Why are rubidium frequency standards so expensive now?

[JBaughb]

These were like $30-50 a couple years ago, pulled from old test equipment. Now they're selling for $200+ Why has demand increased so drastically? Has something specific happened, like a novel new use for them, or maybe supply is limited compared to a few years ago?

[PedroDaGr8]

These were like $30-50 a couple years ago, pulled from old test equipment. Now they're selling for $200+ Why has demand increased so drastically? Has something specific happened, like a novel new use for them, or maybe supply is limited compared to a few years ago?

Much of these rubidium standards were pulled from decommissioned cell phone cabinets. There was a flood on the market due to the change over from 3G to LTE, many cabinets were upgraded and the old rubidium standards were replaced. These found there way on ebay. With a huge supply, prices dropped. Now that most of the conversion work is done there is less supply and prices have gone back up.

[JBaughb]

That makes so much sense. Thank you. Maybe I should wait till 5G comes out....see if prices drop again 

[edpalmer42]

There were a few reasons why the Rb units were cheap.

There was the glut of units that Pedro mentioned.  Some of them had been in service for a number of years and might have been approaching end of life.  It's also worth remembering that the old equipment was treated like scrap metal so there's no telling how badly it was abused.  Hard to believe that anything survived being dropped, having things dropped on them, being left out in the rain for who knows how long, etc.

Actually, compared to buying a new Rb standard, paying a few hundred dollars is still a good deal.  I guess we got spoiled!

Ed

[RickBrant]

I got a Trimble Thunderbolt while they were $95. 

The monitor program shows it's running at a few tens of ppt.

The most fun I've had with it so far is to run its 10 MHz output into the "external reference" input of my freq counter. Now the freq counter has a timebase accurate to a few tens of ppt.

[ivan747]

These were like $30-50 a couple years ago, pulled from old test equipment. Now they're selling for $200+ Why has demand increased so drastically? Has something specific happened, like a novel new use for them, or maybe supply is limited compared to a few years ago?

I say it's Dave's fault! He made a video about it and they became popular and now they are charging a lot for them 

 

[mswhin63]

I say it's Dave's fault! He made a video about it and they became popular and now they are charging a lot for them 

 

Yeah co-incidental that at the release of one of the earlier EEVBlogs on the Rubidium the price when up very soon after. 

[JBaughb]

As far as I can tell, there are three units that keep popping up. the Symmetricom SA22c, Stanford Research PRS10, and the FEI model FE5680A.

I pulled up the data sheets for each and Im reading through them but does anyone have any info on which of these are preferable. Im sure a lot of it has to do with the specific user and their needs but are there any general things to watch out for or interesting features available on one but not the others?

Thanks.

[atferrari]

Why do they have end of life?  Aging of what, actually?  Nothing that you could overcome?

[Nuno_pt]

It's easier and cheaper to build and GPSDO.

[VK5RC]

There are some articles about how to 'rejuvenate' Rb cells that have deposited the Rb badly.
eg http://www.vk3um.com/Rubidium%20rejuvination.pdf
This is from VK3UM's excellent website this is his page on Rb Clocks alone http://www.vk3um.com/Rubidium%20Standard.html

[atferrari]

Thanks blueskull and VK5RC for replying.

After reading two hours ago the user manual of one, the mumbo jumbo of information is overwhelming. 

The way it is described is a quite convoluted implementation. Wow.

[bingo600]

As far as I can tell, there are three units that keep popping up. the Symmetricom SA22c, Stanford Research PRS10, and the FEI model FE5680A.

I pulled up the data sheets for each and Im reading through them but does anyone have any info on which of these are preferable. Im sure a lot of it has to do with the specific user and their needs but are there any general things to watch out for or interesting features available on one but not the others?

Thanks.

I'd chose this way anytime:

1: PRS10         (I actually have one, with 1PPS diciplining)
2: SA22c
..
..
..
X: FE5680      (I actually have 4, and killed the 5'th reversing power  )

I also have an X72 , quite a nice small unit. But unfortunately w. v4.x firmware so no 1PPS diciplining.

/Bingo

[pez]

Why do they have end of life?  Aging of what, actually?  Nothing that you could overcome?

The Rb clocks works as a PLL. The Rb cell resonates at a certain frequency (for CPT clocks, 850.2nm+6.8GHz). At resonant frequency, the transparency of metal vapor goes maximum, therefore, by exciting Rb with such a coherent light source (laser) with 3.4GHz modulated on it, you can observe CPT effect with a photo diode.

When it ages, Rb atoms embeds itself into glass envelope, making it less dense, until the CPT effect is weak enough to be undetectable, then the Rb clock fails.

Hello,

Thank you very much for this most useful explanation!

Well, would you additionally clarify for me at least,
please, if this embedding of the atoms into the glass
also happens when the device is off ?

Sincerely,

[Bud]

It's easier and cheaper to build and GPSDO.

GPSDO has a hassle of requiring an antenna, likely an outside one and longer locking time.

[thinkfat]

Well, would you additionally clarify for me at least,
please, if this embedding of the atoms into the glass
also happens when the device is off ?

Sincerely,

What usually fails in these rubidium clocks is the lamp. Well, other components certainly fail more frequently due to the heat, but the lamp is what ultimately defines the lifetime. These compact rubidium clocks work through a rubidium vapor discharge lamp. This is a no-filament lamp which basically consists of a glass bulb with some metallic rubidium inside. The glass is evacuated. Probably there's also an additional noble gas in the lamp that facilitates the plasma ignition. If you heat the lamp to, like 110°C and feed it with enough RF from a 80-100 MHz oscillator, the rubidium vapor ignites and emits light at a certain wavelength. So, these lamps run pretty hot, obviously.

While they burn, rubidium atoms embed into the glass and the rubidium vapor pressure slowly drops until the energy emitted at the characteristic spectral line (768nm) is too small to feed the detector inside the microwave cavity. Note that there will still be quite a lot of metallic rubidium inside the lamp, but you'd need more heat to increase the vapor pressure to a point where the "physics package" works again. You could try tweaking the heater circuit of course, but since the heater elements are typically transistors, the maximum junction temperature and the thermal coupling resistance puts a limit to what you can achieve.