[time-nuts] Assorted replies, and request for info

ew ewkehren at aol.com
Tue Feb 27 06:58:04 EST 2018

Allow me to make some follow on comments on John’s post. Much of the equipment we use was never intended for what we use it for and developed at a time that certain parameters where not tested, and equipment now available for our efforts is getting rare.
We are working on 1 E-13 frequency at one second, performance the big boys practiced forty years ago with tens of million budgets and lab environment we can only dream of.
Last year we did some work with a Tbolt and M100 using 40 000 seconds and 7 E-17 per uV. This was done with the support of Warren that I updated regularly. One set of data he responded on with: you are getting 1E-13 but what kind of Lab, 0.1 C over two weeks? The answer Juerg,s basement Lab in Switzerland while every body was on vacation. All critical tests are run over night.
What we now spend extensive time on is AC power. I have not seen much on time-nuts on this subject. Latest effort replacing vacuum fluorescents with LED some thing I have not found in the US.
On A9, we have completed two boards one with the LT1793 and one with the LT1021 ready to be ordered today. With the low cost of boards, today will lay out a board with a 741 and plan to do an actual test once we have the boards. Will talk to Corby about it today, he is the only one that can do the test.
In a message dated 2/26/2018 9:46:22 PM Eastern Standard Time, john at miles.io writes:

 Request for data from an off-list friend who is looking for info on an Ovenaire OCXO and/or the instrument it was used in:

> Could you send a post on my behalf asking if anyone else has an 
> Ovenaire 42-15 or if not, a Spectracom 8131 Frequency Standard Oscillator. 

This is from Dennis Tillman, who can be reached directly at dennis (at) ridesoft.com.

> Fun fact -- there's a wide spur at ~2 Hz on the 5065A phase noise plot. What
> do you think that is? On a hunch I opened the front panel and reset the
> blinking amber battery alarm lamp, and voila, that noise went away. Makes
> sense when you think of the power variations associated with a blinking
> incandescent lamp.

I hadn't heard of that one, but some other examples include 1-pps crosstalk on the 10 MHz output of some of the HP GPSDOs (which I've run into myself), and the inadvertent ~5 MHz comb generator that drives the indicator LED on the HP 5370's reference clock interface PCB that Bruce Griffiths noticed several years ago. Presumably neither of these faux pas were bad enough to be noticed by the original designers or their paying customers, but they probably would have been fixed if they had come to the attention of the people involved. Goes to show how improved instrumentation can be a curse as well as a blessing.

Poul-Henning's observation on the 5065A integrator cap highlights the risk of erring too far in the opposite direction:

> When I experimented, I could hardly find *any* property that mattered
> for that capacitor, not even the exact capacitance, because the
> adjustment procedue handles that.

I have a feeling this is true of most of the components in that circuit. The nice thing about an integrator is that it's also a low-pass filter. And the nice thing about a closed loop is that it's, well, closed.

Before spending too much time arguing about whether the opamp should be replaced with an LT1012 or an AD797 or a cryocooled tunnel diode or whatever, I'd suggest replacing it with a 741 and seeing how much *worse* the performance gets. It is easier to measure the effect of that kind of change. If there is little or no harm in using the crappiest opamp you can find, that means that you can safely stop worrying about what the best one might be. 

> One of the TimePods that I had access to in the past was particularly good
> at telling you it was sitting on top of a power transformer. It didn’t matter a
> lot which instrument the power transformer was in. For some weird 
> reason it was a good magnetometer at line frequencies. I never bothered to 
> send it back for analysis. Simply moving it onto the bench top (rather than 
> stacked on top of this or that) would take care of the issue.
> As far as I could tell, it was just the one unit that had the issue. None of the
> others in the fleet of TimePods seemed to behave this way. Given that they
> normally are very good at rejecting all sorts of crud and ground loops, it was
> somewhat odd to see.

I've seen similar behavior here, not only with respect to units responding differently to 60/120 Hz magnetic interference, but also at higher offsets in the absence of an obvious coupling mechanism. There was one case where a TimePod I was working with picked up an unstable low-level spur near 25 kHz from an LED aquarium light fixture several meters away. Other units swapped into the same position did not show the spur at all, and I was never able to narrow down the cause with any certainty. I don't have a good explanation for any of the above, unfortunately.

That being said, Phil Hobbs posted something on sci.electronics.design the other day that I thought was subtly insightful, even though he was just stating an obvious point. Namely, ground loops are inherently very low impedance phenomena, often occurring in the milliohm range. Especially when dealing with anodized aluminum hardware like the TimePod's enclosure, the difference between a test setup where all the coax shields act as a near-perfect shorted transformer turn versus one with significant loss might come down to small differences in fastener torque, or perhaps a missing star washer. So it's possible to envision a scenario where tightening up all the proverbial loose screws actually makes a magnetically-coupled spur worse. 

Lifting a coax shield is usually not the best solution to ground loops, but Phil's offhand comment made me wonder about the effects of deliberately adding just a few ohms of series R. It's on my list of things to look into when I have time.

-- john, KE5FX
Miles Design LLC

time-nuts mailing list -- time-nuts at febo.com
To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
and follow the instructions there.

More information about the time-nuts mailing list