Hi Folks,

Does anyone know a simple way (using digital logic) that has an input a periodic clock and the outputs are:

(1) a 180 degree phase shifted clock

(2) a 0 degree phase shifted clock so that it is in phase with (1).

Thanks!

yeah it's called an inverter...

If the phase accuracy isn't critical one inverter and you are done. If the problem is more demanding than that, you need to specify a bit more what you want.

I know this is an old post. Iam currently trying to generate a mult phase clock. Trying to us a PI (two inverters with output shorted).Any help on how to go about trimming this?My clock frequency is 2G.Currently able to reach 5ps within target across process and 5ps across monte.Any help will be appreciated.

loose-electron wrote on Dec 22^nd, 2009, 9:49pm:

Ha, that is what I was thinking. In addition, to line up the non-inverted signal run it through an "on" t-gate with the same sized devices as the inverter. See figure 12.18 of the switched cap chapter of Razavi's book.

When I first saw it I figured it was an horrible hack, but it actually works well to align the clock and clock-bar signals. Not sure if it will get you 5 ps, but it is a trick worth remembering to get you close. I suppose you could trim the tgate delay by trimming the gate drive. I don't understand the PI cricuit described by analog_rf.

I use cross coupled differential circuit, you can invert the input  and delay match it with a TGate to drive the two diffpair inputs, then buffer out the output. Parasitic matching in important in your layout, also fast rise and fall time with weak load devices in your diff.

Duty Cycle might get F-ed up tho. but you will be almost 100% phase shifted at 0 and 180 with the two outputs.