UK Vintage Radio Repair and Restoration Powered By Google Custom Search Vintage Radio and TV Service Data

Go Back   UK Vintage Radio Repair and Restoration Discussion Forum > Other Discussions > Homebrew Equipment

Notices

Homebrew Equipment A place to show, design and discuss the weird and wonderful electronic creations from the hands of individual members.

Closed Thread
 
Thread Tools
Old 30th Mar 2018, 4:49 pm   #61
MrBungle
Dekatron
 
MrBungle's Avatar
 
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
Default Re: Compensating for VFO drift

Quote:
Originally Posted by G0HZU_JMR View Post
You can often predict phase noise performance quite well using Leeson's equation. In the past I designed a simple 'trainer' oscillator to show people some basic oscillator theory. This 10.7MHz oscillator used an LC resonator and a cheap 50R MMIC gain block.

It's possible to guess the phase noise just from the NF of the MMIC, the power it puts into the resonator, the loaded Q (and loss) of the resonator can be simulated on something like RFSIM99 and the flicker corner frequency can be guessed. This can all be added to Leeson's equation to predict the phase noise at carrier offsets of 1kHz through to maybe 200kHz. The exact frequency of oscillation can be predicted with a quick VNA measurement or you could just simulate it to get a fairly accurate estimate.

Even a modest design like this can produce low phase noise at offsets of 25kHz to 200kHz that would totally outclass something like a classic HP8640B sig gen. i.e. an oscillator like this could be used for critical receiver tests at offsets of 100kHz. I've still got it somewhere, it wasn't designed for low drift but I suspect it will be a few Hz/minute after a warmup. But it won't be stable if the ambient temperature changes a few degrees. From memory it achieved -173dBc/Hz at 100kHz offset and this is within a few dB of theory. Probably 15-20dB better than an HP8640B at this offset
Never seen Leeson's equation before. I will do some research into that. Thank you for the tips. Much appreciated. My only experience with this entire space is watching a spectrum analyser many moons ago and reading books on the subject, particularly EMRFD at the moment (glorious book!). I was trying to build an SA as per the other thread but this hasn't got very far yet for lack of a serendipitous appearance of the right filter.

Quote:
Originally Posted by Terry_VK5TM View Post
LTC1799's are as drifty as anything, temperature sensitive, voltage sensitive and of no use for a stable VFO.

Been there, tried that and gave it up as a bad joke.
Indeed. I think the LTC1799 is fundamentally an RC oscillator. I don't think the datasheet actually explains how it works but I bet there's a transistor in there that is being used as a miller integrator with the external R. Temperature drift all over the shop I reckon.

Quote:
Originally Posted by G8HQP Dave View Post
I found this. Useful introduction to oscillator noise. I remember reading Leeson's book on oscillators about 15 years ago and found that it opened my eyes to things I had never really understood before.
Thanks for the link - will definitely read this

Quote:
Originally Posted by G6Tanuki View Post
Interesting that you mention VXOs and the "Super VXO" - I must get round to playing with one of them someday.

http://electronics-diy.com/electroni...tic.php?id=930 for those who don't know what we're talking about.
Super VXO is marvelous. I got some 7030KHz crystals from ebay. Put two in parallel and put a 1SV149 varactor (aliexpress) in series with a simple BJT colpitts, then an emitter follower for output (not ideal but it works). I can pull it down to 7015-7025 KHz which is enough to pull some CW in. If you go below about 7015, it loses stability and goes mental so I put a trimmer at the top and bottom end of the pot controlling the varactor voltage to keep it in range.

I reckon it's pretty easy to get a QSO around that range.

Intermediate soon so I'll build a transmitter and find out (once I'm happy with my fist which sounds like someone beating a wet sock on the key at the moment)
MrBungle is offline  
Old 30th Mar 2018, 5:23 pm   #62
G6Tanuki
Dekatron
 
G6Tanuki's Avatar
 
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 13,953
Default Re: Compensating for VFO drift

Quote:
Originally Posted by MrBungle View Post
Super VXO is marvelous. I got some 7030KHz crystals from ebay. Put two in parallel and put a 1SV149 varactor (aliexpress) in series with a simple BJT colpitts, then an emitter follower for output (not ideal but it works). I can pull it down to 7015-7025 KHz which is enough to pull some CW in. If you go below about 7015, it loses stability and goes mental so I put a trimmer at the top and bottom end of the pot controlling the varactor voltage to keep it in range.

I reckon it's pretty easy to get a QSO around that range.

Intermediate soon so I'll build a transmitter and find out (once I'm happy with my fist which sounds like someone beating a wet sock on the key at the moment)
Evil Me is wondering just what would happen if you built a Super VXO but used a pair of rocks let's say 10KHz apart.

My thoughts being that one rock would predominate when being pulled LF, the other when being pulled HF. It would seem intuitively expedient to use a pair of rocks with similar levels of excitability.

Of course it *might* just result in a truly horrible wibbly mess at some frequencies where the two rocks were fighting over who had precedence.

Are ceramic resonators easier to 'pull' ?? A while back i remember using a 455KHz ceramic resonator and a CMOS chip as an add-on BFO and it would pull over +/- 5KHz without problems, which in proportional terms would imply a 10.7MHz ceramic resonator should be pullable over at least 50KHz.
G6Tanuki is offline  
Old 30th Mar 2018, 6:18 pm   #63
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Compensating for VFO drift

I managed to find some old phase noise plots of my 10.7MHz MMIC oscillator. These were taken about 5 years ago at work and I didn't filter the power supply very well. So there are lots of spurious below about 1kHz but you can see the general trend of the phase noise. See also the phase noise prediction using an excel spreadsheet I use at work for Leeson's equation.
The agreement is very good.

I think the loaded Q was about 30 and the power in the resonator was about +10dBm. The resonator was quite lossy and so this degrades the noise figure. But you can see how closely the graphs agree.

Somewhere, I've got some newer and cleaner phase noise plots but I can't find them. This theory works well up into the GHz region as well. The second set of plots are for a simple 4.4GHz printed resonator oscillator using a 50R MMIC on a small PCB. The design aim was to try and match the phase noise of my HP8566B spectrum analyser to check its phase noise at a 100kHz offset when I first bought it. The measured plot and the Leeson prediction are quite close although I didn't quite get the phase noise of the prediction. It only takes a tiny bit of phase shift in the PCB layout to take the zero phase point away from the peak in loaded Q and when this happens the phase noise won't be as good as the prediction. But you can see that the simulation graph for loaded Q (QL) shows just over 30 for the loaded Q. This simulation uses the s2p data for the MMIC and also does an EM simulation of the PCB layout to predict the gain and phase around the loop.
Attached Thumbnails
Click image for larger version

Name:	10_7_MMIC_OSC.gif
Views:	89
Size:	31.9 KB
ID:	160222   Click image for larger version

Name:	10_7_MMIC_OSC_LEESON.jpg
Views:	118
Size:	86.1 KB
ID:	160223   Click image for larger version

Name:	4400_osc.gif
Views:	114
Size:	32.1 KB
ID:	160224   Click image for larger version

Name:	4400_Leeson.jpg
Views:	80
Size:	86.1 KB
ID:	160225   Click image for larger version

Name:	4400MHz_SIM.jpg
Views:	96
Size:	93.9 KB
ID:	160226  

__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 30th Mar 2018 at 6:38 pm.
G0HZU_JMR is offline  
Old 30th Mar 2018, 8:31 pm   #64
MrBungle
Dekatron
 
MrBungle's Avatar
 
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
Default Re: Compensating for VFO drift

Quote:
Originally Posted by G6Tanuki View Post
Evil Me is wondering just what would happen if you built a Super VXO but used a pair of rocks let's say 10KHz apart.

My thoughts being that one rock would predominate when being pulled LF, the other when being pulled HF. It would seem intuitively expedient to use a pair of rocks with similar levels of excitability.

Of course it *might* just result in a truly horrible wibbly mess at some frequencies where the two rocks were fighting over who had precedence.

Are ceramic resonators easier to 'pull' ?? A while back i remember using a 455KHz ceramic resonator and a CMOS chip as an add-on BFO and it would pull over +/- 5KHz without problems, which in proportional terms would imply a 10.7MHz ceramic resonator should be pullable over at least 50KHz.
Good question with the two crystals. Worth a punt to find out what happens. Next time I do a GQRP order I'll order a 7020KHz as well and see what happens.

The resonators certainly are much more pullable. There are 3.58, 3.69, 7.37, 7.14, 14.32MHz available as well which work and are in the 80/40/20m bands. I tried to pull a 7.37 down to 7MHz but it actually drifted quite badly. At this point it was worse than a VFO. I reckon 150KHz is possible and safe. I didn't have any 7.14's to try out.
MrBungle is offline  
Old 30th Mar 2018, 8:35 pm   #65
MrBungle
Dekatron
 
MrBungle's Avatar
 
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
Default Re: Compensating for VFO drift

Quote:
Originally Posted by G0HZU_JMR View Post
I managed to find some old phase noise plots of my 10.7MHz MMIC oscillator. These were taken about 5 years ago at work and I didn't filter the power supply very well. So there are lots of spurious below about 1kHz but you can see the general trend of the phase noise. See also the phase noise prediction using an excel spreadsheet I use at work for Leeson's equation.
The agreement is very good.

I think the loaded Q was about 30 and the power in the resonator was about +10dBm. The resonator was quite lossy and so this degrades the noise figure. But you can see how closely the graphs agree.

Somewhere, I've got some newer and cleaner phase noise plots but I can't find them. This theory works well up into the GHz region as well. The second set of plots are for a simple 4.4GHz printed resonator oscillator using a 50R MMIC on a small PCB. The design aim was to try and match the phase noise of my HP8566B spectrum analyser to check its phase noise at a 100kHz offset when I first bought it. The measured plot and the Leeson prediction are quite close although I didn't quite get the phase noise of the prediction. It only takes a tiny bit of phase shift in the PCB layout to take the zero phase point away from the peak in loaded Q and when this happens the phase noise won't be as good as the prediction. But you can see that the simulation graph for loaded Q (QL) shows just over 30 for the loaded Q. This simulation uses the s2p data for the MMIC and also does an EM simulation of the PCB layout to predict the gain and phase around the loop.
That's a rather good outcome actually. Very interesting info as well. Thank you for writing it up.

I'm going to take a look at MMICs as well. Some of them have excellent performance in the HF to VHF spectrum as well I understand. I saw an interesting oscillator someone used which used a small length of coax and a varactor as the resonant circuit for 2m.
MrBungle is offline  
Old 31st Mar 2018, 2:16 am   #66
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Compensating for VFO drift

I could have a go at making a copy of this 10.7MHz MMIC oscillator on a bit of sheet metal instead of a PCB. Hopefully this will regulate the heat better and it might be quite stable. Also, I could use an air core inductor rather than the T50-6 (yellow) powdered iron toroid it currently uses. It's not really designed to be a stable VFO but it might be fun to see how stable it can become with a few tweaks.

One neat feature of this 'trainer' oscillator is that it has very little change in phase around the loop from 15mA MMIC bias right through to 40mA bias. The phase shifts about 1 degree. This allows Leeson's equation to be demonstrated at different power levels without compromising the loaded Q or the centre frequency too much. It has an output test port that shows the same RF power that is being input to the resonator so the relationship between power and phase noise can be demonstrated in real time on a signal source (phase noise) analyser like the E5052A used to make the earlier phase noise plots. Also it has a pair of 50R ports to allow the loop to be broken and examined in open loop on a VNA to measure loaded Q and the phase response and gain around the loop at various bias settings.
It's old and tarnished and fragile because it's built on some scrap PCB material and it really deserves to be put in a formal test box because it is quite an elegant design in my opinion
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 31st Mar 2018 at 2:21 am.
G0HZU_JMR is offline  
Old 31st Mar 2018, 10:00 am   #67
G6Tanuki
Dekatron
 
G6Tanuki's Avatar
 
Join Date: Apr 2012
Location: Wiltshire, UK.
Posts: 13,953
Default Re: Compensating for VFO drift

Quote:
Originally Posted by MrBungle View Post
I saw an interesting oscillator someone used which used a small length of coax and a varactor as the resonant circuit for 2m.
Yes I've seen such things done using the very small solid coax as used in microwave gear. It confused me at first because the solid-coax was coiled up and I thought it was an airspaced coil.

Historically, some of the first VHF amateur-band oscillators [on 56 and 112MHz in the 1930s] used the same idea but were implemented using copper pipe - giving them the obvious name of "organ-pipe oscillators".
G6Tanuki is offline  
Old 31st Mar 2018, 1:02 pm   #68
G8HQP Dave
Rest in Peace
 
Join Date: Sep 2008
Location: Solihull, West Midlands, UK.
Posts: 4,872
Default Re: Compensating for VFO drift

Quote:
Originally Posted by G6Tanuki
I've always thought that minimising power-dissipation throughout the LO stage was important for stability, but clearly this is not always the case!
Minimising power is important for stability, but stability is not the only issue.
G8HQP Dave is offline  
Old 31st Mar 2018, 3:37 pm   #69
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Compensating for VFO drift

With my 4.4GHz oscillator I had to get a balance of phase noise and stability otherwise it would drift too much on the analyser. It had to be fit for purpose. You can get some idea of the stability in the youtube video below. Had I used a better box (rather than a small plastic container and a sock!) to shield/damp it I think it would have been much better.

But once I'd stopped moving you can see a sweetspot in the video of about 20-30seconds where it only wobbles/drifts about 1kHz or so at 4.4GHz. It only needed to be this good for a few seconds to let me make a phase noise measurement once I grabbed/stored a trace on the screen.

This drift would be the equivalent of 1Hz at 4.4MHz. To get it this stable I used some exotic Rogers PCB laminate for the resonator and some decent ATC porcelain caps. The phase noise you see on the analyser display is a mix of the oscillator and the analyser's own internal phase noise. The HP8566B is spec'd at about -115dBc/Hz (typical) at 100kHz at 5.6GHz but I think mine is slightly better than this. The little MMIC oscillator managed -119dBc/Hz at 100kHz offset according to the E5052A phase noise analyser.

https://www.youtube.com/watch?v=bQpzzYau3sM
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 31st Mar 2018, 4:18 pm   #70
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Compensating for VFO drift

Whoops wrong youtube link in the previous post! That could have been embarrassing but luckily it was a random RF related video from someone rather than something dodgy.

Here's the link to my 4.4GHz MMIC oscillator tests.

https://www.youtube.com/watch?v=PMZyMLC2swI

Can I ask a mod to remove the incorrect link in post #69 for me please?
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 31st Mar 2018, 9:05 pm   #71
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: Compensating for VFO drift

I managed to find the 10.7MHz MMIC oscillator board and after a few repairs to squashed connections it was up and running OK. I think I've played with it a few times over the years so it might not be the same as it was in terms of the coupling to the resonator.

I logged it over 1500 seconds (25 minutes) from a cold start and got the result below. This is better than I remember and I think there's an element of fortune here in that the drift is cancelling in my favour in some way. This was biased at the regular 36mA.

But after about 15 minutes the drift settles down and it only drifts maybe 20Hz in the last 6 minutes. But for the first 5 minutes it would be useless as a receiver LO because of the drift as the MMIC warms up everything around it.

I'm sure it would drift a lot if I changed the ambient temperature at the end of the test because of the T50-6 toroid for one thing.
Attached Thumbnails
Click image for larger version

Name:	MMIC_10_7_stability.gif
Views:	142
Size:	9.8 KB
ID:	160264  
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 13th Apr 2018, 9:44 pm   #72
MrBungle
Dekatron
 
MrBungle's Avatar
 
Join Date: Jun 2016
Location: London, UK.
Posts: 3,687
Default Re: Compensating for VFO drift

Interesting stuff with the MMIC oscillators. Going to read into this further over the next few days as I have some time. SPRAT arrived this morning and jogged my memory on this thread as there's an article in there from GM4JTJ on MMIC oscillators. Unfortunately have been rather busy over the last couple of weeks and haven't had time to play.

Had an interesting bout of laziness and decided to jump forwards to "cheat mode" and knock up a simple VFO with a QRP Labs Si5351A synth board ( http://qrp-labs.com/synth.html ) and a Arduino Pro mini. Now that's stable. Hasn't drifted 1Hz in the last 5 hours. Also has three independent outputs so I can generate a BFO as well.

Synth board is $7.75 and the Arduino Pro Mini and USBasp cable were a bank breaking purchase of £5.56 from Bitsbox.

Code is simply:

Code:
#include "si5351.h"
#include "Wire.h"

Si5351 si5351;

void setup() {
  si5351.init(SI5351_CRYSTAL_LOAD_8PF, 27000000UL, 208150);
  si5351.set_freq(703000000ULL, SI5351_CLK1);
}

void loop() {
}
This does however cost approximately 30mA. I entirely understand why DDS and synthesizers are so popular now!

Some pictures:

Click image for larger version

Name:	si5351a.jpg
Views:	115
Size:	102.0 KB
ID:	161084

Click image for larger version

Name:	si5351b.jpg
Views:	98
Size:	69.3 KB
ID:	161085

I got it up to 225MHz before it hit a wall!
MrBungle is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 8:54 am.


All information and advice on this forum is subject to the WARNING AND DISCLAIMER located at https://www.vintage-radio.net/rules.html.
Failure to heed this warning may result in death or serious injury to yourself and/or others.


Powered by vBulletin®
Copyright ©2000 - 2024, vBulletin Solutions, Inc.
Copyright ©2002 - 2023, Paul Stenning.