For anyone interested in what an 18 nano-H inductor looks like under a
magnifying glass -

3 turns of 0.33 mm diameter wire, wound on a 2.2 mm diameter former, over a
length of 2.0 mm.

Its self-resonant frequency is 4.6 GHz and it is tuned to 1 GHz with a
capacitor of 1.3 pF. Q = 330 at 1 GHz.
----
Reg

"Reg Edwards" wrote in message
...
An 18 nano-henry inductor seems, to me, to be an impractical inductor
value.
The length of the connecting wires will form an appreciable part of the
inductance. There must be another way of acheiving the desired circuit
functions.

However, to find the length, diameter and number of turns to make such a
small coil you can use program SOLNOID3 which can be downloaded in a few
seconds, free of charge, and run immediately from the the website below.
Try it and see what happens.
---
.................................................. .........
Regards from Reg, G4FGQ
For Free Radio Design Software go to
http://www.btinternet.com/~g4fgq.regp
.................................................. .........
Reg,
A 1/2 turn loop of the proper dimensions might be more practical. I recall
using one that was about .5 in high between .5 centers, made of #14 wire.
(it carried 10 amps). Don't remember the exact value, but it was probably a
little more than what he is looking for.

At any rate, as you say, it is important to add the inductance of the
circuit board traces to the inductor value. BTW, the last job I was on used
402 SM inductors down to about 10 nH, but that was at 2+ GHz.

Tam

"Reg Edwards" wrote in message
...
For anyone interested in what an 18 nano-H inductor looks like under a
magnifying glass -

3 turns of 0.33 mm diameter wire, wound on a 2.2 mm diameter former, over
a
length of 2.0 mm.

Its self-resonant frequency is 4.6 GHz and it is tuned to 1 GHz with a
capacitor of 1.3 pF. Q = 330 at 1 GHz.
----
Reg



Thanks Reg,

I could not run the program solnoid3 on my computer for some reason. I am
using winblows xp.

I did improve the performance of my receiver tho. I checked the graph on
page 10.10 of the handbook and rigged up a straight piece of 20 ga wire for
my inductor. I don't know what the inductance actually is, but it improved
the performance greatly, which is all I was trying to do.

Best Regards,

Joe
KB1KVI

I did improve the performance of my receiver tho. I checked the graph on
page 10.10 of the handbook and rigged up a straight piece of 20 ga wire
for
my inductor. I don't know what the inductance actually is, but it improved
the performance greatly, which is all I was trying to do.

One reason the receiver improved may be because the filter was cutting out
all the wide band noise (other signals).

"Joe" wrote in message
k.net...


Hi John,

That's what I have been doing. Just using a piece of 20 ga wire for an
antenna, it is about 15cm in length (a vertical antenna). On the
transmitter, I was using a loop, a piece of 20 ga wire with a 1 pf cap
between the hot end of the antenna and the transmitter pin, other end
grounded.

Last nite after I read all the responses, I went to the Handbook and
looked
at page 10.10. There is a graph at the bottom of the page that shows
inductances of a straight piece of wire at that frequency. Anyway, I
rigged
up a 0.9" piece of 20ga wire between two 8pf ( I did not have 8.2pf) caps
and the reception increased dramatically. My test method was: Receiver in
my
house on a bench. TX outside about 80 feet away. In the old configuration,
the receiver would sometimes receive the TX pulses. Once I put the TX
inside
my tool shed (about) the same distance away, the receiver would not pick
up
the signal. So after I rigged up my "almost" 18nH inductor and the 2 8pf
caps, I repeated the experiment and the receiver was able to reliably
receive the transmissions in both places. So I definitely saw an
improvement.

As far as FCC goes, this frequency is supposed to be for ISM, and you can
use it unlicensed subject to part 15, that's why my first attempt was to
work on the receiver, not the TX, but I may experiment on the TX now also.

Joe
KB1KVI


Good work, Joe. Keep experimenting. It pays off.

John

"Ralph Mowery" wrote in message
k.net...


I did improve the performance of my receiver tho. I checked the graph on
page 10.10 of the handbook and rigged up a straight piece of 20 ga wire
for
my inductor. I don't know what the inductance actually is, but it
improved
the performance greatly, which is all I was trying to do.

One reason the receiver improved may be because the filter was cutting out
all the wide band noise (other signals).



Hi Ralph,

Yes, that could be it. I have not looked at it on my scope. I will probly do
that tonite. Thanks for the idea. I know that before, the receiver was noisy
as hell. I don't know if they are all like that, but I'll find out. What I
was seeing was just a lot of what looked like random noise, till the TX was
activated, then all the noise seemed to go away and just the code (about
680hz square waves) from the TX was visible. I am using the holtek ht12d on
the receiver and holtek 12E on the transmitter.


Joe
KB1KVI

"Joe" wrote in message link.net...
"Reg Edwards" wrote in message
...
An 18 nano-henry inductor seems, to me, to be an impractical inductor
value.

;-) You just aren't working with small enough, high enough frequency
circuits, Reg.

3 turns of 28AWG on a 4-40 (USA) screw (remove the screw, of course)
will give you about 18 nH. Unloaded Q at 440MHz will be around 200.

The link to the schematic is:
http://www.radiotronix.com/downloads/433_pair.pdf

Easy enough to use SMT components soldered to a piece of copper-clad
for a ground plane. Just put down the two caps and the shunt resistor
"tombstone" style, one end soldered to the ground plane and the other
up in the air, separated by enough distance to put the series coil and
resistor between them. You can keep the parasitic L and C very low
indeed this way. You can also use tiny squares of copper-clad glued
down to the ground plane as tie points.

It's also possible to mechanically make a simple circuit like this on
copper-clad using surface mount parts, by scribing lines with
something like an X-Acto knife and removing strips of copper. In this
case, I could see scribing four lines across the board, separated by
about 1/16 inch, and removing two copper strips, one from between the
first and second scribe, and one from between the thrid and fourth
scribe. That leaves a trace about 1/16 inch wide. Cut gaps in the
trace, about 1/16 inch long, where you want to solder down series
components. Solder shunt components across from the trace to the
copper ground plane. Bridge across the trace with a copper strap to
connect the two gound plane sides, with the bridge up in the air over
the trace. Nice to have a ground plane on the opposite side, with
some holes and wires to connect to the top-side ground planes, but not
necessary. I'm thinking 0805 parts with the sizes I mentioned...make
things slightly larger for 1206. You can make the trace be a 50 ohm
transmission line if you want, by sizing things right.

Cheers,
Tom

"Tom Bruhns" wrote in message
...
"Joe" wrote in message
link.net...
"Reg Edwards" wrote in message
...
An 18 nano-henry inductor seems, to me, to be an impractical inductor
value.

;-) You just aren't working with small enough, high enough frequency
circuits, Reg.

3 turns of 28AWG on a 4-40 (USA) screw (remove the screw, of course)
will give you about 18 nH. Unloaded Q at 440MHz will be around 200.

The link to the schematic is:
http://www.radiotronix.com/downloads/433_pair.pdf

Easy enough to use SMT components soldered to a piece of copper-clad
for a ground plane. Just put down the two caps and the shunt resistor
"tombstone" style, one end soldered to the ground plane and the other
up in the air, separated by enough distance to put the series coil and
resistor between them. You can keep the parasitic L and C very low
indeed this way. You can also use tiny squares of copper-clad glued
down to the ground plane as tie points.

It's also possible to mechanically make a simple circuit like this on
copper-clad using surface mount parts, by scribing lines with
something like an X-Acto knife and removing strips of copper. In this
case, I could see scribing four lines across the board, separated by
about 1/16 inch, and removing two copper strips, one from between the
first and second scribe, and one from between the thrid and fourth
scribe. That leaves a trace about 1/16 inch wide. Cut gaps in the
trace, about 1/16 inch long, where you want to solder down series
components. Solder shunt components across from the trace to the
copper ground plane. Bridge across the trace with a copper strap to
connect the two gound plane sides, with the bridge up in the air over
the trace. Nice to have a ground plane on the opposite side, with
some holes and wires to connect to the top-side ground planes, but not
necessary. I'm thinking 0805 parts with the sizes I mentioned...make
things slightly larger for 1206. You can make the trace be a 50 ohm
transmission line if you want, by sizing things right.

Cheers,
Tom

Hi Tom,

Being so new at this, I don't even know what you mean by 'tombstone style',
and it sounds like an awful lot of work to be scoring pcboard with an Xacto
knife to create transmission line effects.

I talked with a tech rep from radiotronix today. They called me because I
ordered a half dozen modules from mouser. He explained that the pi circuit
on the TX is so the TX sees 50 ohms, which is pretty critical. I have not
done anything with the transmitter modules yet (they are from Velleman
because they are thru hole), but I am building a receiver board on pc board
using pcbexpress and single sided copper pcboard. It will be the first
prototype. Now that I have it working on a breadboard I am going to build
one and see how it works. I am planning on using a straight piece of (20ga)
wire for the 18nH inductor and using the T network of 2 8pf caps beside it.
The tech said this is just a low pass filter, but it seems to make a
difference in receiver sensitivity. The noise on the receiver is supposed to
be there according to him. It is between 1Khz and 1.5Khz. he said some
people use the data slicer pin with a holtek ht12D decoder, which i am doing
or you can also use a micro to look for the bit pattern and decode the
signal. I am just using the TV (transmission valid) to set off the rest of
my circuit which, right now, consists of a latch and an LED.

Joe

Joe wrote:

Being so new at this, I don't even know what you mean by 'tombstone
style', and it sounds like an awful lot of work to be scoring pcboard
with an Xacto knife to create transmission line effects.

You can easily build what you need using single-sided board: a
continuous groundplane, components soldered on top, and connections made
using sticky-backed copper tape (available at craft stores for
stained-glass work).

If you cut strips 0.1in wide, these will be a good approximation to
50-ohm microstrip, for the relatively low frequencies and short lengths
that you will need. On the reverse side, use broader the copper tape to
link the groundplanes of your extension and the Radiotronix module.

Simply lay the SMD components against each other, and solder them
together... but before you do, mark where the connections through to
ground will need to be, for the two capacitors in the pi-network and
possibly other places. Then drill the board and insert links of thick
wire (14AWG) so that the ground connections will have zero length beyond
the thickness of the board - but I really do mean ZERO, or else your
network will not perform correctly. The ground links must be touching
the ends of the capacitor chips before you solder them.

Likewise, use ZERO lead lengths inside the pi-network. Butt the ends of
the two Cs and the L chips directly together, and solder. Any
"connecting leads" must be restricted to the places that can use 50-ohm
microstrip.

This construction method is very quick and simple. The SMD parts are
well-enough anchored to the board for all normal conditions of use.

I talked with a tech rep from radiotronix today. They called me
because I ordered a half dozen modules from mouser. He explained that
the pi circuit on the TX is so the TX sees 50 ohms, which is pretty
critical.

- so remember what I just said about zero lead lengths.

I have not done anything with the transmitter modules yet (they are
from Velleman because they are thru hole), but I am building a receiver
board on pc board using pcbexpress and single sided copper pcboard. It
will be the first prototype. Now that I have it working on a breadboard
I am going to build one and see how it works. I am planning on using a
straight piece of (20ga) wire for the 18nH inductor and using the T
network of 2 8pf caps beside it. The tech said this is just a low pass
filter, but it seems to make a difference in receiver sensitivity.

On the face of it, this pi-network has a 1:1 impedance transformation
(two equal capacitors) so it shouldn't make any difference. However,
there may be stray capacitances and/or inductances in the module that we
don't know about. Bottom line is, this network is what the Radiotromix
techs have found to work.

Since you don't have the design experience and the test equipment to
measure how the whole thing performs, you would be very wise to copy
EXACTLY what Radiometrix did, which includes building the whole network
using the best possible construction practice.

Also, a straight piece of wire is actually a short length of
transmission line: it does not have quite the same electrical properties
as an inductor. So use an 18nH chip inductor like Radiometrix said.

This is really part of your "contract" with the module manufacturer. It
isn't fair or reasonable to expect their tech support to second-guess
what the effects of every individual user's network might be, if they're
all slightly different. You might get away with it this time, but some
of the worst tech-support problems are caused by people who decla

"I did exactly what you said..."

[Now follows 20 minutes of intensive cross-examination over the phone]

"... oh, apart from all these things that are different."

[Deleted: strange sounds on the phone line. The tech has the curly cord
gripped in both hands. He's puling it out straight, and he has a red
glint in his eyes. Be very thankful that you're several hundred miles
away...]



--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)
http://www.ifwtech.co.uk/g3sek

Ian White, G3SEK wrote:
You can easily build what you need using single-sided board: a
continuous groundplane, components soldered on top, and connections
made using sticky-backed copper tape

Sorry, that wasn't quite clear (not enough coffee yet this morning).

The continuous groundplane is on the underside of the board. All the SMD
components are on the bare fibreglass side. The SMD components are held
in place by soldering them directly to each other, to the ground links,
to the striplines, and to various other fixed points.

No board etching is required, and this technique will is good to
frequencies well above 1GHz.

The 0.1in stripline width was for standard 0.064in FR-4. You can
obviously refine that calculation if you want, but there's little point
for this particular application.


--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)
http://www.ifwtech.co.uk/g3sek