I have this crystal oscillator driving the clock input of a DSP (12MHz, 3,3v): http://www.cmac.com/mt/12.8MHz IQXO-71(I)B.pdf
Problem is, that the rise/fall times for this XO isn't fast enough. The DSP requires 3ns! Therefore I need a 12MHz square clock signal with rise/fall times better than 3ns. As I see it, there are 2 soloutions to meet the DSPs timing requirements:
1. Find a XO with the similar footprint and rise/fall times of 3ns.
2. Glue a 1:4 clock distribution IC (belly upwards) or some other logic with fast output logic and then connect its input to the XO that I already have and its output to the DSP. All this will be beatifully connected with wirewrap wire (arghhh).
I haven't been able to get another XO with better specs so I'm leaning towards 2. But which logic part would you recommend and do you foresee any problems with the wire connections (noise) for a 12 Mhz clock?
Kasper
Problem is, that the rise/fall times for this XO isn't fast enough. The DSP requires 3ns! Therefore I need a 12MHz square clock signal with rise/fall times better than 3ns. As I see it, there are 2 soloutions to meet the DSPs timing requirements:
1. Find a XO with the similar footprint and rise/fall times of 3ns.
2. Glue a 1:4 clock distribution IC (belly upwards) or some other logic with fast output logic and then connect its input to the XO that I already have and its output to the DSP. All this will be beatifully connected with wirewrap wire (arghhh).
I haven't been able to get another XO with better specs so I'm leaning towards 2. But which logic part would you recommend and do you foresee any problems with the wire connections (noise) for a 12 Mhz clock?
Kasper
3 nS, eh?
That means you need something with a -3 dB point over 115 MHz. Or to put it another way.....you will have to pass at least the 7th harmonic of your clock.
(Tr = 0.35/F)
Jocko
That means you need something with a -3 dB point over 115 MHz. Or to put it another way.....you will have to pass at least the 7th harmonic of your clock.
(Tr = 0.35/F)
Jocko
A good choice for decreasing square wave rise time is a Schmitt trigger. Take a look at the 74F14 from your favorite IC manufacturer.
Thank you for your replies
I'm going to try the CY2305SC-1H from cypress (http://www.cypress.com/cfuploads/img/products/38-07140.pdf).
It has rise/fall times below 1,5ns (0,8-2,0v) which should be just fine.
F-logic is only specified for 5v operation but I guess there is some 3,3v schmitt trigger with fast rise/fall times. I stopped looking though because the cypress part was available from my local supplier.
Kasper
I'm going to try the CY2305SC-1H from cypress (http://www.cypress.com/cfuploads/img/products/38-07140.pdf).
It has rise/fall times below 1,5ns (0,8-2,0v) which should be just fine.
F-logic is only specified for 5v operation but I guess there is some 3,3v schmitt trigger with fast rise/fall times. I stopped looking though because the cypress part was available from my local supplier.
Kasper
- Status
- Not open for further replies.