EMRFD Message Archive 40

Message Date From Subject
40 2006-07-28 20:41:34 Stan What are you going to build before XMAS?
There are many good starts to building some minimal circuits in the
EMRF book. Lets face it only rarely is one of us going to
build/design a complete multi-band solid state transceiver like a
CDG2000 or K2 from scratch.

That said, for example there are numerous opportunities to build
small simple SSB exciters using either Phasing or Crystal filters.
The basis are in the book. Some of the old timers will remember
the SSB-Jr from the GE Ham News days. Now with your EMRF text at
your side, your PC connected to the Internet; any ham who
understands the basic LC and ohm law equations can build an solid
state equivalent that will work even better than the SSB-Jr.

Programs like LT Spice from Linear Technology (free), DL5SWB mini
ring core calculator program "mini RC-Calculator (free), the RF CD
disks from EMRF and the ARRL handbook are the tools. Down load the
data sheets for 2N3904, 2N3906, 2N4401, 2N5484, 2N7000 and J310
which are all cheaper than a case of water or your favorite brew and
you are set to have a entire winter of fun. Even use an old glowbug
6AG7, 5763, etc as your final and you got it man.

There are enough fellows on this list to answer your questions if
you get in to trouble. So what is keeping you from getting
started. The tools are Free, the transistors are cheap, the
technical help is available for the asking.

Their should be at least 50 entries (different designs) in an
homebrew ARRL contest just from this list alone.

Need more to convince you – take a look at Hans Summers G0UPL – look
at all the things he has designed and built. Clear thinking, a few
low cost parts, a desire to exceed is what it takes.

Who will be First to post pics of their new design, even if it does
not work and you want help. DO IT

My two cents - Stan AK0B
51 2006-08-01 16:14:05 kilocycles Re: What are you going to build before XMAS?
Stan, Roger & group,
My apologies if this got posted twice...I receive an error "Cannot
retrieve Message 40" the first time I replied.

Here's what I've got so far for an almost all-EMRFD, 2-band CW/SSB
transceiver design...

First, a caveat...since I plan on using many of these modules with a
Belthorn SSB IF module-based 17 and 12 m SSB transceiver as well, I'm
CADding most everything in EAGLE, and I will use my Harbor Freight
catalog pages-toner transfer method of making PC boards. I had to
make several of the components for the rig in EAGLE, such as the MMICs
used in the transmit RF chain, and the mixers.

For the EMRFD transceiver, I've made a PCB and am installing the parts
for the IF system in Fig. 6.50. All that's needed is to plug in the
L-network components for the final crystal filter(s) design. This
will be what I consider to be the foundation module for a 5 Watt 20
and 15 meter CW/SSB transceiver. I have chosen the IF at 9 MHz (or as
an alternative, 9.216 MHz; later), with LO injection via mixing HFO's
of 25 and 18 MHz with a VFO of 5 to 5.500 MHz. These are all cheaply
available crystals. After my characterization of the 9 MHz ones, I
have some concerns about them...they were Jameco surplus, and no
longer listed, unlike all of the others I've used, which are Mouser
components. I have tested just enough good ones for a 6-Pole
Chebychev SSB filter, and a 4-pole CW filter. After discussing this
with Jim, K8IQY (we both use Jim's PVXO design for crystal
characterization), I've obtained as a backup several 9.216 ECS series
resonant crystals as a possible IF frequency, with a corresponding
shift in VFO frequencies.

As it stands now, I'm planning on using the VFO and HFO described in
Fig. 6.148, modified for the 5 MHz range, switching the HFOs with a
74HC7453 analog multiplexer IC. That's been CADded in EAGLE. I'm
planning on rolling my own diode mixers used throughout the rig, and
I've obtained several FT-23-43 cores, Agilent/Avago HSMS-2822 RF mixer
diode pairs, and some trifilar wire to wind them with. I designed the
mixers to fit on machine-pin 14-pin DIP sockets, with every other pin
removed. This is similar to the SBL-1, except the DIP is wider. I've
made a "flex-mixer component in EAGLE to accommodate either the
homebrew mixers or the SBL-1 footprint. The LO will feed the transmit
and receive mixers through triple-tuned BPFs. Post mixer amplifiers
and terminations are planned per the guidance in EMRFD.

The transmit RF stages are based on Fig. 11.14, and I've obtained some
surplus MAR-3's, and the Avago MSA-0686 MAR-6 equivalents from
Digi-Key. The PA design that I am currently working on is actually
from "Solid State Design", Fig. 39, pg. 66, substituting some on-hand
2SC2078s.

My preliminary working design for the product detector and audio
stages is based on Fig. 6.72 (from the EZ-90 receiver). That's still
in the unfinished CAD stage.

While I am not planning on using a bidirectional IF design, I intend
to use using the IF filters in the transmit chain, following the
schema of Fig. 6.152 for the SSB and CW generators.

I haven't decided on the receive RF stages yet. Also, I haven't
detailed out the control and switching circuits for the power and
shared modules. My intent is to use inexpensive ICs such as the
74HC4053 and 4066, but some low-current, sensitive 12V DPDT relays of
the same form factor as the Omron G6A series that I used for the EMRFD
Chapter 1 transmitter last Christmas are available for around $2, half
the price of a G6A-234; but that's still much more expensive than the ICs.

I've been working on this project since late last Fall, on and off, as
I've done other projects as well. I'm not sure if I can finish it by
this Christmas, but I'd surely like to. I also have to finish several
important pieces of test equipment, including the W7ZOI versi
57 2006-08-02 11:48:28 Wes Hayward Re: What are you going to build before XMAS?
Hi Ted, et al,

Wow, you have a strong plan. That sounds like a wonderful little
transceiver and we will all be anxious to see it.

A couple of things came to mind as I read your posting while
flipping through the books. First, you seem to be on top of the
problems of the crystal filters. On the other hand, you can live
with a fairly low crystal Q when building SSB filters. Crystal
loss may degrade the corners of a Chebyshev filter, but that's
rarely a problem. So you may be OK.

I have no experience with the mosfet analog switches for RF
switching applications. What I would probably do is to put the
switches that I wanted to use in a breadboard and measure their
performance. The parameters of interest would be the loss when in
the "on" status, the attenuation when "off," and the IMD (IIP3) for
both situations. The Fairchild part that W7AAZ, et al, have used
in the "Triad" receiver of EMRFD Ch 6 is a very low resistance when
on and performs well as a mixer. I would guess it would be great.
The 4066 is, I believe, a smaller geometry part and may not work as
well. Some of these switches may work better at higher impedance
and that is worth investigating too. IMD is a hard thing to model
with SPICE, for the switch models may not always be good enough.

That amplifier from SSD is a pretty old design. It would be
interesting to compare performance with a similar design with
transformers wound on some high permeability balun cores. Also,
the biasing in that design looks a little weak. With that 2K
resistor, you are only stuffing 14 mA through the bias diode. That
will turn it on to perhaps 0.6 volt or so, which is barely enough.
It may take some playing, which could be a lot of fun and
educational. For this chore, I would pull out the set up described
in Ch 7 that allows you to do IMD measurements with the PC and a
sound card.

The homebrew mixers sounds like an instructive exercise. I would
want to evaluate the mixers alone for IIP3, gain, and isolation.
Generally the newer TUF mixers have better isolation than the old
SBL-1 parts, and the TUF-3 is preferred over the TUF-1 for HF
applications.

Good luck with the project. It sounds like a real gem.

73, Wes
w7zoi
60 2006-08-02 14:38:16 kilocycles Re: What are you going to build before XMAS?
Hi Wes,
Thanks for the input...it's always welcome from you.

On the variuos IC switches, the FST3125M from the Triad receiver has
by far the better specs for On resistance, by a factor of 10, but
Digi-Key and Mouser bot have min order quantities of over 2,000 units.
That is really too bad, because they look ideal for my purposes. I
got he idea of the 74HC4053 from Steve, KD1JV's Minimalist SSB rig,
where he uses it for all bidirectional switching. No direct
experience, though, because that project is still in the "Zip-lock
bag" stage...PCB and all components, waiting to be built! The
74HC4066 from the applications in EMRFD could perform the same
function, and it has 35 ohms Ron vs 50 for the 4053, and both have
similar leakage currents in the .1 uA range. The Icc quiescent
current of the 4053 is only 4 uA compared to 40 uA for the 4066.
Those are my best ability interpretations of the data sheets, anyway.

I've seen specs on some of the relays that are useable well above UHF,
and they have isolation specs better than 60 dB. Of the inexpensive
$1 to $2 surplus 720 ma/12V sensitive DPDT relays, the only one I've
seen published RF specs for is a Tyco/Axiom FT2/FU2 with
-30dB@100MHz/1.02 VSWR. The Omron G6Y SPDT is 65 dB isolation and
0.2dB insertion loss at 900 MHz is about $4 from Mouser.

On the mixers, I hope to be able to test them separately as you say.
My PC-based O-scope/FFT spectrum analyzer only goes up to 16 MHz, but
hopefully Mike, KO4WX will get his homebrew SA back working soon, and
I can come up with a plan for doing that.

A couple of other comments, Wes. I was intrigued by the 18 MHz
transceiver RF chain in Fig.11.14 because it uses MMICs, something I
knew little about. So, I started doing the reading on those
fixed-gain devices, and decided that I'd use that part of the circuit
for the transmit stage. I could go ahead and use the IRF511
(actually, I have a bunch of 510s) as shown in the same figure as the
PA. Another of my theoretical weaknesses is impedance matching of
these low impedance circuits via baluns or RF transformers. I have a
simple single IRF510 PA circuit for 80m on the test bench that I'm
experimenting with now, trying to get a feel for manipulating the
parameters empirically. So far, I've been able to get 3.5 Watts out
for a (resistively-attenuated) drive of 1.7 Watt. I bought a variety
of balun cores and other ferrites from CWS Bytemark to experiment with
in these circuits, to try to duplicate some of the EMRFD designs as
practice.

In a nutshell, my plodding progress
61 2006-08-02 15:21:54 Dan Rae Re: What are you going to build before XMAS?
kilocycles wrote:

> the FST3125M from the Triad receiver has
>by far the better specs for On resistance, by a factor of 10, but
>Digi-Key and Mouser both have min order quantities of over 2,000 units
>
Digikey has the TI version in small quantities:
296-1165-1-ND
<http://www.digikey.com/scripts/DkSearch/dksus.dll?Detail?Ref=395132&Row=67043&Site=US>
56 cents each
I have used them in my cdg2000 interchangeably with the Fairchild version...
73
dan
ac6ao / g3ncr
62 2006-08-02 15:43:11 brainerd@wildblue... Re: What are you going to build before XMAS?
The FST3253 is available from Mouser in single quantities for .39(it is currently "On
Order"). It has the same on resistance (4 ohms) as the FST3125 but is a 1 of 8
demux like the 74HC4053. I am using it as a mixer in my Software Defined Radio
and works well. Check out the softrock40 and the soft_radio Yahoo groups.

Dave - WB6DHW

On 2 Aug 2006 at 21:37, kilocycles wrote:

> Hi Wes,
> Thanks for the input...it's always welcome from you.
>
> On the variuos IC switches, the FST3125M from the Triad receiver has
> by far the better specs for On resistance, by a factor of 10, but
> Digi-Key and Mouser bot have min order quantities of over 2,000 units.
> That is really too bad, because they look ideal for my purposes. I
> got he idea of the 74HC4053 from Steve, KD1JV's Minimalist SSB rig,
> where he uses it for all bidirectional switching. No direct
> experience, though, because that project is still in the "Zip-lock
> bag" stage...PCB and all components, waiting to be built! The
> 74HC4066 from the applications in EMRFD could perform the same
> function, and it has 35 ohms Ron vs 50 for the 4053, and both have
> similar leakage currents in the .1 uA range. The Icc quiescent
> current of the 4053 is only 4 uA compared to 40 uA for the 4066.
> Those are my best ability interpretations of the data sheets, anyway.
64 2006-08-02 19:11:10 kilocycles Re: What are you going to build before XMAS?
Dan and Dave,
Thanks for the info on both of the ICs. I added the FST3253s to an
outstanding Mouser order, and I see that Digi-Key now has the elusive
Panasonic 2SC5739 in stock now, so I put in an order with them as
well, for a bunch of the transistors and the FST3125s.

I'll check the two yahoo SDR groups to see how the FST bus switches
been employed. From looking at the data sheets, they look like
work-alikes to the 4066; I guess the "bilateral" nature is implied.
It was stated specifically