external 10 MHz reference phase noise requirements ?

Paul - WB5AGF

( I am way-out-of-my-league technically in this area so if I ask some really dumb questions please make allowances .... )

As I write this there is .... somewhere out over the vast Pacific Ocean ... a little package with my name on it ... containing a 10 MHz Piezo oven-ized crystal oscillator ('OCXO' - Model 2920136) that I purchased a few days ago on eBay.

My intent is to build-up a secondary frequency standard using the Piezo OCXO with input from articles that I've been able to locate (some I remembered from years ago) and also from exposure I had to this type of device long ago when in the Air Force (I was stationed at a base overseas and, in the next aisle from my duty station, there was a rack with a Sulzer double-ovenized crystal frequency standard).


My question has to do with the phase noise requirements that are
applicable  to the FLEX-6x00 external 10 MHz frequency reference :

  - What are they ?

  - What do they need to be so that the phase noise characteristics
     of the radio are not compromised ?


My understanding of how the radios use the external reference (based on Steve's & Tim's comments in other news-threads) is that on power-up the radios do a search for a frequency reference : first checking for a 10 MHz external reference, then (if there wasn't an external reference) looking to see if the GPS internal reference is installed and finally (if neither of the previous were found) the radio looks to its on-board TCXO (temperature compensated crystal oscillator).

Now my presumption is that the (very impressive) phase noise specs, which FLEX gives to the 6x00 radios, is based on the use of the on-board TCX0.

Does the external 10 MHz reference directly 'drive' the radio or instead is there a heavily-filtered PLL that is 'locked' to the external reference ? It occurs to me that a very 'slow-tracking' PLL would serve to filter out phase noise coming in on the external 10 MHz reference .... but I'm only guessing that such a filtering action is applied.

Information please.

- Paul, WB5AGF
  Garland, Texas

Steve-N5AC

Yes, everything 10MHz goes through the jitter cleaner.  We want the 10MHz reference's opinion on how often 10MHz would tick, but little else ;-)

Steve-N5AC

Hey, Paul, here's the scoop.  Since I know you're an EE, I'm going to use that to my advantage ;-)  The primary oscillator in the FLEX-6500/6700 uses what is commonly referred to as a "jitter cleaner".  This is a TCXO in a PLL with a very small loop bandwidth (think 10Hz range).  The phase noise characteristics of the crystal (in the TCXO) are where the radio derives it's phase noise characteristics.  The only place where the jitter cleaner is susceptible to reference noise is inside of the loop bandwidth.  So, if the phase noise characteristics of your reference are worse that the radio inside of 10-20Hz, your reference will dominate.  But, this is really not an issue, generally, as the phase noise characteristics we care most about are in the 500Hz-100kHz range anyway.

Paul - WB5AGF

Howdy Steve;

You don't say so explicitly but from your comments it sounds as though a 10 MHz external reference (when detected during the radio's start-up process) is run through the same 'jitter-cleaner ' as the internal OCXO would be ( if it was used ) - yes ?

- Paul, WB5AGF

PS -
The Piezo OCXO was in my mailbox when I went out and checked a few minutes ago. Now I've got to be busy designing the secondary frequency standard (a very stable DC source for the OCXO's control Voltage is required and I may put in a buffer amp to ensure load isolation).

Paul - WB5AGF

Thanks Steve;

      I gotta' get building ....

- Paul

Peter K1PGV

You know... I followed Steve's whole post, or, I thought I did. Right up until the last sentence: "But, this is really not an issue, generally, as the phase noise characteristics we care most about are in the 500Hz-100kHz range anyway." Can somebody take pity on a poor software engineer, and explain just a bit further? I get that if the phase noise within 10hz or so is greater than the radio's inherent phase noise, the phase noise of the reference will dominate. OK. But, WHY is this not really an issue? Sorry for what might be a remedial question. I promise I'll be extra-nice to the next person that asks what I consider a simple question about Windows drivers ;-) Peter K1PGV

Paul - WB5AGF

Hi Peter;

Think about why we have concerns about phase noise (in this situation it's when a source is being used in a piece of gear that's transmitting).

When another station (call it 'Station B') is listening, oh say more than a few kHz away from a transmitting station (call it 'Station A'), then it will be Station A's phase noise that 'smears' energy onto the chunk of spectrum where Station B is trying to receive that causes trouble (presuming that Station B's receiver has decent adjacent channel selectivity).

However if Station B is foolish enough to tune its receiver right up on Station A's transmitting frequency then it won't matter what Station A's near-in phase noise is because Station B's receiver is tuned right to Station A's transmitting frequency.

See ... it was easy !

- Paul, WB5AGF

Peter K1PGV

Thank you, Paul. Yes... Easy when you explain it so well. Clear now! Many thanks, Peter K1PGV