I would like to make a flexible CW TX with a broad coverage of HF,
something like:

VXO or BB BB POWER
AMP TUNABLE BPF
VFO === DRIVER === (2-5W, derated)
=== PLUGIN LPFs
(50-100 mW) AMP (.5 to 1W) RF
TRANSFORMER R/T SWITCHING


I'd like the driver-PA chain to be no-tune and unflappably stable
over, say, 7-15 or 3-30 MHz, with 50 ohm I/O. I guess the driver
stage(s) need not be nondistorting, as a C class PA will produce much
in the way of harmonics to be accurately filtered out, but it would be
nice-to-have. I don't know if a more linear power amp would help, but
I'm willing to use large a derated final and go balanced and/or class
A, with negative feedback, if it helps stability and bandwidth. That
would help spectral purity too, but I see it as a side effect not to
be relied on too much.



I looked around, found plenty of professional ready made gear, but not
much in the way of tips.

I'd be tempted to start as with two low-distortion audio-like stages
with negative feedback and broad bandwidth-limiting, such as small
caps to ground, and large coupling caps + VHF beads in the signal
paths, but it would be trial and error.

Any suggestions on where I should have looked???



tx es 73 de n1jpr/i2

Ooops! I see lines were cut short. I meant this:


VXO, VFO, DDS.... (50-100mW)
(interchangeable alternatives)
¦¦
¦¦
¦¦
BROADBAND DRIVER (.5-1W)
¦¦
¦¦
¦¦
BROADBAND PA (2-5 W, highly derated devices)
Z-MATCHING RF TRANSFORMER
¦¦
¦¦
¦¦
SWITCHING ================= RX
TUNABLE BPF
PLUG-IN LPFs
¦¦
¦¦
ANTENNA

wrote in message
ups.com...
Ooops! I see lines were cut short. I meant this:


VXO, VFO, DDS.... (50-100mW)
(interchangeable alternatives)
¦¦
¦¦
¦¦
BROADBAND DRIVER (.5-1W)
¦¦
¦¦
¦¦
BROADBAND PA (2-5 W, highly derated devices)
Z-MATCHING RF TRANSFORMER
¦¦
¦¦
¦¦
SWITCHING ================= RX
TUNABLE BPF
PLUG-IN LPFs
¦¦
¦¦
ANTENNA

I sincerely think you would have a lot more fun trying to clone a HW8.

There is no need to pursue much linearity in a cw transmitter ... unless you
try to build an almost unpractical one.

On Sep 11, 8:06 pm, "Philo" wrote:


I sincerely think you would have a lot more fun trying to clone a HW8.

There is no need to pursue much linearity in a cw transmitter ... unless you
try to build an almost unpractical one.

Granted, Philo...

Here's how the argument goes:
- goal: minimal part count + ability to put clean QRP signal anywhere
- DC drain no problem - class C not indispensable
- linear broadband keeps part count down and adds little spurs
- do all filtering after PA, if possible with multi-tuned tanks or
plug-in LPFs

However, my request is now moot.... I don know what possessed me when
I did my previous searches, but the aforesaid info CAN be found on the
web.

Here is one example:

http://www.qsl.net/ik4auy/article_4_english.htm

Plenty of usable ideas.

I am now experimenting with a simple 2-stage drive (2nd stage push-
pull) with plenty of negative feedback. May be enough to go on air.

wrote in message
ups.com...
On Sep 11, 8:06 pm, "Philo" wrote:


I sincerely think you would have a lot more fun trying to clone a HW8.

There is no need to pursue much linearity in a cw transmitter ... unless
you
try to build an almost unpractical one.

Granted, Philo...

Here's how the argument goes:
- goal: minimal part count + ability to put clean QRP signal anywhere
- DC drain no problem - class C not indispensable
- linear broadband keeps part count down and adds little spurs
- do all filtering after PA, if possible with multi-tuned tanks or
plug-in LPFs


There are a few items in your project I have doubts about.

The linear amplifier, for instance. You could build a very linear broadband
amplifier with paired CATV transistors in push-pull. Those transistors
became old stuff and were trashed the moment they were replaced by IC's
(TRW, Motorola, etc.). You still can find some, but a matched pair will be
hard to come by. With some transistors you may even have problems to get
beneath 5 MHz. And forget about a power supply of 12 volts: you will need at
least 2 x 24 volts DC. Anyway, the maximum output will be around 2 volts if
you require quality. So you'll need another 10 to 12 dB to get an output of
one watt or so. This will be an even greater challenge if you want to keep
spurious/harmonics down to -40 or -50 dB.

I also doubt filtering at the output only is practicable. Once garbage got
in an amplifier chain, it is very hard to get rid of it. Since you are
considering using plug-in filters, build sets of three: use the first
between the oscillator and the driver stage, the second one between the
driver and the final stage the third in the antenna circuit.

Have fun!

On Wed, 12 Sep 2007 13:46:32 -0700, wrote:

On Sep 11, 8:06 pm, "Philo" wrote:


I sincerely think you would have a lot more fun trying to clone a HW8.

There is no need to pursue much linearity in a cw transmitter ... unless you
try to build an almost unpractical one.

Granted, Philo...

Here's how the argument goes:
- goal: minimal part count + ability to put clean QRP signal anywhere
- DC drain no problem - class C not indispensable
- linear broadband keeps part count down and adds little spurs
- do all filtering after PA, if possible with multi-tuned tanks or
plug-in LPFs

However, my request is now moot.... I don know what possessed me when
I did my previous searches, but the aforesaid info CAN be found on the
web.

Here is one example:

http://www.qsl.net/ik4auy/article_4_english.htm

Plenty of usable ideas.

I am now experimenting with a simple 2-stage drive (2nd stage push-
pull) with plenty of negative feedback. May be enough to go on air.



You might want to check out the Mistubishi RD16HHF1 RF MOSFET. They
are available from RF Parts at $4.20 apiece, and have similar specs to
the 2SC1969, except with 4 dB more gain. The data sheet tests them at
12.5VDC @ 30 MHz. W7ZOI did a post to the EMRFD Yahoo group about
them. I bought 35 for our QRP club at $3.10 each from RF Parts.

By the way, CB shops advertise them as "drop-in replacements" for the
2SC1969. Grain of salt there........

Hi, one that would possibly be quite close your needs has been described in
the VHF Communications 1 / 1982. It's a linear 3 stage wide band HF driver 4W
output and 44db power gain. Supply voltages are 28V 0.25A and 15V 0.08A.
The article is on four pages.

br oh6io


wrote:
I would like to make a flexible CW TX with a broad coverage of HF,
something like:

VXO or BB BB POWER
AMP TUNABLE BPF
VFO === DRIVER === (2-5W, derated)
=== PLUGIN LPFs
(50-100 mW) AMP (.5 to 1W) RF
TRANSFORMER R/T SWITCHING


I'd like the driver-PA chain to be no-tune and unflappably stable
over, say, 7-15 or 3-30 MHz, with 50 ohm I/O. I guess the driver
stage(s) need not be nondistorting, as a C class PA will produce much
in the way of harmonics to be accurately filtered out, but it would be
nice-to-have. I don't know if a more linear power amp would help, but
I'm willing to use large a derated final and go balanced and/or class
A, with negative feedback, if it helps stability and bandwidth. That
would help spectral purity too, but I see it as a side effect not to
be relied on too much.



I looked around, found plenty of professional ready made gear, but not
much in the way of tips.

I'd be tempted to start as with two low-distortion audio-like stages
with negative feedback and broad bandwidth-limiting, such as small
caps to ground, and large coupling caps + VHF beads in the signal
paths, but it would be trial and error.

Any suggestions on where I should have looked???



tx es 73 de n1jpr/i2

(still me, N1JPR/I2 Filippo, different account...)


Thank you folks, lots of food for thought!

An update:

- I found 3 2SC3502 video finals in the PCB next to the tip of a CRT
in a PC VDU: fT 150MHz, but low max Ic 0.1A and about 1 watt
dissipation. Might be OK as oscillator or buffer, but I'm not sure it
would make a good driver. I'll try it.

- Decided to try a seemingly unsuitable device, a 2SK2545 low-
frequency power switching MOSFET from some industrial automation
contraption, with a humongous 1300pF input capacitance.

I drove the MOSFET DC coupled (!) at 0V DC with a common-collector
BD135 driver in order to swamp the base with current, and I am getting
out over 2W at 10MHz at about 12 ohm measured impedance, in class C,
with 12V supply.

Sounds cool considering that this MOSFET can switch 600V, handle 6A
and dissipate 40W. Actually, the MOSFET indeed STAYS rather cool,
while the dummy load of 3x 4 ohm 1 watt resistors get quite hot.


The common-collector driver + common-source final combination is quite
promising. My next challenges are

1) Making sure that no AF,LF, VHF spurs ('cept "clean" harmonics) and
clicks are generated across a range of voltages and frequencies. Now I
do get strange vibes at the lower DC supply voltages, probably via the
supply bus. Negative feedback seems to make a difference even in
class C.

2) Reduce driver current! For some strange reason I can't convince
drivers to run in class C, neither as common drain nor as common
source: even if I raise emitter voltage above base, gain drops (and
the MOSFET final goes to sleep), but the driver output still seems
undistorted and the driver dissipates like a stove, more so than the
final... more later when I get some proper data.


To those that advise putting tank circuits between stages: go see the
many 2- or 3-transistor QRP transmitters you find in magazines, kits,
etc.: putting all the filtering after the C-class final is very
common. If the VFO or VXO puts out just fundamental + harmonics, then
a LPF is all that's needed.


It would be cool to stitch together a 2-transistor thingie that works
on 5-15 MHz, gives 2W RF at 12V and ???W when fed say 30Vdc - into a
higher impedance, of course.


Stay tuned, & keep reminding me to be sane. :-)