EMRFD Message Archive 327
Message Date From Subject 327 2007-01-11 13:26:35 Allison Parent VFO drift on power up The VFO is Hartly using J310, circuit is identical to
UVFO or 12.72 in EMRFD. Frequency is in the 7MHz range.
Inductor is wound on t50-6 material. Tuning is air
variable and RIT is via varicap (RIT tuning range 3khz).
Problem is powerup drift downward about 1khz that
stabilizes after several minutes. I have temperature
compensated the VFO for external drift and drift from the
Varicap with a number of N150 and N750 caps. This has
not had much effect on the startup drift.
Thoughts on why the initial drift is there?
328 2007-01-11 19:22:14 Shawn Upton Re: VFO drift on power up Startup drift is normal: any transistor used in an
oscillator will dissapate some amount of power, even
if it's only microwatts. This will serve to heat up
the transistor, even if just a small amount. The
parasitic capacitances inside the transistor are very
temperature dependant, and so any temperature shift
will cause some amount of capacitance shift also.
329 2007-01-12 10:04:59 Allison Parent Re: VFO drift on power up While that is true in all designs I've worked with the directi 330 2007-01-12 10:44:22 Dave Re: VFO drift on power up Try localized heat on the components and see if you can sort out which
one is most temperature sensitive. I made a quarter-inch cube of
drywall compound, sealed with white paint, stuck to a toothpick. I
lay it on top of my soldering iron's barrel until it's too hot for me
to touch, then use it as a "hot spot generator". I used to try this
with a metal block, but the presence of a conductor made a bigger
impact on the circuit than the dielectric of the drywall compound.
331 2007-01-12 11:32:54 John Marshall Re: VFO drift on power up All true, Shawn. The other frequency determining parts also
dissipate tiny amounts of power and are subject to some drift from
the resulting temperature increase.
I haven't done any experiments to verify this but I think the
difference between startup drift and longer term drift from
externally induced temperature changes has to do with the way the
added heat propagates through the different components. At turn-on,
the transistor begins generating heat and its temperature rises much
faster than that of the other components. Once all the parts come up
to the same temperature, the compensation can regulate externally
induced changes that affect all the parts more uniformly. (I'm still
struggling with how to verify this experimentally).
Just thinking out loud about how to avoid this start-up drift...
Maybe it would be possible to compensate the active device separately
by thermally bonding a separate Nxxx capacitor to it. Experimentally
determining the values for both long term and startup drift would be
quite a challenge. How about building the oscillator in a hefty
diecast box and thermally bonding all the major dissipators to it?
Sounds attractive but I'm afraid it might tend to stretch the the
"two slope" drift into something too variable to compensate.
It might be better to work on a "precognitive startup module" that
would turn the oscillator on a few minutes before you want to use it.
On Thu Jan 11, 2007 7:22 pm ((PST)), "Shawn Upton" firstname.lastname@example.org
> Startup drift is normal: any transistor used in an
> oscillator will dissapate some amount of power, even
> if it's only microwatts. This will serve to heat up
> the transistor, even if just a small amount. The
> parasitic capacitances inside the transistor are very
> temperature dependant, and so any temperature shift
> will cause some amount of capacitance shift also.
332 2007-01-12 12:14:53 Luiz Amaral Re: VFO drift on power up Remember that coils normally have positive thermal coefficient and, then,
when heated, they produce down drifts.
Luiz - PY1LL
----- Original Message -----
337 2007-01-12 20:00:17 Allison Parent Re: VFO drift on power up Thanks everyone.
Here's the deal. I built my first transistor Hartly osc back
around 1964 so none of this is new. Built many since then and
usually I get an osc that is stable on startup but drifts with
ambient temp. One has to remember the active device is only
dissipating a few milliwatts. Also all the tricks like an
annealed tank coil were used.
After finally pulling the board out I studied it in a themal
box with temperature probes. In an insulated environment the
temperature changes for the parts are unmeasurable (less than
.1degree F change)! What I did find is the RIT circuit
(SEE EMRFD FIG12.72) 9V regulator was not stable. Apparently
just a bad part. Very frustrating as the assembly was box
in box for shielding (For MiniR2 RX).
The 78L09 Three terminal regulator would start at 9.05V and
drift down to 8.8! and become stable there. It was not
temperature as heating/cooling it had little effect but even a
momentary power off would restart the cycle. If I disabled the
RIT the drift on startup was +200hz and settled very quickly.
That was the clue. It was also the source of confusion as most
Varicap tuned circuits the drift is very bad in the positive
direction with temperature but not at startup and this was
negative and only at startup. A new 78L08 (didn't have a 09
on hand) was used and the VFO now drifts +150hz the first
minute and settles.
I've done a lot of work with a foam lobster shipping box
with a Peltier cooler, a heater and small fan in it for
temperature testing. Some time with a VFO in that
kind of box makes it easier to get a frequency stable VFO.
Then murphy throws a curve.. This time it was a part that
failed after bench testing was done.
339 2007-01-13 08:57:19 Shawn Upton Re: VFO drift on power up Ah, drifty voltage regulation. I haven't understood
why voltage references aren't used more often in radio
projects like this; I guess 78xx's eventually settle
down enough, and are cheap--but references aren't that
expensive. At work I often use LM4040's for easy
references, but I doubt they are the cheapest
solution--I use them as they don't require a capacitor
(usually use a 0.1uF anyhow) and just a resistor, and
have reasonable tempco (100ppm/C for the better
parts), for up to 20mA. There used to be 2.5, 4.096,
5 and 10V versions available.
Shawn Upton, KB1CKT
Expecting? Get great news right away with email Auto-Check.
Try the Yahoo! Mail Beta.
340 2007-01-13 13:32:19 Allison Parent Re: VFO drift on power up 341 2007-01-13 14:19:52 Lasse Re: VFO drift on power up Some of these 3-legged stabs do want a certain minimum
current to operate properly, maybe if you waste 5-10 mA through a
resistor you might see
>Most three terminal regulators employ the same voltage reference
>as the LM4040 and many others. The usual three terminal regulator
>performance is very good and I'd expect a few millivolts at
>worst not several tenths of a volt like this one exhibited. FYI,
>I checked the 78l08 I used and it was less than a 10 millivolts
>startup drift (load is only 1.5ma!). Also the startup drift was
>for less than 10 seconds where the offending part drifted around
>What was annoying is the problem looked like a typical VFO problem
>and it was not. It was compounded by a fairly elaborate essembly
>to shield it and mount. I really didn't want to disassemble that.
>For things that need better I still have a few tubes of LM723s.
342 2007-01-13 14:55:48 Loren Moline WA7S... Re: VFO drift on power up Hello,
I have had great luck using a Zener diode for the VFO. No real current so
didn't need anything more. My 40 meter DC receiver runs on 9 volts and VFO
has 6 volt regulated with zener. It is really stable.
----Original Message Follows----
343 2007-01-14 09:14:33 Allison Parent Re: VFO drift on power up