[time-nuts] Photodiodes for high frequency OPLL

David McQuate mcquate at sonic.net
Sat Mar 30 14:00:08 EDT 2013


You'll need a photodiode that can detect photons at your lasers' 
wavelength.  You may be able to use a photodiode at a shorter than its 
design wavelength as long as there are not coatings (eg anti-reflection) 
that block the wavelength of interest.  You'll need to make sure that 
both lasers illuminate the same photodiode area, or you won't get a 
signal at the difference frequency.  The difference frequency power 
level is generally pretty low, so operating above the photodiode 
bandwidth is difficult.  My work was at HP and Agilent, who manufactured 
the photodiodes we used.

The photodiode frequency response is primarily limited by the depletion 
region's capacitance.  The circuit model is simply a current source 
shunted by a capacitor (perhaps with bond wire inductance to the RF 
connector) so the RF output current falls with increasing signal 
frequency.  I don't see much possibility of improvement thru 
matching--though using an amplifier with very low input impedance might 
help.

Dave

On 3/30/2013 3:48 AM, Attila Kinali wrote:
> Moin,
>
> I'm currently reading up some stuff on optical PLLs and am stuck with some
> details i cannot find any data on.
>
> The goal is to make two lasers locked with about 7GHz of offset to eachother.
>
> So far, i figured out that PIN photodiodes can go up to several 100MHz
> transition frequency and avalanche photodiodes are available up to 2GHz.
> The only photodiodes that go higher are those for the telecom range
> at 1-1.5um, which is a bit low for my needs.
>
> But, i have seen descriptions of such OPLL that work with multi GHz
> offsets in the 700-800nm range.
>
> Now the question is: Do these use special made photodiodes with a very
> high transition frequency or do they just treat the transition frequency
> as a 3dB point and amplify the signal after detection?
> (They do not describe the detection circuit at all, so i guess it must
> be something readily available)
>
> If it's the former, any idea where i could get such photodiodes?
>
> If it is the later, how much damping can i assume with PIN or avalanche
> photodiodes per decade?
>
>
> Thanks in advance!
>
> 			Attila Kinali
>
>

-- 
Clear Stream Technologies



More information about the time-nuts mailing list