How many 10 Mhz outputs do you REALLY need?
Our usual method in the Ultrafast Laser business is to buy a commercial four channel Digital Delay generator, drive it from the the system master clock, delay ALL the outputs as needed. This way we would have programmable lead or lag as needed. In rare cases we could end up using a simpler box from Allen Avionics which is a programmable analog video delay built to mimic a coax line delay.
I just checked Minicircuits and they do not have a stock voltage controlled phase shifter module at 10 Mhz. Although they do have them at lower and higher frequencies. Even if they did, I doubt it would reach your accuracy requirements.
I'd take a look at the Highland Technologies Inc, Stanford Research Systems Inc, and Quantum Composers Inc, web sites.
Highland Technologies specializes in that sort of low jitter, multiple output timing. John Larkin at Highland makes a lot of custom stuff that does not show up on the web site, so you might wish to speak to him.
If you need to roll your own take a look at these two FPGA based designs:
A simple, precise, and low jitter delay/gate generator
J. Kalisz1, A. Poniecki1 and K. Ró?yc1 Rev. Sci. Instrum. 74, 3507 (2003);
Accurate and low jitter time-interval generators based on phase shifting method.
P. Kwiatkowski1, Z. Jachna1,a), K. Ró?yc1 and J. Kalisz1 Rev. Sci. Instrum. 83, 034701 (2012)
And the best one, but it uses a ASIC:
Low-jitter wide-range integrated time interval/delay generator based on combination of period counting and capacitor charging K. Klepacki, M. Paw?owski, and R. Szplet, Review of Scientific Instruments 86, 025111 (2015); doi: 10.1063/1.4908199
Usually you can find a few used Digital Delay Generators on Ebay...
I'm curious why you would need 400 precise clock stations? Are you working on a kilometer ring accelerator or something similar?
I used to deal with this sort of stuff in a former career.
If we had to do so something really nasty, like sink the laser mode locking to the master bunching clock of a accelerator, generally they would have a stable house clock distribution at 380 Mhz or 1 Ghz on hard coax to align local edges with. Each station at a ring or Linac would have a clock bench divided down, synced, and preadjusted, for delay. Generally the details of such systems get published, so you might want to look for research papers from that area. If you really needed 400 outputs, that is the way to go.
Start with the search term "Clock Distribution" on Google and you'll get lots of hits, including chipsets from companies other then AD.
Steve
Home
Help
Search
Login
Register
