[time-nuts] 9390 GPS RX

Ed Palmer ed_palmer at sasktel.net
Mon May 27 09:32:37 EDT 2013


The first thing I'd do is let it cook for a few days.  Monitor it 
carefully to see if it's drifting and in which direction.  If it's 
moving, just leave it alone and see how far it will go.  Be sure to 
document things so you know if it's gradually slowing down.  Once it 
really is stable, you can take stock of where you are. Unfortunately, 
when you're playing with errors this small, it takes a long time to make 
measurements that you can have confidence in. You might have to look for 
a drift in the order of 1e-10 per day. It shouldn't be drifting, but I 
wouldn't put much faith in it until I'd tested it.

This is one of the times that you want to check two or three ways to 
make sure that the frequency really is .05 Hz low.  You mentioned that 
your counter reads .07 Hz high so you're getting near the resolution 
limit of your counter.  A .05 Hz error is an error of 5e-9 which is huge 
for a Rb standard.  Is it giving you a locked indication?

Is your house standard a GPSDO or another Rb?  If so, and if you haven't 
already done so, use your 5370 in time interval mode to compare the 
frequency of the FRK against your house standard.  This will 
(relatively) quickly confirm the direction and size of the frequency error.

I'm not familiar with the 9390, but it wouldn't hurt to plug the FRK 
back in and see what happens.  Obviously, nothing's going to break. If 
the 9390 complains, that will be more confirmation that the FRK still 
has a problem.

Once you're sure that the unit is stable, if the frequency is still off, 
take a look at the C-field coil.  The range of the C-field adjustment is 
2e-9 which is comparable to the error you measured. According to the 
manual, more C-field current gives you a higher output frequency.  They 
actually mention that the most common C-field fault is an open coil.  
This would cause the frequency to be low, but the manual doesn't say by 
how much.  Check the C-field circuit to see if it's working properly.  
The manual tells you how.

Did you notice that taking the thing apart was the LAST thing I 
suggested if nothing else worked? :)

Ed


On 5/27/2013 3:13 AM, Mark C. Stephens wrote:
> Hi Ed,
>
>
> All looking good with EFC reconnected.
>
> I will chuck it on time-lab shortly.
>
> Except the output frequency is -0.05Hz out.
>
> Do you think the 9390 EFC will pull it in O/K?
>
>
> -marki
>
>
> -----Original Message-----
> From: time-nuts-bounces at febo.com [mailto:time-nuts-bounces at febo.com] On Behalf Of Ed Palmer
> Sent: Monday, 27 May 2013 1:33 PM
> To: Discussion of precise time and frequency measurement
> Subject: Re: [time-nuts] 9390 GPS RX
>
>
> On 5/26/2013 8:24 PM, Mark C. Stephens wrote:
>> Ed, I'd unplugged the Rb from the counter and into the timer for measurement.
>>
>> During this time it had dropped out of lock again.
>> That's why the Allan Deviation is so bad, it was sweeping.
>> But I didn't realise as there was no counter attached!
>>
>> I need to connect something up to the 'lock' pin for a visual indication of lock to avoid further embarrassments..
> I've got manuals for both the FRK and the 'FRK (H or L)'.  The FRK says that the lock input can handle up to 70V @ 50 ma.  The other one doesn't say.  You should be able to run an LED with a dropping resistor from the
> +24 supply.
>
>> I have cracked it open again and revisited the crystal board.
>>
>> I found that the centre frequency was a tad low so I upped C12 to 27pf
>> (from 22pf) With 6v EFC and the trimmer cap in mid position its dead on 10Mhz.
> Remember that it doesn't have to be dead center.  After all, that's why the thing sweeps in the first place.  It's looking for the signal and it will find it if it's anywhere in the sweep range.
>
>> And I slowed down and waited (thanks Ed..) overnight to see if it drifted off anywhere else.
>> But at 10Mhz it has stayed.
> Very nice! :)  More thought, less action.  Move as though you're defusing a bomb.  Ideally, you know what you want to do, why (with at least two seperate tests to confirm your reasoning), and what should happen when you make the change.  Adjustments are bad enough, be triple sure before you touch the soldering iron.  Yes, it's overkill for most situations, but if major dollars or an irreplaceable device are at stake, it will pay dividends in the long term.
>
>> I will reconnect EFC later on today (being a work day) when I get some time to move.
> Work?  What is this Work thing you speak of?  Oh, yes, I think I do have some vague recollection from years past.  I think I tried it but didn't care for it.  :)
>
>> I also have the sneaking suspicion the external adjustments have been 'got at'
>>
>> Looking at the trimmer pot through my mag lamp for c-field adjust, I can see a lot of scratches and wear.
> I wouldn't be surprised to find that someone played with the external adjustment in a misguided attempt to compensate for the crystal drift.
> Totally ineffective.  The C-field is at the wrong end of the system.
>
>> Hopefully the Oscillator current capacitor C11 and the temperature adjustment hasn't been touched.
> You mean for the lamp?  They're pretty hard to get at.  I had to adjust the lamp temperature on one unit and had a devil of a time finding a screwdriver that could reach it.  The reading on the lamp voltage monitor will tell you if those adjustments are correct. Remember, move slowly and carefully.  :)
>
>> Anyway, Wish me luck ;)
> Always :),
>
> Ed



More information about the time-nuts mailing list