Quadrature phase out of PLL with extended PLL range? Idea

[neazoi]

Hello I have an idea which I would like to discuss with you.

I am thinking of a way to produce two accurate 90 degrees out of phase signals, out of a pll, for the radio amateur bands up to 30MHz. I need this, to drive a set of quadrature mixers. I do not want to use a dds since it is hard to solder and program.
The range of my pll is 8KHz to 16MHZ in 1KHz steps.

Dividing down the 16MHz with a 74F74 will give a maximum of 8MHz in two phases or a maximum of 4MHz if I wanted four phases (for other purposes).

The Idea I have to extend this range is this one:

Do not include a low pass filter after the pll. Instead include a high pass filter that will be able to pass the second harmonic of the fundamental frequency. Then apply this second harmonic signal to a 74f74 and divide it by 2. The result will be two signals 90 degrees out of phase at the fundamental frequency.

Is my consideration valid? What do you think?

 

[FvM]

A symmetrical signal has only odd harmonics. You need to use a frequency doubler circuit before you can generate usable amounts of second harmonic.

 

[torbai]

Your idea is very interesting.Before viewing yours,my opinion was that using a PLL which generate the double time frequency directly.FvM is right.There has only odd harmonics.You should either use a doubler or a double time frequency PLL.

 

[neazoi]

The problem is that PLLs do not have so big range, that is why I was thinking of the doubler. To generate a 2-phase signal from 1-30MHz you need a 2-60MHz PLL and to get 4-phases you need a 4-120MHZ, which does not exist.

by the way, can the doubler be used as a quadrupler, or do I need to connect 2 doublers in series?

A doubler can be easily made from a pair or quad of diodes and a transformer, or an mc1496 balanced mixer.. It helps in filtering as well, because the strongest signal will be the doubled one and there will not be any fundamental one, I think?

Considering frequency stability, which after doubling will be worse than the fundamental, this stability will be "recovered" as the doubled signal will be divided by 2 again.

Also if incrementing the fundamental say by 1khz, the doubled frequency will be incremented by 2khz. But when divided down by 2, you recover the same increment (1khz).

Considering the setting of the pll again, there is nothing to be changed, as the set frequency will theoretically appear after the doubler and the divider as well, but in dual phase now.

My considerations are based on the fact that the 74f74 and other such IC dividers can maintain good phase throughout the whole frequency range, within their limits.

If you have anything else to propose I will be glad to hear it before I try testing the setup.

 

[chuckey]

I was thinking of using a high quality PLL (PLL1) on one oscillator. Now you want a second oscillator(PPL2), feed this with the DC from the first PLL1, this is to get it into the right range (8KHz- 16MHz??). Now feed PLL1 into a quad diode product detector along with PLL2. Repeat using -PLL1 and PLL2. Subtract the outputs of the two quads detectors, this will give you a DC signal corresponding to the phase error of PPl and PLL2 from 90 degrees - if they are at 90 degrees the output will be zero. Hopefully this output will be enough to finally pull PLL2 to frequency AND set its phase to 90 degrees.
Frank

 

[neazoi]

It seems to me a bit complicated, like the dds solution, but I am not sure

 

[torbai]

An ADF4360-9 can generate a signal covering 1.1MHz-200MHz.However,it can be also hard to program.

 

[torbai]

Oh,it reminds me that ADI has a document which uses two AD9850s to generate a quadrature phase signal.

 

[neazoi]

interesting, instead of having a phase network to have two dds programmed differently by the same micro. But still too complicated to my taste...

 

[BigBoss]

The best way is to use "Polyphase Filter" to obtain Quadrature signal pairs.
With PF, you will get +90,-90,+180,-180 combination.

 

[neazoi]

The best way is to use "Polyphase Filter" to obtain Quadrature signal pairs.
With PF, you will get +90,-90,+180,-180 combination.

You mean these capacitor/resistor filters?
They are not able to maintain phase accuracy over a wide range. They are acceptable for 3khz audio though