Topics

raduino crystal question

ashok.das81@...
 

Dears,
I am planning to build a clone of raduino. I needed to know what type of crystal is used in original raduino that used through hole parts. Si5351A datasheet says only smd crystals with capacitance less than 10pf are recommended, i am wandering what crystal raduino used ? I wanted to make using through hole crystal and no through hole crystal i could find having less than 10 pf capacitance. So can i make raduino using ordinary microprocessor crystals?

Regards
Ashok

Jerry Gaffke
 

Yes, the 25mhz crystal on the Raduino from hfsignals is throughole,
has a parallel capacitance of around 20pf.
That causes the reference oscillator to be off by a few hundred hertz,
but this is easily corrected by calibrating the frequency in software.

Jerry, KE7ER

On Wed, Aug 21, 2019 at 09:02 PM, <ashok.das81@...> wrote:
I am planning to build a clone of raduino. I needed to know what type of crystal is used in original raduino that used through hole parts. Si5351A datasheet says only smd crystals with capacitance less than 10pf are recommended, i am wandering what crystal raduino used ? I wanted to make using through hole crystal and no through hole crystal i could find having less than 10 pf capacitance. So can i make raduino using ordinary microprocessor crystals?

ashok.das81@...
 

On Wed, Aug 21, 2019 at 10:37 PM, Jerry Gaffke wrote:
Yes, the 25mhz crystal on the Raduino from hfsignals is throughole,
has a parallel capacitance of around 20pf.
That causes the reference oscillator to be off by a few hundred hertz,
but this is easily corrected by calibrating the frequency in software.

Jerry, KE7ER

On Wed, Aug 21, 2019 at 09:02 PM, <ashok.das81@...> wrote:
I am planning to build a clone of raduino. I needed to know what type of crystal is used in original raduino that used through hole parts. Si5351A datasheet says only smd crystals with capacitance less than 10pf are recommended, i am wandering what crystal raduino used ? I wanted to make using through hole crystal and no through hole crystal i could find having less than 10 pf capacitance. So can i make raduino using ordinary microprocessor crystals?

 Thanks KE7ER, thats enough for me to start building.  May I ask for some insight into calibration process ?

Regards
Ashok

Jerry Gaffke
 

To calibrate the si5351, just repeatedly adjust the value of the calibration variable si5351bx_vcoa,
until the uBitx is operating at the correct frequency (or you measure one of the three output clocks to be correct).
You must call si5351bx_setfreq() to update the output so the new si5351bx_vcoa value can take effect.

The above method says how to calibrate iteratively, without any math.
Alternately you can calculate the correct value for si5351bx_vcoa.
That variable is nominally 875000000, which is the frequency of the internal oscillator, 875.0 mhz.
All output frequencies (and the operating frequency of the uBitx) are strict ratios of that internal oscillator.

Example: 
Assume you are just starting out, and si5351_vcoa is at the nominal value of 875000000
You tell clk2 to be at 30mhz with     si5351bx_setfreq(2, 30000000);
But your frequency counter says clk2 is at   30.001000mhz
That means that the internal oscillator is actually    875mhz * 30.001/30.000 = 875.029167mhz
Adjust the si5351bx_vcoa value to be 875029167, and the next call to si5351bx_setfreq() should be dead accurate.

In my own Raduino code I adjust si5351bx_vcoa directly, and save it to EEPROM when happy with the calibration.
Farhan added a level of indirection to this, his routine si5351_setcalibration() adds an offset to the nominal 875000000 value,
and stores that offset to EEPROM.

There have been lots and lots of posts in this forum about trouble getting the si5351 calibrated.
Which suggests to me that all available firmware out there is doing this wrong.
See posts 35235, 54501 44278, 44515

Regarding the crystal loading capacitance:
The crystals are manufactured to have a parallel resonant frequency that is on target
when there is a specific capacitance (perhaps 20pf) across that crystal due to 
traces and other parts connected to the crystal (parts such as the si5351).
The Raduino apparently has significantly less to 20pf,
so the outputs of the si5351 are consistently too high in frequency by a few hundred ppm.
See posts  35275, 61287.

Jerry, KE7ER
 


On Wed, Aug 21, 2019 at 10:45 PM, <ashok.das81@...> wrote:
Thanks KE7ER, thats enough for me to start building.  May I ask for some insight into calibration process ?

ashok.das81@...
 

Thanks KE7ER thats helps a lot.

Gordon Gibby
 

Crystals are specified by their manufacturer to oscillate to correct frequency when placed in a circuit that has a certain amount of parallel capacitance, which is a property of the circuit, not the capacitance per se of the crystal

If you need a few extra picofarads of capacitance and your circuit to meet the manufacturers requirement, that is easily added with a small capacitor.  

The reason that the bitx series tends to oscillate  just a little bit higher than where it should’ve, is because there apparently isn’t enough capacitance in parallel— which makes the crystal oscillator be a little higher than expected 

You can easily fix that in the source code of the software by just changing the designated frequency of the Crystal to match where it actually oscillates, and then all of your adjustments should be simple from then on     A good communications receiver will allow you to measure the actual oscillation frequency.  Dip a  wire near the raduino and you’ll pick up a  signal

Even easier is to command the read you a note to go to the specified frequency of the Crystal, and then measure the amplified output!   That’s a very loud signal!

Add 10-20 picofarads in parallel with the capacitor and you may find that you get it exactly on frequency without doing any software at all

These are wonderful things to learn in 
ham radio


On Aug 23, 2019, at 00:09, "ashok.das81@..." <ashok.das81@...> wrote:

Thanks KE7ER thats helps a lot.

David Wilcox
 

Thank you Gordon.  Helps my understanding of crystals and Bitx adjustments.

David J. Wilcox K8WPE’s iPad

On Aug 23, 2019, at 3:31 PM, Gordon Gibby <ggibby@...> wrote:

Crystals are specified by their manufacturer to oscillate to correct frequency when placed in a circuit that has a certain amount of parallel capacitance, which is a property of the circuit, not the capacitance per se of the crystal

If you need a few extra picofarads of capacitance and your circuit to meet the manufacturers requirement, that is easily added with a small capacitor.  

The reason that the bitx series tends to oscillate  just a little bit higher than where it should’ve, is because there apparently isn’t enough capacitance in parallel— which makes the crystal oscillator be a little higher than expected 

You can easily fix that in the source code of the software by just changing the designated frequency of the Crystal to match where it actually oscillates, and then all of your adjustments should be simple from then on     A good communications receiver will allow you to measure the actual oscillation frequency.  Dip a  wire near the raduino and you’ll pick up a  signal

Even easier is to command the read you a note to go to the specified frequency of the Crystal, and then measure the amplified output!   That’s a very loud signal!

Add 10-20 picofarads in parallel with the capacitor and you may find that you get it exactly on frequency without doing any software at all

These are wonderful things to learn in 
ham radio


On Aug 23, 2019, at 00:09, "ashok.das81@..." <ashok.das81@...> wrote:

Thanks KE7ER thats helps a lot.