Damien Teney wrote:

Hello all,
I 've built PLL in order to command the VCO of a VHF receiver. Unfortunately
it doesn't work very well: the VCO output is on average equal the same as
the frequency programmed in the PLL, but it is constantly varying a bit
around it. I guess the problem comes from the output filter, but I've
re-calculated it and it should be okay. Where could the problem come from ??

FYI, the PLL is build from a schematic of the Motorola application note 980
(http://www.mcarsweb.com/_divers/sche...cation-note_98
0.gif).

Thank you in advance for your answers ;-)

Some other points. Do you have a filter to remove the reference
frequency from the loop. A notch filter will clean out a lot of noise.
I had to straighten out and re-certify the design of a commercial
synthesizer after the purchasing department bought different capacitors
used in the VCO circuit, and in fixing that, I found several other
problems.

Does your power supply use a linear or a switching regulator? To
achieve a very low phase noise we had to use a pair of low dropout
linear regulators inside the shielded case.

How do you bring signals in and out of the synthesizer case?
feed-thru capacitors and coaxial connectors are a must. The Microdyne
synthesizer I mentioned used a DB 25 to connect the power and logic
inputs, and had a two rows of feed-thru capacitors behind the connector
to filter all signals, including DC. Also, make a spare top cover with
small holes to reach test points. Screw it down to the case, and make
any adjustments. Working without a cover can change the frequency, and
in some cases, take the VCO to the edge of its range where you will get
a lot more noise.

A big part of building a clean VCO is electronics, but a lot is
mechanical, too. A solid case, good shielding, and board layout. Mount
the PC board to the case with short spacers, preferably permanently
attached to the case. For homebrew work you can use brass spacers and
solder the them to the case after they are screwed in from the back. It
will make a big difference.

A clean ground system in the RF circuits is critical. We had to
replace a couple cases where a PEM standoff didn't make good contact
after it was pressed in, but moved the three circuit boards to a new
case, and they passed on the first try.
--


Michael A. Terrell
Central Florida

Michael Black wrote:
And of course, one reason three terminal regulators were invented was
so small sections could have their own regulators.

They are cheap enough that it's hardly worth not having a separate
regulator for the VCO. And maybe even a separate regulator for the
phase detector (and any amplifier) section.

Absolutely. For example, if you need really clean 5V supplies in various
places, a good technique is to run a pre-regulated "semi-clean" 12V rail
around the system, and then regulate locally to 5V.

Don't rely totally on the IC to clean-up the noise on the input rail.
Remember that it's a feedback system, so there is always a
high-frequency corner, above which the regulator can't keep up. Above
that corner frequency, the regulator becomes increasingly transparent
to high-frequency input noise and sharp spikes.

If low output noise is important, include some good old-fashioned R-C
filtering between the 12V line and the input to the regulator IC. Don't
forget to leave more than the minimum required voltage at the input of
the regulator... and that's not the average DC level - it's the voltage
at the *bottom* of any noise spikes at the input, so you have to allow a
little extra input voltage.

Even so, it's still only R = V/I, like we used to do routinely for zener
diodes (I'll swear that half the attraction of IC regulators is that
they let people avoid that simple calculation). If you want to use two
poles of R-C filtering, simply split the R into two parts.

You're still faced with other sources of noise, for example noise
originating in the regulator itself, and things like induced voltages
and ground loops... but all those are actually separate problems.

In different applications, R-C input filtering combined with a voltage
clamp device is a good way to protect regulator ICs and their downstream
components from the very sharp spikes induced by lightning. The
pre-filtering reduces the peak amplitude of the incoming spike and also
attenuates its high-frequency components, down to levels that the
voltage clamp and the regulator IC can safely handle between them.


--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)
Editor, 'The VHF/UHF DX Book'
http://www.ifwtech.co.uk/g3sek

Michael Black wrote:
And of course, one reason three terminal regulators were invented was
so small sections could have their own regulators.

They are cheap enough that it's hardly worth not having a separate
regulator for the VCO. And maybe even a separate regulator for the
phase detector (and any amplifier) section.

Absolutely. For example, if you need really clean 5V supplies in various
places, a good technique is to run a pre-regulated "semi-clean" 12V rail
around the system, and then regulate locally to 5V.

Don't rely totally on the IC to clean-up the noise on the input rail.
Remember that it's a feedback system, so there is always a
high-frequency corner, above which the regulator can't keep up. Above
that corner frequency, the regulator becomes increasingly transparent
to high-frequency input noise and sharp spikes.

If low output noise is important, include some good old-fashioned R-C
filtering between the 12V line and the input to the regulator IC. Don't
forget to leave more than the minimum required voltage at the input of
the regulator... and that's not the average DC level - it's the voltage
at the *bottom* of any noise spikes at the input, so you have to allow a
little extra input voltage.

Even so, it's still only R = V/I, like we used to do routinely for zener
diodes (I'll swear that half the attraction of IC regulators is that
they let people avoid that simple calculation). If you want to use two
poles of R-C filtering, simply split the R into two parts.

You're still faced with other sources of noise, for example noise
originating in the regulator itself, and things like induced voltages
and ground loops... but all those are actually separate problems.

In different applications, R-C input filtering combined with a voltage
clamp device is a good way to protect regulator ICs and their downstream
components from the very sharp spikes induced by lightning. The
pre-filtering reduces the peak amplitude of the incoming spike and also
attenuates its high-frequency components, down to levels that the
voltage clamp and the regulator IC can safely handle between them.


--
73 from Ian G3SEK 'In Practice' columnist for RadCom (RSGB)
Editor, 'The VHF/UHF DX Book'
http://www.ifwtech.co.uk/g3sek

"Michael Black" wrote in message
...
"Michael A. Terrell" ) writes:
Damien Teney wrote:

Hello al
I 've built PLL in order to command the VCO of a VHF receiver.
Unfortunately
it doesn't work very well: the VCO output is on average equal the same
as
the frequency programmed in the PLL, but it is constantly varying a bit
around it. I guess the problem comes from the output filter, but I've
re-calculated it and it should be okay. Where could the problem come
from ??

FYI, the PLL is build from a schematic of the Motorola application note
980

(http://www.mcarsweb.com/_divers/sche...cation-note_98
0.gif).

Thank you in advance for your answers ;-)

How clean is the power supply? Any noise, ripple or hum will appear
in the output of the VCO.

And of course, one reason three terminal regulators were invented
was so small sections could have their own regulators.

They are cheap enough that it's hardly worth not having a separate
regulator for the VCO. And maybe even a separate regulator for
the phase detector (and any amplifier) section.

Michael VE2BVW

In addition to proper loop filter parameters, separate supply regulation,
etc., a VCO needs to be placed and usually shielded to keep outside
noise from perturbing the control voltage ... you could be getting stray
pickup of RF, digital noise, etc. onto the control voltage ...

Carl - wk3c

"Michael Black" wrote in message
...
"Michael A. Terrell" ) writes:
Damien Teney wrote:

Hello al
I 've built PLL in order to command the VCO of a VHF receiver.
Unfortunately
it doesn't work very well: the VCO output is on average equal the same
as
the frequency programmed in the PLL, but it is constantly varying a bit
around it. I guess the problem comes from the output filter, but I've
re-calculated it and it should be okay. Where could the problem come
from ??

FYI, the PLL is build from a schematic of the Motorola application note
980

(http://www.mcarsweb.com/_divers/sche...cation-note_98
0.gif).

Thank you in advance for your answers ;-)

How clean is the power supply? Any noise, ripple or hum will appear
in the output of the VCO.

And of course, one reason three terminal regulators were invented
was so small sections could have their own regulators.

They are cheap enough that it's hardly worth not having a separate
regulator for the VCO. And maybe even a separate regulator for
the phase detector (and any amplifier) section.

Michael VE2BVW

In addition to proper loop filter parameters, separate supply regulation,
etc., a VCO needs to be placed and usually shielded to keep outside
noise from perturbing the control voltage ... you could be getting stray
pickup of RF, digital noise, etc. onto the control voltage ...

Carl - wk3c

Yes but the VCO, if I control it manually (with a potmeter instead of the
PLL), it is quite stable (enough for me ;-). So the main problem I have now
would not be solved with a shield, I think.

Yes but the VCO, if I control it manually (with a potmeter instead of the
PLL), it is quite stable (enough for me ;-). So the main problem I have now
would not be solved with a shield, I think.

and your filter values Damien?
i use those pll chips myself in various tx's..6.25khz to 100khz steps and
have no problems so will dig out the values for 6.25khz for you tonight to
compare.(using LF351/(6) (i think) as filter.
also pll built on double sided pcb (1w exciter.)
no seperate screening apart from vco coil,vco regulated with 78L08,( which
also supplies the 145152) noise is not a problem.sounds like as you suspect
really bad vco overshoot.
"Jim Pennell" wrote in message
nk.net...
This link is to a webpage which shows quite a few PLL design program
links. I am not sure which will best fit your application, but it is a
place to start. Actually, many of them may be able to help sort things
out.


http://www.circuitsage.com/pll.html



Jim Pennell
N6BIU

and your filter values Damien?
i use those pll chips myself in various tx's..6.25khz to 100khz steps and
have no problems so will dig out the values for 6.25khz for you tonight to
compare.(using LF351/(6) (i think) as filter.
also pll built on double sided pcb (1w exciter.)
no seperate screening apart from vco coil,vco regulated with 78L08,( which
also supplies the 145152) noise is not a problem.sounds like as you suspect
really bad vco overshoot.
"Jim Pennell" wrote in message
nk.net...
This link is to a webpage which shows quite a few PLL design program
links. I am not sure which will best fit your application, but it is a
place to start. Actually, many of them may be able to help sort things
out.


http://www.circuitsage.com/pll.html



Jim Pennell
N6BIU

Okay thank you. I don't know well what values you want; I took what was
advised in the motorola application note
loop damping factor = 0,707
VCO gain = 25 132 741 rad/V
VCO center frequency = 77 MHz
PLL supply voltage = 5V
etc.

The values of the components of the filter are given in a previous post.


"zindazenda" a écrit dans le message de news:
...
and your filter values Damien?
i use those pll chips myself in various tx's..6.25khz to 100khz steps and
have no problems so will dig out the values for 6.25khz for you tonight to
compare.(using LF351/(6) (i think) as filter.
also pll built on double sided pcb (1w exciter.)
no seperate screening apart from vco coil,vco regulated with 78L08,( which
also supplies the 145152) noise is not a problem.sounds like as you
suspect
really bad vco overshoot.