EMRFD Message Archive 7592

Message Date From Subject
7592 2012-07-14 08:31:49 Reginald Beardsle... Broadband RF design
After several days of reading, I find my library lacks a good reference on broadband RF design.

My immediate interest is a spectrum analyzer frontend covering 0-300 MHz, but what I'm really looking for is a Phd level monograph on the design of broadband (i.e. 10 octaves or more) RF sections. All the ugly stuff where theory meets reality.

Does anyone have any suggestions? I'm living in a small town 75 miles from a rather modest university, so professional journal references are difficult to get. I only recently starting taking QST & QEX. Books are the easiest to get, so long as they're not out of print.

In the absence of a monograph I am prepared to make a longer trek and spend a day collecting papers. So a good set of citations from a recent paper would be very helpful.

Individual design papers are welcome, but I need to understand the problem better before I try to evaluate particular solutions.

Thanks.

Have Fun!
Reg
7595 2012-07-14 11:11:22 Chris Trask Re: Broadband RF design
>
> After several days of reading, I find my library lacks a good reference on
> broadband RF design.
>
> My immediate interest is a spectrum analyzer frontend covering 0-300 MHz,
> but what I'm really looking for is a Phd level monograph on the design of
> broadband (i.e. 10 octaves or more) RF sections. All the ugly stuff where
> theory meets reality.
>
> Does anyone have any suggestions?
>

For that range of octaves you'd be hard pressed to use any form of
transformer feedback amplifier. There are resistive feedback amplifier such
as series/shunt and multi-loop feedback. But, for really good low noise and
wide bandwidth such as you're interested in you should take a serious look
at distributed amplifiers.

These go back more than half a century and were first realised using
pentode vacuum tubes. They were used extensively by Tektronix for the
vertical amplifiers in the 500 series oscilloscopes.

There are many patents, professional papers, and graduate theses on
these. Generally, they are presently designed using FET devices. The
performance can be improved using a FET/bipolar cascode and by tapering the
impedance of the input and output transmission lines.

You can find an extensive bibliography on these at:

http://www.home.earthlink.net/~christrask/distamp.html

Chris Trask
N7ZWY / WDX3HLB
http://www.home.earthlink.net/~christrask/
7596 2012-07-14 13:38:12 ha5rxz Re: Broadband RF design
As a start you could speak to the good folk at Minicircuits. Off the shelf they can supply you with an amplifier chip that can work from DC to 3 GHz and a mixer where, if you use the IF port as an input, will work over a similar range. Attenuators, bias tees, oscillators, they're all there and the tech support department are willing to help.

Disclaimer: I have no connecti
7597 2012-07-14 13:59:09 lasse moell Re: Broadband RF design
OK, so you want start building a spectrum analyzer... Usually a spectrum analyzer is not *that* sensitive, often noise figurers >20 dB can be found. This ease things considerably, as you may only need a low pass filter and maybe an attenuator for isolatoin in front of a mixer. Once you have passed the mixer you are dealing with "narrow band" signals and things should be fairly easy.

Do read about Wes W7ZOI's simple spectrum analyzer, it should give you some ideas!

Building wide band amplifiers often means having loads of feedback and low gain in each stage.

Or as some suggested, use MMIC which runs from DC to GHz and are very cheap too!
7598 2012-07-14 14:15:37 RW Mail List Re: Broadband RF design
Hi,

You may want to have a look at the:
GALI-39 WIDEBAND AMPLIFIER KIT

http://www.minikits.com.au/kits2.html#gali39

Roderick Wall, vk3yc.



7599 2012-07-14 18:14:28 Russell Shaw Re: Broadband RF design
7600 2012-07-14 18:14:58 Kerry Re: Broadband RF design
> You may want to have a look at the:
> GALI-39 WIDEBAND AMPLIFIER KIT
>
> http://www.minikits.com.au/kits2.html#gali39
>
> Roderick Wall, vk3yc.


That's a co-incidence Roderick; I am just finishing one of these.

I just came in for a coffee-break from the workshop where I'm machining the housing and saw your post.

The PCB has provision for two GALI-39s; that seemed useful as I wanted about 40dB of gain.

Mark at MiniKits has never built one with two MMICs so I'm the guinea pig! :)

With the component values I used it gives about 40dB gain from 5 MHz to 1 GHz in a quick test with the PCB "open" on the bench.

It oscillated at first but that was because I used clip-leads; when I tacked a couple of SMAs to the PCB and connected it properly it was excellent.

I will do final tests when it's properly cased in an RF-proof box and send them to Mark at MiniKits.

So I can recommend the MiniKits kit for a good wideband amplifier at reasonable cost.

Kerry VK2TIL.
7601 2012-07-14 18:49:22 RW Mail List Re: Broadband RF design..new thread Wideband MMIC amplifiers.
Hi Kerry,

I have one on order and I’m hoping to get it Monday in the post. I want to use it with the N2PK VNA behind the receiver antenna to measure S21 radiated field strength. This is to test a 10 meter vertical antenna and ground plane system as Rudy has in part 1 of his articles:

http://www.antennasbyn6lf.com/2009/12/series-of-qex-articles-on-ground-system-experiments.html

You’re lucking being able to machine a housing for your MMIC amplifier.

Regards,

Roderick Wall, vk3yc.

7603 2012-07-15 08:50:51 Michael Dunn Re: Broadband RF design
Does anyone know of wideband (~1GHz) amp/buffer chips with high
resistance (a la op-amp) inputs?

Everything I've found so far has been 50 ohms, or relatively low R
(1-100K)...


thanks
7604 2012-07-15 08:50:51 Reginald Beardsle... Re: Broadband RF design
First, thank you to everyone.

I'm looking for a book if one exists.  I have a 5000+ volume technical library.  I like having the answer on my shelf.  It's hell to move, but a real joy the rest of the time.

The Softrock Ensemble II RX uses an Si570C to provide the LO for the Tayloe mixer. By selecting the appropriate bandpass filters it can cover 0 to 3 MHz or 1.8 to 30 MHz. The Si570C can generate clocks up to 280 MHz. With modification that should allow getting an arbitrary segment of spectrum anywhere from 0 to 70 MHz. The B part goes to 810 MHz which would make possible a range up to 200 MHz.

FWIW The 300 MHz was the result of my not reading the data sheet closely enough :-(
The A part goes to 1.45 GHz, but w/ gaps.

The Si570 generates the output by dividing a master clock. It's claimed that it's free of the spurs common in a DDS. Whether that's actually true or not I don't know.

The Ensemble II RX is controllable via USB. So a program can set a frequency, record the passband signal to disk and then repeat the process at a number of frequencies.

This makes possible a DFT based spectrum analyzer. The signal at each LO frequency is transformed via FFT and stored in the appropriate part of the "sweep" spectrum.
After all the steps, the output can be written to disk as ASCII text and plotted w/ any convenient plotting program. It requires writing a very simple program.

This is significantly different from the normal superhet approach because the frequency resolution is limited only by the length of time you're prepared to wait for the result. It also avoids having to build narrow IF filters.

Based on the $70US price of the Ensemble II RX, a 70 MHz analyzer should be about $100 for the parts. I don't yet know what the B part needed to reach 200 MHz costs or whether the C part can be pushed beyond the 280 MHz spec. It's not unreasonable to expect that the many C parts can go a good bit higher. One could also build a hybrid using a superhet frontend to convert to a range the C part can handle.

The motivation for all this is precisely the difficulties W7ZOI and many others who have built hobbyist spectrum analyzers write about. I've collected probably a dozen such papers over the years. If you don't have access to appropriate test gear during construction, it seems to me an impossible undertaking.

A lot of Softrock receivers have been built w/ very limited test gear. That makes me think the design might allow an end run around the problems that make a conventional spectrum analyzer design difficult to build.

I'll know more after my soldering station arrives and I can start building the Ensemble II.

Have Fun!
Reg
7605 2012-07-15 08:50:51 Reginald Beardsle... Re: Broadband RF design (candidate?)
Has anyone here read this?

Design of Ultra Wideband Power Transfer Networks
Binboga Siddik Yarman

http://www.amazon.com/gp/product/0470319895/sr=1-3/qid=1342317238/ref=olp_product_details?ie=UTF8&me=&qid=1342317238&sr=1-3&seller=

It was published 2010 and appears to cover what I'm looking for. Unless someone has read it and pans it, I'll probably get it.

Have Fun!
Reg
7606 2012-07-15 09:44:03 Ivan Makarov Re: Broadband RF design
> The A part goes to 1.45 GHz, but w/ gaps.



I believe one part does not. One would need to order 3 separate parts for
each of the 3 specified frequency ranges.



Regards,

Ivan

VE3iVM



7607 2012-07-15 10:26:07 Chris Trask Re: Broadband RF design
>
>First, thank you to everyone.
>
> I'm looking for a book if one exists.  I have a 5000+ volume
> technical library.  I like having the answer on my shelf. 
> It's hell to move, but a real joy the rest of the time.
>

That certainly eclipses my 3500+ volume technical library. Mine is limited to analogue, passive, and RF design plus the associated math and physics. Not a single microprocessor book to be found anywhere.


Chris Trask
N7ZWY / WDX3HLB
Senior Member IEEE
http://www.home.earthlink.net/~christrask/
7608 2012-07-15 11:31:16 William Carver Re: Broadband RF design (candidate?)
Reginald, let us know your opinion of the book (pricey, but what isn't?)
for practical application.

W7AAZ


On Sat, 2012-07-14 at 19:13 -0700, Reginald Beardsley wrote:
>
>
> Has anyone here read this?
>
> Design of Ultra Wideband Power Transfer Networks
> Binboga Siddik Yarman
>
> http://www.amazon.com/gp/product/0470319895/sr=1-3/qid=1342317238/ref=olp_product_details?ie=UTF8&me=&qid=1342317238&sr=1-3&seller=
>
> It was published 2010 and appears to cover what I'm looking for.
> Unless someone has read it and pans it, I'll probably get it.
>
> Have Fun!
> Reg
>
>
>
>
>
7609 2012-07-15 22:22:04 victor Re: Broadband RF design
About the Ensemble based spectrum analyzer:
You will have some serious shortcomings: the obvious one is the finite opposite image rejection, second is the fact that the input mixer has low conversion loss for odd harmonics. For example, tuning the receiver for 10MHz will show a 30 MHz signal attenuated only by 10dB + any input filter that you use. It is not practical to use switched multiple sub-octave filters with high enough rejecti
7610 2012-07-15 23:31:10 Ashhar Farhan Re: Broadband RF design
reggie,

a lot depends upon what you want the spectrum analyzer to do. if your idea
is to simply measure the harmonic output from your transmitter then even
alan's 555 based spectrum analyzer will do. here is the link
http://www.vk2zay.net/article/256 (it fits into an altoid box). the
construction is simpler than a simple superhet (which is what a spectrum
analyzer is).

you could look at something more complex. like hans summers analyzer
http://www.hanssummers.com/spectrumanalyser.html

however, when you want to measure IMD, IP3 and other stuff, then you need
far far better shielding, dynamic range etc. so, apart from the electronic
design, the fabrication itself becomes a challenge. How do you keep a
signal that is 100db stronger from your tracking generator out of the
analyzer's input?

a good spectrum analyzer is as complicated to make as good receiver
(neither of them are trivial). hence, the difficulties faced by homebrew
builders is precisely because of the the high performance expected.

a problem with SDR based spectrum analyzers that use 'zero-IF' is that
almost all mixers (including the tayloe mixer) respond to the signal
harmonics as well. hence, there are bound to responses at the harmonics of
the received frequency, one can eliminate it by low pass filters, but then,
it will become impossible to sweep more than an octave at a time.

i have built two version of spec ans by now using the si570, and i am onto
my third (hopefully, i will get it right this time). the si570s have a
quirk. their phase noise increases at lower frequencies. i guess this is
due to the increasing jitter among the digital frequency dividers. when i
say that the phase noise increases, i meant it by si570's own standards. it
is a very very quite chip about 50 MHz. hence, a better idea is to beat
down the output of the si570 to HF range by mixing it with a low noise
crystal oscillator.

- farhan

7611 2012-07-16 07:14:01 Roelof Bakker Re: Broadband RF design
Hello Michael,

The MAX 497 is a dip 16 package with 4 buffer amplifiers that goes to 375
MHz.
The input is 1M / 2 pF.

I have one on the bench used as a 4 port distribution amplifier for 0 -
100 MHz.

73,
Roelof, pa0rdt
7616 2012-07-16 21:20:01 Michael Dunn Re: Broadband RF design
Thanks, though not the response I need. Probably because it's a DIP!!!!! :-)

I'll have to do a more exhaustive search - will post any interesting findings.


At 4:13 PM +0200 7/16/12, Roelof Bakker wrote:
>Hello Michael,
>
>The MAX 497 is a dip 16 package with 4 buffer amplifiers that goes to 375
>MHz.
>The input is 1M / 2 pF.
>
>I have one on the bench used as a 4 port distribution amplifier for 0 -
>100 MHz.
>
>73,
>Roelof, pa0rdt
7618 2012-07-16 21:20:17 Reginald Beardsle... Re: Broadband RF design
Farhan,

Thanks for the links.

The Softrock will just be to prototype the concept and write the software. It clearly doesn't have the shielding needed for serious work.

There are a great many construction issues.
I plan to use PCB boxes for each section w/ feed throughs and coax connections, etc. All in a larger box w/ lots of room to allow changing sections. As noted, I've read quite a few articles on home brew units over the years.

For what I have in mind, I don't think a tracking generator makes sense because the LO is not being swept. Not sure yet what would take its place though.

I'd love to learn more about your work w/ the Si570.

Have Fun!
Reg


---