Message	Date	From	Subject
2311	2008-10-24 15:49:17	timshoppa	Multiband LO chain final incarnation
I spent some time yesterday melting solder and have had made good progress on my multiband LO chain. It's evolved a lot over the past month. Receiver and TX have IF at 4.433 MHz. (Actually TX BFO has a knob for rubbering transmit offset on CW - to me XIT makes much more sense than RIT, as long as I don't go out of the band that is!). I use a nice Millen old dial drive/pointer on a Hartley VFO with output from 4.933 to 5.033 MHz. The 0.1MHz is spread across about 5 inches of dial face, which is great. The multiband scheme is inspired by G3ROO's schematic in recent ARRL handbooks. Each of the bands (currently 80M, 40M, 30M, 20M) has its own narrow- range VCO. Each of the bands has a crystal oscillator. They are activated by DC on a control line from the bandswitch, and diodes make sure that only one has its signal coming out. The VCO's are similar to those in the G3ROO schematic, but I have further restricted each tuning range by making the fixed capacitance be a bigger part of the tank capacitance. And I use some SMD varactors because through-hole varactors seem to have completely disappeared from Mouser/Digikey catalogs. Everything is dead bug, with the exception of the SMD varactors which I carve up pads for. The VCO is mixed with crystal oscillator with a NE602 followed by a low-pass. So frequencies work out as: 80M band: 3.5-3.6 MHz 80M VCO: 7.933-8.033 MHz 80M Xtal: 3.0 MHz 80M NE602 out: 4.933-5.033 MHz 40M band: 7.0-7.1 MHz 40M VCO: 11.433-11.533 MHz 40M Xtal: 6.5 MHz 40M NE602 out: 4.933-5.033 MHz 30M band: 10.1-10.2 MHz 30M VCO: 14.533-14.633 MHz 30M Xtal: 9.6 MHz 30M NE602 out: 4.933-5.033 MHz 20M band: 14.0-14.1 MHz 20M VCO: 18.433-18.533 MHz 20M Xtal: 13.5 MHz 20M NE602 out: 4.933-5.033 MHz I have not yet built any other bands, but there's positions for 4 more bands on my dead-bug board. What's nifty about this are that the crystals for 80M, 40M, 30M, and 20M are stock items out of the Digikey catalog, the tuning direction is the same on all the bands, and the VFO is synchronized with lower band edge in every case. For 15M I'd need a 20.5MHz crystal, and while this isn't in the catalogs there are 20.48MHz crystals. For 10M I'd need a 27.5MHz crystal, and while that isn't in the catalogs there is a 27.4688MHz. Where we are in the solar cycle, I'm not too worried about 15M and 10M (although on Field Day I worked a lot of stations in the mid-Atlantic and upper midwest on 10M CW). For 160M I'd need a 1.3MHz frequency, and even though I don't see that in the catalogs I could divide 6.5MHz by 5 or 13MHz by 10 and get there. The NE602's mixer output goes to a 74HCT9046 frequency/phase detector input. The Hartley 4.933-5.033 MHz VFO goes to the other input. The output of the phase detector drives the low-pass loop filter, which I copied verbatim from the ARRL handbook, which sets the tuning voltage to the VCO's. I did abandon the buffers suggested in the ARRL handbook. I tried them and they just don't work. W7EL-style feedback pairs are what I use instead. The end result is pretty nifty. The VCO outputs work just fine in my receiver. The whole "synthesizer box" with room for 8 crystal oscillators, 8 VCO oscillators, and buffers and phase detector loop filter is a box with a jack for VFO in and a jack for synthesizer out, and some DC lines for selecting band. So I think this is pretty nifty, especially for no microprocessors. The K2's LO scheme is clever too, but it'd take a microprocessor. Not that I have a whole lot against them, just trying to do without in the interest of simplicity. But I look at what's in the K2 and wow, is that straightforward. They don't have a VCO per band, they're far more clever with their little relays and switching varactors/capacitors in and out! Tim N3QE
2321	2008-10-26 12:13:59	richj_focus	Re: Multiband LO chain final incarnation
Very nice work. Can i get schematic of this vfo project? I was wondering, not having a ticket to operate, is 100khz per band too limiting? Thanks for the education. rich
2326	2008-10-27 09:57:42	timshoppa	Re: Multiband LO chain final incarnation

2327	2008-10-27 10:40:46	Wes Hayward	Re: Multiband LO chain final incarnation
Hi Tim and group, Re the LO system that you built: I looked in both of my current copies of the ARRL Handbook, which are 2005 and 2008, and in both cases, Figure 10.56 is a "G3ROO/GM4ZNX summing loop PLL." It is a so called one-on-one loop. That is, one voltage controlled oscillator is locked to another oscillator on a 1 to 1 basis. The phase noise of such a system can be quite good if they are done with care. We treated 1 on 1 loops a little in EMRFD. See page 4.22. One of the readers ask for a schematic. The next step in this progression is to determine the phase noise and spurious output content of these designs. Unfortunately, the 1 on 1 loop example in EMRFD, figures 4.43, 4.44, and 4.45, plus the VFO, are no longer in place as a system. The modules are around somewhere. If I can find them, I can build up the system again and do a phase noise measurement using the crystal notch filter methods that were presented earlier this year. That would be a useful thing. Good luck with the experiments. 73, Wes w7zoi

I spent some time yesterday melting solder and have had made good
progress on my multiband LO chain. It's evolved a lot over the past
month.

Receiver and TX have IF at 4.433 MHz. (Actually TX BFO has a knob
for rubbering transmit offset on CW - to me XIT makes much more
sense than RIT, as long as I don't go out of the band that is!).

I use a nice Millen old dial drive/pointer on a Hartley VFO with
output from 4.933 to 5.033 MHz. The 0.1MHz is spread across about 5
inches of dial face, which is great.

The multiband scheme is inspired by G3ROO's schematic in recent ARRL
handbooks.

Each of the bands (currently 80M, 40M, 30M, 20M) has its own narrow-
range VCO. Each of the bands has a crystal oscillator. They are
activated by DC on a control line from the bandswitch, and diodes
make sure that only one has its signal coming out.

The VCO's are similar to those in the G3ROO schematic, but I have
further restricted each tuning range by making the fixed capacitance
be a bigger part of the tank capacitance. And I use some SMD
varactors because through-hole varactors seem to have completely
disappeared from Mouser/Digikey catalogs. Everything is dead bug,
with the exception of the SMD varactors which I carve up pads for.

The VCO is mixed with crystal oscillator with a NE602 followed by a
low-pass.

So frequencies work out as:

80M band: 3.5-3.6 MHz
80M VCO: 7.933-8.033 MHz
80M Xtal: 3.0 MHz
80M NE602 out: 4.933-5.033 MHz

40M band: 7.0-7.1 MHz
40M VCO: 11.433-11.533 MHz
40M Xtal: 6.5 MHz
40M NE602 out: 4.933-5.033 MHz

30M band: 10.1-10.2 MHz
30M VCO: 14.533-14.633 MHz
30M Xtal: 9.6 MHz
30M NE602 out: 4.933-5.033 MHz

20M band: 14.0-14.1 MHz
20M VCO: 18.433-18.533 MHz
20M Xtal: 13.5 MHz
20M NE602 out: 4.933-5.033 MHz

I have not yet built any other bands, but there's positions for 4
more bands on my dead-bug board.

What's nifty about this are that the crystals for 80M, 40M, 30M, and
20M are stock items out of the Digikey catalog, the tuning direction
is the same on all the bands, and the VFO is synchronized with lower
band edge in every case. For 15M I'd need a 20.5MHz crystal, and
while this isn't in the catalogs there are 20.48MHz crystals. For
10M I'd need a 27.5MHz crystal, and while that isn't in the catalogs
there is a 27.4688MHz. Where we are in the solar cycle, I'm not too
worried about 15M and 10M (although on Field Day I worked a lot of
stations in the mid-Atlantic and upper midwest on 10M CW). For 160M
I'd need a 1.3MHz frequency, and even though I don't see that in the
catalogs I could divide 6.5MHz by 5 or 13MHz by 10 and get there.

The NE602's mixer output goes to a 74HCT9046 frequency/phase
detector input. The Hartley 4.933-5.033 MHz VFO goes to the other
input. The output of the phase detector drives the low-pass loop
filter, which I copied verbatim from the ARRL handbook, which sets
the tuning voltage to the VCO's.

I did abandon the buffers suggested in the ARRL handbook. I tried
them and they just don't work. W7EL-style feedback pairs are what I
use instead.

The end result is pretty nifty. The VCO outputs work just fine in my
receiver. The whole "synthesizer box" with room for 8 crystal
oscillators, 8 VCO oscillators, and buffers and phase detector loop
filter is a box with a jack for VFO in and a jack for synthesizer
out, and some DC lines for selecting band.

So I think this is pretty nifty, especially for no microprocessors.

The K2's LO scheme is clever too, but it'd take a microprocessor.
Not that I have a whole lot against them, just trying to do without
in the interest of simplicity. But I look at what's in the K2 and
wow, is that straightforward. They don't have a VCO per band,
they're far more clever with their little relays and switching
varactors/capacitors in and out!

Tim N3QE