Hello Andy,
Thanks for the excellent link, work and documentation! This helps a
lot. It confirms some steps and thoughts and adds a few one.
So what is missed in my circuit is the allpass network that makes a 90
deg delay of the input signal. The R and C values must be adapted for a
frequency range of 11...12.7 khz, to get a proper image frequency
cancellation. I even found another circuit in the web but haven't
realized that the network makes a delay, i just thought it is a LPF for
harmonic supression...
I downloaded your excel sheet and played with the values. Which values
do you recommend for R and C to come in the range where i can find the
optimum by adjusting the potentiometer? Not so easy to find the best
values by playing with R and C. And why do i need 2 allpass filters? A
90 deg shift could be done by using just one, or not?
BTW, is there an advantage using 74HC... devices instead usual CD40...
CMOS devices regarding better image and LO rejection? Switching times
at 137 khz should be no problem for the standard parts, isn't it? I
cannot find the 74FST3253 at my usual distributors. Is there a problem
using the CD4052 instead? The low output impedance is not needed since
i rather what to get a rectangular output to drive a switch mode PA. So
i will add further CMOS gatters at the output after filtering and
amplification...
Best 73 es tnx agn.
Stefan/DK7FC
Am 18.07.2011 14:14, schrieb Andy Talbot:
For an IQ upconverter, you need two mixers fed with SIN and COS
(90 deg separated waveforms) as well as 90 degree shifted baseband
inputs from a PC or audio phase shift network. Then by adding or
subtracting the outputs, one sideband is cancelled and the other
reinforced.
Andy
'JNT
2011/7/18 Stefan Schäfer <[email protected]>
LF,
Maybe someone can help me with an open question about IQ Mixers:
I want to build a mixer that generates a rectangular output at 137.7
kHz (to steer my LF PA) by using SL's digimode terminal at 12.7 kHz and
adding 125 kHz that i get from a down divided 8 MHz xtal (divided by a
74HC4060 and locked into a PLL of my GPS 10 kHz Ref.).
There are several ways to do that of course. It should be a cheep
solution without special parts so i choosed a CD4052 analog switch /
multiplexer and built a mixer that switches the positive and negative
wave at the LO frequency to the output. I am using 2 AF transformers
(1:4 types, 1 on the soundcard side) at the input to achieve a galvanic
decoupling and a +/- voltage.
When looking to the output spectrum i can see a peak at 137.7 kHz at 45
dB above noise. The LO peak is about 20 dB below that. But the image
frequency at 112.3 kHz is just 10 dB below the 137.7 peak. So i need
filters of course. But the ditance between LO+RF and LO-RF is just 25.4
kHz and somehow it is difficult for me to achieve a suitable Q and i
have had problems with temp drifting Cs and do not like to make to much
effort on the filtering.
I do not know much about IQ mixers but i had in mind that the reject
the image frequency by the design and do not need much filtering at the
output. So i read a bit about the mixers in wiki and at some webpages.
I learned that i need 90 deg shifted drive signals for such a mixer and
found a suitable circuit. Since i down divide 8 MHz in a 4060 i am
easily getting a 500 kHz output and so i built that circuit that gives
me the shifted drive signals by using a CD4013 (2x D flipflop). Now i
already have the 125 kHz IQ drive signals and found a circuit how to
arrange a suitable mixer at http://www.elexs.de/iq1.htm
and/or http://www.yu1lm.qrpradio.com/DR2B%20HF%20SDR%20RECEIVER-YU1LM.pdf
But the question is where do i get my 137.7 kHz signal now? There are 4
"cables" that come out of this mixer type, 2 pairs that are fed to a
summer. Then i have the I and Q output that is fed to the soundcard in
these examples. Somehow it begins to dawn on me that i will not get the
wanted signal on the mixers output, right?`The signals can be
calculated in a suitable PC software, which is no solution for me here
of course. I would go on experimenting and will find the solution on
the topic of course but anyway i want to ask here before since i know
that there is someone who knows the answer.
Is there no chance to easily reject the image frequency or do i have to
spend the filtering effort? If there is a complete hardware solution
(no programmable devices, no software), please tell me :-)
The TX mixer should then allow me to remote steer my LF TX and even
make QSY, i.e. decide to call CQ at 137.71 kHz and if no one replies,
make QSY to beacon mode at 136.17x kHz for example. At the output i
want to add a OP device or comparator that forms a rectangular wave to
drive the PA. Later i want to add a AM circuit that filteres the 12.7
kHz and provides a variable DC voltage that will steer the PA's SMPS
step down converter to allow QRO/QRP by varying the volume at the PC...
Tnx in advance.
73, Stefan/DK7FC
|
|
