Found home made rubidium reference that seems to really work well

[edpalmer42]

James,

I couldn't resist - I had to try it myself.  But, unlike you, I threw caution and common sense out the window and stuffed my victim into a Dewar flask!   

I only have a few OCXOs with seperate oscillator and oven power.  I chose an Ovenaire OSC 59-52 oscillator from about 1982.  It's an old style 2x2x4 inch size.  I think it came from an HP 3330B synthesizer as the High Stability option.  1e-9/day aging, 0 - 55C.  Power drain for the oscillator was ~32ma @ 15V.  The oven draws ~75ma @ 29V in a 22C ambient.  I measured the internal temperature by taping a Type K thermocouple into the hole for adjusting the frequency.  It measured about 74C.

After a two hour sauna inside the Dewar flask, the frequency only rose by ~0.016 Hz (5 MHz nominal) and part of that would be normal drift.  When I took it out of the Dewar, the frequency rose by 0.032Hz over a period of minutes.

Over the two hour test, the internal temperature rose by ~5.4C and the oven current dropped from 74ma to 26ma.  Both surprised me.  I didn't think the temperature would change that quickly and I expected the temperature change to drive the oven current to zero.  I don't know if the design won't allow the current to go to zero or if it's just old and leaky.

The biggest surprise was the case temperature.  It started out at a nice cozy 36C.  After two hours it was 60C.   

Obviously, this wasn't a practical thing to do.  But it was interesting.  I doubt that a bit of extra insulation is helpful, but I agree that it's probably safe.

I also agree that you should never add extra insulation to a Rb standard.  A common parameter is to keep the base of the unit below 65C.  But don't go crazy.  A huge heatsink is not required or beneficial.

Ed

P.S.  It's been about 30 minutes since I removed the oscillator from the Dewar.  In that time, everything has returned to the same values as before except for the frequency.  It's now 0.03 Hz higher than the initial value.  It appears that the rise in temperature while in the Dewar partially counteracted the oscillator's attempt to drift up.  Once things cooled off, the drift was restored.  Previous tests have shown that this oscillator tends to drift up in frequency.  I don't think I can draw any conclusions regarding the effect of the temperature change on frequency.

[notfaded1]

How big is your Dewar flask?  I've wondered about putting electronics inside them.

Bill

[ArthurDent]

Any of the older style Rb oscillators that take over 10 minutes to lock or any of the newer ones that take over 5 minutes to lock I would be concerned about. Where yours takes over ½ hour to lock I would have reservations about using it unless you can find that there is some obvious problem you can correct.

I have an old FRK-HLN made in 1993 that takes 4 minutes 15 seconds to lock and has a distribution amp inside as well and I use this as my primary Rb standard. I also use a LPRO-101 that locks in 3 minutes that also has a distribution amp built in. Both units are very stable and I use a GPSDO to check them.

Here are some photos of those two units and the LPRO-101 unit’s frequency displayed on two different counters. They typically vary about .0005 max between readings and the graph shows the long-term stability of the LPRO-101 over a 12 hour period after it was adjusted to be as close as I could adjust it against my GPSDO.

[edpalmer42]

How big is your Dewar flask?  I've wondered about putting electronics inside them.

I have a couple of Pope 8600 Dewars (350 ml volume, 70 dia x 125 mm deep internal), but the ones that I experiment with are just Thermos bottles that I bought from the local second-hand store.  Look for wide-mouth bottles with plastic liners.  Keep the liner.  You can use it to fill up the space if necessary and to protect the glass from scratches.

Ed

[notfaded1]

Well it looks like maybe the 010 option in my 53131A isn't exactly rock solid.  When I get my new cables I'll feed the Datum 8040 in as the frequency standard for the counter and then measure the unknown box.  I'm glad that timelab works so easily.  Great suggestion Ed!  I'm definitely on the right track now.  When my new GNSSDO get's here I'll have a good GPSDO to check against the rubidium but the fact that both the Datum 8040 and unknown Rb are really close seems to be good.  Now I need to feed a better standard into the counter than it's upgraded oscillator option 010.  I still need to add the output connectors to my Symmetricom Syncserver as well.  That's going to require drilling out the back of one of them  to make room for the new 3 outputs and 3 inputs.  This is fun stuff.  With help from timenuts like @Ed and @AuthurDent I'm figuring a LOT more out.  I can also hook the two Rb standards to the 53131A... tell me this is it ok to use video 75Ω BNC video cables to connect these?  I have a bunch of nice 75Ω BNC cables.  I bought the Pomona cables for this reason thinking it's a no no to use 75Ω cables with test gear with 50Ω inputs because of the reflections it can cause.  I have some nice 50Ω terminators but this won't substitute for the actual cables will it?  I'm sorry for the questions I'm a CIS engineer not an EE or RF engineer but I'm learning fast.  Most of my world is crypto, networks, and digital compute, storage, and virtualization.

If it won't hurt anything I'll try with the 75Ω cables.  I understand the BNC connector may or may not be a little different with the 75Ω having smaller center pin possibly?  One thing I know... the 75Ω male BNC will fit on any 50Ω BNC female.

Bill

[edpalmer42]

Yes, the 75 ohm cables will be okay - at least for now.  For BNC connectors, 50 ohm and 75 ohm are mechanically compatible.  If you have true 75 ohm connectors (many video cables use 50 ohm connectors) there's no white insulator around the male pin or the female socket.  Once you get your 50 ohm cables, you can compare the results with both cables and see if there's any difference.

I think the 53131A has an option for 50 ohm or 1M ohm termination so you don't need the external termination.

So that ADEV graph is the frequency of the unknown Rb as measured by the 53131A?  There seems to be a slight problem here!    In the 'Acquire' screen, did you tell Timelab that the 'Data Type' is Frequency rather than Time Interval?

Ed

[ArthurDent]

Oops, in post #27 I forgot to add the photos of the two counters I mentioned in the post. The photos are now added if you're interested.

[notfaded1]

Thanks @ArthurDent... I also have a BG7TBL.  I just ran the Datum 8040 right after it locked over with the HP5313A and got this... something looks maybe not so good.  Do I need to feed the counter a better frequency standard than it's 010 option high stability oscillator to get better data?  Also does the BG7TBL work with timelab???  That would be nice but I don't think it does maybe... not sure.  So where in the picture it says ADJ you adjusted the frequency by adding some components next to the LPro-101?

Bill

[notfaded1]

Yes, the 75 ohm cables will be okay - at least for now.  For BNC connectors, 50 ohm and 75 ohm are mechanically compatible.  If you have true 75 ohm connectors (many video cables use 50 ohm connectors) there's no white insulator around the male pin or the female socket.  Once you get your 50 ohm cables, you can compare the results with both cables and see if there's any difference.

I think the 53131A has an option for 50 ohm or 1M ohm termination so you don't need the external termination.

So that ADEV graph is the frequency of the unknown Rb as measured by the 53131A?  There seems to be a slight problem here!    In the 'Acquire' screen, did you tell Timelab that the 'Data Type' is Frequency rather than Time Interval?

The 75Ω cables BNC have white insulator at the base of the PIN but NOT around it like my 50Ω cable has.

Ed

Yes that's the unknown measured by the 53131A Ed.  I used the defaults in the Acquire screen.  I'll try starting it again and check... still learning this software but it's really nice for free!  I'll check the Frequency vs. Time Interval.  I assumed since it supported the 53131 that it would configure it for the sampling?

[notfaded1]

Yes, the 75 ohm cables will be okay - at least for now.  For BNC connectors, 50 ohm and 75 ohm are mechanically compatible.  If you have true 75 ohm connectors (many video cables use 50 ohm connectors) there's no white insulator around the male pin or the female socket.  Once you get your 50 ohm cables, you can compare the results with both cables and see if there's any difference.

I think the 53131A has an option for 50 ohm or 1M ohm termination so you don't need the external termination.

So that ADEV graph is the frequency of the unknown Rb as measured by the 53131A?  There seems to be a slight problem here!    In the 'Acquire' screen, did you tell Timelab that the 'Data Type' is Frequency rather than Time Interval?

Ed

Ed-

The 75Ω cables BNC have white insulator at the base of the PIN but NOT around it like my 50Ω cable has.  I'll run the test again and check the frequency vs. time interval as I assumed the timelab would configure it.  I'm running the sample with the correct setting now... I'm an idiot and can't read apparently  Which is better way to measure frequency or time interval?  I suppose both give you information but they correlate correct?  Also yes the 53131 does have 50Ω or 1MΩ.

Bill

[notfaded1]

@ArthurDent

I really like your Rb standard... it's nice.  I've done a lot of research on the various Rb standards LPRO-101 is legit... I was hoping maybe the one in the wood box would be one actually!

Bill

[notfaded1]

Ok much better now  here's the comparison with both Rb measured by 53131.  I turned the unknown Rb off finally.  I'll let it cool down tonight and tomorrow can dig into it some more and also time the lock period better.  I'll probably open the wood box too... it's killing me not knowing what's in there.

Bill

[edpalmer42]

Yes that's the unknown measured by the 53131A Ed.  I used the defaults in the [Timelab] Acquire screen.  I'll try starting it again and check... still learning this software but it's really nice for free!  I'll check the Frequency vs. Time Interval.  I assumed since it supported the 53131 that it would configure it for the sampling?

No, the author decided not to do any configuration.  Timelab just accepts data.  You have to tell it what the data is.  Considering Timelab's price (free!), I'm not complaining.

Which is better way to measure frequency or time interval?  I suppose both give you information but they correlate correct?

Frequency and Time Interval (phase) are two sides of the same coin.  You can convert phase to frequency with no issues, but there are some problems going the other way.  When it comes to measuring ADEV, there are a few differences.  Most notable is the difference in slope of the resulting graph.  In your graph, it takes about two decades of frequency for the graph to drop one decade in ADEV, i.e. the slope is -0.5.  If you measure Time Interval, the slope is -1 which causes the graph to drop faster.  This gets you into the low values that we're really interested in sooner.  I've seen graphs of OCXOs that didn't look impressive because the measurements were frequency based and by the time the graph got down to the good stuff, the crystal aging was already causing the results to rise.

The reason for the difference in slope is hidden in the fog of theory.  The value of the slopes suggests that it's related to different types of noise.  The slope of -1 suggest white or flicker phase noise while -0.5 suggests white frequency noise.  I don't know why the different measurements would exhibit different types of noise.  I'm more of an application guy. 

Ok much better now  here's the comparison with both Rb measured by 53131.

Yes, much better.  But you you might be able to do even better.  You should measure the noise floor of your 53131 to see if it's working up to spec and to see what results you can expect from it.  This measurement is for Time Interval.  I don't know if the noise floor in frequency mode is the same or not.  That's a question for the theory guys. 

All you need for this test are two pieces of BNC cable and a BNC T adapter.  Connect the T adapter to Channel 1.  Set the impedance of Channel 1 to 1M ohm.  Connect a BNC cable from one side of the T adapter to Channel 2.  Make sure the cable can't flop around and make sure that the temperature won't change during the test.  The length doesn't matter.  A meter or two is fine.  Set the impedance of Channel 2 to 50 ohm.  Feed a square wave from your 8040 to the open connector on the T adapter.  I think the 8040 has a 1 PPS output.  That would be perfect.  Set the triggering on both channels appropriately for the input signal.  Now measure Time Interval and collect the data with Timelab.  Since there's nothing between Channel 1 and 2 but a piece of cable, the delay should be very stable.  Any variation will be due to noise or limitations inside the counter.  The resulting ADEV graph will show you the best results you'll ever see from that counter.

A good rule of thumb is that the ADEV @ 1 sec. should be about the same as the minimum time interval resolution of the counter.  For the 53131A, that would be about 5e-10 @ 1 sec.  Since most good quality counters exceed their specs, you might do a bit better.

If your results are much better than spec, it means your counter is making multiple measurements and averaging them.  You have to disable that 'feature' before any of your measurements will make sense.  If your results are much worse than spec, you may have a setting wrong on your counter or a fault in your counter or your 8040.

This isn't a long test.  If you watch the ADEV graph as the data is being collected, you'll see that the 1 sec. value flops around to begin with, but stabilizes within a few minutes.  You can stop the test there.  Sometime when it's convenient, you can let the test run longer to see how far down you can go.  Eventually, the graph will become a flat, horizontal line, but that's usually so far down that there isn't any point in measuring it.

Ed

[ArthurDent]

The problem with comparing the Rb oscillator you got against your HP53131A is that you don't know how close either one is to being correct. My HP53132A also has the 010 high stability timebase like yours but I almost always use the 10Mhz from one of my GPSDOs as the external timebase. You will notice that in the photo of my HP53132A there is a pushbutton switch under the channel 3 input connector that I installed to replace the switch on the back because I got tired of reaching around back to turn the counter off. Just using the standby switch leaves the OCXO and the power supply on which I don’t like. When I bought the counter I had to replace the filter caps in the power supply that were starting to bulge and the fan that had gotten noisy due to being left on for a long time.

The OCXO you have will age(drift) at the rate of 1.5x10E-8/month so the longer it has been since you last calibrated the timebase, the more inaccurate it will be. For checking an Rb oscillator you really need a standard that is more accurate and stable than the option 010 timebase in the counter and more accurate than the Rb you’re testing. The easiest solution is a GPSDO.  If you have a GPSDO you can easily calibrate the option 010 timebase following the directions in the service manual but using that calibrated internal timebase that still won’t give you the stability and accuracy you need to check an Rb oscillator.

Both Rb standards I posted photos of have been modified and the Quartzlock unit originally had an OCXO which I replaced with the LPRO-101. The adjustment is the standard external pot arrangement shown in the LPRO-101 manual which is much easier to use than the internal adjustment. That the Rb unit you bought doesn’t show any means of adjustment make me wonder what level of accuracy the builder was requiring from the unit as they should occasionally be recalibrated against a higher standard.  Here is a photo of another LPRO-101 Rb oscillator I built showing the adjustment locations.

[notfaded1]

@ArthurDent Did you look closely at my pictures... The unknown Rb has a bunch of pots for adjustments mounted on top of the wood box in a strip.  Since right now I have no idea what they do I'm not touching them since the unit does lock.  My new enhanced GNSS GPSDO is on its way from Hong Kong now from BG7TBL.  It should give me a good idea how close I am.  The two stand alone Rb are very close to each other.  I haven't done it yet but I'm planning to upgrade one of my symmetricom sync servers with the inputs and outputs.  I have two of them one has Rb osc and I  already received the stacked BNC females for the addition from digikey... They weren't cheap at all the double BNC.  I was surprised at the price.  See pic with red circled pots.  Two of the wires are labeled heater???

Bill

[notfaded1]

Yes, much better.  But you you might be able to do even better.  You should measure the noise floor of your 53131 to see if it's working up to spec and to see what results you can expect from it.  This measurement is for Time Interval.  I don't know if the noise floor in frequency mode is the same or not.  That's a question for the theory guys. 

All you need for this test are two pieces of BNC cable and a BNC T adapter.  Connect the T adapter to Channel 1.  Set the impedance of Channel 1 to 1M ohm.  Connect a BNC cable from one side of the T adapter to Channel 2.  Make sure the cable can't flop around and make sure that the temperature won't change during the test.  The length doesn't matter.  A meter or two is fine.  Set the impedance of Channel 2 to 50 ohm.  Feed a square wave from your 8040 to the open connector on the T adapter.  I think the 8040 has a 1 PPS output.  That would be perfect.  Set the triggering on both channels appropriately for the input signal.  Now measure Time Interval and collect the data with Timelab.  Since there's nothing between Channel 1 and 2 but a piece of cable, the delay should be very stable.  Any variation will be due to noise or limitations inside the counter.  The resulting ADEV graph will show you the best results you'll ever see from that counter.

A good rule of thumb is that the ADEV @ 1 sec. should be about the same as the minimum time interval resolution of the counter.  For the 53131A, that would be about 5e-10 @ 1 sec.  Since most good quality counters exceed their specs, you might do a bit better.

If your results are much better than spec, it means your counter is making multiple measurements and averaging them.  You have to disable that 'feature' before any of your measurements will make sense.  If your results are much worse than spec, you may have a setting wrong on your counter or a fault in your counter or your 8040.

This isn't a long test.  If you watch the ADEV graph as the data is being collected, you'll see that the 1 sec. value flops around to begin with, but stabilizes within a few minutes.  You can stop the test there.  Sometime when it's convenient, you can let the test run longer to see how far down you can go.  Eventually, the graph will become a flat, horizontal line, but that's usually so far down that there isn't any point in measuring it.

Ed

Ed here's the results from your suggested setup with the BNC T and time interval with 75Ω cables for now... my new cables came today I just have to pick them up.  That looks pretty good on internal 010 oscillator a little better than 5 right?

Another thing I have noticed... until the Rb locks it's frequency moves around a lot... I can see it being beneficial to watch the process with timelab on startup although ADEV wouldn't seem like the measurement for that???  Whatever the deal with Rb is... it seems it's a temperature related thing.  Considering there are wires labeled heater maybe it's getting up to temp slowly and that's why it takes longer to lock?  That makes me think something may not be adjusted the way it should be... since there are so many pots on top until I open the wood box I won't know anything.  I'm timing the lock right now.

[ArthurDent]

@ArthurDent Did you look closely at my pictures... The unknown Rb has a bunch of pots for adjustments mounted on top of the wood box in a strip. 
Bill

Those adjustments aren't part of the oscillator but are a filter on the output. They look like the 10.7Mhz i.f. transformers that they used in F.M. radios. There is just a string of them in series, adjusted to 10Mhz to filter the output. They have nothing to do with the actual oscillator.

https://www.minikits.com.au/image/cache/catalog/components/ift455-03-a-320x320.jpg

[notfaded1]

It looks like right around 2 hours and 15 minutes the Rb started going green lock and then back to red unlocked.  Now around 2.5 hours it's staying green.  It went back and forth for while between red/green until it's now just solid green.  Strange behavior but it seems like some temperature thing to me.  The frequency is now stable again and I can feel the outside of the metal case is just mildly warm like it was after running for two days.  That gauge says 84.4°F.  I think the highest it goes to is around 87 after being on for longer.

My 8040 on the other hand locks in less than 5 minutes so yeah... there's a difference but from what I've seen so far the big box once it locks is pretty stable.

Bill

[james_s]

So a final update on the insulating the OCXO experiment and then I'll stop polluting this thread with it.

I let it runs all day after my last update and it eventually settled at 136.2F and stopped increasing, actually it was slowly oscillating between 136.2 and 136.1 so I assume that's as high as it would go and terminated the experiment. I very much suspect that is the internal setpoint and what happened is the added insulation allowed the outside of the casing to gradually reach the temperature inside it as heat did what heat does and flowed from the hotter region to cooler regions. I've never taken one of these apart myself but I've seen a few teardowns of similar units and it doesn't seem they have anything exotic inside, there is only so much insulating you can do without making the can excessively large.

I think it would be interesting to compare the frequency stability between OCXOs run in various conditions. Free air in a still and climate controlled room, free air in an uncontrolled space like a garage or outside, inside a closed box, inside a well insulated box, inside a DIY second oven, etc. Of course I'm sure manufactures of these oscillators have already studied all this at length.

[notfaded1]

Here's the ADEV of the time interval on the unknown Rb... looks a little better.  I'm going to overlay the 8040 time interval next. 

[FriedLogic]

Frequency and Time Interval (phase) are two sides of the same coin.  You can convert phase to frequency with no issues, but there are some problems going the other way.  When it comes to measuring ADEV, there are a few differences.  Most notable is the difference in slope of the resulting graph.  In your graph, it takes about two decades of frequency for the graph to drop one decade in ADEV, i.e. the slope is -0.5.  If you measure Time Interval, the slope is -1 which causes the graph to drop faster.  This gets you into the low values that we're really interested in sooner.  I've seen graphs of OCXOs that didn't look impressive because the measurements were frequency based and by the time the graph got down to the good stuff, the crystal aging was already causing the results to rise.

I'm not quite sure what you mean there - the way that frequency counters normally work is to measure time and then convert it to frequency anyway. Any difference is just down to how they do it.

The 53131A calculates the frequency from a large number of measurements taken across the gate time, which effectively filters the result, affects the apparent stability (and therefore the ADEV), and often results in an offset which will also change with the gate time.

Counters that use single start and stop measurements at each end of the gate time would give results much closer to time interval measurements.

Time interval measurement is used a lot for precision applications, but how much of a difference it makes all depends on the particular application.

The relative effect of time interval errors goes down with time. An error of 1 nanosecond over 1 second is equivalent to 1E-9, but only 1.157E-14 over 1 day.
With a frequency error the time accumulates. An error 1E-12 is only 1 picosecond over 1 second, but 86.4 nanoseconds over a day.

[edpalmer42]

Ed here's the results from your suggested setup with the BNC T and time interval with 75Ω cables for now... my new cables came today I just have to pick them up.  That looks pretty good on internal 010 oscillator a little better than 5 right?

Yes, that looks good.  The result isn't due to the Option 010 oscillator, it's inherent to the counting mechanism of the counter - specifically the resolution of the timing circuit - 500 ps in this case.  So since that matches up with the counter's specs, you now have confidence that the counter is working properly.  It's also interesting to note that when you were measuring frequency, the ADEV at 1 sec. was about the same.  Looks like the noise floor is the same whether measuring frequency or phase.  The only difference is the slope.  You also know that anytime you see a number at 1 sec. that's better than 5e-10, something's wrong.  (Hint, hint) 

Another thing I have noticed... until the Rb locks it's frequency moves around a lot... I can see it being beneficial to watch the process with timelab on startup although ADEV wouldn't seem like the measurement for that???  Whatever the deal with Rb is... it seems it's a temperature related thing.  Considering there are wires labeled heater maybe it's getting up to temp slowly and that's why it takes longer to lock?  That makes me think something may not be adjusted the way it should be... since there are so many pots on top until I open the wood box I won't know anything.  I'm timing the lock right now.

What you should see is the frequency sweeping up and down over a few hundred Hertz.  As the unit heats up, the sweeps will get closer to the nominal output frequency.  Once the sweep reaches the right frequency, lock becomes possible if everything else lines up.  Until it locks, ADEV doesn't make sense, but watching the frequency graph during the warmup will show the sweeping behaviour.

It looks like right around 2 hours and 15 minutes the Rb started going green lock and then back to red unlocked.  Now around 2.5 hours it's staying green.  It went back and forth for while between red/green until it's now just solid green.  Strange behavior but it seems like some temperature thing to me.

Yeah, something's not right.  Once it locks, it should stay locked.  It's almost like one of the heaters is dead and it's just absorbing heat from around it until it kind of gets warm enough.  I predict that Rb is either headed for the repair bench or the scrap heap.  But the rest of the unit should be okay.  Just drop another Rb into the case and you'll have a nice bench standard.

Ed

[edpalmer42]

Frequency and Time Interval (phase) are two sides of the same coin.  You can convert phase to frequency with no issues, but there are some problems going the other way.  When it comes to measuring ADEV, there are a few differences.  Most notable is the difference in slope of the resulting graph.  In your graph, it takes about two decades of frequency for the graph to drop one decade in ADEV, i.e. the slope is -0.5.  If you measure Time Interval, the slope is -1 which causes the graph to drop faster.  This gets you into the low values that we're really interested in sooner.  I've seen graphs of OCXOs that didn't look impressive because the measurements were frequency based and by the time the graph got down to the good stuff, the crystal aging was already causing the results to rise.

I'm not quite sure what you mean there - the way that frequency counters normally work is to measure time and then convert it to frequency anyway. Any difference is just down to how they do it.

The 53131A calculates the frequency from a large number of measurements taken across the gate time, which effectively filters the result, affects the apparent stability (and therefore the ADEV), and often results in an offset which will also change with the gate time.

Counters that use single start and stop measurements at each end of the gate time would give results much closer to time interval measurements.

Time interval measurement is used a lot for precision applications, but how much of a difference it makes all depends on the particular application.

I'm not quite sure what you don't understand.  If it's how there can be problems converting frequency into phase, it's because, as you state, frequency is an average over the gate time, ie. a filter, and so, destroys information that prevents a complete, precise conversion to phase.  Phase however, is a series of instantaneous measurements that, if you make enough of them, can be precisely converted into instantaneous frequency measurements.  When measuring frequency, if you measure every cycle, you have all the information necessary to do a precise conversion, but that's almost never done.  What this all means in practice is another matter.

Ed

[notfaded1]

Ed-
Here's the time interval of both Rb overlaid.  It looks pretty dang good I think!    It's also nice to know the counter is working correctly.  When my GNSSDO comes I'll have more to compare to.  I need to decide which of the two syncservers I want to upgrade too.  The one without the Rb actually works better.  I'm tempted to try and move the Rb from the one to the other too or maybe better swap the good working one's motherboard into the Rb model and move the Rb onto it.  This way it'll have a label on the back that shows the -Rb in the model number.  Then I'll have two GPSDO's to compare to.  I like the idea of feeding the GPSDO 10MHz into the counter replacing it's internal osc to better measure the Rb's.  Then I'll better know how exact the Rb's are.  I'm thinking they're pretty close because they agree with each other.  Thanks for your help.

I turned the unknown Rb off again to cool down.  Next I should open it up since something must not be right for it to take 2.5 hours to lock solidly.  It would be nice if it would lock in under 5 minutes like the Datum 8040 does.  I'm picking up my new cables too so I can ditch the 75Ω BNC cables for this.

Bill

[FriedLogic]

Another thing I have noticed... until the Rb locks it's frequency moves around a lot... I can see it being beneficial to watch the process with timelab on startup although ADEV wouldn't seem like the measurement for that???  Whatever the deal with Rb is... it seems it's a temperature related thing.  Considering there are wires labeled heater maybe it's getting up to temp slowly and that's why it takes longer to lock?

No, ADEV is not the tool for that, but I think that Timelab can also show a normal frequency plot which is more suitable. ADEV is better for showing some things, and standard linear plots of frequency, temperature, etc., against time are better for showing others. You need both to get the best idea of what's going on - it's not a case of one or the other.

Some of the manuals for rubidium oscillators give a good explanation of what's going on. There are some Efratom ones here:
http://ftb.ko4bb.com/getsimple/index.php?id=manuals&dir=02_GPS_Timing/Efratom
Basically, the frequency is swept back and forth looking for the resonance frequency to lock on to. This won't happen until it's all got to temperature.