Hello everyone.

开云体育

开云体育

Hi Roberto and Andrew

I have a part built ubitx as well.? I ran Raduino software on Arduino nano, using all 3 clocks, and found the large number and strength of birdies on all bands made the rx unusable.? I assumed the si5351 was just too dirty when using 3 clocks.? However I moved onto another project and did not spend time debugging my build.? Did you use si5351 and if so, is it clean enough?

73 Paul VK3HN.?

开云体育

think about the fact, that three squarewaves are generated within a tiny chip, for clocking digital circuits crosstalk normally is not a bit issue, but for analog circuits this must be an issue!

This means that to use three SI5351 for individualt frequecies seems to me the best solution. The chips are cheap!!

For best freqeuncy stability only one xtal osc ( TXCO or even OCXO 25 MHz or 27 MHz) should be used, but its three outputs must be "perfectly decoupled" meaning mimimize the crosstalk between the outputs of the buffer stages.

Another thought:

The second osc for the conversion 45 MHz - 12 MHz should not be blaced within the range of the first LO i.e. 45 + 12 = 57 MHz) but 45 - 12 MHz = 33 MHz. This is quite close to the maximum input frequency of the uBITX (30 MHz).? Lowering the second IF to say 5 MHz would ease the situation.

At least: the phase nosie performance of the SI5351 is not at its best and degrades the higher the output frequency is.

Henning Weddig

DK5LV

Hello all,

As said in datasheet of SiLabs, decoupling and ground plane is very important (see §6)
On raduino the GND via is too far from chip !! ..and it causes a common impedance to ALL outputs !
May be with arranged layout and better GND we can improve spurs...

I did several tests and trails to improve bridies and spurious of the Raduino module :
/g/BITX20/message/33741
- with one activated output : it is clean !
- loading all 3 outputs at 50 ohms instead of one makes more bridies !
- grounding the Xtal casing : no improvement
- having seperate 50 ohms coax is better, but there is still coupling on tracks from the chip to Raduino's terminals (puting the finger on it increase more spurs)
- reducing the driver output current and power improve much more better : at level 1 (4 mA) I have +8.5 dBm, with a 2 dB attenuator pad it is just sufficient for +7dBm mixer
with 0.1 dB added loss compared to +10 dBm drive at level 2 or 3
- I will try at level 0 (2mA) +4.5dB without attenuator to mixer...
- using industrial mixers with -50 to -60dB LO rejection are better than self made mixers : TFM-2, SBL-1, SRA-1 ... the coupling LO to RF or LO to IF is reduced

I aggree with Henning notes and will try to change 2nd osc to 33 MHz (but 33-11.997... = 21.003.. MHz !),or find a 33 Mhz xtal...

For Dimitar, I tried activating PLL B on Farhan's software : no change on spurs ! I found on library that PLL B is not affected on CLK0,1,2? anyway !
Now I'm using Allard's subroutines to configure the Si5351 instead of the library, I spare about 10 ko of ROM in the Nano !!! but no change on bridies? !

I'm rebuilding the Ubitx software from Fahran with corrections, improvements and? Allard's tricks : real dual VFO, A=B, RIT, new Smeter, adding voltmeter, tuning more stable,...
Now I'm facing to CW mode and review the compete function...
I'm not an expert in C++ , I started to learn since september with this project (I have a stong backgrouned in assembler)

?To EB7ME, is your receiver clean ? Did you try with antenna connected to a dummy load ?

73 de F5RCT
Jean-Matthieu

Hello Paul:
Well, yes, I used Si5351, since I am following Farhan's instructions as best I can.
Basically, my set corresponds to the idea of Farhan, except in the control, since I am using an Arduino Mega 2560 board and mechanical encoders.
The main difference, apart from the previous one, is that in the design of my PCB I have included the Si5351, so the connections of this IC to the mixers are very short.
In the attached pdf you can see the design.
In my receiver? some birdies appear, but it works acceptably well. In fact, I get clear signals in all the bands (you can see it in the video)

73 础苍诲谤é蝉 EB7ME

开云体育

The project is on the right track.
Everything is already in a metal box (cookie), it is already possible to transmit, getting about 12W in 40m. and approximately 7 in 20m. For the other bands I do not have antennas.
QSO's in 40m are relatively easy to do, depending on the propagation, but even with my precarious antenna I have been able to perform QSO's with EA8 and HB9, in addition to practically all EA. Keep in mind that I use a mic preamp that has some compression.
The front-end has changed, being governed by an arduino mega2560 pro and a 2.4 "touch screen.
It is possible to adjust the frequencies of BFO, IF, CW Tone and an offset if necessary for the VFO. For the Si5351 the calibration factor and the outputs can be adjusted.

As a disadvantage, this Arduino works at 16MHz and it seems that, together with the screen, it introduces some noise. Working now to eliminate it.
I just uploaded a small video showing all that: ? ??
or search for "ubitx EB7ME" on YouTube

Thank you very much Farhan for your great idea.

Regards?

EB7ME? 础苍诲谤é蝉