Keeping equipment turned on

[tkamiya]

I now have such things as multiple RF generators, frequency counters, spectrum analyzers and oscilloscopes.  I also have GPSDO and distribution amplifiers.  No DVMM yet, and my RF generators are synthesized type.  (Yes, I'm grinning ears to toes)

What I am wondering is, considering all of what I have are time based and GPSDO is always on, is there any need to keep rest of the system (distribution amplifier included) turned on all the time to avoid warm up time?   

When each equipment has its own reference oscillator but also has an external reference, does external reference takes over or does it discipline internal oscillatior?  (a reason to keep them on)

[cdev]

If you want the best longevity and highest stability, it makes sense to leave a GPSDO on all the time. Also, where you place it matters too, try to put it somewhere where the ambient temperature doesn't vary a lot. That means don't put it (as I did with mine at first), by a window next to a hydronic heater/radiator and outside wall.

Attaching it to an inner wall might work out okay. An inner closet away from exterior walls might be the best place for it in a typical home. Its more likely to be kept in a fairly comfortable temperature range without wild swings in some kind of enclosed insulative container within the more occupied areas of a house.

Note- I am not going to speculate about equipment longevity below, i am just going to speculate on accuracy of the frequency and time measurement aspect of GPSDOs versus other sources of time.

With the caveat that 1. Other people here know a lot more about this than I do. and 2. My own equipment isn't this good so I am just repeating what Ive read..

its my understanding that the following usually applies but only up to a point..
Not as much when your application starts to get into the range where nanoseconds matter..

A GPSDO is a sort of intermediate standard that excels at some kinds of things and isnt the best at others.

Instruments that use an external GPSDO should usually rely on that.. assuming its a good GPSDO, its likely it's stability is superior to most other oscillators of the kinds most people have access to, (unless they are "time nuts" or calibration labs..) Also, its verifiable.. its capable of being NIST certified.. which may or may not matter..

So then, in most contextx it would usually become immaterial how stable their own reference oscillators were, if it was pretty clear that although good they were unlikely to be as good as a GPSDO.

 unless that oscillator for some reason was preferable to the GPSDO  A free running oscillator may be more stable than a GPSDO at a very fine level of detail, basically.

In some critical applications where the free running oscillator is very good, the GPSDO might not be as stable when taking measurements that relied on a short time constant.

A rubidiium frequency standard or perhaps a very very good OCXO.. with exceptionally low jitter might be better, substantially better for those things than an average quality GPSDO.

NIST has free publications that do an excellent job of explaining this.

[tkamiya]

My GPSDO is an ovenized Xtal kind, and antenna is placed outside in full view of most of the sky.  (so it's always locked).  It's a GPSTM, Nortel Trimble NTBW500AA GPSTM.  As I understand it, these were used as base station timing device for now-defunct 2G cellular networks.  I know some of my gear uses ovenized XO, some are TCXO, and some are just plain XO.  (why would someone make 4GHz frequency counter with 10^-6 time base??)

As I understand it, Ovenized XO can be more "stable" in short term wise and jitter wise when compared to Rubidium.  So my plan was to attach all of this to distributed lab standard.

Now, the "lab" is for my personal use, so I don't have any established standard, but I kind of got bit by this test gear collection thing..... so...

[G0HZU]

I've worked in RF labs all my career and it's rare to leave anything switched on 'long term' in our labs. This includes high end test gear with decent OCXOs. Also, it's rare to connect  the 10MHz reference ports to a common reference unless there is a requirement for controlling relative phase or producing a zero beat across several pieces of kit. But even when this happens, in almost every case, it really doesn't matter which OCXO gets used as the 'daddy' and it doesn't matter how accurate it is (within reason). So a GPDSO or Rb standard isn't needed.

Here at home I have never left equipment on 24/7 because that would be a waste of time and money and also a potential smoke/fire hazard. Maybe I'm just not deluded enough to think I 'need' a fleet of GPSDOs and a fleet of DMM calibrators and all the baggage that comes with all this stuff. We generally don't need it at work either. I do sometimes connect the 10MHz ref ports on my sig gens here at home but this is when I need phase lock between the generators so I can tweak the phase between signals from each generator. But I don't need to leave test gear on 24/7 to achieve this. I just turn it on, leave it to warm up a bit and then use it. It doesn't add ANY value using a 24/7 GPSDO to any of this stuff. But I guess it's easy to delude oneself that it does add value...

There's plenty to learn about electronics without needing to get obsessive/nutty about calibration. I kind of feel sorry for the OCD types I see with a bench stuffed with calibration gear and numerous GPSDOs and counters and megadigit DMMs. Every day should be a school day if you are interested in electronics but don't spend that time fretting over volt nut and time nut wankery

[tkamiya]

Hi.... Thank you for the wake up call.  Yup.  I have an OCD that was actually diagnosed by a physician.  I kind of use it to my advantage sometimes as mentally forced QA control.

I am/was sort of aware all this wasn't necessary but as a hobby, it's fun to pursue "the best" and it's really not the result that matters (as they are already good) but all the research and thoughts that goes with it that counts.  I had few times in different area that extra knowledge became pretty handy.  Of course, it's a matter of degree as well.  I have no reason to want a Rb or Hydrogen Maser reference standard.  But do I like seeing long strings of zeros.

I've long held a belief that repeatability (precision) counts more than absolute value (accuracy).  Having a known single standard does that for me.

Anyway, thank you for the reality check!

[G0HZU]

The awful truth is that most people I see who are into this stuff just buy more and more commercial gear to feed the OCD. They seem to plug it in, play with it, declare it an accuracy upgrade, cram it next to the older gear on the workbench and then talk about it on internet forums as if they have achieved something like gaining a merit badge. But the only person achieving anything is the person who designed or sold the commercial gear.
Commercial DMMs and counters/OCXOs have been boringly brilliant for decades now so why devote a huge chunk of hobby time to this niche area of electronics that has been 'cracked' for decades?

[cdev]

The reason I leave my GPSDO on when I can is that my particular unit, like the OP's, is actually designed to be left on all the time.

When its turned off for a few weeks, and then back on it takes quite a long time, many hours, to regain its accuracy. I get the feeling that it will last longest if its just left on and left alone to do its thing.

As far as power draw, although its fairly warm when its running, still the GPSDO only takes its full current draw (around 2.5 A) for less than a second when its first turned on and then afterward it seems to just chug along at around 300-400 ma. (although I have never done a systematic analysis of if it draws more power at some moments, it may draw more power sometimes, but when Ive looked at the current meter its never high except at the moment of startup.)

I also have a micro-PC with a connected GPS that I leave on as a time server. The SBC runs on a cell phone charger and the GPS with its antenna takes only around 20 ma so it is using the 3.3 volts that's present on the SBC's GPIO header.

[tkamiya]

Muahahahaha.....  more pretty zeros! 

[Hydron]

GPSDOs should be left on continuously - this is what they are designed for!

Note that OCXOs can take quite some time after turn-on for the "ageing", i.e. frequency drift over time (separately to temperature), to settle down to the (normally lower) long term value. It can easily take WEEKS for this to happen, especially if they have been off for quite a while.

An OCXO in a piece of equipment which isn't left on continuously should still be better than a TCXO/XO, but if you want the absolute best accuracy/stability then it should be left on, and ideally disciplined by GPS. I believe that some equipment can keep the OCXO on while the rest of the unit is off?

As for GPSDOs vs other oscillators, if the oscillator being disciplined is an OCXO then ideally the output would be the best of both worlds - the good short term stability of the OCXO and long term stability of the GPS clock disciplining the oscillator. This requires care designing the internal loop filters which adjust the OCXO tuning voltage/command to track the long term GPS frequency reference.

[G0HZU]

Saying that GPSDOs or OCXOs should be left on all the time doesn't really demonstrate WHY this adds any real world value to a typical hobbyist sat at home in a shed or spare bedroom. Even if it improves the long term stability of the oscillator it won't matter a jot to virtually everyone. Why not put a cardboard cutout of a 10.00000000001 MHz display on your bench and save a lot of pointless faffing about? You could even pretend the gear is powered all of the time. It won't add or detract ANYTHING from your workbench but it will never drift and it might save space and allow you more free time to do/learn other things.

[cdev]

Clocks have been the bane of all humanity, I agree.

[cncjerry]

For a lower power reference, just get an HP 10811 oscillator, tune it to your GPSDO twice a year, and shut the GPSDO off.  The 10811 will have better phase noise performance and be more than stable enough between calibrations.  A 10811 requires about 2.5W of power, maybe less.  My Lucent GPSDOs take around 60W once warmed up and locked, higher when first turned on, maybe 75W.  An Rb frequency standard varies between 12W and 20W, maybe a little more. 

I turn my GPSDO on when I get new equipment, to check an oscillator or FRS, or for the bi-annual ARRL Frequency Measurement Test.  Now that I have a Cs Beam I doubt I'll need the GPSDO at all. The Cs can also be put in standby, switched on and within 30 minutes be close enough.

[texaspyro]

Saying that GPSDOs or OCXOs should be left on all the time doesn't really demonstrate WHY this adds any real world value to a typical hobbyist sat at home in a shed or spare bedroom.

Oscillators have a problem called "retrace"  If you turn them off, they don't come back on at the old frequency.  Generally the longer they are off, the greater the retrace error.  A GPSDO can compensate for the retrace error. 

Also, when an oscillator is turned it takes time for it to stabilize.  An oscillator that has been off for a long time can take several weeks or months to fully stabilize.   The one I put in my HP-53132A counter had been off for a few years and took around 6 weeks to stabilize.  The counter is designed to keep the OCXO always powered on, even if the power switch is turned off.  HP would not have implemented this feature if was not important.

Whether to turn one on or off depends upon your requirements, but if you are bothering to use OCXOs it sort of implies stability and accuracy are important for the application.
 

[nctnico]

I'd just leave the GPSDO on (if you use it regulary) because it needs at least a couple of minutes to get a good fix.

[texaspyro]

I'd just leave the GPSDO on (if you use it regulary) because it needs at least a couple of minutes to get a good fix.

Most GPSDOs take up to a week to learn the characteristics of the device so they can compensate for temperature and drift.    They also have disciplining time constants in the thousands of seconds and it can take 10 time constants to fully stabilize.

[cdev]

How do they behave when their ambient temperature environment is abnormally chilly?

Is it informed in any way by their latitude and the time of year it is?

I guess I mean, do they have any means of determining "hey, its the winter where I am now, better adjust for this chilly ambient air in a different way than I would if it was summer.." ?

[texaspyro]

How do they behave when their ambient temperature environment is abnormally chilly?

Most GPSDOs have temperature sensors in them.  The learning process attempts to correlate oscillator frequency corrections with temperature and also attempt to determine the oscillator drift rate.   These factors get fed to the disciplining algorithm and are most important when the GPSDO is in a "holdover" state and is unable to use the GPS signal to keep the oscillator on freq.  The HP GPSDOs take a week to fully learn the characteristics of the system.  They have a message to report the estimated holdover uncertainty.

[TheUnnamedNewbie]

In our university labs a lot of gear is never turned off. The I think our VNAs have only ever been turned off when they are sent back to Keysight or R&S for calibration. Some other stuff is turned off for storage because the lab changes a lot (every time a new chip comes back from tapeout, the person testing will just get what gear they need at their bench). However, usually once it's turned on for verification at the bench it's left on until we need to take the setup apart again because we are done.

Saying that GPSDOs or OCXOs should be left on all the time doesn't really demonstrate WHY this adds any real world value to a typical hobbyist sat at home in a shed or spare bedroom. Even if it improves the long term stability of the oscillator it won't matter a jot to virtually everyone. Why not put a cardboard cutout of a 10.00000000001 MHz display on your bench and save a lot of pointless faffing about? You could even pretend the gear is powered all of the time. It won't add or detract ANYTHING from your workbench but it will never drift and it might save space and allow you more free time to do/learn other things.

How is the hobbiest quest for a better voltage reference or a better clock stability any different from a stamp collectors desire to get the most rare and precious stamps, or a florist wanting to get the most beautifull flowers in their garden? Not everything has to have a practical purpose to it - if these people enjoy "knowing" they have this clock, or get joy out of their pursuit for accuracy, why not let them enjoy it and help them with your knowledge when they need help?

[texaspyro]

One good reason to turn equipment off is if it uses VFD, OLED, gas discharge, custom, etc displays.  These have finite lifetimes and are usually custom, unobtainable devices.  A lot of perfectly good test equipment is rendered useless because the display has failed and the manufacturer no longer supplies them (or sells them for insane prices).

As an example, HP used to sell replacement displays for the 531xx counters for less than $100.  Now to replace the display you must pay them around $1800 dollars for a complete front panel board.

[cdev]

Many- surprisingly many, of the most important inventions never had anything to do with a university or corporation!

[joeqsmith]

How is the hobbiest quest for a better voltage reference or a better clock stability any different from a stamp collectors desire to get the most rare and precious stamps, or a florist wanting to get the most beautifull flowers in their garden? Not everything has to have a practical purpose to it - if these people enjoy "knowing" they have this clock, or get joy out of their pursuit for accuracy, why not let them enjoy it and help them with your knowledge when they need help?

Totally agree.  If I had to look at the practicality of my hobbies to justify them, I would have no hobbies.  I doubt I would do much of anything at all.

Almost forgot, my reference clock (old telcom unit)  has been running 24/7 for over 14 years now but I do turn off the other equipment when it is not in use.   

[G0HZU]

How is the hobbiest quest for a better voltage reference or a better clock stability any different from a stamp collectors desire to get the most rare and precious stamps, or a florist wanting to get the most beautifull flowers in their garden? Not everything has to have a practical purpose to it - if these people enjoy "knowing" they have this clock, or get joy out of their pursuit for accuracy, why not let them enjoy it?

Yes, but you can apply the same argument to audiophoolery even if the products they buy are snake oil.  Yet they get ridiculed on here a lot. At least the audiophool gets to listen to some music via the snake oil products!  As a comparison, does looking at 10.0000000001MHz get more interesting after a few hours/days/weeks/months? Do you get upset if it changes to 10.0000000002MHz? If it does change, do you go out and buy something better? It sounds like a very blinkered hobby.

I'd argue that hobbies based on stamps and flowers ought to more diverse and interesting compared looking at a commercial display showing 10.0000000001MHz every day. Maybe I'm wrong?

[mtdoc]

Perhaps the definition of "Hobby" is needed:

Hobby
: a pursuit outside one's regular occupation engaged in especially for relaxation

Volt-nuttery or Time-nuttery are hobbies.  As is collecting vintage test equipment, etc.


Audio foolery on the hand  is when one is using technically false explanations (i.e. lying) to sell a product (or buying such product based on technical ignorance).

That is different than being an audiophile as a hobby.  Audiofools are a subset of audiophiles

[tkamiya]

There are so many things that makes absolutely no logical sense but gives individuals immense pleasure.  Heck, when we get right down to the essentials, there aren't that much that we really NEED.  Discussing something similar with a sales person I know....  he says, he doesn't cater to needs.  He specialty is wants.

As to pleasure, if it gives someone a purpose and joy, who am I to judge?

I, too, dislike lies and false science.  But that's not the most insane things I've seen people do either.