Thanks for the input...........................I wish I had some ability
programming microcontrollers. I really need to learn how to work with at the
very least, PICs. Anyway, this generator would sell for the under 150 dollar
range, putting it a couple of steps above some of the offerings in this
range.
I do like that HP equipment...........I had an 8650B for awhile and it was
very good. I replaced it with a Boonton 103D and a Wavetek 2407. I've got an
older spectrum analyzer (8558B with an 853 mainfraime). An older 8444
tracking generator allows me to measure the response of filters. I did build
the N2PK Vector Network Analyzer last year, and it does serve me very well
when it comes to doing sweeps of very narrow bandwidth filters. It is pretty
nice, since I can look at the forward transfer function and the phase
response at the same time. Using a directional coupler, I can view forward
response and return loss at the same time................I use this checking
antenna resonance.
It would be nice to come up with a different audio oscillator for the
modulation section. As long as I can keep the distortion fairly low, that
would be ok. Let's see....................the Exar chip...........that
would be an XR2206. I originally used this device for an FSK generator that
I designed for a customer. They wanted to have something testing caller ID
units. I also had some Burr-Brown 4523s these were a nice Quadrature
function generator chip. I don't know what happenned to those.
So.........you worked for Exar. That must have been a pretty cool job.

Pete

wrote in message
oups.com...
The problem with delivering something that appears to be a bit of test
equipment, is you are competing with used test equipment. When it comes
to test equipment, 20 year old HP wins over brand spanking new brand X
every time. Most of the used gear I've bought over the years is still
used in real labs. I could get it calibrated if I had the need. I have
to think really hard if I ever bought any new test equipment.

If your box was digital through and through, then maybe you might get
some sales. I'm not sure which Exar chip you are using (BTW, I used to
work there), but the Twin-T is an analog topology and thus has all the
bad features of analog as well as the good. It should be possible to
make a completely digital AM or sideband modulator with a DSP chip and
high speed DAC. The coordic is how the sine wave is computed, though at
the time nobody really talked about it since there were still patents
on coordic processors from the scientific calculator manufacturers. One
of the best books I've seen on the subject was a masters or PHD thesis
by a Stanford grad whose full name escapes me. First name Ahmend, which
might as well be Joe or Bill nowadays.

A coordic is a lookup table technique that converges on the the sin and
cos. You make a sine wave generator by accumulating phase, then taking
the sin/cos of this phase value. If the phase accumulator were 8 bits,
you would use a scale where 256 is one revolution. This way as you
accumulate phase, the sine function automatically wraps around after
reaching one revolution. You control the amplitude of the sine wave by
the initial value of the coordic. It's really much easier than it
sounds, at least the algorithm, Programming the DSP is another story.

I suspect the Wellbrook has a jfet amp because of the positive ground.
I think today most CATV circuits are bipolar to simplify the biasing.
For certain frequency ranges JFET has lower noise.You can parallel them
up for lower noise. There are a few companies still serious about
Jfets, but they are becoming less and less since the money is in high
bandwidth.
http://www.calogic.com/ comes to mind as a jfet company

Another idea would be to make VHF/UHF amplifiers using SiGe.

I do have one design that uses a JFET as a source follower............this
way, one can take advantage of the unloaded Q of the circuit. This is
followed by an RF Micro Devices RF2044. This device is a GaAs MMIC that has
20dB of gain with a NF of 4dB. Frequency range is from DC to 6GHz and linear
power output of this device is +17dBm. 1dB compression point is
+18.5dBm.

Pete

No such thing as a dumb question.................that is how we learn!
Anyway, Signal Generators, at least in the context of test equipment, are
tunable RF sources that have modulation capabilities. If they are
synthesized, they can be accurate enough sources for calibrating the tuning
dials of receivers. They can also be used in conjunction with signal tracers
by injecting the signal through the different stages of a system and using
the signal tracer for tracing the path of that injected signal. A very
useful technique for locating a defective stage in a receiver. And yes,
signal generators are external pieces of test equipment.

Pete

wrote in message
...
Another dumb question from me.What are Signal Generators and how do they
work? Are they external devices.I don't know unless I ask? I own one or
two old Signal Tracers,I don't know anything about them or how they are
suppose to be used.
cuhulin

The trick with the Wellbrook loop is that the loop isn't tuned. Tuned
loops make sense if your radio itself has poor filtering, but anyone
who forked out money for a Wellbrook already has a decent radio. You
have to admit the Wellbrook marketing is clever. They don't have to
compete with Kiwa, Quantum, etc.

In a good amplifier, you set (or blow) the noise figure at the input
stage. [You can do worse of course if you aren't careful.] Thus
starting with a source follower means you are inserting a noise source
right that the front end. Since noise is "input referred", starting
with a source follower (i.e. no gain) means that you don't have the
gain of the transistor to reduce the noise. Thus you should start with
a gain stage. Now this is where I don't know what Wellbrook does. For
low noise, you want a single ended input stage (one less elemen to
contribute to the noise). For the loop to work, it has to float about
ground as you don't want the antenna to be the loop relative to ground,
i.e. you want the loop to be the antenna. Thus Wellbrook puts in a
differential to single ended converter, aka the transformer. Now what I
don't know is the turns ratio of the transformer. Are they picking the
turns ratio fo impedance matching? Is the goal voltage gain?

I don't think the Wellbrook is complicated in terms of the number of
components, but it requires good planning to make a low noise amp that
works with the loop.

You may want to investigate North Hills transformers. They often show
up at swap meets or on ebay. They are very wide band.

That's just what I was thinking about, concerning the transformers. I
understand the concept of the input stage determining the system
NF..............for a MW loopstick, this doesn't seem to be an issue,
especially with a tuned circuit that has a loopstick with a measured Q of
400............I am using ferrite material that has a permeability of 900.
Anyway, a source follower had a voltage gain of approximately 0.95, so there
shouldn't really be any noise of any magnitude inserted into the system.
There is, however quite a bit of current gain; this is good when driving a
lower impedance load such as a MMIC. This link should provide some insight:
http://www.qrp.pops.net/probe1.htm
This person has some example schematics.
This is also a good link (PDF file): www.analogzone.com/hft_0930.pdf
Since I am using a very low value source resistor, about 62 ohms, this
should be the main contributing factor as far as noise voltage is concerned.
The MMBF5486 has a 3dB to begin with, so once again, I don't believe noise
to be an issue here.
I can make some NF measurements to verify this, but as you say, the
differential mode seems to be the way to go, especially with a good JFET
pair. I do have some Siliconix E431 and U430 dual JFETs on hand, but since I
have less than 30 of each it wouldn't make any sense to build anything
unless it was a one-of custom project. Ulrich Rohde has a good design
example in his communications receiver design book.

wrote in message
oups.com...
The trick with the Wellbrook loop is that the loop isn't tuned. Tuned
loops make sense if your radio itself has poor filtering, but anyone
who forked out money for a Wellbrook already has a decent radio. You
have to admit the Wellbrook marketing is clever. They don't have to
compete with Kiwa, Quantum, etc.

In a good amplifier, you set (or blow) the noise figure at the input
stage. [You can do worse of course if you aren't careful.] Thus
starting with a source follower means you are inserting a noise source
right that the front end.


Since noise is "input referred", starting
with a source follower (i.e. no gain) means that you don't have the
gain of the transistor to reduce the noise.

The only noise that there should be here is antenna noise.

Thus you should start with
a gain stage. Now this is where I don't know what Wellbrook does. For
low noise, you want a single ended input stage (one less elemen to
contribute to the noise). For the loop to work, it has to float about
ground as you don't want the antenna to be the loop relative to ground,
i.e. you want the loop to be the antenna. Thus Wellbrook puts in a
differential to single ended converter, aka the transformer.

This is how they achieve their very high IP2 and IP3 specifications. A good
way to go, but after a certain input level, the ferrite material will become
saturated. Now, if they are using the type of material that is used in the
broadband transformers in transmitter output stages, that is a different
story.


Now what I
don't know is the turns ratio of the transformer. Are they picking the
turns ratio fo impedance matching? Is the goal voltage gain?

I am not sure about this myself. Dale, W4OP would understand this one.

I don't think the Wellbrook is complicated in terms of the number of
components, but it requires good planning to make a low noise amp that
works with the loop.

I have seen a couple of designs that use an op-amp at the front end. I have
thought about using an AD797 as the input stage. With a 30 volt bipolar
supply, dynamic range could be very good since you would have approximately
28 volts of swing at the output stage.

You may want to investigate North Hills transformers. They often show
up at swap meets or on ebay. They are very wide band.


Pete

Differential pairs have more noise than a single ended design, which is
why the transformer is the way to go. An ideal transformer has no
noise. I real life transformer has the noise of the resistance in the
windings. This is why the transformer choice isn't trivial.

If you buffer (source follow) then use an amplifier, the noise of the
source follower gets amplified. This is why you wouldn't normally do
any buffering before amplification if low noise was the issue. However,
in the case of a tuned loop, you need the high impedance to keep the Q
high. Wellbrook gets around that problem by using a loop that is not
tuned. The assumption is the purchaser has a real radio and doesn't
need front end filtering. So cource having a high dynamic range like
the 7030 is a plus.

Loops using ferrite will not work as well as a Wellbrook IF the
Wellbrook uses a large enough loop. What you are gaining with the
Wellbrook is aperture. It soaks up lots of RF due to it's size.

Aperture is exactly the point I was going to get at.........this is the
reason that the receivers I have built that use the Mini-Circuits mixers as
the first stage have worked very well with an 8 foot diameter untuned loop.
No IM products anywhere in the LW/MW bands.
wrote in message
oups.com...
Differential pairs have more noise than a single ended design, which is
why the transformer is the way to go. An ideal transformer has no
noise. I real life transformer has the noise of the resistance in the
windings. This is why the transformer choice isn't trivial.

I understand this, but take a look at the AD797, and other op-amps that are
used at very low input signal levels. This device is designed for ultrasound
transducer applications, strain gauge amplifier service, etc. There are
quite a few very low noise op-amps that are suitable for RF service.

As far as transformers, I don't think there is such a thing as an ideal
transformer. In my experience, the input Xl must be at least ten times the
impedance of the expected driving source's impedance at the lowest frequency
of interest.There is going to be a certain amount of DC resistance in the
primary winding if we are going to be able to achieve the required
inductance to provide the proper load. If we are talking about a single
turn loop that has a relatively low reactance in this application, that
would be ok.

If you buffer (source follow) then use an amplifier, the noise of the
source follower gets amplified. This is why you wouldn't normally do
any buffering before amplification if low noise was the issue. However,
in the case of a tuned loop, you need the high impedance to keep the Q
high.

Exactly my point.....................in the LW/MW ranges, this is not an
issue. When you are using a high Q loopstick, the output voltage out of the
secondary winding is relatively high. This is observed from my own
experience. In direct comparison with the Palomar active loopstick, my own
units have better performance in the areas of output voltage, and small
signal pickup is slightly better because of the higher selectivity of my
implementation.
For higher frequencies where the ambient noise is lower, a different
technique would be required.

Wellbrook gets around that problem by using a loop that is not
tuned. The assumption is the purchaser has a real radio and doesn't
need front end filtering. So cource having a high dynamic range like
the 7030 is a plus.

That isn't such a good deal......................anybody can build that type
of antenna for relatively cheap................the clincher is the amplifier
stage that Wellbrook uses. I would like to take a look at a schematic of
their circuit and see how they do it.

Loops using ferrite will not work as well as a Wellbrook IF the
Wellbrook uses a large enough loop. What you are gaining with the
Wellbrook is aperture. It soaks up lots of RF due to it's size

Pete KE9OA wrote:
Aperture is exactly the point I was going to get at.........this is
the
reason that the receivers I have built that use the Mini-Circuits
mixers as
the first stage have worked very well with an 8 foot diameter untuned
loop.
No IM products anywhere in the LW/MW bands.
wrote in message
oups.com...
Differential pairs have more noise than a single ended design,
which is
why the transformer is the way to go. An ideal transformer has no
noise. I real life transformer has the noise of the resistance in
the
windings. This is why the transformer choice isn't trivial.

I understand this, but take a look at the AD797, and other op-amps
that are
used at very low input signal levels. This device is designed for
ultrasound
transducer applications, strain gauge amplifier service, etc. There
are
quite a few very low noise op-amps that are suitable for RF service.

I'm more familiar with the LT1028, but I took at look at the 797, which
fortunately is an easier op-amp design to follow. It does have a
differential input, but that is because it is an op-amp, which requires
such an input. However, it is worse for input noise. It should be 41%
more noise than a single ended design.[Uncorrelate noise sources add in
a RMS fashion IIRC).

As far as transformers, I don't think there is such a thing as an
ideal
transformer. In my experience, the input Xl must be at least ten
times the
impedance of the expected driving source's impedance at the lowest
frequency
of interest.There is going to be a certain amount of DC resistance in
the
primary winding if we are going to be able to achieve the required
inductance to provide the proper load. If we are talking about a
single
turn loop that has a relatively low reactance in this application,
that
would be ok.
The ideal transformer was just thrown out there for discussion, since
the noise source in a model of the transformer would be the resistance
of the windings.


If you buffer (source follow) then use an amplifier, the noise of
the
source follower gets amplified. This is why you wouldn't normally
do
any buffering before amplification if low noise was the issue.
However,
in the case of a tuned loop, you need the high impedance to keep
the Q
high.

Exactly my point.....................in the LW/MW ranges, this is not
an
issue. When you are using a high Q loopstick, the output voltage out
of the
secondary winding is relatively high. This is observed from my own
experience. In direct comparison with the Palomar active loopstick,
my own
units have better performance in the areas of output voltage, and
small
signal pickup is slightly better because of the higher selectivity of
my
implementation.
For higher frequencies where the ambient noise is lower, a different

technique would be required.

Wellbrook gets around that problem by using a loop that is not
tuned. The assumption is the purchaser has a real radio and doesn't
need front end filtering. So cource having a high dynamic range
like
the 7030 is a plus.

That isn't such a good deal......................anybody can build
that type
of antenna for relatively cheap................the clincher is the
amplifier
stage that Wellbrook uses. I would like to take a look at a schematic
of
their circuit and see how they do it.

Ah, but that is the idea! You sell the amp and people come up with
their own loops. That is also the beauty of the tuned loop in that you
don't have to know exactly what the end user is going to use.




Loops using ferrite will not work as well as a Wellbrook IF the
Wellbrook uses a large enough loop. What you are gaining with the
Wellbrook is aperture. It soaks up lots of RF due to it's size

I've got the drift of what you are saying................good idea!
wrote in message
oups.com...

Pete KE9OA wrote:
Aperture is exactly the point I was going to get at.........this is
the
reason that the receivers I have built that use the Mini-Circuits
mixers as
the first stage have worked very well with an 8 foot diameter untuned
loop.
No IM products anywhere in the LW/MW bands.
wrote in message
oups.com...
Differential pairs have more noise than a single ended design,
which is
why the transformer is the way to go. An ideal transformer has no
noise. I real life transformer has the noise of the resistance in
the
windings. This is why the transformer choice isn't trivial.

I understand this, but take a look at the AD797, and other op-amps
that are
used at very low input signal levels. This device is designed for
ultrasound
transducer applications, strain gauge amplifier service, etc. There
are
quite a few very low noise op-amps that are suitable for RF service.

I'm more familiar with the LT1028, but I took at look at the 797, which
fortunately is an easier op-amp design to follow. It does have a
differential input, but that is because it is an op-amp, which requires
such an input. However, it is worse for input noise. It should be 41%
more noise than a single ended design.[Uncorrelate noise sources add in
a RMS fashion IIRC).

Some day, there will be a noiseless op-amp. Not in my
lifetime..................


As far as transformers, I don't think there is such a thing as an
ideal
transformer. In my experience, the input Xl must be at least ten
times the
impedance of the expected driving source's impedance at the lowest
frequency
of interest.There is going to be a certain amount of DC resistance in
the
primary winding if we are going to be able to achieve the required
inductance to provide the proper load. If we are talking about a
single
turn loop that has a relatively low reactance in this application,
that
would be ok.
The ideal transformer was just thrown out there for discussion, since
the noise source in a model of the transformer would be the resistance
of the windings.

Understood.



If you buffer (source follow) then use an amplifier, the noise of
the
source follower gets amplified. This is why you wouldn't normally
do
any buffering before amplification if low noise was the issue.
However,
in the case of a tuned loop, you need the high impedance to keep
the Q
high.

Exactly my point.....................in the LW/MW ranges, this is not
an
issue. When you are using a high Q loopstick, the output voltage out
of the
secondary winding is relatively high. This is observed from my own
experience. In direct comparison with the Palomar active loopstick,
my own
units have better performance in the areas of output voltage, and
small
signal pickup is slightly better because of the higher selectivity of
my
implementation.
For higher frequencies where the ambient noise is lower, a different

technique would be required.

Wellbrook gets around that problem by using a loop that is not
tuned. The assumption is the purchaser has a real radio and doesn't
need front end filtering. So cource having a high dynamic range
like
the 7030 is a plus.

That isn't such a good deal......................anybody can build
that type
of antenna for relatively cheap................the clincher is the
amplifier
stage that Wellbrook uses. I would like to take a look at a schematic
of
their circuit and see how they do it.

Ah, but that is the idea! You sell the amp and people come up with
their own loops. That is also the beauty of the tuned loop in that you
don't have to know exactly what the end user is going to use.




Loops using ferrite will not work as well as a Wellbrook IF the
Wellbrook uses a large enough loop. What you are gaining with the
Wellbrook is aperture. It soaks up lots of RF due to it's size


You aren't making it easy for me, are you? That's ok......................it
keeps me on my feet!

Pete

Pete KE9OA wrote:

Aperture is exactly the point I was going to get at.........this is the
reason that the receivers I have built that use the Mini-Circuits mixers as
the first stage have worked very well with an 8 foot diameter untuned loop.
No IM products anywhere in the LW/MW bands.

What kind of preselection design do you use ahead of the first mixer
stage?

----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+ Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption =----

It is just a single tuned circuit, where I am tuning the primary of the
loopstick, with the secondary winding driving the amplifier.
With the MMIC I use as an RF amplifier, overload isn't an issue. The RFMD
RF2044 can put out +17dBm of linear power into 50 ohms............1dB
compression doesn't occur until +18.5dBm. It looks like this device is
discontinued, so I need to find another one. I've got around 400 of the
2044s on hand, but for a production run, that isn't quite enough.
For the input stage, I am going to switch to a differential JFET design.
This should help in achieving deep nulls with the antenna. Although I have
been building different iterations of these antennas, the balanced design is
new for me.............I do see advantages with this approach, so I will be
going in this direction for a marketable item.

Pete

"starman" wrote in message
...
Pete KE9OA wrote:

Aperture is exactly the point I was going to get at.........this is the
reason that the receivers I have built that use the Mini-Circuits mixers
as
the first stage have worked very well with an 8 foot diameter untuned
loop.
No IM products anywhere in the LW/MW bands.

What kind of preselection design do you use ahead of the first mixer
stage?

----== Posted via Newsfeeds.Com - Unlimited-Uncensored-Secure Usenet
News==----
http://www.newsfeeds.com The #1 Newsgroup Service in the World! 120,000+
Newsgroups
----= East and West-Coast Server Farms - Total Privacy via Encryption
=----