[time-nuts] 10 MHz Band-Pass Filter Needed

[Bruce Griffiths]

John Ackermann N8UR wrote:
> Dan Rae wrote:
>   
>> Richard W. Solomon wrote:
>>     
>>> The GPSDO I want to use has an output rich in harmonics. In some
>>> cases that is good, but Murphy rules and in the application I have
>>> today, it is not good.
>>>
>>> I need a 10 MHz Band-Pass Filter, Bandwidth is not critical, something
>>> small with SMA connectors would be ideal, but I can live with BNC.
>>>
>>> Anyone have such a beast or know where I can get one ? I checked Mini-Circuits
>>> and choked on the price !!
>>>
>>>   
>>>       
>> Dick, it is really easy to build one.  The wonderful (free!) filter 
>> design program ELSIE will give you all the help you need.  I would have 
>> thought an hour or so with a couple of toroids would do the trick.  Even 
>> just a Low Pass filter would usually do to turn a square wave into a sine...
>>     
>
> It's actually much better to use an LPF than a bandpass filter if you 
> don't have subharmonic energy to deal with.  An LPF with a cutoff midway 
> between the fundamental and the 2nd harmonic will show much less tempco 
> (in the form of phase shift over temperature) than a bandpass filter.
>
> John
>
>   
John

Another approach is to use relatively high Q traps/bandstop filters to
eliminate/minimise the unwanted harmonics together with a low pass filter.
The high Q traps contribute little phase shift and phase shift tempco at
the fundamental.

Bruce