Wiring for distribution of a 108MHz sine wave

Colleagues,

How could one distribute a 108MHz sine wave (generated by a PLL) across several boards? The are 4x identical recipient boards. These boards live in one a narrow tube 20mm across, and they are spaced 0.5m apart (total length 2m).

If I could use coaxial cables and connectors, then this would be a trivial question. Unfortunately, I can?t use coax because of cost and ease of installation (this is for developing countries). I can do with some signal loss (I think, but don't know how much loss).

Any suggestion, insight or reference is really appreciated!

Cheers,
- Nick

You can e.g. use an unshielded twisted pair.

Cheap suggests an off-the-shelf solution. Ethernet and USB cables come to mind because both are twisted pairs and both give you options in terms of shielding and general cable quality versus cost.

I would get a differential digital cmos gate, and generate a differential 108 MHz signal on the driver board, and then share it differentially with the other boards. Since the differential receiver/gate will have a relatively high impedance, you can load the line at each gate with a resistor placed differentially. OR you can just use the gates for the shorter paths, and a resistive load at the end of the longest path.

Should add maybe 30 cents to each board to do it that way.

IF you have ground loop issues, you can capacitively couple the signal with two series caps at each new receiver board (you would use a resistive bias method to DC bias the gate's +/- inputs at the halfway point).

I agree that frequency is easier to distribute as a digital signal. But ultimately, I need to have a sine wave. What method can you recommend for getting a sine wave from the digital signal?

Low-pass filtering is the first thing that comes to mind, of course. But my frequency will not always be 108MHz, in some situations it will be 8MHz ¹. An LPF for 108MHz will pass too many harmonics of 8MZh. Are there other methods for getting a sine wave from a digital signal?

¹ Apologies for not mentioning the 8MHz until now: I was too focused on the upper bound of my frequency range.

Employ LVDS drivers and receivers over unshielded twisted pair

well, coax would be the simplest way, since no harmonic filtering will be needed to return it to being a mostly sine wave.

You could still do things differentially, i.e. an ANALOG differential output amplifier, driving shielded twisted pairs. At the receiving end, an ANALOG differential input op amp.

But now you are starting to approach the cost of a coax cable. I mean, what does a coax cable with RCA connectors on it cost, maybe 50 cents in production quantities?

you say the boards are "in a tube". you could try something out of the box, like linearly modulating an LED with the sine wave, then using photodiodes on each board to receive the modulated light. It would not be super linear: i.e. would generate harmonics, but not as many as a true digital signal would. if the light will not propagate well inside the tube, you could use a plastic light bar....i.e. a piece of clear solid plastic tube inside of the bigger cylinder....distributing the light, a flexible light pipe.

Indeed, I agree ....
I don't see why the OP has a problem with coax, I wouldn't use anything else .... there are a number of miniature coax cables that are available
that are under 3mm in diameter


Dave