Circuit for the generation of UWB signal
(UWB = Ultra Wideband)
We are two students working on a project, with a UWB
Signal to be generated.
We want to use a classic circuit for it.
See the appendix (Picture 1)
But we do not have typical values of the components [R, C, SRD (Step
Recovery Diode) and Schottky Diode]
Picture 2: The output should come out one Monocycle Pulse
Frage1: What are the values of the components [Rs, RL, RB, CD, CF] are taken
be?
Frage2: What properties should SRD and Schottky diodes have?
What is the value of Trr = (Reverse Recovery Time) =?
should we use?
Thanks in advance
Regarding Pic.1
Rs is simply 50 Ohm line impedance - for example 1.15mm stripe width on Rogers 4003 substrate (with 0.5mm thickness and 17um copper coating).
RL (L stands for "load") hast to be the same as Rs, and because almost all measurement equipment is standardized to 50 Ohm You should use this one as well.
You can use SRD from Aeroflex-Metellics- for example this one:
http://www.aeroflex.com/AMS/Metelics...D-beamlead.cfm
go for the one with the shortest transition time and smallest parasitic capacitance/inductance.
They have a European sales office in Germany (near Munich) - ask there for samples, but it will take some time (I went throu this once).
Forget the Schottky - all which I found on the market are to slow to cancel the negative voltage swing of the generated pilse (Trr is to big). Despite of this you want to get the monocycle, so you don't need to cancel out the negative voltages.
If you need a monocycle, you have to use the shorted stub (short by via) - in theory one stub would give you only gaussin pulse, but in reality it will be already monocycle :)
Lokk in the web for other architectures, because you have chosen the hardest one for this task. If you'll not find anything else (but it's not probable) I can search for it in my old projects.
I hope it helped.
Regards!
Added after 27 minutes:
Sorry - one correction - the reason why the Schottky will not help in this case is not the switching time (which is in order of 100ps) but the parasitic capatitance of the package. Like this the capacitance will cancel the DC and you will get a small (or not - depending on C value) negative voltage swing after main pulse.
Sorry once more for the lack of precision i the previous post :)