[time-nuts] Phase measurement of my GPSDO

Tue Apr 14 15:52:51 UTC 2020

Hi

> On Apr 14, 2020, at 11:01 AM, Tobias Pluess <tpluess at ieee.org> wrote:
> 
> Hi Bob
> 
> yes I totally agree, you can save a lot of space by having those USB boxes
> for almost every measurement.
> However I am happy that we have at work the "old school stuff" (by which I
> don't mean we have old equipment, actually it is pretty modern) because in
> my opinion the user experience is much better if you have buttons to press
> and knobs to rotate. I don't like the scopes where you have to set up
> everything via touchscreen, and I don't like the USB scopes. Therefore I am
> actually quite happy with the kind of equipment that sits in its own box,
> the disadvantage is that it needs space. However I think this is only a
> concern for the homelab, where space is limited. A place where they don't
> have the space to install a proper network analyzer is perhaps a bit ...
> funny, I'd say. But for home usage you're probably right. However many of
> the USB things have lots of drawbacks the "pro" equipment doesn't have. But
> perhaps we are wandering off the subject a bit now :-)

I have more space on my home bench than I did on any bench at any
place I ever worked. I have never worked at a location (or visited one) where
full sized / old style network analyzers were deployed to all dozen or three 
engineering benches. Same thing with 3048 style phase noise test sets. 
For phase noise, the issue was spur pickup as much as anything else. The
USB stuff actually is *better* than the old gear in that regard. 

> 
>> Lots of projects started and never fully completed. I have piles and
> piles of them ….
> 
> yeah I know. I have them as well. Not piles, but certainly quite a few. But
> I have also piles of stuff I have finished. For instance from time to time
> I try to manufacture RF waveguide components like directional couplers and
> bandpass filters and the like. While I don't actually need them I simply
> want to see whether I could make them and how precise it is possible. (it
> is possible quite accurately up to perhaps 80GHz, above that precision
> becomes an issue). And as you know especially the waveguide stuff is bulky
> :-) I even once made my own N connectors just to see whether it was
> possible. (it is, as well. When gold plated or at least polished, they look
> very similar to the commercial ones.)
> 
> But this is definitely seriously off-topic ;-) To ask a back-to-topic
> question: you said you wrote your own programs to extract ADEV from
> timetagged data. Could you give a hint on how you "massage" (as you said)
> the data from the time tagger to get proper phase info and ADEV?

I wrote the first stuff back in the 70’s. It got redone several times as I
moved from plant to plant over the years. By the late 90’s these systems
began to fade out in industry. 

Normally the code looks for jumps and then makes arbitrary decisions
about how to “patch” them. With a time tagger, most of the problems are
actually dropouts in the data. How you fill those depends on what you
are doing ( you may re-do the run ). Simply putting in the phase data
from the previous point is an alternative. 

Bob

> 
> Tobias
> 
> 
> On Tue, Apr 14, 2020 at 2:51 PM Bob kb8tq <kb8tq at n1k.org> wrote:
> 
>> Hi
>> 
>> The whole “volume of clutter” on the desk / bench / around the lab /
>> basement
>> is one of the many drivers getting me to move over to some of the more
>> modern USB based measurement gear. You can swap out 400 lb of cranky
>> (but useful )  gear from the 1960’s for what would fit in a small
>> backpack.
>> 
>> Even at work, the same sort of drivers applied. A USB based box was
>> something
>> that you could have a dozen of out on many benches. The big beast setup
>> hulked
>> away on it’s own bench over in the corner. A lot more got done with the
>> little boxes.
>> 
>> There are a lot of grades of USB gear. I’m not suggesting that my $100
>> network
>> analyzer competes for accuracy with a shiny new R&S or Keysight device. It
>> just
>> *might* do as good a job as the broken down pile of gear from 1968 though.
>> For
>> a bit more than $100, I could get a better USB version ….
>> 
>> =====
>> 
>> Can you build your own TICC / scope / network analyzer / PC / television?
>> Sure you can. It’s just parts put together in this or that fashion.
>> Getting one to the
>> point that it measures up with what you can buy will take a while (likely
>> a couple
>> of passes).
>> 
>> Your time is “free" since this is a hobby. Does that really include 40
>> hours a week
>> for a year to get this or that pass done? At least to me … nope. How about
>> the lab
>> full of application specific test gear to get this piece of gear debugged
>> *before* you
>> can use it ? Hmmm….
>> 
>> Simply getting the board laid out is the easy part of any of this. On a
>> normal
>> product design, the work to get to that point is maybe 10% of the effort
>> put in
>> to get the job done. A home project may be even more lopsided since it
>> likely
>> is the first time you have done this or that.
>> 
>> Lots of twists and turns. Lots of projects started and never fully
>> completed.
>> I have piles and piles of them ….
>> 
>> Bob
>> 
>>> On Apr 14, 2020, at 8:15 AM, Tobias Pluess <tpluess at ieee.org> wrote:
>>> 
>>> Hi Bob
>>> yes sure I know. At least my homebrew GPSDO and the STAR4 GPSDO I have
>> are
>>> always powered, but since they don't take much space on my desk that's
>>> okay. But for the signal generators (or SpecAn etc) it is a completely
>>> different story; I usually take them from the shelf when I need them and
>>> put them back afterwards, because they simply take too much space away.
>> And
>>> my home lab has only limited space, unfortunately. I am thinking since
>>> months about a better setup (how arrange everything, how to stack my
>>> equipment and such) but it is not so easy because simply stacking all
>>> equipment on top of each other is unsatisfactory (in that case one needs
>> in
>>> general the instrument at the very top, which is simply too far away
>> :-)).
>>> The 8663A is a nice machine, but it would fill my desk already by 50%, so
>>> no way to leave it there (and therefore I cannot keep it powered).
>>> 
>>> Tobias
>>> 
>>> 
>>> On Tue, Apr 14, 2020 at 1:54 PM Bob kb8tq <kb8tq at n1k.org> wrote:
>>> 
>>>> Hi
>>>> 
>>>>> On Apr 14, 2020, at 2:31 AM, Tobias Pluess <tpluess at ieee.org> wrote:
>>>>> 
>> 
>> 
>> _______________________________________________
>> time-nuts mailing list -- time-nuts at lists.febo.com
>> To unsubscribe, go to
>> http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
>> and follow the instructions there.
>> 
> _______________________________________________
> time-nuts mailing list -- time-nuts at lists.febo.com
> To unsubscribe, go to http://lists.febo.com/mailman/listinfo/time-nuts_lists.febo.com
> and follow the instructions there.