[time-nuts] 10MHz to 32MHz?

[Magnus Danielson]

From: "Bruce Lane" <kyrrin at bluefeathertech.com>
Subject: Re: [time-nuts] 10MHz to 32MHz?
Date: Sun, 22 Apr 2007 09:05:46 -0700
Message-ID: <200704220905460850.45BD3255 at 192.168.42.129>

> Good day,
> 
> *********** REPLY SEPARATOR  ***********
> 
> On 22-Apr-07 at 21:11 Lester Veenstra M0YCM K1YCM wrote:
> 
> >And if the 32 mhz source is a sealed unit, I suspect you will find that
> >once
> >you mechanically set it as close as practical, that varying the Vcc will
> >allow you to use it as a narrow VCO.
> 
> 	Fair enough. However, I need some clarification. Magnus used some terms that are utterly unfamiliar to me. See below.
> 
> 	<snippety>
> 
> >The most straight-forward fashion is naturally to lock it through a PLL.
> >
> >Highest common comparator frequency is 2 MHz, so divide 10 MHz by 5 and 32
> >by
> >16, compare with SR-gate, use either a passive lag or better yeat, and
> >active loop. Three digital ICs and an op-amp. If the 32 MHz reference is a crystal
> >then hooking into the caps for a varicap is easy enought for most hams.
> 
> 	I suspect a full-blown design engineer would have no trouble following this, Magnus, and I thank you for your post, but I'm afraid you lost me on a couple of things.
> 
> 	What is meant by 'passive lag?'

That is PLLish slang for a standard RC lowpass filter.

> 	Likewise: 'active loop?'

That is PLLish slang for an integrating loop in which you have an op-amp based
integrator, such that you have a capacitor in the negative feedback path of an
op-amp and feed the phase comparator output to the negative input through a
resistor. This is equalent to the integrator term which Bruce was refering to,
it can also be referred to as the DC path since it is is almost quiet.
You also want to balance this to control the damping, which you do by bypassing
it with a resistive path and add the result of them both, but this can be
acheived by inserting a resistor in series with the capacitor. This way you end
up having a loop filer setup which is built from two resistors, a capacitor and
an op-amp. You want a low-leakage, high impedance op-amp, so try something a
thad better than the 741...

> 	On the IC's: I can guess at two of the digital side (divide-by-5 and divide-by-16 counters), but what would the third one be?

While there are many good phase detectors, such as the AD9901, you can get by
using a 'HC00 (or better) wired up as a secured SR-flip-flop.

Ref -> pin 13
Osc -> pin 1
pin 11 -> pin 10
pin 3 -> pin 4
pin 6 -> pin 12
pin 6 -> pin 9
pin 8 -> pin 2
pin 8 -> pin 5
pin 6 -> active loop filter (inverted output)
pin 8 -> passive loop filter (non-inverted output)

It is a quick-hack but can be made to work. The high comparator frequency of
2 MHz is to your advantage for such a quick-hack.

The drawback of the passive loop filter (RC lowpass) is that as you lower the
filter cutoff, you lower the capture & track range, which is not very good.
It is however fairly simple to get it to work.

With an active loop filter you need to be a bit more careful in how you set
your parameters.

> 	The op-amp I can guess at -- probably high frequency/high slew rate, I'm thinking something in the LF series. And I'm also guessing that it would be used to compare the two divided-down signals, and produce a difference voltage if they're out of phase?

The phase comparator does that, the op-amp is there for filtering and mostly
LF signal works actually.

> 	The part about "hooking into the caps for a varicap" means nothing to me in its current form. Are you trying to say I'd need a varactor effect for the 32MHz source? (That's the only idea that makes sense to me).

Yes, unless you already have a CV control, in which case it is just a matter of
hooking up.

I think you can have use for a low-tech solution like this to start with, just
to get your toes wet. After that there is an endless path to perfection!

Cheers,
Magnus