U3s Frequency drift / stability #u3s #problem #wspr


JC - EA2KV
 

Hi ¡
I know that there are already lots of post about but I have been reading quite a lot these days and Im still not able to find out what happen with my U3S.  I have been playing during few days with several configurations and the U3S is beating me up.  The unit was on my drawer during long time because of this issue, but Im back again encouraged to make that working properly...

- Problem: 
I'm only working on WSPR mode.  After certain time (depending on the Cal steep config.) the TX freq gets out of the RX window; what is more, without stopping at any point inside RX winding,  Guess the the goal an stabilization inside the RX window.
I start always with a Ref. Freq 27.004.000 with good results......
Also, a reset of the Atmega has been done.
 
- Set up: 
     U3S board ver. 2. // Atmega  V 3.08
     PSU 5V from regulated feed.
     Si5351A (including small cooler on the crystal.
       6x  LPF.
     GPS controlled (SKM52)
     U3S config.
          Frame Start: 10 - 02
          Cal Step Time: 10 - 120 (tryed with several configs here and final result is teh same,  out of RX winding)                  Park mode: 2 - 150.000.000
          Sys Frq: 20.000.000
          Ref Frq: 27.004.000
     Monitoring my TX frq. with SDR

What Im missing?  thanks for any light you may give.

73 de EA2KV, José


Bruce K1FFX
 

Hi, Jose -

I might suggest lowering the Park frequency.  With my U3S unit (on 20m and 30m), I found it was initially drifting downward in
frequency.  When I set Park to 150,00,000, it was then drifting upward in frequency.  I experimented with 
different settings of Park, and ended up at something like 141,300,00 ... that worked well for my setup
(I also put a heat sink on the crystal on the synthesizer board and put some foam around the whole
synthesizer assembly ... have a look at Hans' Application Note on this topic.

73 -

Bruce K1ffx


Curt wb8yyy
 

Jose

Another aspect is needing several calibration cycles before the PLL is well calibrated.  in my case it was 5 or more. Since then my U3S is sufficiently stable from a cold start, with none of the advanced features you have. Do try parking up at vhf as suggested. I did not learn enough about this in the manual, but from messages here. Somewhere there is a message from Hans suggesting higher cal parameters for a newly installed unit or controller. It would be nice to have this instruction in the manual.

Curt


Alan G4ZFQ
 

I'm only working on WSPR mode.  After certain time (depending on the Cal steep config.) the TX freq gets out of the RX window; What Im missing?  thanks for any light you may give.
Frame Start: 10 - 02
Cal Step Time: 10 - 120

José,

Tell us your full configuration. What TX frequencies?
How many active frames, number of transmissions in 10 minutes?
Is it really doing a calibration cycle?

73 Alan G4ZFQ


JC - EA2KV
 

Hi Bruce,
Tks.  Will try with Park freq.
The heat sink is already installed on the crystal and the U3s is inside his black Al box.
The point is that the drifting seems never ending even outside RX windings.


Hi Curt,
Tks to you too.  The Unit has been running during three days. ,means lots of calibration cycles.
I even tried with Cal Step Time: 10 - 240,   or 30 - 240, you know.... faster freq ref change but never findind a point arround where to be +/- stable inside the RX window.

 73 - ea2kv


JC - EA2KV
 

Hi guys,
Back again,  let me explain the case in a better way.
The drift Im suffering is, as per this technical note: http://qrp-labs.com/images/appnotes/AN001_A4.pdf a "long term drifting",  my drift during transmision is decent.

More in detail:  
Quote from technical note.
  • Ultimate3S transmission sequence is configured such that there is a pause in transmissions at the end of each cycle, during which the GPS 1pps is used to measure the reference frequency of the Si5351A Synthesiser (27MHz crystal) and compute a correction in software.....".
Unquote.

In my case, the "computing correction in software" ocurring with every single calibration, which is normal in order to get the +/- right reference BUT, never stop moving the Ref. Freq value puting finally my TX out of any RX winddow.

Could this issue be a software problem with the ATMEGA??

Tks - EA2KV


Alan G4ZFQ
 

Quote from technical note.
* Ultimate3S transmission sequence is configured such that there is a
pause in transmissions at the end of each cycle, during which the
GPS 1pps is used to measure the reference frequency of the Si5351A
Synthesiser (27MHz crystal) */and compute a correction in
software/*.....".
José

Yes, this is how it SHOULD work.
But if it is not configured correctly it may not calibrate at all.

Do you see it calibrate?
What message is shown after the calibration?
Have you set the transmit frequency correctly, 1500Hz above "Dial" frequency?

If we know it is setup correctly then we can consider a fault.

73 Alan G4ZFQ


JC - EA2KV
 

 Hi Alan,
Answering your questions:

--> Do you see it calibrate?
Yes, U3S is calibrating.

--> What message is shown after the calibration?

After calibration what I get on the screen is the the Ref. Freq. result,

--> Have you set the transmit frequency correctly, 1500Hz above "Dial" frequency?

Yes,  and I see my reports on WSPRnet.

Could be the corruption of the Atmega,  I mean, a softwre problem rather than U3S configuration?.... Setup is above on my 1st post.

73 José, EA2KV


Alan G4ZFQ
 

--> What message is shown after the calibration?
After calibration what I get on the screen is the the Ref. Freq. result,
José,
Maybe your firmware does not also show +/-/= to show if it shifts the reference?
I guess the reference moves every time? Always one way?

Yes,  and I see my reports on WSPRnet.
? Could be the corruption of the Atmega,  I mean, a softwre problem rather
than U3S configuration?
Are you sure the GPS is connected correctly? Is the PPS going to the correct place? Is there a bad connection? Is the PPS and Data shorted?
Turn off the GPS correction and set the reference to 27.004 so that you get spots. Does it move a lot then?

I think it is unlikely that corrupted firmware would do this but I guess it must be a possibility.

73 Alan G4ZFQ


JC - EA2KV
 

Morning Alan,
With GPS correction set to "on", after calibration I can always see the +/-/= correction routine, and it is always one way, heading to "-".
GPS checked and connected correctly.
GPS correction is turned off and Freq.Ref set to 27.004, with this setup Im getting very good results, works perfect with very small drift and seeing lot of spots.
Will try to find some one here to load again the Atmega.
Tks for your help.
73 - EA2KV.


Alan G4ZFQ
 

With GPS correction set to "on", after calibration I can always see the +/-/= correction routine, and it is always one way, heading to "-".
José,

Have you tried GPS mode 3 and 4?

You have shown there is nothing wrong with the U3S, provided the PPS setting is correct and the GPS is operating correctly then that does just leave the Atmega firmware.

73 Alan G4ZFQ

GPS checked and connected correctly.
GPS correction is turned off and Freq.Ref set to 27.004, with this setup Im getting very good results, works perfect with very small drift and seeing lot of spots.
Will try to find some one here to load again the Atmega.


JC - EA2KV
 

Alan,
You hit the nail ¡¡
From instruction manual firmware V3.08, GPS Mode settings:
Quote:
#0 GPS is not used at all. The kit is in free-running mode, it will run from it's 20MHz crystal oscillator.
#1 For backward compatibility, this is how the kit worked in firmware version v2.03. The kit triggers on a change in the voltage level on the 1pps input. A timer is included to prevent it triggering on both rising and falling edges and to correctly detect the right edge, theoretically. This mode would probably be useful if the pulse width from the GPS is very short.
#2 The kit triggers on the rising (positive) edge of the 1pps signal. This is appropriate for most GPS modules, where the pulse width is 100ms for example.
#3 The kit triggers on the falling (negative) edge of the 1pps signal.
Unquote:

I was set #2,  now set to #1 I'm able to see the Ref. Freq +/-/=, meaning no negative drift all the time.

Guess that the SKM52 module has a different pulse behavior than other modules.

Alan, again, tks a lot and much appreciated your support.

73, José EA2KV