EMRFD Message Archive 6738

Message Date From Subject
6738 2011-10-15 08:48:09 Tim 10M VFO ideas
10M is wide wide open this morning round the world. Time to homebrew a QRP 10M transmitter.

For the finals I have some RRDHF16's. One will do 15W on CW pretty easy I figure which will put me way above most QRP definitions :-)

But the VFO... any simple ideas for a sweet and stable 10M VFO?

Quadrature outputs for use with a R2Pro would be a real plus.

I'm thinking a 112 MHz PLL, divided by 4 with 74AC74 with very simple LC low pass to make quadrature 28MHz for the R2Pro.

PLL would track a 4.0-4.4 MHz LC VFO going into one input of a phase comparator. To do that I could use a 28 MHz divide by 4 output from the PLL, mixed with a 24 MHz crystal in a NE602, then low-pass to yield the other 4.0-4.4 MHz input to the phase comparator.

Any other ideas? Clever VXO mixing schemes using microprocessor crystals?

Tim.
6739 2011-10-15 09:31:45 Ed - K9EW Re: 10M VFO ideas
How about a DDS-type VFO? Several good kits available, some in QST also.

Or... a simple 5MHz VFO and a mixer/filter.

Keep us informed of your progress - I'm not aware of any 10m monoband kits
available... yet.

73,
ed - k9ew



6740 2011-10-15 12:18:19 Dave Re: 10M VFO ideas
An Si570 or a Si598. Frequency is prorammed by an I2C interface, so
you need a microprocessor to select frequency. Both are available from
Silicon Labs.

Dave - WB6DHW
<http://wb6dhw.com>

6741 2011-10-15 16:16:50 Tim Re: 10M VFO ideas
VCO efforts for a 112 MHz PLL were interesting. I got something working, a Hartley with a J310, but the fractional range is way too broad using my first configuration of back to back varactors and only through some manual influence do I get the LF part of the PLL loop to come into lock. Ideally the tuning range will not be tens of MHz but only 0.5 MHz plus some slop.

Spent some of the afternoon trying to make a 112MHz VCO with the center frequency set with a resonant length of coax. Tried a J310 with transformer coupling to 50 ohm coax on the output and a L match on the far end, and it oscillated for sure, but not
6742 2011-10-15 17:28:44 Chris Trask Re: 10M VFO ideas
>
> 10M is wide wide open this morning round the world. Time to homebrew a
> QRP 10M transmitter.
>
> For the finals I have some RRDHF16's. One will do 15W on CW pretty easy
> I figure which will put me way above most QRP definitions :-)
>
> But the VFO... any simple ideas for a sweet and stable 10M VFO?
>
> Quadrature outputs for use with a R2Pro would be a real plus.
>
> I'm thinking a 112 MHz PLL, divided by 4 with 74AC74 with very simple LC
> low pass to make quadrature 28MHz for the R2Pro.
>
> PLL would track a 4.0-4.4 MHz LC VFO going into one input of a phase
> comparator. To do that I could use a 28 MHz divide by 4 output from the
> PLL, mixed with a 24 MHz crystal in a NE602, then low-pass to yield the
> other 4.0-4.4 MHz input to the phase comparator.
>
> Any other ideas? Clever VXO mixing schemes using microprocessor crystals?
>

What you've described here is workable, though I ould be inclined to
simplify it more by dividing the 28MHz be 8 and then using a stable 80M VFO
for the phase comparator inputs.

Chris
6743 2011-10-15 18:40:01 Tim Re: 10M VFO ideas
6744 2011-10-16 08:11:11 Jason Burke Re: 10M VFO ideas
Hi Tim,


I've been playing around with some circuits that may be of interest...


First, I created a VCO based on the schematic in Fig. 4.13(b) using a pair of J310s which worked very well to 200MHz and beyond.  Before this I tried every oscillator circuit I could find a schematic for, but in the end I found the negative resistance configuration with JFETs to offer the best performance in terms of giving me a clean signal (harmonics all >35 dB down on my spectrum analyzer with ugly bug construction) and reliable performance across a large frequency range.


Next, I got it working with some old 145170 PLLs I had laying around and an AVR Tiny85 to control the 145170 (though any uC would do).  A couple of notes:  My reference frequency was 10kHz, although I think it would still work fine at 5kHz or lower.  The loop filter cutoff was around 200Hz (could go higher for CW) and it locked up reasonably quickly (I didn't measure the lock time, but it can't be more than 100-200mS and is probably quite a bit faster).  The VCO was configured to have a narrow tuning range with the coarse tuning being done with NP0 capacitors and the fine tuning done by a pair of varactors.  As I recall the KVCO was about 100kHz/volt and with a 1-5 volt tuning range that gave me only 400kHz to work with on 15M.  This could easily be increased if needed (and it would be for divide-by-4 operation at on 10M).


For finer tuning steps than about 5kHz, I was thinking it should be possible to use a cheap DDS for the reference frequency of the PLL, something like the AD9831.  However, I'm not sure how badly this would compromise phase performance of the VCO, although I think that this is what some commercial transceivers do to provide 1Hz tuning resolution.  Alternatively, you could make a VXO circuit for the PLL reference oscillator and get some "fine tuning" range of maybe +/- 5kHz at 28Mhz with a suitable crystal, though this would compromise any digital frequency display based on the programming of the PLL.


I don't know about the R2Pro, but a DC receiver based on a sampling quadrature mixer is appealing as it is a tried and proven design which eliminates the primary disadvantage of a regular DC receiver while retaining most of the advantages.  If you don't mind working with SMT components (I don't so long as I have a PCB), then there are a number of logic families which are fast enough to handle the switching at 112MHz, though you can forget about any of the 74xC parts in this role.


I was a little unclear on what Chris meant by dividing the 28MHz by 8, but maybe he is talking about a subsampling approach?  If you built a 112MHz VCO, divided it by 32 (4 to get I/Q signals at 28MHz, then 8 to subsample at 3.5MHz), you could use the output to build a subsampling quadrature mixer for 10M with a bandpass filter ahead of it.  I haven't seen this done before, but it seems like it should work.  With this approach, you may even be able to get away with 74HC parts for the sampling quadrature mixer, which would make for easy ugly-bug construction.


The Si570 is a popular chip for building VFOs, but despite its awesome phase noise performance, it falls short of achieving the "holy grail" of a wideband VFO because it has a non-sinusoidal output requiring a separate set of filtering for every band you want to use it on.  I like the AD9851 solutions out there (e.g. DDS-60) since they provide sinusoidal output across the entire band and require only a single low pass filter around 1/3 of the clock frequency.  If you operate at a low fraction of the clock frequency, the spurious responses are minimal.  Both of these solutions push the cost up quite a bit since the ICs alone cost over $20 each and are only available in SMT packages which are quite difficult to use with ugly bug construction though I'm sure someone has done it.


Anyway, best of luck with your project and let us know how it progresses.


--
Jason

[Non-text portions of this message have been removed]
6746 2011-10-17 07:17:47 Harold Smith Re: 10M VFO ideas
Might I suggest a 'super VXO' and, if necessary, frequency multipliers.
They're not nearly so high tech, but they're fairly simple and about crystal
stable. The downside is that they usually take some tweaking to optimize.

See: http://www.qsl.net/7n3wvm/supervxo.html
http://py2wm.qsl.br/VXO/VXO-e.html
http://www.amqrp.org/projects/NB6M%20contestvxos/VXO%20Experiments.html

de KE6TI, Harold