As you only have a sub 1G speccy you can verify the LO by using another lnb at 9.75G and holding it near the output with an open pcb. You should see an IF of around 10.3xx - 9.75G on the sniffer LNB IF output. The bullseye is good for this as it was optimised for sub 1G IF signals and the internal VCTCXO is pretty good. Its VCTCXO as the MCU internal aplies a pwn voltage offset to the VCTCXO. It was calibrated in the factory to sub 1kHz but these were built 5 yrs ago the vctcxo drifts. It can be recalibrated, and I released the hardware tondo to the British amateur radio guys. Its on github somewhere
Yes, you are right. I should get another Bullseye LNB to use for that kind of testing. I don't know what the phase noise performance of the LNB is so it may have some limitations in seeing stuff like spurs close to the frequency of my desired signal but it would still be very useful. I will take your advice and get another one to use in this way.
@@n1bug its about -90dBc/Hz at 10k offset. I know as I designed it. The VCO on the bullseye is very wideband and has a fixed division based on the Tone/no tone. The leakage out of the F-connector is about -40dBm. So you can use the Leo source to lock the VCO on your modified unit and get your 10.3xxG out of the F-connector. Use tone to select 10.6G LO as this is closer to your desired, and the PLL division will be 424. Thus to get 10386.360 use a ref into the modified bullseye of 24.45368 MHz (ish). Then you should see 10368.36032 MHz at the fconnector output at about -40dBm. You can filter and amplify. If you use no tone then the LO will be 9.75G so the PLL division ratio changes to 390. You can try a reference freq of 26.585589MHz to get your 10368.360 etc
as long as the input reference is clean and a decent phase noise you should have a good beacon LO without having to use a multiplier. There are also a number of low cost PLL+VCO IC's in the market that can easily do this, but not as cheap as this method. Remember that the phase noise of the Leo reference needs to be good. The LNB phase noise will be be the same as the LEO - 20log(424) or limited to about -90dBc/Hz @ 10k (in loop BW)
@@jonwalker8999 Let me see if I understand. Are you saying modify the LNB by removing the VCO, feed reference into the red F connector and get 10 GHz at -40 dBm out of the green F connector? I will have to try that when I get another Bullseye. It sounds like I should get a few more to experiment with.
@@n1bug Yes, thats correct, you have already done that in a previous video where you used the reference from Leo's unit to lock the LNB. So just test it on that unit with a different reference freq. The internal LO of the LNB downconverter chip is nominally 9.75/10.6G (universal) with a 25M reference. The way it changes frequency is to alter the internal fixed divider within the chip. The internal VCO in the downconverter (not the VCTCXO) is very wideband and can cover about 9.2 to 11.7G. The LO leakage from the downconverter IF out is about -30 to -50dBm . So you can take the downconverter LO from the F-connector, filter and amplify. You can then use another lnb to sniff off the leakage and verify with your sub 1G Speccy. The LNB will have about 60dB gain; just be careful that you don't compress the LNB by holding it too close as it can develop spurs when it's in compression.
It's good to be working on projects and doing videos again! There should be two or three more videos fairly soon, then another pause. I would be doing this all the time if I could.
I will try to look into the spurs generation around 576MHz. Same output frequency can be achieved with a lot of different configuration settings. I can confirm that I also see +-20kHz spurs on 576.020 MHz using default software settings. I have left the comment here before but TH-cam removed it. Let's see if this one lasts. Thanks Leo Bodnar
I think they removed your previous comment before sending a notification. I did get notified of this one so let's hope it stays. I will be curious what you might find with alternate settings. I know of at least one 10 GHz beacon that is on the air that way, probably with spurs. I don't think the spurs bother most people, but my project goal was to create the cleanest beacon possible given budget. I am now using the new LBE-1420, which also had spurs at 576.020 or 1152.040 MHz but is working great as a 10 MHz reference shifted upward. By setting the reference appropriately and using a synthesizer programmed for 1152.000000 I can get just about any 10 GHz frequency I want with no spurs and good phase noise. This is a much more frequency agile setup than any previous method tried and will probably be my final solution.
Hi. Very interesting. I have a Leo Bodnar LBE-1420 and have run a few phase noise tests on it and it is pretty good, especially for the price. I will do some more at different frequencies to see if I can see any spurs. Drew VK4ZXI
Hi Drew. I would be interested in your findings. I need to try to do more phase noise checks on the beacon. Of course phase noise at 10 GHz is much higher than it is at the reference or the 1152 MHz synthesizer output but my beacon is looking better than most that have gone on the air in recent years. Many of them are only 40 to 50 dB down at 10 kHz. Mine looks to be at least 60 dB, possibly a little better than that. I don't have a proper measurement system but if I calibrate the dB scale on the SDR and S meter on the conventional receiver I can get a "no worse than" figure. Of course any estimate made this way is limited by receiver phase noise.
I see another comment from you below about using a second LNB as a sniffer for testing. I replied to that. Ultimately I am trying to generate a clean signal on 10368.390 MHz. I use a x9 multiplier from 1152 MHz to 10 GHz, so I need 1152.043333333333 MHz into the multiplier. It is easy for most synthesizers or GPS reference clocks to generate 1152.000 cleanly but when asked to do 1152.043 or such they have spurs on the output. My solution to this is to move the 10 MHz GPS locked reference into the synthesizer up a few Hz and program the synthesizer for 1152.000. This moves the final 10 GHz frequency to where I want it without meaningful spurs. So it seems I have a solution that works for me.
As you only have a sub 1G speccy you can verify the LO by using another lnb at 9.75G and holding it near the output with an open pcb. You should see an IF of around 10.3xx - 9.75G on the sniffer LNB IF output. The bullseye is good for this as it was optimised for sub 1G IF signals and the internal VCTCXO is pretty good. Its VCTCXO as the MCU internal aplies a pwn voltage offset to the VCTCXO. It was calibrated in the factory to sub 1kHz but these were built 5 yrs ago the vctcxo drifts. It can be recalibrated, and I released the hardware tondo to the British amateur radio guys. Its on github somewhere
Yes, you are right. I should get another Bullseye LNB to use for that kind of testing. I don't know what the phase noise performance of the LNB is so it may have some limitations in seeing stuff like spurs close to the frequency of my desired signal but it would still be very useful. I will take your advice and get another one to use in this way.
@@n1bug its about -90dBc/Hz at 10k offset. I know as I designed it. The VCO on the bullseye is very wideband and has a fixed division based on the Tone/no tone. The leakage out of the F-connector is about -40dBm. So you can use the Leo source to lock the VCO on your modified unit and get your 10.3xxG out of the F-connector. Use tone to select 10.6G LO as this is closer to your desired, and the PLL division will be 424. Thus to get 10386.360 use a ref into the modified bullseye of 24.45368 MHz (ish). Then you should see 10368.36032 MHz at the fconnector output at about -40dBm. You can filter and amplify. If you use no tone then the LO will be 9.75G so the PLL division ratio changes to 390. You can try a reference freq of 26.585589MHz to get your 10368.360 etc
as long as the input reference is clean and a decent phase noise you should have a good beacon LO without having to use a multiplier. There are also a number of low cost PLL+VCO IC's in the market that can easily do this, but not as cheap as this method. Remember that the phase noise of the Leo reference needs to be good. The LNB phase noise will be be the same as the LEO - 20log(424) or limited to about -90dBc/Hz @ 10k (in loop BW)
@@jonwalker8999 Let me see if I understand. Are you saying modify the LNB by removing the VCO, feed reference into the red F connector and get 10 GHz at -40 dBm out of the green F connector? I will have to try that when I get another Bullseye. It sounds like I should get a few more to experiment with.
@@n1bug Yes, thats correct, you have already done that in a previous video where you used the reference from Leo's unit to lock the LNB. So just test it on that unit with a different reference freq. The internal LO of the LNB downconverter chip is nominally 9.75/10.6G (universal) with a 25M reference. The way it changes frequency is to alter the internal fixed divider within the chip. The internal VCO in the downconverter (not the VCTCXO) is very wideband and can cover about 9.2 to 11.7G. The LO leakage from the downconverter IF out is about -30 to -50dBm . So you can take the downconverter LO from the F-connector, filter and amplify. You can then use another lnb to sniff off the leakage and verify with your sub 1G Speccy. The LNB will have about 60dB gain; just be careful that you don't compress the LNB by holding it too close as it can develop spurs when it's in compression.
Great to see you doing videos again! The new bench looks great too. 73 W1RMD
It's good to be working on projects and doing videos again! There should be two or three more videos fairly soon, then another pause. I would be doing this all the time if I could.
I will try to look into the spurs generation around 576MHz. Same output frequency can be achieved with a lot of different configuration settings.
I can confirm that I also see +-20kHz spurs on 576.020 MHz using default software settings.
I have left the comment here before but TH-cam removed it. Let's see if this one lasts.
Thanks
Leo Bodnar
I think they removed your previous comment before sending a notification. I did get notified of this one so let's hope it stays. I will be curious what you might find with alternate settings. I know of at least one 10 GHz beacon that is on the air that way, probably with spurs. I don't think the spurs bother most people, but my project goal was to create the cleanest beacon possible given budget.
I am now using the new LBE-1420, which also had spurs at 576.020 or 1152.040 MHz but is working great as a 10 MHz reference shifted upward. By setting the reference appropriately and using a synthesizer programmed for 1152.000000 I can get just about any 10 GHz frequency I want with no spurs and good phase noise. This is a much more frequency agile setup than any previous method tried and will probably be my final solution.
Thanks for posting good progress. Hope to catch you guys on 6M soon. 73 Lou KT1R
I will call you if I hear you Lou. I don't get on 6 much since all the DX went to digital which isn't near as much fun for me.
Hi. Very interesting. I have a Leo Bodnar LBE-1420 and have run a few phase noise tests on it and it is pretty good, especially for the price. I will do some more at different frequencies to see if I can see any spurs. Drew VK4ZXI
Hi Drew. I would be interested in your findings. I need to try to do more phase noise checks on the beacon. Of course phase noise at 10 GHz is much higher than it is at the reference or the 1152 MHz synthesizer output but my beacon is looking better than most that have gone on the air in recent years. Many of them are only 40 to 50 dB down at 10 kHz. Mine looks to be at least 60 dB, possibly a little better than that. I don't have a proper measurement system but if I calibrate the dB scale on the SDR and S meter on the conventional receiver I can get a "no worse than" figure. Of course any estimate made this way is limited by receiver phase noise.
TH-cam has deleted my last comments.
What frequencies are you looking for?
I see another comment from you below about using a second LNB as a sniffer for testing. I replied to that. Ultimately I am trying to generate a clean signal on 10368.390 MHz. I use a x9 multiplier from 1152 MHz to 10 GHz, so I need 1152.043333333333 MHz into the multiplier. It is easy for most synthesizers or GPS reference clocks to generate 1152.000 cleanly but when asked to do 1152.043 or such they have spurs on the output. My solution to this is to move the 10 MHz GPS locked reference into the synthesizer up a few Hz and program the synthesizer for 1152.000. This moves the final 10 GHz frequency to where I want it without meaningful spurs. So it seems I have a solution that works for me.