vco tuning voltage
I have questions about tuning voltage line of a VCO in a discrete PLL. How long it can be? Is there any consideration that must be in mind in designing this line?
In theory the line adds delay which turns up in your feedback loop. And if you remember right from your control theory introductory course, delay is very hard to be controlled.
Now back to reality: in practice your loop bandwidth of the PLL is in the order of magnitude of a few kHz. The delay (even if several centimeters) will only add a fraction to the open loop phase response. I wouldn't worry to much about that.
Another thing is the interference pickup. The longer the line, the higher the chances for picking up any disturbing signal from a coupled loop. This is important.
Long lines introduces an additional Resistor, capacitor and Inductance. As consequence this will change your PLL behaviour. You have to take into account these additional parasitics when you design the filter.
Uh, delay? If your tuning line was 1 km long, that would give you only 3 nS delay--in other words it is not a problem. You get 100 to 10,000 time more group transport delay in the digital divider used in the PLL.
The problem is really twofold:
1) noise pickup. As the VCO tune input is a high impedance point, you will pick up all sorts of electrical noise along the length of that line. Also, since the ground of your VCO is now attached to the ground of your PLL filter board by a long length of line, you can easily develop many millivolts of ground loop noise. On some systems I have seen that were rack mounted, I could measure 1 V of ground loop noise going from one rack mounted enclosure to another, for instance.
2) The input to the VCO tune likely looks like a purely capacitive load. If you try to terminate a long line with a reactance, you can get odd multiple reflections. Also, your op amp may go unstable with certain reflected load impedances.
Since there isn't any current , the length can be long.But if you use 3rd order PLL Loop Filter, last R-C components should be connected as close as possible to the Vtune input of the VCO.There won't be any delay or loss because the frequency in the loop filter is only reference frequency that is max few MHz.
To circumvent any noise pickup, you need to ensure that no noisy signals are close to the VCO tune trace and that it has a well defined return path, preferably uninterrupted ground planes above/beneath the trace (packed). Also guard traces can help here, but be careful where to pick the ground connection point for this (close to the VCO tune input pin and nowhere else).
anybody has a schematic diagram of VCO?