Fldigi Calibration

Jim W3IPO

During a Contesita 8/250  QSO I was told my transmit was off frequency, I am using Win10, SSDR 1.6.17  and latest version of Flidigi.  On a 6500 with GPS.

Which of the following is the correct calibration procedure:

1. Focusing on Fldigi software, I used CheckSR.exe with
     a)  sound card input "DAX Audio 1"
     b) output "DAX Audio TX."  

     CheckSR.exe returned:
     a) input error of -31071 ppm
     b) output error of -14 ppm.  

     Should I enter these values in the sound card setup in Fldigi and be done with it?

2. Since I have GPS setup and functioning on my 6500, and the frequency error (under the Setup-GPS tab) is nearly zero, it seems the frequency of the hardware is spot on.

So is there a "sound card" error in my 6500 that needs correction in the Fldigi software under sound card setup. 

Ken - WA8JXM

How do you know that the report was accurate?  Maybe the other station was wrong?  Have you checked with other stations?

Was  your fldigi audio carrier centered on 1500 Hz?  Did the little block at the bottom show 1500?

Did you check to make sure RIT/XIT were turned off?

-- Ken WA8JXM

Jim W3IPO

Ken,

All good points.
I had never considered calibrating Fldigi until now. Then I looked into it.
Just for the sake of discussion assume that all of your questions are answered.

Is calibrating Fldigi with a Flex 6500,  different from setups where the sound card is within the computer or an external sound card?

Some Fldigi users utilize the small program called CheckSR.exe which provides a measure of sound card error that is then corrected in the Fldigi sound card setup.

CheckSR.exe can be downloaded here:   
http://www.k8jtk.org/drive/ham_radio/digital_modes/checksr/CheckSR.exe

Is this a valid procedure in a Flex with no "sound card?"
Have other Flexer's used CheckSR.exe in this manner?

Also I had one typo in my first message. Here is the correction:

     a). sound card input "DAX Audio Rx 1"

---Jim W3IPO

CAUTION: my ideas are sometimes wrong or even dangerous.  


   

Tim - W4TME

The CheckSR is measuring the drift and clocking of the sound card interface and has nothing to do with the frequency of the radio.  The RX and TX should be close to the same, so something is off with the TX channel.  I would check the sample rate of the DAX TX and DAX TX RESERVED sound devices in Windows.  Both should be 2 channel, 48 kHz with a 16-bit depth.

Also make sure that in the Fldigi sound card setup, the sampling rate is changed for native (44.1) to 48k for both the RX and TX channels.

Jim W3IPO

Tim,
Thanks.

CheckSR returns "correction" numbers that are then entered in the Fldigi sound card setup.  My understanding is that this then shifts the frequency of the audio signal sent from Fldigi to the transmitter thus correcting any transmit frequency error.
Am I wrong?

W3IPO

k3Tim

Hi Jim

The correction values do work with FLDIGI as I've confirmed with WWV and the 6500 (with GPS option) is dead on frequency +/- propagation effects. (3ppm correction)

I tried the program CheckSR and it can give some rather large numbers depending on the Flex channels chosen; an odd result. When working properly it looked like the correction would be in the 20's or less for both Xmt and Rcv.
  
I had calibrated the receive side a different way using the method outlined in FLDIGI using the floating o'scope. Using this procedure, the station here had +3ppm for receive. I shall assume Xmit offset is close to the same.

One quick way to check is fire up FLDIGI in frequency analysis mode, set rig to 14,999.400 Kc on USB and WWV carrier will be generating a 600 cycle tone. Fldigi should show the frequency as 15,000.000 Kc  +/-.  Currently I am observing a +/- 0.05 cycle bounce.

How accurate does Contessa frequency have to be?

I suspect we need to calibrate the virtual (digital) sound card in the Flex 6K as well as our sound card.  Unless the DSP / CPU / GPU that generate the signals are driven from the GPS corrected clock. Be that as it may, 3ppm is splitting very fine hairs!


k3Tim

Jim W3IPO

Tim,
You are right that CheckSR gives some unusual results.  After running it about 2 hours it settled around -3100 for input correction (to be inserted in the Rx ppm correction in Fldigi) and -14 for output correction.  These numbers did not seem reasonable.

I will use the frequency analysis mode in Fldigi.

I do not know if Contessa requires more accuracy than the other modes.

Thanks for you input.

Jim
W3IPO

Jim W3IPO

Tim,
I did it.
On air test confirms accurate frequency within a 1-2 hz.
No compensation needed in Fldigi.
Flex rocks!
Jim
W3IPO