[time-nuts] Fury - Rubidium - PIS

Scott Mace smace at intt.net
Tue Jul 27 23:10:03 EDT 2010

I found that on the FRS-C and the X72 the frequency output was not at 
all linear with respect to EFC, which made things worse with large 
dacgains.  After _MANY_ hours of watching and playing with different 
settings, I came to the conclusion that you just need to be patient when 
using Fury+Rb, don't expect it to converge like an OCXO, but it will 
over a few hours and give you a good result as long as you have stable 
power and shield it from air streams.  High dacgain and efcs give you 
the false hope that it will converge fast, but as the unit settles the 
loop will fail, and it will start fighting with itself.

The SYNC:IMME can be useful when trying different loop settings.

I found some of my notes for various things I ran with the Fury.

LPRO-101 currently running
DAC GAIN: 1000
temp co: 0.00
aging co: -0.00558
phase co: 35

lpro orig
lpro cd 92
dac gain 1000
efc scale 1.30
efc damping 2
efc slope pos
phase co 30

lpro 2
coarse dac 95
dac gain 1300
efcs 1.30
efcdamping 5
ocxo slope pos
agin comp 0.05333
phaseco 35

datum ocxo #1 (From RFG-XO)
coarse dac 64
dac gain 80
efcs 4
efcdamping 10
ocxo slope pos
agin comp 0.27086
phaseco 35

datum ocxo #2
coarse dac 64
dac gain 100
efcs 2.8
efcdamping 10
ocxo slope pos
agin comp 0.14901
phaseco 35

datum ocxo #2 from RFG-XO In an RFG chassis
cd 69
dg 30.00
efcs 12.00
efcd 5
ocxo slope pos
aging comp -0.06559
phaseco 40
tempco was a major issue with the datum ocxo
keeping it in the RFG-XO chassis helped

Efratom FRS-A
DAC GAIN: 5000.00
TTL output, pad to below +13dBm, FRS-A needs to settle, open bench
air current problems.

FRS-A #2
dac gain: 4000
efcs: 1.2
efcdamping: 25
phaseco: 15
hard time recovering, from coarsedac bumps

dac gain: 2000
efcs 1.2
efcdamping 25
phaseco: 15
jumps, tempco issues.


On 07/27/2010 09:16 PM, SAIDJACK at aol.com wrote:
> Hi John, Brian,
> actually we set D to 0, and use P and I gains.
> Yes, the DACGAIN is an overall gain of the loop output - to normalize for
> different oscillator voltage/frequency sensitivities.
> I forgot that the DACGAIN is limited to 10,000, so instead of setting it to
>   20,000 one can set it to 10,000 and multiply the P and I parts both by 2x,
> that  gives pretty much the same effect as setting DACGAIN to 20,000.
> Please note that aging and tempco compensation may then take a couple of
> days to fully settle again.
> I have not tried such high gains since we don't have any oscillators  here
> with such small F/V range, please advise what you find out.
> bye,
> Said
> In a message dated 7/27/2010 19:04:36 Pacific Daylight Time, jfor at quik.com
> writes:
>>   That part I understand (your drawing), its a basic phase lock loop.
>>   What I am having trouble with is the Fury's commands relationship.
> OK.  Sorry for the BW.
> Basically you tune a loop by starting with the P, I,  and D set to zero.
> You slowly crank up the P until it starts to become  unstable. (Put in a
> small step perturbation and look at the response for  ringing) Then crank
> up the D until it stabilizes, then crank up the P  again. When you have got
> a stable fairly well performing loop, you  introduce some I. You may have
> to tweek P and D to keep  stability.
> It looks like your system has an overall gain (DACG) and a P  and D
> controller gain. This is not uncommon to avoid switching a bunch of  caps.
> _______________________________________________
> time-nuts mailing list -- time-nuts at febo.com
> To unsubscribe, go to https://www.febo.com/cgi-bin/mailman/listinfo/time-nuts
> and follow the instructions there.

More information about the time-nuts mailing list