GPSDO brands?

[edpalmer42]

I reached out to another Time-Nut to get another opinion on this topic.  After reviewing the thread, he agreed with Texas and Tomato that a short test isn't enough to draw any conclusions regarding the performance of a GPSDO because there are too many things that can go wrong.

As I thought about that, I realized that my experience with GPSDOs isn't that great.  Because of that, I just haven't seen the problems that you guys have.  To me, a GPSDO is just as stable and reliable as a quartz oscillator.  They just work!  So I couldn't understand why you were being so negative.

And since they just worked, and basically worked the same, I got bored with them and stopped buying them.  That left me with older, perhaps higher quality units that were more likely to be stable and reliable.

So the end result was that my idea of a possible quick way to rank GPSDOs was based on a limited data set that didn't reflect reality.  Sorry if I caused any confusion or bad feelings.

Homedad, after totally hijacking your thread, it looks like there's no easy way to rank GPSDOs.  Personally, I've always thought that my Z3801A was the best one I've got and slightly better than my Trimble Thunderbolt.  But they've both gotten rather expensive lately.  The only two 'newer' GPSDOs I've got are the Trimble UCCM, and the NEC unit, both of which are still available on ebay.  Both units tracked well, but I find that the outputs have more frequency/phase noise than the Z3801A although they both look good on my jitter test - which I still think has value even though it's not a definitive test.

I find that other features aren't very important on a GPSDO because, for me, they're just install and forget.  YMMV    One other thing I'll say is that you can cripple a good GPSDO by having a poor antenna setup.  Work hard to get a location clear of obstructions and with good visibility towards the equator.  That's the direction where you'll see more satellites.  It becomes more important as you move away from the equator.

Ed

[cdev]

How can one test a GPSDO 'device under test' without a better one to compare it to?

[texaspyro]

How can one test a GPSDO 'device under test' without a better one to compare it to?

Assuming you mean the output signals, you can't.   You need to compare it to some higher-speced device.  An HP-5071A with the high performance tube works fairly well.  Or a maser.   

[David Hess]

How can one test a GPSDO 'device under test' without a better one to compare it to?

A comparison to a local atomic standard including a rubidium will reveal problems in the GPSDO disciplining.

[texaspyro]

A comparison to a local atomic standard including a rubidium will reveal problems in the GPSDO disciplining.

That depends upon the quality of the atomic standard and of the GPSDO.   A lot of rubidiums (particularly surplus telecom ones) are worse than a lot of GPSDOs.   The HP-5065A rubidium is an exception...

[cdev]

Oooooooo

 

A comparison to a local atomic standard including a rubidium will reveal problems in the GPSDO disciplining.

That depends upon the quality of the atomic standard and of the GPSDO.   A lot of rubidiums (particularly surplus telecom ones) are worse than a lot of GPSDOs.   The HP-5065A rubidium is an exception...

[awallin]

How can one test a GPSDO 'device under test' without a better one to compare it to?

Assuming you mean the output signals, you can't.   You need to compare it to some higher-speced device.  An HP-5071A with the high performance tube works fairly well.  Or a maser.

you can try with three similar-ish units, and three-cornered-hat analysis. with only one or two GPSDOs you can't do much - get three or more
http://www.wriley.com/3-CornHat.htm

[edpalmer42]

You guys are so pessimistic!!

You can confirm basic functionality by comparing two GPSDOs to each other.  If they are both locking to GPS, the average drift between them will be zero.  The signal will wobble around a bit due to noise and differences between the two devices.  Since both devices are moving around, noise measurements won't be nice, but the average frequency will still zero out.  You need a decent time interval counter and the ability to capture the data to confirm this.  You might be able to use a counter with less functionality, but it would be painful.

If your maser is in the shop and you loaned your 5065A to your brother-in-law (what were you thinking??), the Stanford PRS-10 is worth considering.  Even the Efratom FRK or its military cousin the M-100 give surprisingly good results.  All three will give you ADev results that dip well into the e-13 range.

There are no reasonably priced FRK units on ebay right now, but the M-100 and PRS-10 are, or have recently been, available for a not too outrageous price.

Ed

[texaspyro]

You guys are so pessimistic!!

Nope,  we're time nuts.   We want our shineys to be all shiney down to the atto-hertz.

[edpalmer42]

You guys are so pessimistic!!

Nope,  we're time nuts.   We want out shineys to be all shiney down to the atto-hertz.

Yeah, I've noticed....    I can't handle that.  I start to get lightheaded when I go below 1e-13.   

[texaspyro]

You guys are so pessimistic!!

Nope,  we're time nuts.   We want out shineys to be all shiney down to the atto-hertz.

Yeah, I've noticed....    I can't handle that.  I start to get lightheaded when I go below 1e-13.   

Poor thing...  time to climb back into your crib with your 555 timers and RC oscillators until you feel better...     Take two cesium capsules with a glass of rubidium hyrdroxide and consult TVB in the morning.

[Housedad]

You guys are so pessimistic!!

Nope,  we're time nuts.   We want our shineys to be all shiney down to the atto-hertz.

Does that fall into the category of Obsessive-Compulsive Disorder?