EMRFD Message Archive 1361
Message Date From Subject 1361 2008-01-23 18:26:30 bkopski DIODE RING MIXER DISCUSSION FOLLOWUP I have to say Thank You Very Much to the many folks who reacted both
here and privately to my original post. I've been following up on
your inputs / suggestions with more testing (and mixer-making!) and
now have an interim report as some have requested.
One area of question concerned the core material I used. I have
confirmed with test winds that I indeed have the FT37-43 parts I
first described. I have also confirmed that all trifilar windings
and associated winding connections (polarities) were (are) correct.
HOWEVER I did find one assembly miswire namely re: "which wires
goes to which connector" in the case of the HB DBM. Here I had an
error. While this mixer "worked", this error accounted for the
reported high losses (~13 dB) and minimal LO rejection. It also
accounted for the difference in output amplitude between the two main
output frequencies as displayed in the earlier photo for that mixer.
Upon correcting this the DBM losses reduced to the expected ~6 dB
area and the two main outputs are equal. (Previously they differed
As several suggested, I have tried other diodes namely 1N5711 for
the 1N4148 and find that my HB DBM still displays the same nominal ~6
dB loss. However, I find as expected that this nominal performance
can now be achieved with a few dB lower LO drive than with the
With respect to the suggestions re: driving the 1N4148 diodes harder
(more than the +7 I used), I find rather small improvement with
drives up to +12 (all I can muster right now). Conversely, drive
reduction to +5 results in relatively small increases in the
displayed losses. In summary, drives from +5 to +12 to the 1N4148
DBM only move the output amplitude over less than 2 dB total and the
often described +7 value seems to be "good enough" for now at
least. For the 1N5711 DBM, output begins to decrease below about +3,
and does not significantly improve up to +12.
From your inputs, I have come to realize that my choice of test
frequencies was not the best given the relationship between 5 and 15
mHz. I'm changing this.
One input suggested there are better mixers (eg H-mode)but my topic
of interest remains the "simple" diode ring for now, for the project
at hand. (Remember - what was really getting to me was why I was not
getting the routinely reported performance for the ring diode
configuration and not whether it was the best mixer.)
Regarding the HB SBM this one remains a mystery. I have completely
rebuilt this version twice today with both new transformers and both
diode types. The loss however remains ~ 10 dB clearly more loss
than any of the DBMs I now have working. Is this what an SBM does
or could I still have something wrong here?
Interestingly, in all the above, I have found that permutations of
the three ports can result in some interesting results. Swapping
about the RF, LO, and IF functions actually resulted in
several "nicer" output displays over the SA displayed frequency
range with no measurable changes in the amplitudes of the main
outputs i.e. losses remained essentially constant - including
with the SBL-1 and ADE-1. What did vary was the amplitude of the
many products (other than the main outputs) in the SA display. In
general, the SBL-1 and ADE-1 continue to be overall "better"
performers with respect to the displayed output spectrum.
I had one typo in my previously posted text: In "Aside from loss
values of about 10 and 13 dBm respectively " "dBm" should read "dB".
One reader privately expressed some confusion re: the SA scales in
the photos. To clarify, the first grid line down from the top grid
line represents the RF signal input level of -10 dBm my reference
level from which mixer losses are measured. (Note again that each
main grid represents 10 dB.) Thus, a mixer loss of 6 dB would be
displayed at 60% of the next main grid down. A loss of 10 dB would
be shown at one full grid down from that -10 dBm reference line = 2
grid lines from the top.
Anyway, given all your assistance, I'm now "straightened out" re:
building more DBM mixers but as one person commented - building may
not make sense given the low cost and good performance of the ADE-1.
Now if only I could get my hands around that nagging, large HB SBM
loss issue ..hi.
Cordially and 73,
1362 2008-01-23 18:55:47 Allison Parent Re: DIODE RING MIXER DISCUSSION FOLLOWUP Hi,
One last thing to try in the HB mixer. While I've used FT37-43
rings with good results I've found BN43-2402 small binocular cores
to be much better. The difference is better balance though
mechanical balance and better bandwidth of the transformer with
As to the cost of making mixers.. A bag of 25 of 43-2402s are
$7 and 1m4148s run about 2-3 cents each and 1n5711s can be had
for about 10 cents in small quantity. For the price of a few ADE1s
I can make a dozen. Then again if I need really small the surface
mount ADE-1 is the best.
One comment, I rarely match diodes unless the requirement is for
optimum balance. Then I have to also build the transformers for
capacitive balance and optimal balanced to unbalanced performance.
The simple trifiler winding on a ring is easiest but not optimal
in that regard.
As to comments
1363 2008-01-23 22:06:54 kerrypwr Re: DIODE RING MIXER DISCUSSION FOLLOWUP Very interesting experiments, Bob.
W7ZOI suggested in Introduction To RF Design that balance in these
kinds of mixers can be improved by "pseudobaluns" at the RF and LO inputs.
(As ever, W7ZOI was ahead of his time in his understanding; his
"pseudobalun" is now usually called a "current balun" but his
description is more-apt).
1364 2008-01-24 07:09:01 Alberto I2PHD Re: DIODE RING MIXER DISCUSSION FOLLOWUP >Is that what others call the "Guanella" balun, or is that a completely different beast ?
> (As ever, W7ZOI was ahead of his time in his understanding; his
> "pseudobalun" is now usually called a "current balun" but his
> description is more-apt).
73 Alberto I2PHD
1365 2008-01-24 12:27:44 Wes Hayward Re: DIODE RING MIXER DISCUSSION FOLLOWUP Hi Alberto, and group,
The term "Guanella" balun is exactly the correct term for the thing
that we often call a current balun. I've been trying for 40 years
now to get a copy of the original reference. It is one of those
things that was published in an extremely obscure journal, yet it is
a paper that has become a central foundation for much of what we
do. Most frustrating.
The term "pseudo balun" is one that I used for IRFD, for it sounded
suitably formal to fit with a college text, which was the intended
purpose of that book. But this was not my fundamental discovery.
Back in SSD, we used the term "sorta balun," which was a descriptor
that I picked up from an old friend, Will Alexander, WA6RDZ, now
SK. Will was an applications engineer for Fairchild Semiconductor
back in the old days when they were just starting to think about some
things that we now take for granted. He was there when the first
transistors came out of Fairchild, probably in 1964 or so. Will
recognized that a simple bifilar winding on a toroid formed a
transformer type that was much different than a traditional balun.
The usual balun is a circuit that generates two voltages that are of
equal magnitude, but are 180 degrees out of phase. Will didn't
evidently think that publicati
1366 2008-01-24 12:28:47 kerrypwr Re: DIODE RING MIXER DISCUSSION FOLLOWUP Yes; like this;
UNBAL ========= BAL
Hope you can understand my poor ASCII Art! :)
1368 2008-01-24 17:47:45 Alberto I2PHD Re: DIODE RING MIXER DISCUSSION FOLLOWUP Wes Hayward wrote:
> Hi Alberto, and group,Hi Wes,
> The term "Guanella" balun is exactly the correct term for the thing
> that we often call a current balun. I've been trying for 40 years
> now to get a copy of the original reference. It is one of those
> things that was published in an extremely obscure journal, yet it is
> a paper that has become a central foundation for much of what we
> do. Most frustrating.
> --- snip ---
thanks for the info. I have been able to find the following reference to that original article, but I think it is
old stuff for you. But maybe somebody, reading the reference, can come out with something...:-)
Guanella, G., “Novel Matching Systems for High Frequencies,” Brown-Boverie Review,
vol. 31, pp. 327-9, Sept. 1944.
73 Alberto I2PHD
1369 2008-01-24 18:04:00 Bob Larkin Re: DIODE RING MIXER DISCUSSION FOLLOWUP Hi All - The Brown-Boveri paper is tough to get ahold of, but the patent is
on the web:
It, along with other reported work done by Guanella, suggests a brilliant mind!
73, Bob W7PUA
At 11:35 PM 1/24/2008 +0100, Alberto I2PHD wrote:
>Wes Hayward wrote:
> > Hi Alberto, and group,
> > The term "Guanella" balun is exactly the correct term for the thing
> > that we often call a current balun. I've been trying for 40 years
> > now to get a copy of the original reference. It is one of those
> > things that was published in an extremely obscure journal, yet it is
> > a paper that has become a central foundation for much of what we
> > do. Most frustrating.
> > --- snip ---
>thanks for the info. I have been able to find the following reference to
>that original article, but I think it is
>old stuff for you. But maybe somebody, reading the reference, can come out
>Guanella, G., âNovel Matching Systems for High Frequencies,â
>vol. 31, pp. 327-9, Sept. 1944.
>73 Alberto I2PHD
1370 2008-01-24 18:32:09 Leon Re: DIODE RING MIXER DISCUSSION FOLLOWUP ----- Original Message -----
1374 2008-01-25 09:59:37 Ray Anderson Re: DIODE RING MIXER DISCUSSION FOLLOWUP The Brown Boveri Review (now known as ABB Review) is available online
at: http://www.abb.com/abbreview however it appears that the archives
only extend back to 1996.
The real old stuff probably resides in some dusty engineering library
24, 2008 6:32 pm (PST)
----- Original Message -----
1377 2008-01-26 04:40:06 w4zcb77 Re: DIODE RING MIXER DISCUSSION FOLLOWUP 1383 2008-01-29 14:31:55 Wes Hayward Re: DIODE RING MIXER DISCUSSION FOLLOWUP Hi Bob and gang,
After the recent discussion regarding the singly balanced mixers, I did
some experiments and simulations. I summarized the experiments on my
web site. See http://w7zoi.net/sbm.pdf I'll also transfer this up
here as a file.
Many thanks for all of the discussion. It has really been
enlightening for me.
Thanks also to those who helped me to finally read the Guanella paper!
1384 2008-01-29 19:21:46 Allison Parent Re: DIODE RING MIXER DISCUSSION FOLLOWUP