View Single Post
  #17   Report Post  
Posted to sci.electronics.repair
mike mike is offline
external usenet poster
 
Posts: 634
Default Your experience witg function generators

On 3/19/2012 2:33 PM, Antonio I0JX wrote:
Statements like "frequency stability is bad or poor will get you nowhere.
What are your requirements for stability?
Under what conditions? Temperature variability, time variability,
noise??? At what frequency?

My requirements are fairly modest. Just to tune HF receivers and their
IF chains. Say from 200 kHz to 10 MHz. Noise and temperature is not an
issue for my case.


Your requirements are not modest. You are asking the device to perform
a function it was never intended to do.

If you expect to tune receiver IF chains with a function generator,
you need to be willing to wait for everything to warm up, use a secondary
frequency measurement device and do the job quickly.
It does sweep, so you could sweep the IF and use a marker.

What is the stability you are MEASURING?

Frequency measurement results:

- Just turned on. Frequency = 5.000 MHz Delta F = 0
- 15 minutes after. Delta F = -98 kHz (drift = 6.5 kHz per minute)
- 30 minutes after. Delta F = - 116 kHz (drift = 3.7 kHz per minute)
- 45 minutes after. Delta F = -132 kHz (drift = 2.9 kHz per minute)
- 60 minutes after. Delta F = -141 kHz (drift = 2.3 kHz per minute)
- 75 minutes after. Delta F = -147 kHz (drift = 2.0 kHz per minute)


I don't understand your numbers at all.
I see 6 kHz in 15 minutes. By my math, that's ~400 Hz. / minute

Drift is too high for any practical use.


Drift is just fine for the INTENDED uses.

TEK CFG280 11MHz sweep function generator.
Started at 5MHz.
Drifted up to 5.077 in 3 minutes, then started down.
After 52minutes it was back down to 5.022 and drifting
down at about 300Hz/minute.
Pretty similar to yours after initial warmup.
I made no attempt to factor out the startup drift of the
internal frequency counter.

What is the stability spec of the generator?
Stability is a NUMBER, not a vague statement of dissatisfaction.

Tell HP. They do not mention frequency stability in the manual and no
useful information can be found on the web.

HP equipment of that vintage was typically near the better end of
performance for equipment of that vintage. But even HP equipment breaks.
Performance is usually better near the top end of the dial.
Just 'cause you can get 1000:1 range on the knob doesn't mean
you should.

This is just I what I would to like to know: which is people experience
with non-HP generators using an analog frequency-generation circuit.

OK, I looked up the spec. I can't find any reference in the spec
to frequency stability, other than it takes an hour for the frequency
to get within 5% of the dial setting.
That oughta tell you something.

Well, that figure is also related to dial accuracy, but can serve to
broadly figure out frequency stability.

Remember, that's 5% of full scale.
So with a 1000:1 frequency range, that can be up to 5000% off.

In general, if your requirements contain the word "stability",
a function generator is a poor choice.

I know, but if they produce them, they evidently serve to do something.


A function generator is a very nice general purpose instrument.
You're using it for the wrong purpose.

Back in the day, I managed a function generator design group.
I don't remember a single discussion about frequency stability.
It was all about triangle linearity, square-wave transient response,
sine-wave distortion and amplitude flatness with frequency.

In my case there is nothing I can do with such a poor generator. And
that is HP! Again my question is: is my generator faulty or all
instruments based on the same (analog) frequency generation principle
(charging a capacitor at constant current) behave more or less the same?


There's nothing inherently unstable about the topology. But the devil
is in the details. A crystal is stable because you can't change the
frequency.
A function generator is designed for a 1000:1 frequency range.
That makes it exceedingly sensitive to parameter variations.


The HP is more vulnerable because it does include sweep functions
designed to change the frequency.
I'm stating the obvious here, but make sure the modulation functions are
set to
something other than FM.
And the inputs are grounded, including the VCO input on the back.

I'm too lazy to study the schematic, but anything connected to the
frequency path can inject drift. You might be able to cut loose some
of the sweep functions.

Power supply voltage is critical. Direct measurements are difficult because
of the small variations involved. Load the output with 50 ohms and vary
the amplitude from 0 to max. If the frequency changes, you might
want to check out the power supply. The reverse is not true.
If the frequency does not change, that doesn't rule out temperature
variations in the power supply voltages.

You can poke around inside with a tiny heat source. I use a catalytic
butane hot-air gun from Portasol. Freeze mist works too if you can still
find it. But, since there's a lot of temperature compensation going
on in the loop, heating one component in the loop may lead you astray.


Bottom line, I think your generator is working fine. You're using it
for an unintended purpose outside it's performance ability.

Regards.

Tony