|
scwallac Post subject: Below 50MHz Posted: Fri Aug 04, 2006 1:32
am Lieutenant Joined: Fri Aug 04, 2006 1:22 am Posts:
3 Location: Alexandria, VA Hi, I am working on a PCB layout
comprised entirely of MMIC components. I have a silly question regarding
IF circuitry below 50MHz.... Given the long wavelengths at these
low IF frequencies, is it really necessary to bother with microstrips
for the IF portions of my circuitry? A quick calculation shows that
1/20th the wavelength of 50MHz is something like 16", which is well
below the total size of my circuit board or any possible trace lengths.
So, is microstrip a waste of time for IF circuitry? -Sean
Top IR Post subject: Posted: Fri Aug 04, 2006 4:38
am Site Admin Joined: Mon Jun 27, 2005 2:02 pm
Posts: 373 Location: Germany Hello Sean, Microstrip lines
are not a bad idea at all even at low frequency, because they are located
in the top latyer of the PCB - This helps to debug problems with test
equipment. In general, with whatever kind of transmission line
you will use, it is important to keep the right dimesnions of the line
to match the characteristic impedance. *) Another note:
Of course that the lines needs to be as short as possible even for these
lwo frequencies, because as someone mentioned if you have lines which
are connected to the IF port fo a Mixer they transfer relatively high
power levels which can radiate a lot. Hope this helps!
_________________ Best regards, - IR Top
yendori Post subject: Posted: Fri Aug 04, 2006 1:16 pm
General Joined: Thu Sep 25, 2003 1:19 am Posts:
50 Location: texarcana Hi, Immediatly after a mixer for
example, the LO feedthrough signal may be high in power, greater than
0dBm at times. Eventhough your desired signal is below 50 MHz,
it it s important to be aware of you spurious signals. If you traces
are high in impedance you may cause unwanted radiation. Rod
Top languer Post subject: Posted: Fri Aug 04, 2006
3:55 pm Captain Joined: Sun May 14, 2006 8:53 pm
Posts: 17 Location: Earth I am not sure I fully understand the
question, but let's see. At 10MHz, 100MH, 1000MHz the microstrip
width for a specific impedance is all the same (i.e. Zo is not related
to frequency). At higher frequencies where the with may be comparable
to the wavelength (say 1/8) you want to adjust H or Er to avoid this,
but at lower frequencies it's a non issue. So for impedance purposes,
it's all the same. So why would you want to provide microstrip
connections? To minimize reflections, power loss, radiation (for the
later it may even be worth considering CPW). You could use stripline,
but why would you? This may require additional layers (save for suspended
stripline) and you have no access to the traces (it may help in radiation
though). Whatever you do there is one thing you most definetly
want to have, a return path for the signal. So if you do not have differential
lines (unlikely from your description), microstrip or stripline would
guarantee this. And microstrip is pretty much "free", two-sided circuit
with ground on bottom. Top scwallac Post subject:
Posted: Fri Aug 04, 2006 4:40 pm Lieutenant Joined: Fri
Aug 04, 2006 1:22 am Posts: 3 Location: Alexandria, VA Quote:
Eventhough your desired signal is below 50 MHz, it it s important to
be aware of you spurious signals. If you traces are high in impedance
you may cause unwanted radiation. yendori - Interesting point.
Thanks. Quote: I am not sure I fully understand the question,
but let's see languer - Let me phrase the question a little
bit differently: If I am working exclusively with frequencies
below 50MHz, and my trace lengths are all less than 12", then by "textbook"
theory my traces are not acting like transmission lines at these frequencies
anyway. If the traces are not acting like transmission lines, then why
not just pretend this is an audio signal and route the signal around
without controlling the trace impedance at all? (BTW, I'm not
arguing that this is actually true - rather I am wondering if this is
true and soliciting your opinion!) Top languer
Post subject: Posted: Fri Aug 04, 2006 6:29 pm Captain
Joined: Sun May 14, 2006 8:53 pm Posts: 17 Location: Earth
Quote: If I am working exclusively with frequencies below 50MHz,
and my trace lengths are all less than 12", then by "textbook" theory
my traces are not acting like transmission lines at these frequencies
anyway. I guess what I tried to clarify is that no matter what
the frequency they are transmission lines, even a wire is a transmission
line (by "textbook"). Quote: If the traces are not acting
like transmission lines, then why not just pretend this is an audio
signal and route the signal around without controlling the trace impedance
at all? Using microstrip does not imply controlled impedance.
Whenever you use a trace-dielectric-ground you have by definition a
microstrip. If you decided to control its impedance or not, that is
a different matter. Ground "flooding" and other similar RF techniques
were extensively used years ago when double-sided PCBs were leading
edge technology. This was for IF frequencies of 10-20MHz. My point is,
you do not have to use microstrip, stripline or anything similar. But
whatever you use you will be using transmission lines, the key thing
is to ensure the proper return path. You do mention the use of MMICs,
if this refers to MMIC amplifiers and such then careful layout consideration
is important because most devices now have gain up to several GHz. Meaning
that unless you suppress this gain, you have to make sure how the device
will react at these frequencies. All this of course is quite
important if one is making some sort of commercial product, for some
hobby-type endeavor one could take many exceptions. Top
scwallac Post subject: Posted: Mon Aug 07, 2006 11:52 am
Lieutenant Joined: Fri Aug 04, 2006 1:22 am Posts:
3 Location: Alexandria, VA Thanks for the insight everyone!
Posted 11/12/2012
|