Do these Ebay sellers of the rubidium standards have some means of checking the accuracy of the 10mhz output? I assume there is some type of internal cal trimmer and how do I know hasnt been tweaked? Are they just selling surplus pulls from Cell Phone industry and trusting they are accurate and have sufficient remaining life or do they have lab standards to check performance before selling?
I didn't ask the seller; I did my own research. The frequency of the VCO is locked to that of a Rubidium lamp. It takes a few seconds for the lamp to warm up, and then the unit gives a "Lock" output. Upon cold-start, I hear a beat note between it and 10 MHz WWV. In 10 or 20 seconds, it pulled-in and I heard no more beat note. A Google search will find a plethora of information. I have mine enclosed in a chassis with decoupling on the DC input, LED monitoring, and output interfacing.
Originally Posted by K9TW
Bryan: Yes I assumed you could use WWV as your primary standard to compare it to. I have done that over the years with xtal calibrators. However it requires good strong WWV signal and how touchy the trimmer is. Can probably get it within a hz. I have played around using O-scope and audio spectrum analyzer software to look at the beat note.
I think I will email couple of the ebay rubidium standards sellers (I am leaning toward the LPRO-101) and ask them if they have a lab and if they use a primary standard to check the units they sell and what their standard is. Should be interesting.
I guess by using WWV at minimum you could tell if it someone had tweaked it and it was grossly off. That was my main concern.
As I think about it I suspect a OCXO would be accurate enough as a portable standard and less costly and not have to worry about the aging aspect.
Appreciate your response. I have more homework to do. Happy Easter to all. Terry K9TW
Last edited by K9TW; 04-08-2012 at 11:49 AM.
Terry, good post. I didn't even know about rubidium standards until I came across your post! I spent a couple of hours today surfing the web and reading up on both rubidium and gps freq standards. I guess you're taking a gamble with these pulls since you don't know how much life is left in the rubidium lamp. Anyway, I have a HP 8657D signal gen sitting here that I'd like to get a standard for, and I'm not sure if I can wait until Dayton. ;-) Keep us updated on what you end up... 73, Ron NR8O
Ron: Take a look at Iso-Temp OCXO model 134-10. Can find these on Ebay for around $50.00. I plan to look for one of these at Dayton in addition to the Rubidium. I think either would be good for ham service bench, but as always what will be your primary source to ensure the OCXO or Rubidium is accurate? It appears to me you have choice between WWV and GPS. If you get the Rubidium look on Ebay for a "rubidium lock". WA3IAC sells a kit that lights an LED to indicate when it has locked. It comes with PC board for around $6.00. Terry K9TW
Rubidium based sources go back over 30 years so I would think the newer ones, even if they are pulls, should be higher quality than the ones I dealt with 30 years ago.
Around 1979 I was involved in a project that required two coherent frequency standards for use at 18 GHz. We started with the 10 MHz sources but they were not stable enough. We went back to the manufacturer and had them create a 5 MHz version. To test their stability, we then took the 5 MHz sources and added battery operation and the necessary RF circuitry to get the signal to 18 GHz. One source would be used in a transmitter and the other in a receiver. Once the batteries were charged, we would disconnect them, so each unit ran independently of the other. We then took the 18 GHZ signals from each unit and ran them into a Mixer. What we expected was a stable DC note on the output. The 18 GHz output was then connected to a A/D converter and a computer (PDP-11/34) collected I and Q data over a span of 12 hours. We didn't have a problem with them drifting slightly from each other, as long as we could predict the drift. The drift data could then be included in the signal processing.
We found that although the drift was generally no more than 1 Hz, it was very difficult to predict. At the time, these sources cost around $10,000 each and were not very reliable. We had to swap sources many times before we got two that were reliable. Part of the reliability issue was that they had to work in aircraft, so the sets were run with then strapped to shake tables.
The company wanted me in Germany for the next 3 or 4 years so I had to pass my knowledge to someone else. I never really found out how they did. I do know that when I came back from Germany (1984) the program was still active. So either they made a lot of progress or they got really good at dancing around the issues.
I really don't think that many people have the need for anything as stable as these sources, bit, if you can get one for around $50 now, I would say it's a great investment. At those prices, why not buy two? If it goes bad, you can always swap in the spare.
73, Martin, K7MEM
Ash Fork, AZ
In my area, it seems that every pickup truck or SUV comes with one or more dogs. It's so common that I can only assume that the dog(s) must come with the vehicle. So logic tells me that, if you want to keep the truck for a long time, go for the multi-dog option. Otherwise, if the dog dies, you have to buy a new truck. I have five dogs (4 dogs as of 4/4/2013, RIP Katie), so I'm set for a few years.
Rubidium Standard Uncertainty
In reply to a few of the above, you cannot compare a rubidium standard to HF WWV once the rubidium standard has stabilised. There are too many frequency errors introduced by propagation into the over-the-air frequency to approach the native accuracy of the rubidium standards. The only comparisons that can be made are against caesium standards. This can be done either directly or via GPS disciplining. The latter will introduce about a half-magnitude error. However, for most amateur use the difference is inconsequential. I speak to this from having spent 10 years as an engineer in a metrology lab. We used a GPS disciplined frequency standard since we could not afford a caesium standard. It was considered traceable and in no need of any other calibration. It has been 10 years since I left that work but at the time I understood all of the "why" behind the uncertainty of the lab standard.
Thanks for the info guys, I'm learning a lot. So it looks like there are several options. The rubidium pulls on eBay, although from reading the time-nuts mailing list I missed out on the $39 (with shipping) pulls that a bunch of the guys were buying back in January. The least expensive ones now are in the $70-$80+ range, shipping included from China. There is a guy selling the Trimble Thunderbolts for $169 shipped. Another route might be to pick up a Navman jupiter T Tu60 GPS board, being sold by user nhbbobb1985 on eBay with interface board for $65, and phase lock it to an ovenized 10Mhz crystal oscillator. All overkill for what I personally would be using it for, but pretty neat knowing this kind of accuracy is available...
Exactly. I have my FE-5680A mounted in a chassis to allow for interfacing, with some added PS decoupling.
Originally Posted by K7MEM
$70-80 is still dirt cheap, compared to anything else you can find. With the interface board kit that fits in place of the TCXO in my TS850, I've spent less than the cost of Kenwood's TCXO, and have much better precision & stability. The interface board is smart enough to know when the external reference is lost, and enables the OEM xtal oscillator.
Originally Posted by NR8O
I have been eyeing one of these, but haven't pulled the trigger on one yet. My current reference is a HP OCXO and it gets me within 1/2 a Hertz on the Frequency Measurement Tests...
BTW, there is a FMT next Thursday - perfect opportunity to put your new rubidium source to work!
Signature line removed for your viewing pleasure.