Hi all,

finally after some days I was able to find again the original Harry's page,
the link is right: http://web.telia.com/~u85920178/

and in the Menu you have to look at "AF/amp circuits", the "3 W HF QRP" is
one I mean.
I prefer to point the original circuit as it's by Harry.
Anyway I tested it with several different BJT and had always good results.
Take care at the bias trimmer. If someone is interested I can supply some
pics of my last assembly, but I think Harry's page is more than enough. Hope
to be useful  73 to all group

Alex I0SKK


email: i0skk@...

Web: www.eco-lavoro.com/i0skk



--
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.1.405 / Virus Database: 268.11.7/432 - Release Date: 29/08/2006

--- In qrsstxrx@yahoogroups.com, "Stan" <ak0b@...> wrote:
>
>
>
> Harry's (SM0VPO) RF amplifier web page
>
>   http://web.telia.com/~u85920178/tx/rf_pa_cct_00.htm
>
> I am not sure if this is the one Alex was referring, but it is sure
> nice and he has a lot more information on his web site.
>
> Stan ak0b
>

  Hi Stan  ,  and all

  I think that Alex is referring to this schematic:
  http://web.telia.com/~u85920178/tx/qrp-hf1.htm
  Is the same I am using since the beginning of my QRSS
  experiments ,one year ago .

  I have uploaded on "files" my actual QRSS_RX circuit
  with VXO around 10 140 000 HZ .
  I am very curious to replace the VXO
  with my DDS based VFO (AD 9835)
  for having a wider view ( certainly at cost of more noise...).

                                              73 de Paolo I1DFS

Harry's (SM0VPO) RF amplifier web page

   http://web.telia.com/~u85920178/tx/rf_pa_cct_00.htm

I am not sure if this is the one Alex was referring, but it is sure
nice and he has a lot more information on his web site.

Stan ak0b

--- In qrsstxrx@yahoogroups.com, "Alessandro Santucci" <i0skk@...> wrote:
>
> Hallo guys,
>
> I read your post about PA; if I can be useful in past I used a
revisited PA
> circuit by SM0VPO, I tested it with several BJT and several power
(from 500
> mW up to 4 W on 7 and 10 MHz) and found it very stable and clean in
> frequency spectrum. As I'll find it I'll post on the group.
>
> 73's
>
> Alex I0SKK

Hi Alex- that will be very interesting.  Thanks a lot...

Ciao - Andy

Hallo guys,

I read your post about PA; if I can be useful in past I used a revisited PA
circuit by SM0VPO, I tested it with several BJT and several power (from 500
mW up to 4 W on 7 and 10 MHz) and found it very stable and clean in
frequency spectrum. As I'll find it I'll post on the group.

73's

Alex I0SKK


email: i0skk@...

Web: www.eco-lavoro.com/i0skk



--
No virus found in this outgoing message.
Checked by AVG Free Edition.
Version: 7.1.405 / Virus Database: 268.11.5/425 - Release Date: 22/08/2006

Thanks for the comments fellows.

I think I am going to build a couple of I7SWX's circuit Figure 3 from
RedCom June 2004.

While not the top specs, it is better than what I have been using in
my homebrew designs and should not have the problem of trying to build
duplicate transformers.

I will try one as a DC receiver and one driving a 9 MHz IF.

Stan ak0b

--- In qrsstxrx@yahoogroups.com, "Hans Summers" <hans.g0upl@...> wrote:

>
> You are referring to the original CMOS 4066.


This is correct.  It is also the chip that Stan mentioned, to which my
reply was directed.  I agree with him (and you) that there are several
better chips to consider.


Andy

Andy,

> The same could be said of the 4066.  I have the National
> Semiconductors cmos data book in front of me.  It says that with a
> supply voltage of 15v (i.e. the best conditions) the timing delay from
> the control input to the signal output is 50ns on, and 50ns off.  That
> leaves precisely no time for the signal to get through at 10MHz.  Of
> course the definition of timing delays is 10%-90% so some signal will
> in fact get through, but clearly this device is more suitable for
> lower frequencies - its performance is compromised at 10MHz even at 15v.

You are referring to the original CMOS 4066. That is like considering
using the 4013 instead of a 74AC74 as the flip flop.

The 74HC4066 has delays around 12ns at 5V supply.

> The channel 'on' resistance is 80 ohms with 15v supply, and 270 ohms
> at 5v, so if this device is used as a mixer, it should be used in
> circuits with impedance levels considerably higher than 50 ohms if
> substantial additional conversion loss is to be avoided.  A matching
> transformer at the input should be routine.

Similarly the specs for a 74HC4066 are about 45 ohms resistance at 5V supply.

For maximum performance I agree though, a better device would be the
FST3125 or equivalent, having 5nS delay and 4 ohm ON resistance (if my
memory serves correctly).

73 Hans G0UPL
http://www.hanssummers.com

--- In qrsstxrx@yahoogroups.com, "Stan" <ak0b@...> wrote:
>
> After going through the files on H mode mixers I wonder why they are
> not being used in homebrew equipment to a greater extent.
>
> When one looks at the write up on the 4066 one it is superior to
> anything one could ever do with a 602.  And the 4066 is not the best
> one in the lot.
>
> What problems should we be designing out if we want to use H mode?
>
> Switching transits, timing problems, PCB layout, costs ?
>
> appreciate any comments,
>
> Stan ak0b
>
Hi Stan - a couple of observations-

I have never understood why the SA602 is used as a front-end mixer in
receivers.  Sprat often carries such designs, but even casual reading
of the Philips data sheet makes it clear that the device is highly
unsuitable for such an application.  I suppose the fact that it is
easily available and simple to use makes people go for it
unthinkingly.  Perhaps there is indomitable optimism that mere facts,
as carried on data sheets, can be ignored.

The same could be said of the 4066.  I have the National
Semiconductors cmos data book in front of me.  It says that with a
supply voltage of 15v (i.e. the best conditions) the timing delay from
the control input to the signal output is 50ns on, and 50ns off.  That
leaves precisely no time for the signal to get through at 10MHz.  Of
course the definition of timing delays is 10%-90% so some signal will
in fact get through, but clearly this device is more suitable for
lower frequencies - its performance is compromised at 10MHz even at 15v.

The channel 'on' resistance is 80 ohms with 15v supply, and 270 ohms
at 5v, so if this device is used as a mixer, it should be used in
circuits with impedance levels considerably higher than 50 ohms if
substantial additional conversion loss is to be avoided.  A matching
transformer at the input should be routine.

Re designing out problems with commutative mixers (of which the H mode
is a particular type), my feeling is that there is some scope for
simplification at the cost of a small (at present not clear exactly
how much) reduction in 3rd order intercept.  The potential
simplification is to dispense with the 74ac74 bistable, and drive the
antiphase enable inputs from inverters.

The best figure quoted for intercept point I have seen for this type
of mixer is 54dBm.  This should be compared with -13dBm for the 602,
and an estimated 20dBm as a guideline minimum requirement for an hf
receiver.  Clearly there is some scope for judicious trimming of the
design here: it is useful to distinguish between a situation where
designers are exploting the limitations of a technique, and one in
which a practical device is being considered.

I think it is useful to distinguish between H mode mixers and other
commutative types (i.e the general family of mixers using switching
devices). As I understand it the H-mode concept is to arrange the
circuit so that when the switch is on, the signal is grounded (rather
than being switched through to the load).  This avoids the possibility
that the signal will be distorted (introduction of harmonics), which
is normally the case when a large-amplitude signal takes the channel
voltage close to Vdd or Vss when the channel is on.  The 'H' comes
from the form of the circuit - it is H-shaped.

Best wishes to all - Andy G4OEP

Hi Stan and all
I agree with you for the H_MODE mixer.
I did a lot of tests with the NE602 but there is no coparison
with a switching mixer with the 74HC4066.
I am just testing on 30 mt a DSB version of it , with very good
results.
For a single frequency I think is the best solution.

                                      73 de Paolo  IZ1KXQ / I1DFS


--- In qrsstxrx@yahoogroups.com, "Stan" <ak0b@...> wrote:
>
> After going through the files on H mode mixers I wonder why they are
> not being used in homebrew equipment to a greater extent.
>
> When one looks at the write up on the 4066 one it is superior to
> anything one could ever do with a 602.  And the 4066 is not the best
> one in the lot.
>
> What problems should we be designing out if we want to use H mode?
>
> Switching transits, timing problems, PCB layout, costs ?
>
> appreciate any comments,
>
> Stan ak0b
>

After going through the files on H mode mixers I wonder why they are
not being used in homebrew equipment to a greater extent.

When one looks at the write up on the 4066 one it is superior to
anything one could ever do with a 602.  And the 4066 is not the best
one in the lot.

What problems should we be designing out if we want to use H mode?

Switching transits, timing problems, PCB layout, costs ?

appreciate any comments,

Stan ak0b

Hi to all on the list,

   Regarding my recent posting "AF Filter experiments" I decided my
filter was performing well enough and decided to look for a better design.

In my search for the better design I found a much better active filter
but also much more complex. It left me thinking that this kind of
thing often performs much better in software.

Then I remembered, Spectran has a difinable bandpass filter option
built into the software and a "DeHum" mode to remove power line noise
too. I have not use Spectrum Lab very much but it probably has these
features too. So perhaps we dont need a hardware filter with high
spec', just enough filtering/conditioning to remove the risk of
overloading the sound card and keep out the strongest of the unwanted
signals.

I will put my bread-board back in the cupboard :-)

73's to all on the list.

Des (M0AYF)

I have just put the finishing touches to a web-page devoted to the
discussion of signal frequency crystal filters, work undertaken in
collaboration with Des M0AYF.

http://g4oep.atspace.com/crystal_filters/signal_frequency_crystal_filters.htm

A point dropped out of this which caught my attention - a very simple
and well-shaped single crystal LSB (!!!) filter can be made which
might be excellent for the qrsstxrx transmitter, PROVIDED inverted
audio can be generated.  I would be very interested to hear comments
from the Helschreiber and fancy signal buffs (triangulators, etc)
about whether they would find it easy to adapt their techniques to
provide a frequency-inverted af signal or whether, for them, usb
transmission is a must-have.


Andy G4OEP

Hi everyone,

    This group seems to have gone a little quiet so I thought I would
post an item which may be of interest to the group and the QRSSTXRX
project as a whole.

Hans (G0UPL) pointed out that with the simple sa602 based dc-rx the AF
bandwidth is wider than required with the result that Argo's agc
reacts to interference or qrm and looses sensitivity. Hans also
sugested an AF filter to "roll-off" frequencies much above those
required for qrss might be helpful

Paolo (I1DFS) also pointed out that moving LO frequency such that the
audio fed to the sound card corrosponding to 10.140 MHz is moved upto
say 2 or 3 kHz helps to seperate power line harmonics from wanted signals.

My initial thought was to implement one stage of high pass filtering
cascaded with one stage of low pass filteringto attack both problems.
This becomes an audio bandpass filter! So the next step was to adapt
one of the many designs for CW audio filters built around op-amps. The
design I have used has two stages each with a Voltage gain of 3 and a
"Q" of 10 with the filter centered on 1650 Hz which corrosponds to
10.140050 for my dc-rx. The desgn I have chosen is not ground breaking
technology but it does come with the full set of design formulas so we
can chop-and-change as required to suit our requirements.

Results so far:
I dont wish to "jump the gun" but initial tests look.... interesting.
The good news is that any trace of power line noise/harmonics becomes
history. With the filter engaged I can just detect the 50 Hz line in
Argo at full sensitivity but no visible harmonics of 50 Hz. In its
present form the filter clearly "rings" though this has to be put into
perspective. S-l-o-w qrss signals dont appear to suffer because of
this. QRM is much reduced though "close in" signals are still a
problem. At the time of writing I have seen only one qrss signal with
which to test and that was Hans signal but it was of too higer
strength to make meningful conclusions as to the performance of the
filter.

One other thing struck me, while doing initial tests on a quite band
(except for the odd burst of data and lightening crash)I was able to
have my "line I/P" gain and "sensitivity" set to much higher levels
than before (maximum) with no hint adverse effect on Argo.

The only "bad" news so far is that any "click" of interference now
becomes a "bang" due to the gain of the filter. To avoid damage to the
sound card due to static "pops" etc I decided to add a few
back-to-back diodes in the o/p of the filter. This is probably a good
thing to do with any rx connected to a souncard particularly if you
have an expensive soundcard. My soundcrd cost about $15 but I dont
want to blow it up just yet.

Its still early days and I think the filter may benefit from less gain
and perhaps even lower-q but I will continue with the bread board
tests and wait for some "real signals" to evaluate the filter with.

73,s

Des (M0AYF)

On 8/11/06, Paolo Saia <postacarlotta@...> wrote:
> What do you think about my 1/2 DCRX ?
>
>                                        73 all  Paolo IZ1KXQ

Hi Paolo

Yes, this is the kind of thing we had in mind: doubler then
divide-by-2. The best results would be obtained with the 74AC74 since
it is the highest speed of the common logic families.

I note that you have a small error in your diagram: pins 2 and 3 (D
and CLK) of the 7474 are swapped around. The 2xf signal from the 7404
should drive the CLK, and D should come from /Q (pin 6).

73 Hans G0UPL
http://www.hanssummers.com

Hi everyone,

    I have uploaded a couple of items you may find interesting relating
to crystal filters.

The first is located in the folder called "Crystal Banpass Filters"
This is chapter five of the "Radio Communication Handbook" first
printed in 1939 which contains one of the most detailed descriptions I
have seen describing the single crystal (or "crystal gate") filter and
the "lattice" using two crystals.

The second is located in the folder called "Crystal IMD" and covers
the subject of IMD in crystal filters.

I hope you find them interesting.

73,s

Des (M0AYF)

What do you think about my 1/2 DCRX ?

                                         73 all  Paolo IZ1KXQ


--- In qrsstxrx@yahoogroups.com, "ik1zyw" <pcravero@...> wrote:
>
> Hi from IK1ZYW.
> I've browsed through the web archives of past messages, and I
> understood right now you/we/they are concentrating on the TX design,
> plus XTAL filter.
>
> I've also understood the receiver will be a QSD, that is the reason
> for the need of a frequency doubler. Here comes my thought. Please
> acccept my apologies if this option has already been discarded.
>
> Since the QRSSTXRX will operate on a single frequency, there is no
> need to use a doubler in order to produce 4 quadrature driving
> signals: simple networks can be used to provide a 90 degrees shift.
>
> See these designs: http://www.yu1lm.qrpradio.com/homebrew.htm ,
> especially "Monoband SDR HF receiver DR2C-YU1LM.pdf"
> http://www.yu1lm.qrpradio.com/Monoband%20SDR%20HF%20receiver%20DR2C-
YU1LM.pdf
> which shows a very clean solution for a monoband XTAL-driven QSD.
>
> Please disregard this message if it is way too off-topic.
>
> Paolo IK1ZYW
>

Hi all I am a little in a hurry now .
I did something that seems working "fine" in my lab  !
I will insert the not complete schematic in the "files" .
I will be able to complete it , next week (I hope).

  Comments very appreciated.

                              73 all and have a nice weekend

                                                  Paolo IZ1KXQ

Hi Paolo

> I've also understood the receiver will be a QSD, that is the reason
> for the need of a frequency doubler. Here comes my thought. Please
> acccept my apologies if this option has already been discarded.

I think the currently favoured solution for the receive mixer is the
H-mode mixer adapted for baseband output by replacing the two output
transformers with a summing opamp arrangement. Not a QSD mixer. But
the H-mode mixer optimally requires a precise 50% duty cycle drive,
hence the need for the doubler and divider.

> Since the QRSSTXRX will operate on a single frequency, there is no
> need to use a doubler in order to produce 4 quadrature driving
> signals: simple networks can be used to provide a 90 degrees shift.

See above. It is not a quadrature LO drive which we want, but a
precise 50% duty cycle square wave to drive the H-mode mixer.

Your point still stands however: it is possible to use a "squarer"
circuit to obtain such a signal without a doubler and divider. We did
discuss this but I think we came to the conclusion that a square may
be susceptible to some jitter problems and such like, and that for the
real ultimate performance it would be better to use a doubler followed
by a divide-by-2.

> See these designs: http://www.yu1lm.qrpradio.com/homebrew.htm ,
> especially "Monoband SDR HF receiver DR2C-YU1LM.pdf"
> http://www.yu1lm.qrpradio.com/Monoband%20SDR%20HF%20receiver%20DR2C-YU1LM.pdf
> which shows a very clean solution for a monoband XTAL-driven QSD.

YU1LM has done some great work there. But the 90-degree phase shift
from an RC network cannot be perfect: component values will drift with
temperature leading to deviations from 90-degree shift, and I think
this arrangement would be particularly susceptible to noise causing
jitter. More so than a doubler/divide-by-2.

> Please disregard this message if it is way too off-topic.

Keep the suggestions coming!

73 Hans G0UPL
http://www.hanssummers.com

Hi from IK1ZYW.
I've browsed through the web archives of past messages, and I
understood right now you/we/they are concentrating on the TX design,
plus XTAL filter.

I've also understood the receiver will be a QSD, that is the reason
for the need of a frequency doubler. Here comes my thought. Please
acccept my apologies if this option has already been discarded.

Since the QRSSTXRX will operate on a single frequency, there is no
need to use a doubler in order to produce 4 quadrature driving
signals: simple networks can be used to provide a 90 degrees shift.

See these designs: http://www.yu1lm.qrpradio.com/homebrew.htm ,
especially "Monoband SDR HF receiver DR2C-YU1LM.pdf"
http://www.yu1lm.qrpradio.com/Monoband%20SDR%20HF%20receiver%20DR2C-YU1LM.pdf
which shows a very clean solution for a monoband XTAL-driven QSD.

Please disregard this message if it is way too off-topic.

Paolo IK1ZYW

Hi to everyone on the list,

Just to let you all know I have put the files relating to signal
frequency filters all in a single folder in the "files" area of the group.

They had been placed in the files area as individual pages but the new
folder is much more tidy.

73,s

Des (M0AYF)

--- In qrsstxrx@yahoogroups.com, "tigger" <aj-smith@...> wrote:

The AF offset is adjustable in "SpecLab", see ...

Menu SpecLab:
"View/Windows" --> select Digi Mode Terminal

and in the menu of the terminal:
--> select "Digimode Configuration"
--> center frequency TX
--> center frequency RX

I have set both, RX and TX center frequency to 800 Hz. The result is
shown in the file centerfreq.jpg posted to the file area. The chirped
HELL letters appear some Hz above the so called "center" frequency
(?) but when you select --> "Mode" --> "send unmodulated test tone at
TX frequency" in the terminal menu a line is displayed exactly on 800
Hz. Maybe the characters's bottom lines of the font are empty.

Chris, DL6JAN

>
>  If it is from say 990 - 1000Hz then the 120Hz low pass
> > filter will attenuate them. I have never used Multitone so I do
not
> > know. If the offset was 1000Hz then that would still work Ok with
the
> > narrow crystal filter but would require a VXO offset from the
crystal
> > passband by 1kHz.
> >
> Hans - you are right - I went out to do some shopping, and this
point
> hit me as I entered the supermarket, and I hoped no-one would spot
it
> before I got back !  I have removed the filter, and might add
another
> one later, but I actually don't think this is the right place for
it.
>  On reflection I think that audio filtering should be done at the
> source.  I might put the buffer back, since the input resistance of
> the 602 is a bit low - 1.5k.  The circuit looks a lot simpler now.
>
>
> Andy
>

--- In qrsstxrx@yahoogroups.com, Paolo Saia <postacarlotta@...> wrote:
>
> Hi Andy
>   I am testing your 10140 MHZ LO + Doubler.
>   In the schematic there is a drawing error , I think .
>   The base of the first 2SC945 is  not connected to the bias resistors.
>   I am using the 2N2222 instead of your  , because available at the
moment.
>   Now  it is not oscillating , what could I do ?
>
>
                      73 Paolo I1DFS

Hi Paolo - thanks for spotting this error.  I have uploaded a
corrected file.

I suspect that the difficulty you are having making the circuit work
is to do with the core of the first transformer.  I have used 2
ferrite beads binocular-fashion.  I have uploaded a picture
(doublertransformer.gif) of the actual circuit I built, so that you
can see what the transformer looks like.

Also in the picture you can see the core on the right which is part of
the 20MHz tank circut.  The 10MHz transformer is on the left.

The core for the 20MHz tank is similar to ones that I have been using
throughout my experiments for xtal filter transformers, etc.

Thank you for testing this circuit.  Please let us know how you get
on.  It is important that the circuits should be reproducible,and I
need to know if they are not.

73
Andy

I have posted the tx file as .jpg format to make it more accessible.

Andy

I have posted the tx file as .jpg format to make it more accessible.

Andy

If it is from say 990 - 1000Hz then the 120Hz low pass
> filter will attenuate them. I have never used Multitone so I do not
> know. If the offset was 1000Hz then that would still work Ok with the
> narrow crystal filter but would require a VXO offset from the crystal
> passband by 1kHz.
>
Hans - you are right - I went out to do some shopping, and this point
hit me as I entered the supermarket, and I hoped no-one would spot it
before I got back !  I have removed the filter, and might add another
one later, but I actually don't think this is the right place for it.
  On reflection I think that audio filtering should be done at the
source.  I might put the buffer back, since the input resistance of
the 602 is a bit low - 1.5k.  The circuit looks a lot simpler now.


Andy

Andy

About the bandwidth: Assuming multitone transmissions are originated
from PC software via the PC soundcard, are we 100% certain that they
come out at baseband? I mean, when Johan ON5EX sends his Mickey Mouse
pictures with say 10Hz bandwidth, at what actual audio frequency are
they generated? If it is from say 990 - 1000Hz then the 120Hz low pass
filter will attenuate them. I have never used Multitone so I do not
know. If the offset was 1000Hz then that would still work Ok with the
narrow crystal filter but would require a VXO offset from the crystal
passband by 1kHz.

73 Hans G0UPL
http://www.hanssummers.com

On 8/8/06, tigger <aj-smith@...> wrote:
> I sat down and did some serious thinking this morning, calculator in
> hand.  The result is a circuit diagram - qrsstx.doc.  Some component
> values are missing at present.
>
> I am fairly confident about this design, but anything could change.
> The PA and driver is from VK3XU's 'TCF' transceiver, the rest is my
> own. The design power is 1W, reducible.
>
> I have added a 2nd order 0.2dB ripple 120Hz Chebychef af lpf, which
> might be an unnecessary refinement. I am not certain how ssb qrss
> operators control their bandwidth - there is always the xtal filter of
> course.
>
> I intend to test it initially with the dds as VFO, and make the VXO later.
>
> This is a proposal for discussion - so comments are welcome.
>
> Andy
>
>
>
>
>
>
>
>
>
>
>
> Yahoo! Groups Links
>
>
>
>
>
>
>
>

I sat down and did some serious thinking this morning, calculator in
hand.  The result is a circuit diagram - qrsstx.doc.  Some component
values are missing at present.

I am fairly confident about this design, but anything could change.
The PA and driver is from VK3XU's 'TCF' transceiver, the rest is my
own. The design power is 1W, reducible.

I have added a 2nd order 0.2dB ripple 120Hz Chebychef af lpf, which
might be an unnecessary refinement. I am not certain how ssb qrss
operators control their bandwidth - there is always the xtal filter of
course.

I intend to test it initially with the dds as VFO, and make the VXO later.

This is a proposal for discussion - so comments are welcome.

Andy

Hi Folks,

    The IRF510 gets my vote. Re Power level, 500 mW should be ample but
! Watt would be nice to have as reserve for "tiny antennas".

73,s

Des. (M0AYF)

--- In qrsstxrx@yahoogroups.com, "tigger" <aj-smith@...> wrote:
>
> Woud anyone like to suggest a maximum power output for the TX ?  1W?
>
> I suggest that the PA device should eb the IRF 510.  Any comments?
>
> Andy
>

About crystal oscillators: David VK6DI has constructed Butler
oscillators and found them to have extremely impressive stability.
They are also supposed to be the best you do in terms of low phase
noise. This is probably less of an issue for the transmit side than
receive. For the receiver I would suggest that a Butler oscillator
would be pretty optimum.

73 Hans
http://www.hanssummers.com



On 8/7/06, tigger <aj-smith@...> wrote:
> Thanks Hans - I'm glad I am not the only one who has this problem.
> And ok too on the cw/ssb distinction.  I think some of my class C
> 'amps' have actually been phase-locked oscillators !
>
> Anyway, I have been pushing ahead.
>
> I have made a PA lpf which has the right characteristic: it is a 5th
> order chebycheff (2 inductors). These filters have greater slope in
> the stop band than constant-k or butterworth types, and the increased
> in-band ripple is of no consequence given the bandwidth we are using.
>
> I have also started putting together the filter interface between the
> 602 and the 1st driver.  And it is going well in the test circuit.  I
> simply have a primary to the input transformer (see my previous
> posting re the half-lattice) fed differentuialy from the 602's
> outputs.  I tried using the 602 differential outputs directly to the
> filter, but it messed up the fiilter characteristics.
>
> I now have to link the output of the filter to the pre-driver.  I
> really need a high impedance point here (hence the matching
> transformer at the output of the filter).  Would anyone object if I
> replaced Drew's 2N2221 with a 3sk45 2-gate mosfet - I could ditch the
> second transformer then...?  The 3sk45 gives bags of power gain, and
> has a dissipation of 350mW, so make lovely small PAs (and very little
> feedback capacitance, so more likely to be stable).
>
> Does anyone have thoughts re the xtal oscillator for the TX ?
> It could be the 602, but it needs ovening, so this would mean ovening
> the entire mixer/oscillator sybsystem.
>
> We could use a separate Colpitts osc....
>
> Also...  The oscillator has multifunctions.  If the same osc is used
> for tx and rx it must do all of this-
> 1) It must be shifted a kHz or so for tx/rx switching,
> 2) It must be tuneable for tx
> 3) It must be modulated for fm/fsk
> 4) it must be trimmed to place the rx lo in the right freq at the
> right temperature.
>
> I am tempted to simplify this by using a separate unovened osc for the
> rx ( see my file posting re the doubler), and design an op-amp summer
> system with separate inputs for the various remaining functions, the
> output of the op-amp feeding a varicap.
>
> Suggestions, comments ?
>
> Andy
>
>
>
>
>
>
>
> Yahoo! Groups Links
>
>
>
>
>
>
>