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 > Specific Vintage Equipment > Vintage Amateur and Military Radio

Notices

Vintage Amateur and Military Radio Amateur/military receivers and transmitters, morse, and any other related vintage comms equipment.

Closed Thread
 
Thread Tools
Old 14th May 2020, 9:31 pm   #1
Jon_G4MDC
Nonode
 
Jon_G4MDC's Avatar
 
Join Date: Nov 2015
Location: Welwyn Garden City, Hertfordshire, UK.
Posts: 2,013
Default LO for genuine 10m FM10D Pye Cambridge

I have had one of these for some time. I dug out some crystals to be able to test TX and RX somewhere in band - just enough to know it works. And it does.
Now I would like to make some meaningful use of it. I could buy a pair of rocks to put it only on 29.6MHz. To do more would break the bank at today's prices! Not very useful on only the one frequency.

I had the idea to take one of the modern Si5xxx versatile clock chips with a PIC to program it and make a VFO. I got the code working for the frequencies involved, 35MHz RX and 29/12 or about 2.5 MHz for TX. A look at the output spectrum sent me running for cover. I'm not putting that on the air! I have not tried to clean it up. It looked beyond hope.

I'm thinking it might have to be a classic 10kHz step synthesiser of 35MHz (29 +6MHZ IF) and 29MHz divided for the TX feed. MC145106 anyone? (Old hat)

Some years ago I did a 10m SSB rig from a CB. It used the set's original VCO, crystal osc and down mixer, then a VFO from a medium wave tranny fed to a 4046. It worked OK.

Wondering which way to go. Any input very welcome!

Added- 10m is coming back to life the last week or two! CW from Italy and Spain and more exotic stuff on FT8 has been seen.

Last edited by Jon_G4MDC; 14th May 2020 at 9:40 pm.
Jon_G4MDC is offline  
Old 14th May 2020, 11:12 pm   #2
Julesomega
Nonode
 
Julesomega's Avatar
 
Join Date: Dec 2012
Location: Stockport, Greater Manchester, UK.
Posts: 2,085
Default Re: LO for genuine 10m FM10D Pye Cambridge

Trouble is any spurs at 2.5MHz will be increased by something massive when you multiply up to 30MHz (20dB?)
Will be interested to hear how you get on
__________________
- Julian

It's good here

Last edited by Julesomega; 14th May 2020 at 11:23 pm.
Julesomega is offline  
Old 15th May 2020, 2:12 pm   #3
G8BBZ
Pentode
 
G8BBZ's Avatar
 
Join Date: Oct 2005
Location: Hebden Bridge, West Yorkshire, UK.
Posts: 199
Default Re: LO for genuine 10m FM10D Pye Cambridge

Hi Jon,
Have you looked at the "Progrock" from QRP Labs. Intended for precisely the application you have.
cheers
__________________
Peter G8BBZ
G8BBZ is offline  
Old 16th May 2020, 9:22 am   #4
Jon_G4MDC
Nonode
 
Jon_G4MDC's Avatar
 
Join Date: Nov 2015
Location: Welwyn Garden City, Hertfordshire, UK.
Posts: 2,013
Default Re: LO for genuine 10m FM10D Pye Cambridge

Thanks for the suggestion of the prog rock. That is a very similar approach to the Si5351 + PIC that I built. I used the ADC in the PIC to read a pot and make a VFO. It had 512 steps so it tuned 29.2-29.7 in 1kHz steps which I thought ideal for 10m FM. It had a 4 digit LED display similar to the old Sinclair calculators.

I decided it was not nearly clean enough for the job. At the time I had no shack spectrum analyser so the testing of it was done quickly on borrowed gear. Maybe it's worth taking some more time over it, at least to repeat the measurements and see if it might be improved.

It is a nice approach but as it stood I wouldn't put it on the air multiplied up or not. A pity since in all other respects it was ideal. Here it is tuned to 29.600
Attached Thumbnails
Click image for larger version

Name:	FM10D_vfo.jpg
Views:	106
Size:	74.5 KB
ID:	205951  

Last edited by Jon_G4MDC; 16th May 2020 at 9:50 am.
Jon_G4MDC is offline  
Old 16th May 2020, 11:58 am   #5
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: LO for genuine 10m FM10D Pye Cambridge

I keep meaning to have a play with these SiLabs devices but never seem to get around to it. So I can't offer much input other than a few initial guesstimates to predict any system issues.

I'm not familiar with the Pye Cambridge FM10D either but if it uses a x12 multiplier to get to (say) 29.6MHz then this will almost certainly cause some far out noise floor issues when using the SiLab part.

At a guess the original circuit used a ~2.46MHz Tx crystal and I'd expect the far out phase noise of such an oscillator across 100kHz to 1MHz offsets to be -160dBc/Hz to maybe -175dBc/Hz. So when multiplied by N=12 to 29.6MHz this noise level will rise up 20*logN = 21.6dB. This would be better than -140dBc/Hz and would be OK for a PMR set I think.

However, I think the Si5351 phase noise floor at ~2.5MHz Fout is typically about -140dBc/Hz across 100kHz to 1MHz offset. This would mean that the far out phase noise across 100kHz to 1MHz would be worse than -120dBc/Hz when multiplied up by 12. I think most commercial HF amateur radios achieve better than -135dBc/Hz far out Tx phase noise but there are a few lemons out there that are worse than -130dBc/Hz.

These days, it is unlikely there would be another CB or amateur operator within a few miles so maybe this doesn't matter in reality. However, 35-40 years ago a radio that produced a noise pedestal like this would be quite antisocial as there could be a dozen CB operators within a couple of miles in a typical town or city.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 16th May 2020, 12:17 pm   #6
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: LO for genuine 10m FM10D Pye Cambridge

To put things in perspective, I tried measuring an old CB 27/81 UKFM CB radio and on transmit it managed -145dBc/Hz at 1MHz offset. This is close to the limit of my analyser so it might be slightly better than this. The Silabs part may well degrade the Pye Tx phase noise to worse than -120dBc/Hz at a similar offset but it would be best to prove this with a measurement.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 16th May 2020, 3:57 pm   #7
Jon_G4MDC
Nonode
 
Jon_G4MDC's Avatar
 
Join Date: Nov 2015
Location: Welwyn Garden City, Hertfordshire, UK.
Posts: 2,013
Default Re: LO for genuine 10m FM10D Pye Cambridge

Thanks for the further thinking about phase noise. Ideas are turning to the SiLabs part with "clean up" PLLs afterwards. It would need them on both TX and RX. The 35MHz RX LO had obvious discrete spurs and the 2.5MHz TX output masked the discrete spurs in fog (fields of grass). These would be terribly anti-social even before multiplication.

There is no particular reason to do the TX PLL at Fc/12. It could be given direct FM making the original phase modulator and multipliers redundant.

The shame about that is "getting at" an original rig quite so much. That wasn't the idea.... but 6 channels at say £20 per custom made crystal (I'd be lucky) is £240 and I would still be guaranteed not to have the frequency required!

Last edited by Jon_G4MDC; 16th May 2020 at 4:04 pm.
Jon_G4MDC is offline  
Old 17th May 2020, 11:52 am   #8
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: LO for genuine 10m FM10D Pye Cambridge

If it helps I managed to find some Si5351 phase noise plots online by KE5FX. The phase noise at 100kHz offset on the 3MHz plot is about -140dBc/Hz which is similar to what I've seen in the past.

There are plots at 3MHz, 10MHz, 13MHz, 20MHz and 100MHz.

https://nt7s.com/2014/11/si5351a-investigations-part-7/

It looks like the Si5351 should appear quite clean on a typical spectrum analyser even at 100MHz output because a typical lab spectrum analyser LO will be noisier than the Si5351. Even the legendary HP8568B analyser would be too noisy to test the Si5351 below 100MHz and I think a modern low noise signal source analyser (SSA) would be needed.

Is there any way there could have been something else 'new' in the workroom that could have caused spurious pickup on your plots? Something like a new laptop PSU or phone charger nearby or a noisy system clock on your test board?
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 17th May 2020, 12:12 pm   #9
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: LO for genuine 10m FM10D Pye Cambridge

I've just had a rummage in my stash of synth and clock gen boards and I've realised that I've actually got one of these Si5351 devices here. I bought it months ago on ebay but never tried it. I had a vague recollection that I'd bought the Si570 but it turns out it is the Si5351. It is on a little blue PCB with a 25MHz clock crystal and it says 8kHz to 160MHz on the PCB.

I have to confess that the reason I didn't ever try it was that I read the datasheet after it arrived and these things are extremely complicated to program. I've written code for lots of PLL and DDS chips before but the Si5351 seems to be much more involved. So I put it to one side for a while and it has remained unused ever since.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 17th May 2020, 3:44 pm   #10
Jon_G4MDC
Nonode
 
Jon_G4MDC's Avatar
 
Join Date: Nov 2015
Location: Welwyn Garden City, Hertfordshire, UK.
Posts: 2,013
Default Re: LO for genuine 10m FM10D Pye Cambridge

Some of what is written in the Si5351 application stuff is complicated and if i recall some of it is wrong. A few years ago I had some assistance to write the code run on my PIC board. I could do the driving of the ADC + display and the routines to make frequency into binary to send by I2C to the Si5351 registers. The bit I could not do was sequence them to run together - I had some help with that.

It is possible some of the nasties I observed were from the PIC and display rather than being the fault of the 5351 per se. I didn't have the time to try holding it all in reset so that the 5351 was the only thing running. I suppose I should run it up again and have another look.

Happy to discuss programming if you fancy a go. I also have an Si570 to try but that requires different code that I have yet to tackle. I have an Si570 kit VFO running in a 60m transceiver but I have not made any measurements of that so far.

Last edited by Jon_G4MDC; 17th May 2020 at 3:52 pm.
Jon_G4MDC is offline  
Old 17th May 2020, 7:31 pm   #11
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,484
Default Re: LO for genuine 10m FM10D Pye Cambridge

There is a comprehensive step by step guide to using the Si5351 module on Adafruit's website, including an Arduino .ino sketch to control it on the later pages. I realise Jon already has some PIC code running - not sure if that would be assembly language or 'C', but the Arduino material may prove helpful for Jeremy especially if he works in 'C' anyway.

https://learn.adafruit.com/adafruit-...rator-breakout
SiriusHardware is offline  
Old 17th May 2020, 11:45 pm   #12
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: LO for genuine 10m FM10D Pye Cambridge

Thanks. I'll have a look at the datasheet and the programming tutorials when I can get a chunk of free time. I may well still need some advice as this device looks to be very complicated to program!

In the meantime I can describe some typical performance data for a classic single loop synthesiser solution if that would be of interest? I'm a bit rusty these days but many years ago I did have quite a bit of experience designing stuff like this so I can usually predict the noise and spurious performance quite well. This could be compared to the Si5351 solution. It won't be clearly better or worse in terms of performance but it will need a bit more PCB space!

I have quite a few old synthesiser dev boards here including one with a discrete 30MHz VCO and an old school MC145158-2 PLL with Fujitsu MB504 DM prescaler. This can easily be configured to 10kHz steps and I can demo it to show it matches basic theory for phase noise? The chips used in this synthesiser are obsolete but a modern solution with very similar (or better) performance could be achieved with an Analog Devices ADF4001 PLL chip and a discrete VCO. This could be about 50x40mm in size including the VCO. The old MC145158-2 version is about 90x40mm and this includes the PLL, the prescaler, the VCO and the buffer/isolation stages, all the voltage regs and a MMIC output amplifier.

Back in the 1990s I was tasked with getting the very best performance from the MC145158-2 chip and this involved developing our own discrete (active) charge pump circuit that featured low noise and low sideband energy. At the time this charge pump circuit was very special but I think the modern chips have something very similar inside.

It's just a case of powering the old dev board and taking a few plots if you'd like to see a phase noise plot? I have a reasonably good signal source analyser here to look at phase noise down to about -140dBc/Hz.
__________________
Regards, Jeremy G0HZU

Last edited by G0HZU_JMR; 17th May 2020 at 11:50 pm.
G0HZU_JMR is offline  
Old 18th May 2020, 9:10 am   #13
Jon_G4MDC
Nonode
 
Jon_G4MDC's Avatar
 
Join Date: Nov 2015
Location: Welwyn Garden City, Hertfordshire, UK.
Posts: 2,013
Default Re: LO for genuine 10m FM10D Pye Cambridge

Thanks for the offer Jeremy - at this stage I won't put you to the trouble. I have some MC145152 with MB501 prescaler but I think that is overkill here.

For info the programming I did was in C for an 18F45K20 PIC since it was on the development board that came with a Pickit 3 programmer.
Jon_G4MDC is offline  
Old 18th May 2020, 9:58 am   #14
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: LO for genuine 10m FM10D Pye Cambridge

If the aim was to keep the x12 stages the 'old school' solution for both Rx and Tx could have been a 35-40MHz single loop PLL with a 10MHz reference crystal.

For 29.6MHz Rx with a 6MHz IF it would be a straight 35.6MHz PLL for the Rx LO with 10kHz steps.

In Tx, the R counter inside the ADF4001 (or the MC145158-2) would be changed to 750 giving 13.3333kHz steps. This would run at 39.4666MHz and would then get divided by 16 (74HC74 flip flops?) to 2.4666MHz and would then be fed to the existing x12 multiplier to produce 29.6MHz in 10kHz steps.

I think most people would opt for the modern SiLabs Si5351 version instead.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 18th May 2020, 6:56 pm   #15
SiriusHardware
Dekatron
 
Join Date: Aug 2011
Location: Newcastle, Tyne and Wear, UK.
Posts: 11,484
Default Re: LO for genuine 10m FM10D Pye Cambridge

Just listening around this evening (Monday 18th May) and 10m is a little bit open but you would be hard pushed to realise it as there are only a couple of stations working the band. It's a bit more obvious down on 11m where there are some strong FM signals, albeit with a good bit of QSB, rolling in on the CEPT CB frequencies. Conditions must have been quite good earlier, unfortunately I was stuck at work.
SiriusHardware is offline  
Old 20th May 2020, 12:43 pm   #16
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: LO for genuine 10m FM10D Pye Cambridge

I should point out that the reason I described the slightly strange transmit frequency plan with 13.333kHz step sizes and the /16 is that this method avoids having a VCO running at the transmit frequency of 29.6MHz. If the original modulator was used then this would mean that there would be a risk of the modulated transmit signal getting into the (unmodulated) VCO coil in the PLL at 29.6MHz. This may produce some strange effects. Having the VCO running at 39.4666MHz and then dividing by 16 before the x12 multiplier will avoid this issue.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 20th May 2020, 1:33 pm   #17
Jon_G4MDC
Nonode
 
Jon_G4MDC's Avatar
 
Join Date: Nov 2015
Location: Welwyn Garden City, Hertfordshire, UK.
Posts: 2,013
Default Re: LO for genuine 10m FM10D Pye Cambridge

Congratulations on a sneaky way around that problem! In synthesised AM gear PYE always ran the PLL at 2F and divided it down for the same reason. At the time odd division ratios meant more physical chips not just a bit more complexity in the progamming.
Jon_G4MDC is offline  
Old 23rd May 2020, 8:59 pm   #18
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: LO for genuine 10m FM10D Pye Cambridge

Yes, I think the \16 + *12 setup would work but I haven't thought about it in much depth.

After reading the useful link to the Adafruit webage (Thanks, SiriusHardware ) I ordered a simple Arduino board and it arrived this weekend. This evening I loaded up the bundled Si5351 example code and had a look on a spectrum analyser.

So far I've only played with a few test frequencies but I now see why you were so concerned about the spurious signals from the Si5351. In fractional mode there are a lot of spurious terms out to several MHz. It seems quite clean 'close in' but it really does look a mess further out. I think I need to study the clockbuilder GUI provided by SiLabs to see if there is a way to improve things but I'm not too hopeful...
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Old 24th May 2020, 2:23 am   #19
Terry_VK5TM
Nonode
 
Terry_VK5TM's Avatar
 
Join Date: Oct 2010
Location: Tintinara, South Australia, Australia
Posts: 2,324
Default Re: LO for genuine 10m FM10D Pye Cambridge

If you are only using one output on the Si5351, make sure the other two outputs are turned off (my recommendation is to only ever use one output per chip, as all outputs share a common supply rail).

Unfortunately, the Adafruit pcb design is flawed and there is a lot of crosstalk between outputs and supply rails, so you will never get a clean output anyway.

"Cleanest" output is by using integer divider values, but that does mean you may not get the exact frequency you want (although the error may be only a few points of a percent).

Bear in mind these things were never meant to be used as RF VCO's/VFO's, there are better options out there, but not in experimenter friendly packages.

The latest Clockbuilder GUI that I used recently appears to have build errors and the generated register map gives occasional wrong frequencies.
__________________
Terry VK5TM
https://www.vk5tm.com/
Terry_VK5TM is offline  
Old 24th May 2020, 12:02 pm   #20
G0HZU_JMR
Dekatron
 
Join Date: Sep 2010
Location: Cheltenham, Gloucestershire, UK.
Posts: 3,077
Default Re: LO for genuine 10m FM10D Pye Cambridge

Thanks! Turning off the other outputs does help. The other thing I've found is that the output seems to be quite termination sensitive with respect to the spurious terms it produces. At present I'm using a high impedance RF probe on the output with a high impedance resistive attenuator tip fitted to the probe.

I've also tried running the little Si5351 PCB from its DC power cable with the USB disconnected to see if that helped.

At the moment I've only tried commenting out the other outputs on the Arduino sketch and this seems to help. Is there a way to 'really' turn off the other unused output/features?

I really need to spend more time with this so I'm reluctant to post up any plots. At the moment I'm seeing quite a few spurs about -75dBc to -85dBc. For most amateur use this would be OK I think. However, I haven't explored the device enough to know if I can make things better or worse for other frequencies.
__________________
Regards, Jeremy G0HZU
G0HZU_JMR is offline  
Closed Thread

Thread Tools



All times are GMT +1. The time now is 9:01 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.