I have been trying to get an RF application going for a couple of weeks and
am having a terrible time. I am sending data from one PIC to another with
one of these http://www.parallaxinc.com/html_files/products/RF_Mods/27986-27991.asp at
433 MHz. My results are not consistent. Sometime it works just fine,
sometimes it doesn't work at all.
I can debug this project just fine till I get to the transmitter. I can not
tell if the RF unit is actually transmitting well. I have two questions:
#1. Is there some way I can observe RF directly with my scope? Do I need a
special probe?
#2. I am thinking of buying a field strength meter. The field strength
meter I am looking at claims 5 mv sensitivity at 433 MHz. The transmitter I
am using claims 1 mw power output. Would this field strength meter be
useful with such a tiny transmitter?
Jay
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
Assuming your in the us, 433Mhz is very close to on channel 59 of a TV.
Make sure the TV is in "CATV" or cable mode. When you turn it on
you should notce some effect on the picture-- at least that will
let you know your device is transmitting.
Regards,
Steve
> I have been trying to get an RF application going for a couple of weeks and
> am having a terrible time. I am sending data from one PIC to another with
> one of these
> http://www.parallaxinc.com/html_files/products/RF_Mods/27986-27991.asp at
> 433 MHz. My results are not consistent. Sometime it works just fine,
> sometimes it doesn't work at all.
>
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
What kind of antenna setup do you have? At 433Mhz length can be very
critical. Metal objects near the antenna can cause reflections, and
antenna detuning. 1mw isn't much power even at 433Mhz. At that frequency
the receiving antenna is as critical ans the transmitting antenna.
To determine the antenna length (assuming a 1/4 wave whip) the formula
is 300M Meters/Frequency in Mhz x .25 for the quater wave length, then
allowing for the velocity of propogation it that frequency, shorten this
by 5 to 7 %, so as a ballpark figure, multiply the result of the last
equasion by .93 to .95 for the real antenna length.
>
> I have been trying to get an RF application going for a couple of weeks and
> am having a terrible time. I am sending data from one PIC to another with
> one of these
> http://www.parallaxinc.com/html_files/products/RF_Mods/27986-27991.asp at
> 433 MHz. My results are not consistent. Sometime it works just fine,
> sometimes it doesn't work at all.
>
> I can debug this project just fine till I get to the transmitter. I can not
> tell if the RF unit is actually transmitting well. I have two questions:
>
> #1. Is there some way I can observe RF directly with my scope? Do I need a
> special probe?
>
> #2. I am thinking of buying a field strength meter. The field strength
> meter I am looking at claims 5 mv sensitivity at 433 MHz. The transmitter I
> am using claims 1 mw power output. Would this field strength meter be
> useful with such a tiny transmitter?
>
> Jay
>
> --
> http://www.piclist.com hint: PICList Posts must start with ONE topic:
> [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
> I have been trying to get an RF application going for a couple of weeks
and
> am having a terrible time. I am sending data from one PIC to another with
> one of these
How? Serial data or just digital data? If digital, how do you synchronize
it?
> 433 MHz. My results are not consistent. Sometime it works just fine,
> sometimes it doesn't work at all.
Tell me about it! I have been doing RF communication for sometime now, but
I understand those modules much better now. And to answer your question:
> #1. Is there some way I can observe RF directly with my scope? Do I need
a
> special probe?
From what I know, the normal probe typically goes up to 100MHz only. So,
that's out of question. Besides, unless you have real expensive scope, you
can't view 433 MHz signal (the one I have only goes up to 100MHz). But I
find it not very necessary to look at the RF signal. And I can't answer
your 2nd question.
If you are transmitting serial RS-232 data from one PIC to the other through
RF, it's better to do a NOT on both ends. RS-232 standard maintains the
line HIGH where there is no data, and those RF modules don't like them. So,
using a NOT gate on both ends make it a default LOW, and it works much
better in my workbench. Also, those RF modules need some time to "warm-up"
and so it's good to send some garbage bytes in the beginning before actually
sending data (choose 0x55 or 0xaa to create an oscillation).
I am personally working on getting a chip that'll interface digital data
with these RF modules. I want to program a chip that makes the RF
communication transparent (as if it's wire), so that I can send digital data
anytime I want. It has been going pretty well. The test results have been
promising, but still need some more debugging.
Hope that helps!
Rudy
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
> #1. Is there some way I can observe RF directly with my scope? Do I need
a
> special probe?
Depends on your scope, you need at least 500 MHz bandwidth to play.
Your probes also need enough bandwidth.
You can make a rectifying probe with a small signal diode and cap, but you
loose frequency and waveshape info.
Be aware, you are building a device that has to live underneath ham
transmitters at <=1500W.
January has a major contest weekend coming up too.
> #2. I am thinking of buying a field strength meter. The field strength
> meter I am looking at claims 5 mv sensitivity at 433 MHz. The transmitter
I
> am using claims 1 mw power output. Would this field strength meter be
> useful with such a tiny transmitter?
Hard question to answer, depends on the antennas.
I would look at getting at least a decent receiver.
An IC-R8500 or similar with AM, FM, CW, internal attenuator(s) and a signal
strength meter.
At any given impedance, you can calculate the voltage for a given power
output.
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
As far as transmitting data, those inexpensive modules have a problem with
ASCII.
Whatever scheme you use, needs to end up in a 50% duty cycle of high and
low.
This usually means manchester, but you can fake it with a preamble that is
$AA or $55 for a few bytes.
Faster bauds work better here. The absolute time of an $00 or $FF is
shorter, and so the data slicer will be less likely to rail out.
You can't really get to where you need to be with async serial, but you can
get close.
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
>What kind of antenna setup do you have? At 433Mhz length can be very
>critical. Metal objects near the antenna can cause reflections, and
>antenna detuning. 1mw isn't much power even at 433Mhz. At that frequency
>the receiving antenna is as critical ans the transmitting antenna.
The unit has its own antenna on the circuit board. One of the reasons I got
this unit was because I didn't want to have to fool around with antennas.
<G>
>How? Serial data or just digital data? If digital, how do you synchronize
>it?
The transmitter accepts serial data inverted from a PIC uart and sends it to
the receiver.
>If you are transmitting serial RS-232 data from one PIC to the other
through
>RF, it's better to do a NOT on both ends. RS-232 standard maintains the
>line HIGH where there is no data, and those RF modules don't like them.
So,
>using a NOT gate on both ends make it a default LOW, and it works much
[ snip ]
Thanks for the rest of these tips. I will look into them. Sending an extra
0x55 on the front sounds especially promising. One of the pecularities I
noticed was that some bytes got dropped in a long string of zeroes. Many of
these zeroes are just filler bytes and I can change them to 0x55.
Jay
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
THANKS GUYS!! I didn't know these little boards were data sensitive. I put
a 0x55 junk byte on the front of each data block and padded the first block
out with 0x55 instead of zeroes -- and it works perfectly!!
Jay
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
A simple probe consists of a single 1N4148 (better 1N23) diode with a 5cm
wire on one end and a connector mating with the scope probe tip on the
other (turned IC socket). Set the probe to 1Meg. You may have to add a
smaller resistor between the probe tip and ground to make fast pulses
visible. The scope will be used on 10mV/div probably. Do not touch
directly with circuit, 1-2cm away from tx antenna or tx booster (-;)
should work.
The field strength meter will be useful, for the transmitter only
probably. This is also true for the probe.
>
> A simple probe consists of a single 1N4148 (better 1N23) diode with a 5cm
> wire on one end and a connector mating with the scope probe tip on the
> other (turned IC socket). Set the probe to 1Meg. You may have to add a
> smaller resistor between the probe tip and ground to make fast pulses
> visible. The scope will be used on 10mV/div probably. Do not touch
> directly with circuit, 1-2cm away from tx antenna or tx booster (-;)
> should work.
>
> The field strength meter will be useful, for the transmitter only
> probably. This is also true for the probe.
Will work better with an old dot-size germanium diode or a microwave
schottky diode in SOT23 ( small) package.
But the measuring results will be pure informative.
> That's a really clever little tool Peter,
> thank you very much. :o)
> -Roman
>
> Peter L. Peres wrote:
> >
> > A simple probe consists of a single 1N4148 (better 1N23) diode with a 5cm
> > wire on one end and a connector mating with the scope probe tip on the
> > other (turned IC socket). Set the probe to 1Meg. You may have to add a
> > smaller resistor between the probe tip and ground to make fast pulses
> > visible. The scope will be used on 10mV/div probably. Do not touch
> > directly with circuit, 1-2cm away from tx antenna or tx booster (-;)
> > should work.
> >
> > The field strength meter will be useful, for the transmitter only
> > probably. This is also true for the probe.
>
> --
> http://www.piclist.com hint: The PICList is archived three different
> ways. See http://www.piclist.com/#archives for details.
>
>
>Will work better with an old dot-size germanium diode or a microwave
>schottky diode in SOT23 ( small) package.
>But the measuring results will be pure informative.
.........
> > A simple probe consists of a single 1N4148 (better 1N23) diode with a
5cm
> > wire on one end and a connector mating with the scope probe tip on the
If I remember correctly the 1N23 is a microwave germanium diode, and would
probably work better at this sort of level than a schottky. But then I could
be wrong, it is a long time since I dealt with radar components. :)
The problem is the very poor impedance match and coupling, not the diode.
There is another version of the same probe that uses a closed loop
(crocodile clamp to ground of probe. It can have 1 or more turns. This is
usefull in testing RFID systems and such (at low frequency). You can test
the straight probe quickly using a cellular phone. 1N23 is a microwave
germanium diode. The closed loop probe can be tested with a computer
monitor or tv (it picks up the field from the deflection yoke 1-2 meter
away). The probes are useful for seeing modulation and for snooping for
heavy rf leakage. Both probes pick up 50Hz mains hum and requre some
care to avoid that.
#1) When you have a microwave diode detector (like an HP33330C) and the
input consists of two sinusoidal signals at different frequencies, do their
powers add or their voltages? In other words, if the detector gives an
output of 0.1 V when I put in a 0dBm signal at 2.8GHz, and then 0.2 V when
I also add in a 670 MHz signal of unknown power, should I use 0.2-0.1 V =
0.1 V with the voltage-to-power table to find the power in the 670MHz
signal, or should I find the power associated with 0.2 V and then subtract
0dBm from that? I know the answer should be obvious but I don't know
exactly what the internal circuit of the detector is so I want to ask to
make sure (my guess is that I subtract voltage first then find power). The
datasheet doesn't make it clear what the answer is.
#2) Seemingly unrelated to #1, does anyone know what a typical insertion
loss of a 3 section (3 "hairpins", two coupling gaps) Chebychev hairpin
resonator filter at 3.5GHz (with bandwidth from 3.2 to 3.7 GHz) would be
(if built on 30mil GIL-30 microwave laminate)? I know an exact answer would
require calculations and I don't expect that, just wondering if anyone has
a rough figure? I ask because I have built three of them now and can't seem
to get the IL below about 7dB and I am wondering if I am doing something
wrong or if this is roughly what is to be expected.
I think you don't stand much of a chance of getting the loss lower, and
would have thought it to be more than 10 if your tuning for a 3.2 ~ 3.7
pass
Band I would be happy to see only 7db if all the elements were tuned to
the same frequency.
To get less IL you could couple the elements tighter but it wouldn't be
much good as a filter.
Sean wrote
#2) Seemingly unrelated to #1, does anyone know what a typical insertion
loss of a 3 section (3 "hairpins", two coupling gaps) Chebychev hairpin
resonator filter at 3.5GHz (with bandwidth from 3.2 to 3.7 GHz) would be
(if built on 30mil GIL-30 microwave laminate)? I know an exact answer
would
require calculations and I don't expect that, just wondering if anyone
has
a rough figure? I ask because I have built three of them now and can't
seem
to get the IL below about 7dB and I am wondering if I am doing something
wrong or if this is roughly what is to be expected.
Thanks for the quick reply. Is the insertion loss typically 10dB for only a
three element filter? If so, why do I see boards with 5 and 6 sections in
them! Wouldn't the loss be prohibitive?
Sean
At 12:39 AM 3/27/02 +0000, you wrote:
>I think you don't stand much of a chance of getting the loss lower, and
>would have thought it to be more than 10 if your tuning for a 3.2 ~ 3.7
>pass
>Band I would be happy to see only 7db if all the elements were tuned to
>the same frequency.
> To get less IL you could couple the elements tighter but it wouldn't be
>much good as a filter.
"Is the insertion loss typically 10dB
for only a three element filter?"
Ummmm ... typically, generally or always?
Jut off the cuff, this sounds AWFULLY
HIGH - but I don't know anything about
the application of the implementation
of the filter.
I just went though the design process of
several diferent filters (LP, HP, BS) at
low frequencies (1.8 MHz and below) using
E-Syn and didn't anything like the kind of
IL's you're seeing. Since all filters
can be scaled for frequency - I could have
just as easily been doing work at just under
1.8 GHz ...
I could, during the design, specify such
parameters as 'ripple in the passband' so
perhaps your filters were designed for
such high IL's?
>Hi Peter,
>
>Thanks for the quick reply. Is the insertion loss typically 10dB for only
>a three element filter? If so, why do I see boards with 5 and 6 sections
>in them! Wouldn't the loss be prohibitive?
Imho set up a Spice simulation and see what happens ...
At least at HF/IF frequencies I know of 2-3dB loss in a 3 stage Chebyshev
with not so good coils and silver mica caps.
Thanks for the replies. This is a microstrip hairpin filter (uses coupled
resonators made from quarter wave sections of microstrip transmission
line), so I can't simply set up a SPICE simulation. Also, I think that
these DO have typically greater insertion loss than discrete element
filters (of the same order) do, but I'm not certain. Also, this filter has
three sections, I'm not sure if this equates to a regular 3-pole filter or
not. I suspect not, because I think that each section contributes more
than one pole.
I have Sonnet lite, a program for analyzing these types of circuits but I
don't know how to convert from dissipation factor (the spec in the PCB
laminate datasheet) to dielectric loss tangent (the spec that Sonnet wants).
In addition, it is entirely possible that my design was too restrictive in
some way, so I'd like to get an idea for what a typical hairpin filter IL
is, so that, even if I am getting the best for my design, I can see how
much I stand to gain by relaxing the design.
> >Hi Peter,
> >
> >Thanks for the quick reply. Is the insertion loss typically 10dB for only
> >a three element filter? If so, why do I see boards with 5 and 6 sections
> >in them! Wouldn't the loss be prohibitive?
>
>Imho set up a Spice simulation and see what happens ...
>
>At least at HF/IF frequencies I know of 2-3dB loss in a 3 stage Chebyshev
>with not so good coils and silver mica caps.
>
>Peter
>
>--
>http://www.piclist.com hint: To leave the PICList
>RemoveMEpiclist-unsubscribe-requestTakeThisOuTmitvma.mit.edu
Sorry to take so long replying (too much work) Could you me send the pcb
rough layout, gif or whatever
I assume your using a Teflon based board ?
For this freq I would use if at all possible a satellite LNB (C Band I
think) and possibly a rx hacked front end
But of course it all depends on the application
If you need an rf source in the Gig range for TX /LO, check out aventec
I used to use their VTO modules, TO5 case 4 pins (ground, +, control
voltage and OP (10 Dbm +/- 3Db across the entire op range). They are ~
$150 but IMHO well worth it.
Thanks for the advice and suggestions. The application is the front end of
my amateur-radio based weather radar project. I already built and tested
(and it works!) a LO for the system, using Minicircuits VCOs and an analog
devices dual PLL chip to produce the two needed LO outputs (2.8GHz and
640MHz).
Part of the problem with the front end is that I tried to etch the board
myself and the results are good to mediocre. I'm not using teflon, I'm
using one of the alternative microwave laminates (from GIL, GIL-30). I am
going to put a preamp from Downeast Microwave in front of the receiver so
the noise figure isn't too critical, but I do want to keep it low enough
that the preamp will almost overcome it entirely.
For the PLL board I built, I had AP Circuits make the board for me out of
GML-1000 laminate. However, it cost $270 U.S. because I had to buy two full
panels of the expensive laminate (minimum order). I can't afford spending
that again, BUT I just found out that they allow you to share the minimum
order with other customers and they told me they have one who is willing to
share, so if that works out, I will probably re-do it and go that route and
I'm pretty sure the loss should be considerably less because of the lack of
minute voids, etc., in the traces of the filter.
I simulated my filter in Sonnet Lite (after finding out that dissipation
factor and loss tangent are the same thing) and the loss should only be a
dB or two, so something is definitely wrong with mine. After examining it
with a magnifying glass, I can see that along the edges between coupled
lines, the edges are rough and there are many tiny pits in the trace, so I
bet that goes a long way toward explaining the additional several dB of loss.
>Sorry to take so long replying (too much work) Could you me send the pcb
>rough layout, gif or whatever
>I assume your using a Teflon based board ?
>
>
>For this freq I would use if at all possible a satellite LNB (C Band I
>think) and possibly a rx hacked front end
>But of course it all depends on the application
>
>If you need an rf source in the Gig range for TX /LO, check out aventec
>I used to use their VTO modules, TO5 case 4 pins (ground, +, control
>voltage and OP (10 Dbm +/- 3Db across the entire op range). They are ~
>$150 but IMHO well worth it.
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
> Part of the problem with the front end is that I tried to etch the
board
> myself and the results are good to mediocre. I'm not using teflon, I'm
> using one of the alternative microwave laminates (from GIL, GIL-30). I
am
> going to put a preamp from Downeast Microwave in front of the receiver
so
> the noise figure isn't too critical, but I do want to keep it low
enough
> that the preamp will almost overcome it entirely.
Be careful when using a preamp if you've got any HIGH level signals
Of ANY other freq coming from the antenna they will give you problems
(cross modulation and parasitic mixing).
It is usually best to avoid preamps on RX's
I can see that along the edges between coupled
> lines, the edges are rough and there are many tiny pits in the trace,
so I
> bet that goes a long way toward explaining the additional several dB
of
> loss.
>
> Sean
Can't see it making that much difference, you could try silver plating
your board (maybe Silver Dip/Restorer)
Where is the filter placed, after the mixer?
I assume your using a strip line balanced diode ring mixer they should
perform well for the 1st mixer
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
At 12:30 AM 3/29/02 +0000, you wrote:
>Be careful when using a preamp if you've got any HIGH level signals
>Of ANY other freq coming from the antenna they will give you problems
>(cross modulation and parasitic mixing).
>It is usually best to avoid preamps on RX's
Yes, I'm aware that there is quite a tradeoff here. The problem is that I
need to be able to hear very weak signals. This will only be operated
occasionally and it will be rather far away from most buildings and antenna
towers, etc., when operated. The fact that the preamp is designed for only
a narrow frequency range and that the antenna is only resonant in a narrow
band should help. I'm hoping that this should be enough filtering before
the preamp but if it doesn't work, I can just take the preamp out of line,
since it will just be a separate module.
>Can't see it making that much difference, you could try silver plating
>your board (maybe Silver Dip/Restorer)
I already solder coated it but the solder did not fill in these pits, it
faithfully went only where there was copper underneath :-) I guess I could
try one of those conductive pens? I have not heard of Silver Dip/Restorer,
what is it and where can I find it?
>Where is the filter placed, after the mixer?
>I assume your using a strip line balanced diode ring mixer they should
>perform well for the 1st mixer
The filter is before the mixer (to remove the image frequency and also to
help reduce problems with strong signals at other frequencies, although I
realize that diode mixers don't have much of a problem with the latter).
The mixer is indeed a doubly-balanced diode ring mixer (ADE-30W from
Minicircuits). After the mixer comes an amplifier and then another filter,
but that one will be made using discrete components since it will be at a
much lower frequency (670MHz 1st IF).
Thanks,
Sean
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
> >Where is the filter placed, after the mixer?
> >I assume your using a strip line balanced diode ring mixer they should
> >perform well for the 1st mixer
>
> The filter is before the mixer (to remove the image frequency and also to
I'm curious about this statement. How can there be an "image" frequency
yet, if you are still in front of the mixer? Or is there another mixing
stage ahead of this?
> help reduce problems with strong signals at other frequencies, although I
> realize that diode mixers don't have much of a problem with the latter).
> The mixer is indeed a doubly-balanced diode ring mixer (ADE-30W from
> Minicircuits). After the mixer comes an amplifier and then another filter,
> but that one will be made using discrete components since it will be at a
> much lower frequency (670MHz 1st IF).
>
> Thanks,
>
> Sean
>
> --
> http://www.piclist.com hint: PICList Posts must start with ONE topic:
> [PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
>
>
-- http://www.piclist.com hint: PICList Posts must start with ONE topic:
[PIC]:,[SX]:,[AVR]: ->uP ONLY! [EE]:,[OT]: ->Other [BUY]:,[AD]: ->Ads
I was very imprecise in my language. The purpose I had in mind for the
filter was to create some degree of image rejection. The mixer's LO is at
2.816 GHz and the 1st IF is 670MHz. The intended input signal would be
2.816 GHz + 0.67 GHz = 3.486 GHz. However, any signal at 2.816 GHz - 0.67
GHz = 2.146 GHz would also be converted to 670 MHz if there were no filter.
Sean
At 10:46 PM 3/28/02 -0600, you wrote:
>I'm curious about this statement. How can there be an "image" frequency
>yet, if you are still in front of the mixer? Or is there another mixing
>stage ahead of this?
> I was very imprecise in my language. The purpose I had in mind for the
> filter was to create some degree of image rejection. The mixer's LO is at
> 2.816 GHz and the 1st IF is 670MHz. The intended input signal would be
> 2.816 GHz + 0.67 GHz = 3.486 GHz. However, any signal at 2.816 GHz - 0.67
> GHz = 2.146 GHz would also be converted to 670 MHz if there were no
filter.
>
Sean,
I should have guessed that, sorry, I understand now.
I would like to do side-by-side GPS receiver tests where they use the same
antenna and I want equal power to go into each receiver from the amplified
antenna. The method for splitting power is a Wilkinson divider (unless you
can suggest different) where the 50 ohm microstrip line is split with 70.7
ohm traces with a length of lambda/4 (1.135 in. at 1.575 GHz). You can get
the general idea at
It is further complicated by having one of the receivers (hot) power the
antenna while the other does not (cold). The cold receiver will have a
blocking capacitor to block the antenna supply and an inductor with resistor
to load the receiver (approx. 5mA), and think that an antenna is attached so
I don't get a warning.
1) The DC blocking cap C1 is 100pF. Seems that this size is fairly typical.
I will use a smd part. Is it OK?
2) In the re-radiator designs the builder wound his own inductor. It is 6
turns of #26 (0.4 mm) wire wound on a 3/32 inch (2.4 mm) drill bit shaft,
giving a coil diameter of about 3 mm. The turns are spread apart by one wire
diameter or more, giving a total
length of about 6 mm.
When I do the calculation, this works out to be about 0.04 uH.
I simulated the whole circuit with the LTspice program that RM told the list
about. (Great program). I find that 0.04uH doesn't do much to block the RF
and that something more than 0.1 uH would be required and 1uH would be
better. Does this make sense?
If I choose 0.4 uH, this works out to a .27" dia, 10 turns and .312 length.
Is this workable, or a good idea? Will this wire start to work like an
antenna or anything like that? Would a bought inductor be better? Any
suggestions from Digikey?
3) I've layout out the board (3"x1.4"). It will have ground planes top and
bottom connected by vias (not shown) all around and inside. One end has a
MCX connector for the antenna and the other has two RG-174 cables. Any
improvements you can suggest?
4) Will this have to go into a metal enclosure with the enclosure attached
to ground, or is there enough shielding with the ground planes that I can
use a plastic enclosure?
Gordon,
For GPS frequencies I would definately use SMT caps. 100pF seems a bit
large. If you use an 805 package then the series inductance is approx 3nH
giving a series resonace at ~290MHz. A better solution would be ~15pF (Zc
= ~7 jOhms,) made up from 3 x 4.7pF caps in parallel. but watch out for
stray inductance & capacitance effects.
The 100pf cap may work, or may introduce additional loss if the signal
hits a resonance frequency. By moving the resonance above your operating
frequency you should get more consistant results. (A quick simulation shows
a 0.33dB additional loss using the 100pF cap, although the resonances are
damped out by the source & load impedances).
I would use a 1/4 wave line rather than winding an inductor as anything you
wind is will have self resonances below 1.5GHz. A ferrite bead may be a
suitable alternative.
Overall you seem to be looking along the right lines.
Not really what I'd consider a project for someone who has not done much RF
work previously but if you want to give it a try then there is pleanty to
learn about!.
To make consistant A-B mcomparisons you will be best off with a screened
box or a "can" (or cover ?) over the PCB. You will need to swap the GPS
units around also, if possible, as it will be very hard to get both inputs
with equal loss given the blocking cap & DC feed requirements.
Just my 2c worth.
Richard P
Hi All,
I would like to do side-by-side GPS receiver tests where they use the same
antenna and I want equal power to go into each receiver from the amplified
antenna. The method for splitting power is a Wilkinson divider (unless you
can suggest different) where the 50 ohm microstrip line is split with 70.7
ohm traces with a length of lambda/4 (1.135 in. at 1.575 GHz). You can get
the general idea at
It is further complicated by having one of the receivers (hot) power the
antenna while the other does not (cold). The cold receiver will have a
blocking capacitor to block the antenna supply and an inductor with
resistor
to load the receiver (approx. 5mA), and think that an antenna is attached
so
I don't get a warning.
1) The DC blocking cap C1 is 100pF. Seems that this size is fairly
typical.
I will use a smd part. Is it OK?
2) In the re-radiator designs the builder wound his own inductor. It is 6
turns of #26 (0.4 mm) wire wound on a 3/32 inch (2.4 mm) drill bit shaft,
giving a coil diameter of about 3 mm. The turns are spread apart by one
wire
diameter or more, giving a total
length of about 6 mm.
When I do the calculation, this works out to be about 0.04 uH.
I simulated the whole circuit with the LTspice program that RM told the
list
about. (Great program). I find that 0.04uH doesn't do much to block the RF
and that something more than 0.1 uH would be required and 1uH would be
better. Does this make sense?
If I choose 0.4 uH, this works out to a .27" dia, 10 turns and .312 length.
Is this workable, or a good idea? Will this wire start to work like an
antenna or anything like that? Would a bought inductor be better? Any
suggestions from Digikey?
3) I've layout out the board (3"x1.4"). It will have ground planes top
and
bottom connected by vias (not shown) all around and inside. One end has a
MCX connector for the antenna and the other has two RG-174 cables. Any
improvements you can suggest?
4) Will this have to go into a metal enclosure with the enclosure attached
to ground, or is there enough shielding with the ground planes that I can
use a plastic enclosure?
Could you describe the idea behind a 1/4 wave line inductor? I haven't
heard of it before.
My understanding is that the ferrite bead idea starts to appear as a
frequency based resistor rather than an inductor at frequencies above
several 100 MHz. It appears that the equivalent resistance will be several
100 ohms at GPS frequencies. I'm not sure that is really what I need ...
Do I need to worry about self resonance if I am well above the resonance
frequency?
I am referring to a 1/4 wavelength transmission line.
This length of line has the property of transforming the impedance from one
end to another. (The square root of the product of the input and output
impedances equals the line impedance)
So if this line is AC short circuited at the far end (say by 3 x 4p7 caps
to ground) then the reflected impedance at the "open" end is very high.
Effectively you are feeding the DC via a high AC impedance - which is what
you are trying to achieve with an inductor.
The same effect is used by the 1/4 wave 70.7 ohm matching lines used in
the design. In the pwer feed case the acutal line impedance is not
critical, so for ease of design I'd use the same line dimensions as the
matching sections.
Typically you would feed the DC through a convienient inductor/ cap
(100uH/10nF ?) filter and then through a ferrite bead to the decoupling
caps and then to the 1/4 wave line. This minimises signals fed in both
directions.
Since you are operating in a 50ohm system, anything with an impedance much
greater than a couple of hundred ohms will appear more or less open circuit
(ac wise), while any component with an impedance less than ~5ohms or so
will behave somewhat like a short.
Ferrite beads do look resistive at vhf frequencies, but you are only
requiring AC impedance, not reactance so they do the job just as well (If
they added inductance it could introduce additional resonances so this is a
plus). I'm not sure about how good they are / what grades to use at 1575MHz
but I imagine they should be OK, although totally resistive at that
frequency.
The problem with operating well above self resonance is that the behaviour
is not too well defined. Inductors and capacitors can have several self
resonant areas (Depending on size & construction & adjacent metalwork) so
if you happen to hit one you could get very good - or very poor
performance. Or temperature variations etc. It is simpler to keep below the
first resonance unless you have very well defined components.
Richard
(Christchurch NZ)
Gordon
Williams To: TakeThisOuTPICLISTEraseMEspam_OUTMITVMA.MIT.EDU
<g_will@CYBERU cc:
S.CA> Subject: Re: [EE:] RF Questions about GPS Antenna Splitter
Sent by: pic
microcontrolle
r discussion
list
<PICLIST@MITVM
A.MIT.EDU>
20/05/04 14:02
Please respond
to pic
microcontrolle
r discussion
list
Hi Richard,
Could you describe the idea behind a 1/4 wave line inductor? I haven't
heard of it before.
My understanding is that the ferrite bead idea starts to appear as a
frequency based resistor rather than an inductor at frequencies above
several 100 MHz. It appears that the equivalent resistance will be several
100 ohms at GPS frequencies. I'm not sure that is really what I need ...
Do I need to worry about self resonance if I am well above the resonance
frequency?
hi! please help me with my problem. im a novice PIC16F84A user and i want a PIC TO PIC SERIAL PROGRAM. my objective is to output an 8bit random number generated gby my PIC and the said output will be used by another PIC as an input. i already have an 8bit random number generator. i need to convert it to a serial data and transmitt it, send it to another PIC. the latter PIC then will convert the said input to a parallel data for future use. please do help me with my problem ASAP. thanks a lot
PIC16F84A user,
David Christian G. Herrera
+63 - 916 - 426 - 31 - 41
---------------------------------
Do you Yahoo!?
Yahoo! Domains - Claim yours for only $14.70/year
mitvma.mit.edu with SET PICList DIGEST in the body
