Topics

memory manager/alignment


Mitchel Rought
 

is it possible to change the display frequency through memory manager, after alignment this radio is slightly off, web sdr 7.050.000, this radio 7.049.700, I realize it's not much but still bothers me


Curt
 

yes - see hfsignals instructions for your particular version.  BTW I discovered mine is 150 Hz off now - maybe seasonal drift.  great news is that for many modes it may not have much impact, except when you respond to a CW spot (we even do this in SKCC) it makes it harder to find the signal (my I still have my original first receiver HR20 -- I don't know how much it drifts but lucky to resolve 1 kHz on the dial. 

73 curt


Mitchel Rought
 

curt
thanks for the info
I realize it's not much but it still bothers me, getting old I guess


Aaron K5ATG
 

Just wondering, are there any instructions to the uBitX from the KD8CEC website? It looks like a pretty powerful tool if one knew how to use it. 
--
'72
Aaron Scott
QRPARCI# 16443
GQRP# 16389
4SQRP # 1080


Don - KM4UDX
 

Mitchel -- there are several posts on this. You are no alone in trying to minimize the display freq error. I had the same delta, and it drove me nuts. Did it matter? No really. Did it make me crazy? Totally. hahah

The short answer for you is yes.  Just change the Master Calibration in Memory Manager by a bit, save it, reboot, and observe the new Delta.  If the new Delta is going in the right direction, then keep tweeking via memory manger untill you overshoot, then tweek in the reverse direction, etc. etc. 

If you are a WSJT-X fan, then you can use the frequency calibration routine built into that application.  WSJT-X has a wicked smart process of sampleling known frequences and compairing them to expected freq. Then it calculates the delta (error) and computes the correction. Then applies the correction to the freq command WSJT-X sends to your radio. Presto, your error has been minimized in WSJT-X, while your rig has exactly the same error as always. You can read all about it in WSJT-X help doc.

and get this trick. I ran WSJT-X freq calibration routine BEFORE I use MM with the rig. So the WSJT-X freq calibration process told we what the rig freq error was before I tried to fix it via MM.  And it was inline with expected error like you have of 150-300hz off.  After I tweaked and tweaked via MM, I reran the WSJT-X freq calibration process, and the "correction" it applied was down to next to nothing. so small it was at the noise level. So my MM tweaks got the uBITX error to the 1hz level (or so). More than good enough for me. 

Now, after I redid a bunch of stuff and got a new computer, all that freq perfection was lost and I am back to the factory error...hahaha. but for a while it wicked good...ahahah.


Evan Hand
 

N8KZ worked out a spreadsheet to use with KD8CEC Memory Manager for calibration.

https://groups.io/g/BITX20/message/72183

I have not tried it, just passing on the information.

73
Evan
AC9TU


AndyH
 

Mitchel,

   I'll take you in another direction if you don't mind.  Yesterday, for the first time, I've taken a newly built and uncalibrated Raduino clone from first start to fully calibrated.  I used a signal generator to provide a fixed tone to beat against, and used Farhan's audio tool to get the pass band closer and closer, but couldn't get either where I wanted them.  After the usual back and forth to set the BFO and frequency, I was still about 600 Hz off.    I found an easy way to dial in both settings that last little bit.  I used WSJT-X, WWV on 10 MHz, and Farhan's audio tool.

https://en.wikipedia.org/wiki/WWV_(radio_station)
https://www.hfsignals.com/index.php/bfo-tuning-aid/

  WWV changes audio tones every minute that they transmit a tone.  The WWV Wikipedia page shows the frequency plan as they jump around between 440, 500, and 600 Hz.  Each audio tone will have a secondary line on the waterfall 100 Hz up and 100 down.  When jumping into the frequency adjustment settings, the audio frequency changes - disregard that and just move the frequency up or down a little bit by ear.  Then exit the cal process and check the waterfall.  After a couple of iterations, you can get the frequency spot-on.  Once that's set, run the BFO cal again and use the audio tool to find both sidebands.  Put the lower frequency band solidly into the zone on Farhan's tool and you should be done!

  73, Andy, KG5RKP   


On Fri, Oct 16, 2020 at 11:38 AM, Mitchel Rought wrote:
is it possible to change the display frequency through memory manager, after alignment this radio is slightly off, web sdr 7.050.000, this radio 7.049.700, I realize it's not much but still bothers me


AndyH
 

    One more piece is to identify the exact frequency of the 25 MHz (or 27 MHz) crystal used to feed the SI5351.  This requires programming a Nano with the calibration sketch, letting the Raduino get up to operating temperature, and then measuring the output from the Raduino with a signal counter.  The actual frequency is entered into the Raduino software and it's recompiled and burned into the Nano that will run the Raduino.

https://github.com/etherkit/Si5351Arduino/blob/master/examples/si5351_calibration/si5351_calibration.ino
   
The actual crystal frequency is entered into the Raduino code.  This example is from the CEC firmware, V1.2.  It's from ubitx_si5351.ino line 55:

// If using 27mhz crystal, set XTAL=27000000, MSA=33.  Then vco=891mhz
#define SI5351BX_XTAL 25000000          // Crystal freq in Hz
#define SI5351BX_MSA  35                // VCOA is at 25mhz*35 = 875mhz

   There may be other ways to do it, but this is the method I've used.
 
     Best,
     Andy


On Fri, Oct 16, 2020 at 11:38 AM, Mitchel Rought wrote:
is it possible to change the display frequency through memory manager, after alignment this radio is slightly off, web sdr 7.050.000, this radio 7.049.700, I realize it's not much but still bothers me


AndyH
 

Mitchel,

   I found the crystal measurement routine I use - it's a very short sketch from ND6T from Feb, 2017.  As we only need the actual crystal frequency for this part of the cal, this does exactly what we want.  it sets the Raduino to output a 25 MHz signal on the clock 2 line.  Let it run 20 or 30 minutes until temperatures stabilize, then measure the clock 2 output with a counter.  That number goes into the firmware for the actual 25 MHz crystal frequency.

   OM Cantrell's code follows - it's 20 lines.  Source:  http://www.nd6t.com/bitx/Calibration.htm

/*
 Calibration program for Raduino
 Don Cantrell,ND6T  v 1.1          8, Feb 2017
 Compiles under Si535 library v 2.0.1
 This source file is under General Public License version 3.
 Generates the reference clock frequency so that it can be 
 measured and substituted as the corrected frequency of the
 particular oscillator. 
 */
#include <si5351.h>
Si5351 si5351;
 
void setup() {
    
    si5351.init(SI5351_CRYSTAL_LOAD_8PF,25e6L,0); 
    si5351.set_pll(SI5351_PLL_FIXED, SI5351_PLLA);
    si5351.set_freq(25e8 , SI5351_CLK2); 
}
void loop() {
}

    73, Andy


On Fri, Oct 16, 2020 at 11:38 AM, Mitchel Rought wrote:
is it possible to change the display frequency through memory manager, after alignment this radio is slightly off, web sdr 7.050.000, this radio 7.049.700, I realize it's not much but still bothers me


Gordon Gibby
 

This is the kind of system that I used to calibrate my V3 so long ago.   Once you measure the exact frequency, which was typically about 3 kHz off, you just put that frequency into the software and re-compile.   Bingo!   

It’s been a long while since I did it, but that’s such an easy way to do things.  


On Oct 19, 2020, at 14:23, AndyH <awhecker@...> wrote:

Mitchel,

   I found the crystal measurement routine I use - it's a very short sketch from ND6T from Feb, 2017.  As we only need the actual crystal frequency for this part of the cal, this does exactly what we want.  it sets the Raduino to output a 25 MHz signal on the clock 2 line.  Let it run 20 or 30 minutes until temperatures stabilize, then measure the clock 2 output with a counter.  That number goes into the firmware for the actual 25 MHz crystal frequency.

   OM Cantrell's code follows - it's 20 lines.  Source:  http://www.nd6t.com/bitx/Calibration.htm

/*
 Calibration program for Raduino
 Don Cantrell,ND6T  v 1.1          8, Feb 2017
 Compiles under Si535 library v 2.0.1
 This source file is under General Public License version 3.
 Generates the reference clock frequency so that it can be 
 measured and substituted as the corrected frequency of the
 particular oscillator. 
 */
#include <si5351.h>
Si5351 si5351;
 
void setup() {
    
    si5351.init(SI5351_CRYSTAL_LOAD_8PF,25e6L,0); 
    si5351.set_pll(SI5351_PLL_FIXED, SI5351_PLLA);
    si5351.set_freq(25e8 , SI5351_CLK2); 
}
void loop() {
}

    73, Andy

On Fri, Oct 16, 2020 at 11:38 AM, Mitchel Rought wrote:
is it possible to change the display frequency through memory manager, after alignment this radio is slightly off, web sdr 7.050.000, this radio 7.049.700, I realize it's not much but still bothers me