur-radio
applications they've become very popular in commercial use (e.g.
cell-phone handsets). I haven't yet had a chance to play with these
myself but they look like fun!
I have, really nice devices but a challenge to build circuits with
good symetry at higher frequencies like 6M and up. Some of the
layouts can be a bear to drive properly and have the same port
impedence matching considerations as DBMs. They also still
have losses in th e6-8db range so gain distribution requires care.

However singally balanced FET mixers have been around for a while
and can offer good IP3 with simplier design. Over the years several
designs using both active mixers (single and dual gate [mos and
junction] FETS) as well as MOSfets, and transistors in passive modes.
The nadbooks and QST and Ham Radio featured these designs for
bands such as 40M where broadcasters are a problem.

Allison
Paul Keinanen
2006-03-03 07:29:31 UTC
Permalink
On Thu, 02 Mar 2006 22:06:10 GMT,
Post by A***@nouce.bellatlantic.net
The lowly SA/NE602 isn't so bad considering the power it uses and the
15-17db of gain it offers. Like any power power mixer care in use is
important.
Certainly usable for receiving strong international broadcasters, in
which case the input signals can be sufficiently attenuated, however,
trying to receive any weak signals with such an attenuator at the
front end, is not very productive :-)

Paul OH3LWR
Ken Scharf
2006-03-04 03:20:28 UTC
Permalink
Post by Dave Platt
The FET-switch mixers seem to be a really nice alternative, and
although they've gotten relatively little visibility in amateur-radio
applications they've become very popular in commercial use (e.g.
cell-phone handsets). I haven't yet had a chance to play with these
myself but they look like fun!
There was an article in the ARRL HB (1995 and probably others)
about using an SD5000 quad mosfet as a mixer. I bought a few
of these transistors a few years ago (now where did I put them,
got lost in the last move!). They are supposed to make
a very good mixer. The only disadvantage was the circuit
required injection at twice the lo frequency as it used a jk flip
flop to develop the required 180 degree injection. I suppose
a balum could do the same thing. In any case the AD9954 DDS I am
thinking of using can clock to 400mhz and develop output to
160mhz, so getting up to 80mhz (for use with a 9mhz if) would still
be no problem.
Eamon Skelton
2006-03-04 10:05:16 UTC
Permalink
There was an article in the ARRL HB (1995 and probably others) about using
an SD5000 quad mosfet as a mixer.
Also see N6NWP's article from QST June 1993. And the
H-mode mixer by G3SBI in RadCom and various other RSGB
publications.
I bought a few of these transistors a few years ago...
I see the SD5000 is still listed on Calogic's website
http://www.calogic.net/html/dmos.html Is the SD5000
still widely available? The search engines turn up a
lot of data but very few suppliers. Perhaps a ring of
discrete MOSFETs or a fast switch IC like the FST3125
might be a better choice.

73, Ed. EI9GQ.
--
Linux 2.6.15
Remove 'X' to reply by e-mail.
Yes, my username really is: nospam
Harold E. Johnson
2006-03-04 12:48:53 UTC
Permalink
Post by Eamon Skelton
I see the SD5000 is still listed on Calogic's website
http://www.calogic.net/html/dmos.html Is the SD5000
still widely available? The search engines turn up a
lot of data but very few suppliers. Perhaps a ring of
discrete MOSFETs or a fast switch IC like the FST3125
might be a better choice.
73, Ed. EI9GQ.
GM Ed. Certainly a better choice when it comes to conversion loss. I've
built a half dozen SD5000 mixers and although at least one of them managed a
measured +50 dBm Ip3, (G3SBI measured it not me!) I never managed to get the
conversion loss below 9 dB and more often 10. Conversely, the 3125/ family
readily does 4.8 dB conversion loss thanks to the very respectable Rds on.
REALLY simplifies things in the front end department, to at least 30 MHz,
with attention paid to the front end filters, you can get away without an RF
amplifier.

Regards
W4ZCB
Gian
2006-03-18 23:33:45 UTC
Permalink
Post by Harold E. Johnson
GM Ed. Certainly a better choice when it comes to conversion loss. I've
built a half dozen SD5000 mixers and although at least one of them managed a
measured +50 dBm Ip3, (G3SBI measured it not me!) I never managed to get the
conversion loss below 9 dB and more often 10. Conversely, the 3125/ family
readily does 4.8 dB conversion loss thanks to the very respectable Rds on.
REALLY simplifies things in the front end department, to at least 30 MHz,
with attention paid to the front end filters, you can get away without an RF
amplifier.
Regards
W4ZCB
Hi all,

Harold, W4ZCB, has given some "comments" on the H-Mode Mixer that are
REAL and TRUE.
G3SBI has removed the mixer from the stages that are critical on a
receiver and in the CDG2000 project it was demonstrated the IMD due to
passive components like core of coils. The FST3125 fast bus switch has
lowered the conversion loss of the H-Mode Mixer and made the RX front
end "hot", although not reaching the IP3 of +50dBm it still has plenty
of +dBm to give ...around +40dBm!You can permit yourself to add a xtal
filter behind the h-mode mixer and still getting high numbers on IP3.

If you are looking for a low consumption and simple mixer you may go OK
with the classic NE/SA602 and similar ones. BUT ... if you are looking
for high performance you need the H-Mode Mixer in the 3 transformer
G3SBI or the 2 transformer I7SWX configurations.
The H-Mode Mixer has been tested on downconversion (CDG2000, STAR
projects) and also on upconversion projects (I7SWX). For upconversion I
have devised a new squarer using LVDS and test it at IF of 35 to 70 MHz
(and 100MHz) with conversion loss around -5dbm and IP3 between +35 to
+40dBm (3.3V FST3125).
The H-Mode Mixer has been tested for RF input up to 50MHz in RX and
converter to 27-28MHz IF.
At 144MHz it performs like a standard diode db mixer having a conv loss
around -8dB. The input limit is due to the internal gates timing and
unbalancing as the FST3125 may have a bandwidth between 300 to 400MHz.
For those interested I may suggest a visit to JA9TTT web page where he
has reported measurement on both G3SBI and I7SWX versions using the
74AC86 squarer with balance adjustment. The comments are in japanese
but a translator is making the reading understandable :
ja9ttt.homedns.org/. I have posted a copy in my web page
www.qsl.net/i7swx in the subdirectory homebrewing. JA9TTT page permits
to enlarge the spectrum analyzer screen pictures.
I have also developed a 1 transformer double balance
mixer/demodulator/modulator using the FST3125 and 74AC/HC86 squarer
with an IP3 of around +25dBm.
Notes on the H-Mode Mixers and 1T DBM have been reported in RadCom,
G3VA's Technical Topics column.
For those interested on more detailed information I can make available
notes in English and Italian. Please write to my e-mail address:.

73

Gian
I7SWX
F5VGU
W1-I7SWX
G-QRP #10241
I QRP #571
JJ
2006-03-20 18:24:28 UTC
Permalink
What is the IP3 of a typical dual gate MOSFET mixer?
One can also use HC4066 or 4053 as the switching element at lower freqs.

JJ
Giancarlo Gian Moda, I7SWX
2006-03-21 09:41:49 UTC
Permalink
What is the IP3 of a typical dual gate MOSFET mixer?
One can also use HC4066 or 4053 as the switching element at lower
freqs.

JJ


I really do not know what is the IP3 of a DG Mosfet as I never have
seen any value reported. I hope some readers may have data on this.
Certainly we have to look at IIP3 and OIP3 as the DGM mixer may have
high conversion gain 15-20dB and is somewhat limited in dynamic range.
Maybe it could have an OIP3 around +5dBm equivalent to an IIP3 of -10
to -20dBm .... more or less similar to the NE602.
The DGM mixer was a common mixer in the first series of transistorized
RTX with valves PA (FT101Z, TS820 etc). The DGM has a good square law,
better than JFET.

74HC4066 is OK at low frequency conversion but it has a haigh
conversion loss, around 8-10dB and so associated noise figure.

Today the switched mixer should use fast bus switches like the FST3125
family when looking at high performance mixers. The examples are the
CDG 2000 and STAR (Pic-A-Star) projects with IP3 around +40dBm and
+36dBm, where the H-Mode Mixer with FST3125 is used.

73

Gian
I7SWX
A***@nouce.bellatlantic.net
2006-03-21 14:02:13 UTC
Permalink
On 21 Mar 2006 01:41:49 -0800, "Giancarlo Gian Moda, I7SWX"
Post by JJ
What is the IP3 of a typical dual gate MOSFET mixer?
Depening on the fet and circuit values around +3dbm.
Post by JJ
One can also use HC4066 or 4053 as the switching element at lower freqs.
True. Doesn't work well at 6M at all.
Post by JJ
JJ
I really do not know what is the IP3 of a DG Mosfet as I never have
seen any value reported. I hope some readers may have data on this.
Certainly we have to look at IIP3 and OIP3 as the DGM mixer may have
high conversion gain 15-20dB and is somewhat limited in dynamic range.
Maybe it could have an OIP3 around +5dBm equivalent to an IIP3 of -10
to -20dBm .... more or less similar to the NE602.
Much better than NE602 on IP3, lower on conversion gain (6-10db).
The big differnce is the DGFET mixer is lack of balance so feed
through is an issue. However a pair used in a balanced arrangement
can be a very good low noise mixer. Many of the balanced FET designs
in past issues of Ham Radio and were from hams in Europe where
strong SW broadcast is the normal environment.

A thought on comparing mixers:

NE602, Low power needed around 2-3ma at 6V, IP3 -10dbm.
(for 18mW of power this gilbert cell offers good
perfomance). Low power needed is very small
to nonexistant due to internal OSC. Low NF of <5db at
45mhz. High conversion gain (typicall 15-17DB).

DGmosfet power needed around 4-6ma/12V, IP3 0 to 3dbm
Also low to no LO power needed. Conversion gains
of 5-8db typical. Properly matched, low noise.

Balanced DGmosfet(2) power needed ~8-10ma/12V, Ip3>3dbm
Low to minimal LO power needed. Low conversion gains
but low noise as well.

Ad831 high power needs but all the advantages of Gilbert cell
mixer with very high IP3s (variable with power
consumption setting).

DBM power needed zero (however 10mW of LO required!)
IP3 around 0DBM, requires post mixer IF amp or termination
and has 6-8db loss. System power required for mixer is
usually around 100mW or higher.

Fet switching mixers are all over the map on power required
some requiring little and some needed a great deal of LO power
to drive correctly, they may also require supporting pre and
post amps to insure noise figure and system gain.

Generally the more power, the better the IP3. In cases where power
is available this may not be an issue. Balanced designs regardless
of devices are more robust than single ended. For portable systems
where battery lifetime is a consideration that can be a determining
factor.
Post by JJ
The DGM mixer was a common mixer in the first series of transistorized
RTX with valves PA (FT101Z, TS820 etc). The DGM has a good square law,
better than JFET.
JFETS in balanced active mixers do very well. It's also possible to
use two Jfets to replace a DG mosfet and in most designs that will
perform the same as the DGFET.
Post by JJ
74HC4066 is OK at low frequency conversion but it has a haigh
conversion loss, around 8-10dB and so associated noise figure.
Old device and higher series on resistance.
Post by JJ
Today the switched mixer should use fast bus switches like the FST3125
family when looking at high performance mixers. The examples are the
CDG 2000 and STAR (Pic-A-Star) projects with IP3 around +40dBm and
+36dBm, where the H-Mode Mixer with FST3125 is used.
All of that is good but unless the post mixer amp, IF and filters
following it are up to the task and well matched the results can be
very disappointing. When going to that level of performance one
needs to look at the recieving system and examine carefully.

Allison
JJ
2006-03-21 16:18:27 UTC
Permalink
Anyone remember the CA3028 singly balanced mixer?
It was just 2 transistors as difference amp and current source transistor.
What would the IP3 be for it?

JJ
A***@nouce.bellatlantic.net
2006-03-21 17:39:01 UTC
Permalink
Post by JJ
Anyone remember the CA3028 singly balanced mixer?
It was just 2 transistors as difference amp and current source transistor.
What would the IP3 be for it?
JJ
I've used it and have a box load of them (3028, 3026 and related
parts) still. IP3 depends on bias applied and load it's decent and
far above the single transistor mixer. If the input and output are
balanced the IP3 is around the -10 to 0dbm range depending on
bias current and specific circuit. It's advantage is the LO power
needed is quite low as well as low DC power needs.

It's still possible to find the parts or related differential pair
with current source parts (RCA had a whole series of them).

They also make decent AGC If amps.

Allison
JJ
2006-03-23 17:35:41 UTC
Permalink
Post by A***@nouce.bellatlantic.net
I've used it and have a box load of them (3028, 3026 and related
parts) still. IP3 depends on bias applied and load it's decent and
far above the single transistor mixer. If the input and output are
balanced the IP3 is around the -10 to 0dbm range depending on
bias current and specific circuit. It's advantage is the LO power
needed is quite low as well as low DC power needs.
It's still possible to find the parts or related differential pair
with current source parts (RCA had a whole series of them).
They also make decent AGC If amps.
Allison
The singly balanced mixer using 2 DG MOSFETs seem to be better still.
Does it have a higher IP3 than a single ended mixer?

Why use a AD831 if the noise figure is 10dB?

JJ
A***@nouce.bellatlantic.net
2006-03-23 18:51:58 UTC
Permalink
Post by JJ
Post by A***@nouce.bellatlantic.net
I've used it and have a box load of them (3028, 3026 and related
parts) still. IP3 depends on bias applied and load it's decent and
far above the single transistor mixer. If the input and output are
balanced the IP3 is around the -10 to 0dbm range depending on
bias current and specific circuit. It's advantage is the LO power
needed is quite low as well as low DC power needs.
It's still possible to find the parts or related differential pair
with current source parts (RCA had a whole series of them).
They also make decent AGC If amps.
Allison
The singly balanced mixer using 2 DG MOSFETs seem to be better still.
Does it have a higher IP3 than a single ended mixer?
Balanced mixers generally perfom far better than single ended.
FETs as a balanced mixer do well and if the FETs used are high IDSS
types (J310 or others) the IP3 can be excellent.
Post by JJ
Why use a AD831 if the noise figure is 10dB?
It has gain, high IP3 and 10db my mixer standard is not that bad.
A low gain RFamp with low noise figure will override mixer noise.
Of course the RFamp must have a good IP3 as well.

Allison
JJ
2006-03-24 16:36:11 UTC
Permalink
Post by A***@nouce.bellatlantic.net
Post by JJ
Why use a AD831 if the noise figure is 10dB?
It has gain, high IP3 and 10db my mixer standard is not that bad.
A low gain RFamp with low noise figure will override mixer noise.
Of course the RFamp must have a good IP3 as well.
Allison
Ok, I like to try the AD831.
When is dual supply better than single ended?
There must be a socket to fit the package but is it all right to solder
tiny leads to it for prototyping?

JJ
A***@nouce.bellatlantic.net
2006-03-24 21:26:31 UTC
Permalink
Post by JJ
Post by A***@nouce.bellatlantic.net
Post by JJ
Why use a AD831 if the noise figure is 10dB?
It has gain, high IP3 and 10db my mixer standard is not that bad.
A low gain RFamp with low noise figure will override mixer noise.
Of course the RFamp must have a good IP3 as well.
Allison
Ok, I like to try the AD831.
When is dual supply better than single ended?
There must be a socket to fit the package but is it all right to solder
tiny leads to it for prototyping?
JJ
It's better to avoid the socket, unless its a specially designed
to be RF complient.

The circuit is simpler with dual supply, fewer resistors.

Allison
Giancarlo Gian Moda, I7SWX
2006-03-25 17:10:39 UTC
Permalink
Hi all,

Allison has given important information on mixers data.

His comments are very usefull.
Post by A***@nouce.bellatlantic.net
All of that is good but unless the post mixer amp, IF and filters
following it are up to the task and well matched the results can be
very disappointing. When going to that level of performance one
needs to look at the recieving system and examine carefully.

It is right to relook at a receiver design when selecting a high
performance mixer. The limiting problems will show up in other stages
as the improuvements may not be very detectable.... amplifiers,
filters, l.o. phase noise...

Well one need to start from somewhere to improuve a receiver. If the
non mixer stages have a maximum IP3 of around +15dBm, certainly a
+35/40dBm mixer may not make much of a difference, unless it is
replacing a NE602. In this case you may use the I7SWX 1 transformer
double balanced mixer using the FST3125 (half of an H-Mode mixer) with
an IP3 of +25dBm.

As an example, I defined with afriend of mine a mod on the Elekraft K2.
The 1st mixer (TUF) was replaced by an I7SWX 2T H-Mode Mixer followed
by a 10kHz BW 2 xtal roofing filter . Two tones at 7.050 and 7.070 MHz

IP3 Results of the K2 were:

Original mixer: +15dBm
with H-Mode Mixer: +27dBm

as you may see these are important differences. The actual limiting
stage is the original post mixer amplifier.

73

Gian
I7SWx
JJ
2006-03-25 17:48:23 UTC
Permalink
Post by Giancarlo Gian Moda, I7SWX
Well one need to start from somewhere to improuve a receiver. If the
non mixer stages have a maximum IP3 of around +15dBm, certainly a
+35/40dBm mixer may not make much of a difference, unless it is
replacing a NE602. In this case you may use the I7SWX 1 transformer
double balanced mixer using the FST3125 (half of an H-Mode mixer) with
an IP3 of +25dBm.
Gian
I7SWx
That is a very interesting suggestion. Where can I find the 1 transformer
DBM using the FST3125?

JJ
Giancarlo Gian Moda, I7SWX
2006-03-26 09:34:28 UTC
Permalink
Hi JJ,

the note on the 1T DBM was reported in RadCom, Technical Topics column,
June 2004.

I asked a friend to insert all mixer infos in my web page
www.qsl.net/i7swx but I do not know when he will be able to do it. In
my web page in homebrewing you my read the experiences and measuremnts
done by JA9TTT on both versions (3T & 2T) of the H-Mode Mixer. This is
a software package translation from japanese but it is understandable.
You will be able to see Spectrum Analyzer measurents too.

If you send me (callsignatyahoo.com) your e-mail address I can send you
the complete 1T DBM note in english (more stuff than in TT).

A NOTE for ALLISON ... I see you have been fiddling with DEC stuff.
Were you working for DEC or a customer? I was with DEC for a quarter of
a century ...hi

73

Gian
I7SWX
A***@nouce.bellatlantic.net
2006-03-26 18:27:28 UTC
Permalink
On 26 Mar 2006 01:34:28 -0800, "Giancarlo Gian Moda, I7SWX"
Post by Giancarlo Gian Moda, I7SWX
Hi JJ,
the note on the 1T DBM was reported in RadCom, Technical Topics column,
June 2004.
I asked a friend to insert all mixer infos in my web page
www.qsl.net/i7swx but I do not know when he will be able to do it. In
my web page in homebrewing you my read the experiences and measuremnts
done by JA9TTT on both versions (3T & 2T) of the H-Mode Mixer. This is
a software package translation from japanese but it is understandable.
You will be able to see Spectrum Analyzer measurents too.
If you send me (callsignatyahoo.com) your e-mail address I can send you
the complete 1T DBM note in english (more stuff than in TT).
QEX also has a few articles, also check out KD1JVs site as he has a
project that uses similar.
Post by Giancarlo Gian Moda, I7SWX
A NOTE for ALLISON ... I see you have been fiddling with DEC stuff.
Were you working for DEC or a customer? I was with DEC for a quarter of
a century ...hi
Hi Gian,

I was a digit (dec empoyee) for 10 years. I still have and use PDP-8,
PDP11 and uVAX. I show up in the CP/M and PDP11/VAX usenets.

Your name is familiar to me from many places and more than a few
articles.

Just tried out my new homebrew portable 6M SSB ring on the local net
this morning and it works as wished.

Allison
KB1GMX
Giancarlo Gian Moda, I7SWX
2006-03-27 13:48:37 UTC
Permalink
Hi Allison,

thanks for the update. I was with DEC from 69 till 93 ... serial #
7133 ...hi

My #1 machine (FS) was PDP9 , also PDP series and alittle bit PDP11 and
15... I applied several mods in the field that made crazy many people
in DEC ... but we were working in he field without spare parts ...I
replaced transistros, transformers (core mem drivers) ICs...etc.... The
day I was told not to do on my way...I saw something wrong.. the
largest PDP9 installation in the world (I3) went into smoke ... then my
boss changed mind...hiu

Glad you got your 6m SSB rig working nicely.

BTW if you are in Mass and you do hear my good friend Art, K1GBX, on 6
meters please say hallo for me... he will get crazy...hi I also have
some other good friends from the old DEC Ham club and others...

73

Gian
I7SWX
A***@nouce.bellatlantic.net
2006-03-27 15:52:23 UTC
Permalink
On 27 Mar 2006 05:48:37 -0800, "Giancarlo Gian Moda, I7SWX"
Post by Giancarlo Gian Moda, I7SWX
Hi Allison,
thanks for the update. I was with DEC from 69 till 93 ... serial #
7133 ...hi
From '83 to '93. Had a 6digit number as I was part of CSSE and
central engineering for high end printing systems. I really liked the
Mill.
Post by Giancarlo Gian Moda, I7SWX
My #1 machine (FS) was PDP9 , also PDP series and alittle bit PDP11 and
15... I applied several mods in the field that made crazy many people
in DEC ... but we were working in he field without spare parts ...I
replaced transistros, transformers (core mem drivers) ICs...etc.... The
day I was told not to do on my way...I saw something wrong.. the
largest PDP9 installation in the world (I3) went into smoke ... then my
boss changed mind...hiu
Ah, the plastics fire if memory serves. Never played with any of the
15bit hardware or the controls. Just PDP-8, 10, 11 and VAX.
Post by Giancarlo Gian Moda, I7SWX
Glad you got your 6m SSB rig working nicely.
Just finishing up a 85W brick for it now.
Post by Giancarlo Gian Moda, I7SWX
BTW if you are in Mass and you do hear my good friend Art, K1GBX, on 6
meters please say hallo for me... he will get crazy...hi I also have
some other good friends from the old DEC Ham club and others...
Just talked to him this weekend! He's about 30miles NE of me. Next
net I'll mention you name and say HI.. One of his project was to
restart the old AM nets, he's got maybe 6-8 regulars.

Small world, smaller if we get some propagation.

Allison
KB1GMX , I can be contated via email mycall at the ARRL reflector dot
net.
Giancarlo Gian Moda, I7SWX
2006-03-28 20:36:21 UTC
Permalink
Hi Allison,
Post by A***@nouce.bellatlantic.net
From '83 to '93. Had a 6digit number as I was part of CSSE and
central engineering for high end printing systems. I really liked the
Mill.

Ah..OK CSSE ... I sold a PDP11/10 with a card reader from CSS to the
Italian Goverment as a donation to a University in Peru'...the highest
PDP11 installation on the world at that time...hi
Post by A***@nouce.bellatlantic.net
Ah, the plastics fire if memory serves. Never played with any of the
15bit hardware or the controls. Just PDP-8, 10, 11 and VAX.

OK. PDP9 and 15 wer 18bits .. >I did some work on DEC10, 11 and when
the VAX came out I was in sales... I made some of the first sales in
Italy and most of my VAXes were linked in Decnet to PDP11/70 and 45.

Just finishing up a 85W brick for it now.
Post by A***@nouce.bellatlantic.net
Nice power. On 6 meters I have an FT726 with the 50MHz converter (RX) supermodified ... the only original pieces are 1x Jfet and 2 IFTs ...hi.
I also have an FT920 a Xmas gift (a few years ago) from my good FRIEND
(with capital letters !!!) Art (and his wife)...hi
Post by A***@nouce.bellatlantic.net
Just talked to him this weekend! He's about 30miles NE of me. Next
net I'll mention you name and say HI.. One of his project was to
restart the old AM nets, he's got maybe 6-8 regulars.

How nice... Too bad we did not get in touch during my business
quarterly trips in W1 (for confessions with my boss 2000 to 2003...hi).
I was usually at K1GBX house at weekends ... eating lobsters and other
good stuff...hi.When I took my wife and my daugther Luisa for her first
trip we even had a story in a local newspaper and picture...hi
I m sure when you will tell him we had this kbd chatting he will jump
on his chair and get crazy...hi For a jike tell him we had a QSO on 6
meter SSB... than you will tell him the true story...hi
Post by A***@nouce.bellatlantic.net
Small world, smaller if we get some propagation.
Yes, amateur radio makes the world really small, some time ... hi

Sorry for this non Mixers conversation for our friends.... I will
comment on the following...

73

Gian
I7SWX


Allison
KB1GMX ,
A***@nouce.bellatlantic.net
2006-03-28 23:49:06 UTC
Permalink
On 28 Mar 2006 12:36:21 -0800, "Giancarlo Gian Moda, I7SWX"
Post by Giancarlo Gian Moda, I7SWX
Hi Allison,
Post by A***@nouce.bellatlantic.net
From '83 to '93. Had a 6digit number as I was part of CSSE and
central engineering for high end printing systems. I really liked the
Mill.
Ah..OK CSSE ... I sold a PDP11/10 with a card reader from CSS to the
Italian Goverment as a donation to a University in Peru'...the highest
PDP11 installation on the world at that time...hi
CSSS, Somputer Special Systems. CSSE Customer Services Support
Engnieering.
Post by Giancarlo Gian Moda, I7SWX
Post by A***@nouce.bellatlantic.net
Ah, the plastics fire if memory serves. Never played with any of the
15bit hardware or the controls. Just PDP-8, 10, 11 and VAX.
Err, argh, typo on the 15.
Post by Giancarlo Gian Moda, I7SWX
Post by A***@nouce.bellatlantic.net
Just finishing up a 85W brick for it now.
Nice power. On 6 meters I have an FT726 with the 50MHz converter (RX)
supermodified ... the only original pieces are 1x Jfet and 2 IFTs ...hi.
I also have an FT920 a Xmas gift (a few years ago) from my good FRIEND
(with capital letters !!!) Art (and his wife)...hi
;) I have TenTec 6n2 as my commercial radio and 5 homebrew designs
each with a different topography. Just for 6m. The nicest one has
a -137Dbm mds and a 92db dynamic range with a modest sub 4 db noise
figure.

Next project is a small single 4CX250 amp for when 20W just doesn't
cut it.
Post by Giancarlo Gian Moda, I7SWX
Post by A***@nouce.bellatlantic.net
Just talked to him this weekend! He's about 30miles NE of me. Next
net I'll mention you name and say HI.. One of his project was to
restart the old AM nets, he's got maybe 6-8 regulars.
How nice... Too bad we did not get in touch during my business
quarterly trips in W1 (for confessions with my boss 2000 to 2003...hi).
I was usually at K1GBX house at weekends ... eating lobsters and other
good stuff...hi.When I took my wife and my daugther Luisa for her first
trip we even had a story in a local newspaper and picture...hi
I m sure when you will tell him we had this kbd chatting he will jump
on his chair and get crazy...hi For a jike tell him we had a QSO on 6
meter SSB... than you will tell him the true story...hi
Will do, that will get him going. I figure that will take at least 5
minutes or more for him to wind down.
Post by Giancarlo Gian Moda, I7SWX
Yes, amateur radio makes the world really small, some time ... hi
Sorry for this non Mixers conversation for our friends.... I will
comment on the following...
Hey more than a few of us here can help on the subject, it's a pretty
big one and Mixers are used everwhere.

Later,
Allison
Kb!GMX
Giancarlo Gian Moda, I7SWX
2006-04-01 08:42:18 UTC
Permalink
Hi Allison,

OK for everything.

Please keep me updated on Art's reaction...hi

73

Gian
I7SWX
Paul Keinanen
2006-03-31 19:39:29 UTC
Permalink
On 28 Mar 2006 12:36:21 -0800, "Giancarlo Gian Moda, I7SWX"
Post by Giancarlo Gian Moda, I7SWX
Ah..OK CSSE ... I sold a PDP11/10 with a card reader from CSS to the
Italian Goverment as a donation to a University in Peru'...the highest
PDP11 installation on the world at that time...hi
By the way, did you have to install extra fans to compensate for the
low air density (and thus lower heat carrying capacity) at high
altitudes ?

Some electronic systems have quite low maximum altitude ratings due to
this. The low air density might also cause problems for amateur radio
equipment if operated at full power at high altitudes.

Paul OH3LWR
Giancarlo Gian Moda, I7SWX
2006-04-01 08:40:31 UTC
Permalink
Hi Paul,

This installation was somewhere on the Andes were there was a
astronomic site.

We did not need to use extra fans or other "helping" devices. Certainly
this set-up
was not on an open air site, it was in an building. So "air" presence
was OK for humans.

BTW, in 1966 while working in UK for a small specialized company, I was
involved in a project were we were producing the Eidofor Colour TV
Projectors (Philips Group). Philips had sold 3 of these units to be
used in Mexico City for the Olympic Games. To be sure we would not have
problems with the 3 colour tubes for explosion (I forgot M.C. height
a.s.l.) we tested the system in the BAC UK air chamber (vy expensive
tubes...hi).

73

Gian
I7SWX
Paul Keinanen
2006-04-01 11:07:19 UTC
Permalink
On 1 Apr 2006 00:40:31 -0800, "Giancarlo Gian Moda, I7SWX"
Post by Giancarlo Gian Moda, I7SWX
Hi Paul,
This installation was somewhere on the Andes were there was a
astronomic site.
Apparently this was much before the Atacama site in Chile became
popular.
Post by Giancarlo Gian Moda, I7SWX
We did not need to use extra fans or other "helping" devices. Certainly
this set-up
was not on an open air site, it was in an building. So "air" presence
was OK for humans.
At sea level, the air density is about 1.2 kg/m³ and at altitudes of
4000 m (La Paz?), the air density is only 0,8 kg/m³. In a forced
cooled system, the critical parameter is the _mass_ flowing though the
system and the amount of degrees the air is heated.

In order to run the same mass of air through the system, you would
have to use 1.2:0.8 or 1.5 times the air volume in the mountains.
Post by Giancarlo Gian Moda, I7SWX
BTW, in 1966 while working in UK for a small specialized company, I was
involved in a project were we were producing the Eidofor Colour TV
Projectors (Philips Group). Philips had sold 3 of these units to be
used in Mexico City for the Olympic Games. To be sure we would not have
problems with the 3 colour tubes for explosion (I forgot M.C. height
a.s.l.) we tested the system in the BAC UK air chamber (vy expensive
tubes...hi).
Wasn't Eidophores used during the Apollo flights to show the orbits in
Huston ?

Paul OH3LWR
Giancarlo Gian Moda, I7SWX
2006-04-01 19:32:49 UTC
Permalink
Post by Paul Keinanen
Wasn't Eidophores used during the Apollo flights to show the orbits in
Huston ?
Paul OH3LWR
Hi Paul,

I do not know but in middle of 60s that was the only TV color
projectors
designed by a Swiss engineer, assembled by Peto Scott Instrument Ltd
for Philips
in Addlestone, Surrey UK.
It was tested in the central lab were we were mainly working on
military equipments.
Many of those big blocks were sold world wide. I do remember the Mexico
Olympics
because we had to do that special test in the chamber room were
airplanes were
tested for all 3 machines.

For Peru', I believe it was somewhere over 2000 meters asl and it was
not later than 1980.

73

Gian
I7SWX
Giancarlo Gian Moda, I7SWX
2006-04-01 19:32:54 UTC
Permalink
Post by Paul Keinanen
Wasn't Eidophores used during the Apollo flights to show the orbits in
Huston ?
Paul OH3LWR
Hi Paul,

I do not know but in middle of 60s that was the only TV color
projectors
designed by a Swiss engineer, assembled by Peto Scott Instrument Ltd
for Philips
in Addlestone, Surrey UK.
It was tested in the central lab were we were mainly working on
military equipments.
Many of those big blocks were sold world wide. I do remember the Mexico
Olympics
because we had to do that special test in the chamber room were
airplanes were
tested for all 3 machines.

For Peru', I believe it was somewhere over 2000 meters asl and it was
not later than 1980.

73

Gian
I7SWX

JJ
2006-03-21 19:53:33 UTC
Permalink
For completeness what are the IP3 for a single BJT and single FET mixer?
The ARRL handbook recommended a source resistor voltage of 1V as optimum
bias.

JJ
A***@nouce.bellatlantic.net
2006-03-21 20:46:38 UTC
Permalink
Post by JJ
For completeness what are the IP3 for a single BJT and single FET mixer?
The ARRL handbook recommended a source resistor voltage of 1V as optimum
bias.
JJ
Source resistor has to be correct for the device and 1V [often] may
not be it. Also LO drive will require tailoring for best IP3 and
conversion gain.

Single ended BJT mixers are pretty weak but can offer good conversion
gain. Fets in general are much better but tend to be low conversion
gain. However specific numbers are hard to appraise as circuits vary
considerably. Either would only be considered for low end or where
IP3 is not a consideration or are overridden by other considerations.

The mid 80s low to mid range radios did use Jfets as mixers often
with enough preselection to keep out the offending signals.


Allison
Risto Tiilikainen
2006-03-04 10:18:55 UTC
Permalink
Post by Ken Scharf
Post by Dave Platt
The FET-switch mixers seem to be a really nice alternative, and
although they've gotten relatively little visibility in amateur-radio
applications they've become very popular in commercial use (e.g.
cell-phone handsets). I haven't yet had a chance to play with these
myself but they look like fun!
There was an article in the ARRL HB (1995 and probably others)
about using an SD5000 quad mosfet as a mixer. I bought a few
of these transistors a few years ago (now where did I put them,
got lost in the last move!). They are supposed to make
a very good mixer. The only disadvantage was the circuit
required injection at twice the lo frequency as it used a jk flip
flop to develop the required 180 degree injection. I suppose
a balum could do the same thing. In any case the AD9954 DDS I am
thinking of using can clock to 400mhz and develop output to
160mhz, so getting up to 80mhz (for use with a 9mhz if) would still
be no problem.
Hi !

At the same time Signetics included to the same fet family a double fet
SD6000
It was planned for front end applications
Single fets were done under type numbers SD305 and SD306
All three were N-channel enhancement types. Positive bias only.
I wonder whether Signetics still exists. Signetics was bought by Philips
company.

I use those fets successfully in one of my conventional type homebrew
tranceivers.
The first if is having a 10.7 MHz xtal filter. That is mixed down to
455 kHz either with 10.245 MHz or 11.155 MHz xtal
Sideband selection is done with the selection of mixing direction.
455 kHz is using Collins mechanical filter having very good shape factor
and special skirt for lower sideband use.

73, Risto OH2BT
Dan Andersson
2006-03-28 00:10:51 UTC
Permalink
Post by a***@gmail.com
hello,
I'm building my first reciver. I can't choose what kind of mixers
should I use. I have read that diode ring mixers are far superior
compared to dual gate mosfet mixers.
Is this true for both - first (RF / VFO) and second (IF / BFO) stages?
Or is there any real difference at all?
thanks
There are a lot of historical chips and solutions around on the market and
in the litterature. I use the Analog Devices AD8343 mixer, good for DC to
2.5GHz and with a IP3 of 16.5 dBm.

The 602 is now replaced with the 612 but is still an old design.

If you design your mixer ( /doubler ) with AD8343 and use proper RF
transformers from CoilCraft, it just cannot go wrong... ( well, Murphy et
al of course... )

Cheers

Da / M0DFI
Giancarlo Gian Moda, I7SWX
2006-03-28 20:44:09 UTC
Permalink
Post by Dan Andersson
The 602 is now replaced with the 612 but is still an old design.
If you design your mixer ( /doubler ) with AD8343 and use proper RF
transformers from CoilCraft, it just cannot go wrong... ( well, Murphy
et
al of course... )
Post by Dan Andersson
Da / M0DFI
Hi Dan,

Yes the AD8343 is an interesting mixer... but for me it may be too
noisy (10dB). THe IP3 is certainly much better (+15dBm) than the
602/612 (-15dBm), but this one has a lower NF of 5dB.
If you build my 1 homebrew transformer (50c) FST2125 DBM you do get am
IP3 of +25dBm NF 5dB ... adding a post mixer 10dB LNA you probably get
a 3dB NF.
Certainly you have to select the most valid mixer for the specific
project as always we have to compromise and give priority to one of the
many parameters.

73

Gian
I7SWX
Saandy , 4Z5KS
2006-03-29 09:49:32 UTC
Permalink
guys, bear in mind that the intercept point is primarily a function of
the operating quiescent poit. you cam make it high, really high and get
superlative mixers, even single ended. i designed one aroun the now
defunct VMP4 from Siliconix with an Id of 500mA!!! the damned thing had
nearly +50dB IP#. About the same went with the LNA. All in all 24W
power dissipated in the front end. How many of us can afford that or
need that for the matter?
drop it! what the guy needs is some reasonable value of dynamic range
and around 10-12 dB for IP3. he isn't going to build a radio telescope
receiver on his first attempt. Be reasonable!
with FET's the thing to ensure is that the operating point is
setablished with Vgs set at about half of the pinchoff voltage to
ensure maximum dynamic range. a noise figure under 10dB is adequate for
HF. it's not the galactic noise that limits your reception!
Saandy 4Z5KS
Giancarlo Gian Moda, I7SWX
2006-03-29 13:17:03 UTC
Permalink
Hi Saandy,

I have some russian power FET similar to the VMP4 and probably better
but there is no sense to use them as they are no more in production and
diffcult to purchase in East Europe.

I believe a discussion on Mixers is important particularly for a
newcomer. He must know all the different kind of mixers than he will
have to decide.... it is like talking about small car or motobykes
because it is better run with a bicycle...one should not know about
racing car or motobykes.... he can putchase a byke tha add a little
mosquito engine or pss to a bigger motobyke or a small car later ...
but he will know what he has (he will have to push a lot with his legs
and then see to push less or not at all).... it is like for a portable
RTX were consumption is at premium... you cannot put (probably an
h-mode mixer has you will have to spend 30-50mA plus an amplifier with
other 20mA versus a NE612 that needs 3mA !!! To hell the IP3... one can
use his brain as a DSP...hi

73

Gian
I7SWX
Saandy , 4Z5KS
2006-03-29 09:36:06 UTC
Permalink
go with the simplest one! something like a dual gate FET: that's going
to ease problems with the local oscillator.
But build it as a separate module on a small PCB, and wire it in. that
way you'll be able to experimwnt later on by replacing the module. that
way you can go from easy and simple to complex and high perfoemance.
for a first it hard to beat the method. i see you're getting too much
info, especially for a newcomer!
stay in touch, I'll try to help you to the limit of my abilities.
Saaandy 4Z5KS
Loading...