[time-nuts] Frequency counter recommendation

[jimlux]

Magnus Danielson wrote:
> On 12/19/2010 05:13 AM, Bob Camp wrote:
>> Hi
>>
>> In the case of a $200 5370, you have to wait a while to find one. When 
>> you get it, you likely have to do some work to get it running. At the 
>> very least you will need to do a cal.
>>
>> My guess is that a counter project would be very similar. There is an 
>> order and build process that happens every so often. Eventually you 
>> get a set of boards that *might* work. They still need a bit of this 
>> and that to get them running. Once running you need to do a cal.
>>
>> The calibrated and running counter is something you can have tomorrow 
>> (more or less). That's very different than the kit of boards.
> 
> Another issue is that the lifetime of components in home-cooked 
> solutions. For "long time support" such a solution needs to be 
> retargeted as components come and go. You would then have a line of 
> boards doing the same function but not using the same components, which 
> would quite likely cause a not optimum usage of these components as the 
> capability for performance would shift over time. Eventually will some 
> old design decision be a show-stopper for optimum performance.
> 

That's not as big a deal as it is for commercial test equipment.  If you 
invest $50k in a signal generator, you can expect it to work for 10-20 
years, perhaps not with full mfr support but at least with service 
capability.


If you're buying $500 surplus, the risk is the obscure part that's not 
being made anymore.  If you can't find it, or improvise, you figure, 
meh.. another chassis to be used for cannibalization.

But that's the same for a TAPR style project.. you pick some FPGA or 
microcontroller, and the odds of it being available 10-20 years from now 
is zero.  BUT, the big difference is that hopefully, the software is 
published and at least some design information, so 10 years from now, 
someone can pick it up and respin it with whatever part is available. 
or start from scratch.  Either way, you've invested your $500 and got 
5-10 years use out of it.. That's down in the $10/month (beer money) 
range, and a great deal.

Granted, that assumption of design information might be optimistic.