[Damien Teney]

Hello everybody,
I have build a VHF receiver for the airband, published in the magazine
Elektor a few years ago, and I've managed to build a PLL to improve it. The
PLL is build on another PCB, that is connected to the main circuit, and that
replaces the original potmeter, in order to choose the frequency. The PLL is
able to lock on the right frequency, but the receiver is a lot more noisy,
and it seems really less sensitive with the PLL. See these sketches,
I think it is better than a long explaination.

http://www.mcarsweb.com/_divers/1sketch.jpg
http://www.mcarsweb.com/_divers/2sketch.jpg
http://www.mcarsweb.com/_divers/3sketch.jpg

I have also included the schematics of the VCO (with the buffering
transistor, T4), and the basic schematic of the PLL (taken from a Motorola
application note). I don't know where the problem could come from. I would
say from the bad buffering of the VCO output, but I don't see how it could
be improved.

http://www.mcarsweb.com/_divers/4VCO.jpg
http://www.mcarsweb.com/_divers/5PLL.jpg

I'm really getting mad with this problem ! So thanks to all of you who will
answer ;-)

 

[Tim Wescott]

Hello everybody,
I have build a VHF receiver for the airband, published in the magazine
Elektor a few years ago, and I've managed to build a PLL to improve it. The
PLL is build on another PCB, that is connected to the main circuit, and that
replaces the original potmeter, in order to choose the frequency. The PLL is
able to lock on the right frequency, but the receiver is a lot more noisy,
and it seems really less sensitive with the PLL. See these sketches,
I think it is better than a long explaination.

http://www.mcarsweb.com/_divers/1sketch.jpg
http://www.mcarsweb.com/_divers/2sketch.jpg
http://www.mcarsweb.com/_divers/3sketch.jpg

I have also included the schematics of the VCO (with the buffering
transistor, T4), and the basic schematic of the PLL (taken from a Motorola
application note). I don't know where the problem could come from. I would
say from the bad buffering of the VCO output, but I don't see how it could
be improved.

http://www.mcarsweb.com/_divers/4VCO.jpg
http://www.mcarsweb.com/_divers/5PLL.jpg

I'm really getting mad with this problem ! So thanks to all of you who will
answer ;-)

The buffering looks good, unless you're not feeding enough voltage to
the MC12017. Check the data sheet. Low reference frequencies (your
channel spacing) can lead to phase noise problems, but it is usually in
the form of "whistles" and not random noise.

You may want to look at the tuning voltage with an oscilloscope and see
how it's different with the PLL vs. the pot. Are you sure your PLL
isn't oscillating in closed loop? This voltage will tell you. You can
also view the PLL transient response by changing the commanded frequency
and seeing what it does -- if it rings like heck before it settles down
then you have a loop tuning problem that would severely amplify any noise.

It's usually a good idea with a PLL (or any time you're using a tuning
diode) to restrict the VCO tuning range to no more than you absolutely
need, to keep phase noise low -- but if you were going to have phase
noise problems I would expect that using a 20K pot and a 330K series
resistor would do it, no problem!

 

[Damien Teney]

Thanks for your quick answer !

Low reference frequencies (your
channel spacing) can lead to phase noise problems, but it is usually in
the form of "whistles" and not random noise.

I use a 5.12 crystal with a 1024 divider, that makes a 5 KHz channel
spacing; and my VCO goes from ±65 to ± 90 MHz. Do you think it could be too
small ??

You may want to look at the tuning voltage with an oscilloscope and see
how it's different with the PLL vs. the pot. Are you sure your PLL
isn't oscillating in closed loop? This voltage will tell you.

I get on the same frequency with the same voltage, wether the PLL is
connected or not, so that's OK !

In fact, what looks weird for me is, that the receiver works less well even
when I connect only the output of the VCO to it and let the potmeter to
control the VCO. That makes me think that the problem would more liklely be
in the VCO than in the PLL part. What do you think of it ?

 

[Tim Wescott]

Thanks for your quick answer !




I use a 5.12 crystal with a 1024 divider, that makes a 5 KHz channel
spacing; and my VCO goes from ±65 to ± 90 MHz. Do you think it could be too
small ??




I get on the same frequency with the same voltage, wether the PLL is
connected or not, so that's OK !

In fact, what looks weird for me is, that the receiver works less well even
when I connect only the output of the VCO to it and let the potmeter to
control the VCO. That makes me think that the problem would more liklely be
in the VCO than in the PLL part. What do you think of it ?

I missed that comment in your original post. Yes, that is very odd and
would indicate either a lack of buffering between the VCO and the PLL
controller or some coupling problems. Are you sure you don't have a
ground look between your controller and your VCO, or that you're
disturbing the regulated voltage to the VCO with the current drawn by
the controller?

BTW: The VCO becomes part of the loop, so it's not really a "Phase
Locked Loop" until controller and VCO are talking together.

 

[Rene Tschaggelar]

Hello everybody,
I have build a VHF receiver for the airband, published in the magazine
Elektor a few years ago, and I've managed to build a PLL to improve it. The
PLL is build on another PCB, that is connected to the main circuit, and that
replaces the original potmeter, in order to choose the frequency. The PLL is
able to lock on the right frequency, but the receiver is a lot more noisy,
and it seems really less sensitive with the PLL. See these sketches,
I think it is better than a long explaination.

http://www.mcarsweb.com/_divers/1sketch.jpg
http://www.mcarsweb.com/_divers/2sketch.jpg
http://www.mcarsweb.com/_divers/3sketch.jpg

I have also included the schematics of the VCO (with the buffering
transistor, T4), and the basic schematic of the PLL (taken from a Motorola
application note). I don't know where the problem could come from. I would
say from the bad buffering of the VCO output, but I don't see how it could
be improved.

http://www.mcarsweb.com/_divers/4VCO.jpg
http://www.mcarsweb.com/_divers/5PLL.jpg

I'm really getting mad with this problem ! So thanks to all of you who will
answer ;-)

The geometry of the layout is almost more important than the schematic.
By the missing scope and spectrum analyzer images, I tend to think that
you neither have either, nor have access to either of them. It is almost
infintely had to work without either of them. Further is the phase noise
dependent on the the divider, the higher the divider, the more noise.
Those 5kHz channel distance is quite narrow.

Rene

 

[Damien Teney]

I have a scope, but no spectrum analyzer. I've juste done a few tests and
I've noticed that there are two "problems":
- the receiver products more noise and whistles all along the frequency
range, sometimes low and sometimes really loud; that happens when I connect
the parts like this http://www.mcarsweb.com/_divers/3sketch.jpg
- the receiver products that extra noise also when I ONLY connect the power
supply of the PLL (prescaler + divider/phase comp+ filter), and let the
input of the prescaler and the output of filter "open".

I tend to thinks that the PLL part produces interference that causes that
extra noise, and I guess it would be solved with good shielding of this part
alone ?
The whistles are more unexepected, as the output of the VCO seems well
buffered, and that it couldn't be a closed loop oscillation, nor a problem
with the output filter, as I use the potmeter to control the VCO (and there
is thus no real loop).

Thanks again for your answers.

 

[John Woodgate]

I read in sci.electronics.design that Damien Teney

I tend to thinks that the PLL part produces interference that causes that
extra noise,

I thought that someone would have suggested that already. It's very
likely: you have lots of sharp edges being generated in there, making a
broad-spectrum noise with a succession of discrete frequencies as well.

and I guess it would be solved with good shielding of this part
alone ?

You may have a fine time making it good enough.

The whistles are more unexepected, as the output of the VCO seems well
buffered, and that it couldn't be a closed loop oscillation, nor a problem
with the output filter, as I use the potmeter to control the VCO (and there
is thus no real loop).

The whistles are beats between discrete frequencies being generated in
different parts of your gear. They don't need a conducting path;
emission and reception over very short paths works even at surprisingly
low frequencies, considering how inefficient the 'antennas' are.

 

[Joerg]

Hi Damien,

Provided you have a good prototype build that adheres to RF rules it may
also be worthwhile to experiment with larger values for R and C in the
VCO drive path. The drawing states that the values from the text provide
a "fairly clean" VCO control voltage. But how clean is "fairly" clean?
When it comes to our garage my wife and I have very different opinions
about "clean"

Regards, Joerg

 

[Tam/WB2TT]

Thanks for your quick answer !


I use a 5.12 crystal with a 1024 divider, that makes a 5 KHz channel
spacing; and my VCO goes from ±65 to ± 90 MHz. Do you think it could be too
small ??


I get on the same frequency with the same voltage, wether the PLL is
connected or not, so that's OK !

In fact, what looks weird for me is, that the receiver works less well even
when I connect only the output of the VCO to it and let the potmeter to
control the VCO. That makes me think that the problem would more liklely be
in the VCO than in the PLL part. What do you think of it ?

Damien,
A couple of quick comments here. Unless you want to listen to VOR stations,
you only need to cover 118 to 136 MHz. So, your VCO has twice the range you
need. Also, the channel spacing is 25 KHz. Raising the phase detector
frequency by 5x would help. What is your IF frequency. You are implying
something like 40 MHz.

Since your PLL covers the low end of the FM band, you might want to listen
to the VCO output on an FM receiver to see if it has any FM modulation.

Tam

 

[Damien Teney]

The whistles are beats between discrete frequencies being generated in

different parts of your gear. They don't need a conducting path;
emission and reception over very short paths works even at surprisingly
low frequencies, considering how inefficient the 'antennas' are.

I just tried to put the PLL part alone in a grounded closed metal box,
to test it whith a "shield". I connected the parts like this:
http://www.mcarsweb.com/_divers/3sketch.jpg
but that's the same, I get whistles at some frequencies, all along the
the tuning range.

When I connect the PLL "normally"
(http://www.mcarsweb.com/_divers/2sketch.jpg), it locks on the right
frequency (good point!) but it seems really less sensitive
(interferences ??) and I also have a small whistle in the background.

Help me please !! ;-)

 

[Andrew Holme]

- the receiver products that extra noise also when I ONLY connect the power
supply of the PLL (prescaler + divider/phase comp+ filter), and let the
input of the prescaler and the output of filter "open".

What construction method are you using? Do you have a ground plane?
Is your power supply adequately filtered and de-coupled? Are you
powering the PLL from the same supply as the receiver?

I often build RF circuits on single-sided copper clad board, copper
side up, using the copper as a ground plane. All earth leads are
soldered to the top. All other connections are made underneath i.e.
point to point wiring. I mark the layout on 0.1" graph paper or do it
on the computer and use the paper as a drilling template. I have a
mini craft drill. Finally, I clear copper around the holes by hand
using a drill bit or vero tool.

 

[John Woodgate]

I read in sci.electronics.design that Damien Teney <[email protected]>

I just tried to put the PLL part alone in a grounded closed metal box,
to test it whith a "shield". I connected the parts like this:
http://www.mcarsweb.com/_divers/3sketch.jpg
but that's the same, I get whistles at some frequencies, all along the
the tuning range.

Adequate screening is *difficult*. You probably need 1 nF feed-through
capacitors on all the supply and other connections to the PLL box.
Grounding probably won't make any difference; if it does, it indicates
something else is wrong. Consider how many EMC problems there could be
in a 747. Then think when you saw one with a ground lead. (;-)

When I connect the PLL "normally"
(http://www.mcarsweb.com/_divers/2sketch.jpg), it locks on the right
frequency (good point!) but it seems really less sensitive
(interferences ??)
Probably.
and I also have a small whistle in the background.

This indicates that your screening isn't adequate. Can you put the VCO
in a separate screening box? With feed-through caps on the leads?

Help me please !! ;-)

This now isn't the sort of problem that can be solved with words. You
need a skilled RF man (local amateur radio club?) to give you hands-on
assistance.

 

[SioL]

When I connect the PLL "normally"
(http://www.mcarsweb.com/_divers/2sketch.jpg), it locks on the right
frequency (good point!) but it seems really less sensitive
(interferences ??) and I also have a small whistle in the background.

That whistle in the background could be reference feedthrough.
It means that your PLL steps are kind of small or your PLL filter is
not good enough to filter out the reference. The smaller the reference
frequency the harder it is to filter it out. Try to use higher pll steps just
for a test.

Also, if you hear interference on a given frequency that is probably
due to signals from the PLL. For example, 5.12x15 = 76,8. You'll
likely hear a carrier there or at other harmonics. To make things worse,
receivers are not perfect. You will hear a multitude of other signals,
if they're strong enough. Fvco+Fif and Fvco-Fif are just two usually
strongest, there are others.

You need really good shielding, remember to decouple power to the PLL.
Use a battery to power PLL only to test whether this is the problem.

How are you feeding the mixer in the receiver? Is the RF level the same?

SioL

 

[Ian Buckner]

Hi Damien,

Provided you have a good prototype build that adheres to RF rules it may
also be worthwhile to experiment with larger values for R and C in the
VCO drive path. The drawing states that the values from the text provide
a "fairly clean" VCO control voltage. But how clean is "fairly" clean?
When it comes to our garage my wife and I have very different opinions
about "clean"

Regards, Joerg

"Clean" for high multiplication factor PLL's can be extreme.

Do an estimate of what voltage swing on your VCO would
produce the "faint whistle" you report (estimate what deviation
would correspond to that low level interference, then divide by
the VCO sensitivity in Hz/V).

Your logic signals are 5V swings, and you have a following loop
amplifier. What attenuation do you need between the logic signals
and the VCO control voltage pin?

Put ferrites and lots of decoupling to try to isolate the power rails
from each other.

Regards
Ian

 

[Tim Wescott]

I just tried to put the PLL part alone in a grounded closed metal box,
to test it whith a "shield". I connected the parts like this:
http://www.mcarsweb.com/_divers/3sketch.jpg
but that's the same, I get whistles at some frequencies, all along the
the tuning range.

When I connect the PLL "normally"
(http://www.mcarsweb.com/_divers/2sketch.jpg), it locks on the right
frequency (good point!) but it seems really less sensitive
(interferences ??) and I also have a small whistle in the background.

Help me please !! ;-)

We are helping! For an on-topic thread you're getting a lot of attention!

You are right in doing steps one by one to try to isolate the problem.
For some reason this makes me think of power supply coupling (at least
until I'm proven wrong). Are you connecting the PLL controller to the
same power supply as the VFO? If so, try powering the PLL controller
off of a battery & see if this makes the problem go away. If it does
then your problem is coupling through the power supply.

If it still has problems on battery power try connecting the PLL
controller ground and VCO ground through the feedback connection _only_
(this is very smokey at your frequencies, but what the heck). If the
only connection you have between controller and VCO are a single ground
and a single signal feedback, and if they're on a nice controlled
connection (like a coax) then you know you're feeding signal back
through the controller frequency input.

Once you know where the problem is the solution may become obvious, or
you can at least come back here and ask again.

 

[Tam/WB2TT]

I just tried to put the PLL part alone in a grounded closed metal box,
to test it whith a "shield". I connected the parts like this:
http://www.mcarsweb.com/_divers/3sketch.jpg
but that's the same, I get whistles at some frequencies, all along the
the tuning range.

When I connect the PLL "normally"
(http://www.mcarsweb.com/_divers/2sketch.jpg), it locks on the right
frequency (good point!) but it seems really less sensitive
(interferences ??) and I also have a small whistle in the background.

Help me please !! ;-)

Are you using shielded wire between the PLL and VCO, and the VCO and
receiver? You say you get noise and whistles with the VCO driven from the
pot. Is the 145152 powered down when you do this?

Tam