EMRFD Message Archive 655
Message Date From Subject 655 2007-04-18 11:13:30 Ed Almos DIP Oscillators - A Cautionary Tale
These are the small devices originally intended for digital use that
fit a standard DIP (Dual Inline Package) IC socket. I have even seen
some of them available in frequencies that cover amateur bands and
they can be quite useful.
We do however need to be careful when using these devices, especially
when feeding them power. We are all used to an analog world where
feeding 13.1V rather than 13.8V into our circuit will not cause any
problems apart from (possibly) a slight loss of power. Most of these
oscillator modules demand 5V +/- 5%. Below this figure and the device
may not oscillate, above 5.25V and you may let the 'magic smoke' out.
The reason for this tale is a wasted hour spent chasing a fault. I
have a circuit on the bench which uses a 45 MHz oscillator module fed
from a (faulty) voltage regulator. The oscillator was fed by 4.6V and
was dead, the replacement oscillator was also dead and it was only
when I replaced the regulator that the circuit burst into life. The
power rail rose to 5V.
So, if your oscillator module says 5V supply it means exactly that.
Edward Almos HA6SST657 2007-04-19 08:20:44 Allison Parent Re: DIP Oscillators - A Cautionary Tale 658 2007-04-19 11:34:51 R. Snyder Re: DIP Oscillators - A Cautionary Tale
On the other hand, I've run 5-volt crystal oscillator
modules at 3 or 4 volts to get a sine wave output
rather than the Fourier series expansion of the usual
square wave output. Obviously, that only works with
modules that continue to oscillate at those supply
voltages.
73,
Ross N0GSZ
659 2007-04-19 13:29:14 Jason Milldrum Re: DIP Oscillators - A Cautionary Tale
and stability can be improved if you divide down the frequency of the
oscillator. For example, I just picked up a batch of 80 MHz oscillator
cans. If I were to divide the output by 8, I should get a much cleaner 10
MHz output, right? That might be one way to get a decent LO signal out of
one of these cans.
73,
Jason Milldrum, NT7S
Allison Parent wrote:
> I'll extend that to add:
>
>
> DIP oscillators are handy but come with a price and that is they are
> often dirty. By dirty I mean high or unpredictable levels of phase noise
> and drift. An example si a 36mhz osc used for a quickie 6M to 20M
> converter. It was built and used and worked ok. one day I actually put it
> on a calibrated signal generator for MDS measurement and found it was not
> so hot and even at 10db above that level it was noisy. So I substituted a
> simple 36mhz discrete transistor osc and much of the hiss went away and
> the MDS was esily 4-5db better. Since the mixer and RF amp wer pretty good
> devices (mpf102 and tuf-1 DBM) I could only conclude the dip osc was a
> noise source. Testing it was a signal source confirmed it was. A
> different vendor and model at the same frequency was better and an older
> one was really bad with noise and drift. I did note there are models
> deesigned for higer stability or lower noise for use with DDS as
> references available.
>
> So dip osc are handy and potentially useful they are potential
> sources of noise or other problems.
>
>
> Allison660 2007-04-19 13:55:43 Joe Rocci Re: DIP Oscillators - A Cautionary Tale 661 2007-04-19 14:39:22 Kevin Purcell Re: DIP Oscillators - A Cautionary Tale
I would (naively) think: crystal plus an TTL/CMOS inverter ring -- it
won't win any prizes but won't be that bad for phase noise either.
So by higher what do you mean? 28636kHz? 14318kHz?
When you get down to the old favorite (the above are multiples) of
3579kHz are they better? A crystal plus an inverter or two? Or do
they retain the same PLL design?
I've been thinking of building a simple S9 calibrator using a DIP
oscilator.
Run the DIP at 5V. Assume 1:1 TTL square wave out (a good assumption
or run at 2f and divide down?). Then use a voltage divider and pad to
come down to -73dBm into 50 ohms.
A little math gets the correction for the ratio of the fundamental
peak/peak voltage to the square wave peak to peak to be 0.827 (using
up to the 19th harmonic).
Phase noise shouldn't be too big an issue for this application but
perhaps there are other issues I'm not seeing?
Any input welcome.
662 2007-04-19 15:05:00 arv evans Re: DIP Oscillators - A Cautionary Tale
There are all sort of nay-sayers in the world ready to tell you why everything will not work. Just build your calibrator with that canned oscillator and see how it sounds. It is not like this is a massive engineering effort requiring hours of analysis and design work.
Canned oscillators do usually draw a lot of current (50 to 75 ma is typical). Their outputs sometimes contain lots of harmonics (you were building a calibrator were you not?). They are usually not particularly stable (some have a 4th lead that is connected to a DC steering voltage to keep them exactly on frequency). Most un-damaged units will run at from 4.5 to 5.5 volts, but other than 5.0 volts may put them slightly off frequency. VHF and UHF types sometimes (usually...?) have internal PLL circuitry, but that doesn't make them automatically generators of phase noise. These higher frequency units are used in Cell Phones, WiFi, Radar, and GPS systems where high phase noise components would be detrimental to receiver sensitivity. Some people automatically assume that anything with a PLL generates gobs of phase noise...it ain't necessarily so. Phase noise in a PLL depends mostly on control loop delay. Lots of delay (i.e. big capacitors) usually equates to low (or slow) phase noise. That last statement should start at least a week or two of shouting and arguing!
Arv K7HKL
_._
Kevin Purcell wrote:Ah, this half answers my question: why are the DIP oscillators so bad?
I would (naively) think: crystal plus an TTL/CMOS inverter ring -- it
won't win any prizes but won't be that bad for phase noise either.
So by higher what do you mean? 28636kHz? 14318kHz?
When you get down to the old favorite (the above are multiples) of
3579kHz are they better? A crystal plus an inverter or two? Or do
they retain the same PLL design?
I've been thinking of building a simple S9 calibrator using a DIP
oscilator.
Run the DIP at 5V. Assume 1:1 TTL square wave out (a good assumption
or run at 2f and divide down?). Then use a voltage divider and pad to
come down to -73dBm into 50 ohms.
A little math gets the correction for the ratio of the fundamental
peak/peak voltage to the square wave peak to peak to be 0.827 (using
up to the 19th harmonic).
Phase noise shouldn't be too big an issue for this application but
perhaps there are other issues I'm not seeing?
Any input welcome.
663 2007-04-19 15:20:09 Joe Rocci Re: DIP Oscillators - A Cautionary Tale 664 2007-04-19 18:07:53 arv evans Re: DIP Oscillators - A Cautionary Tale
I checked some HF canned oscillators and most are like this one:
http://rocky.digikey.com/WebLib/Epson/Web%20Data/SG-51,531%20Series.pdf
and use real crystals (no PLLs). When you go above a few hundred MHz or so they start to use SAW oscillators. In between there are some crystal based units that do use a PLL tracking oscillator as the frequency multiplier. The manufacturers don't state what type of oscillator the PLL uses, just that the phase noise is quite low, as in this one:
http://www.ctscorp.com/components/Datasheets/008-0284-0_C.pdf
Maybe it depends on the age of the stuff in your junk box? BTW, some PLLs use conventional varactor tuned LC type oscillators with slow control loops and have phase noise characteristics comparable with the best VFO circuits. VCXO based PLL oscillators for narrow range frequency control are even better.
Arv
_._
Joe Rocci wrote:
>>Some people automatically assume that anything with a PLL generates gobs of phase noise...it ain't necessarily so. Phase noise in a PLL depends >>mostly on control loop delay. Lots of delay (i.e. big capacitors) usually equates to low (or slow) phase noise.------------ --------- --------- -------It's not the PLL that causes excessive noise in some can-oscillators - it's the RC VCO that the PLL locks to the crystal. They're horrendous.But for the intended application of an amplitude calbrator, the can-oscillator should be fine.----- Original Message -----666 2007-04-21 14:04:52 Allison Parent Re: DIP Oscillators - A Cautionary Tale 667 2007-04-21 14:19:46 arv evans Re: DIP Oscillators - A Cautionary Tale
Arv
_._
Allis668 2007-04-21 16:47:20 Andy Re: DIP Oscillators - A Cautionary Tale
oscillator is that it is meant as a very low cost clock source where
accuracy, stability, and phase noise don't much matter. It's what you use
to make some digital circuits work and just need a frequency within .1%
or so.
Frankly, I'd be surprised to find a PLL in a cheap DIP oscillator, unless
it's one of those with a selectable output frequency. A PLL is way too much
complexity.
Higher frequencies (beyond ~20MHz? or so?) tend to require an overtone
oscillator, which simply means the xtal vibrates on a harmonic. The trick
is keeping the loop gain less than unity on the fundamental. If not done
right, there is some chance of weak oscillations on the fundamental (or
other harmonic) frequencies, which can add phase noise. But for crude
digital clocking, it doesn't matter.
> I would (naively) think: crystal plus an TTL/CMOS inverter ring -- itPotentially OK for phase noise. But no guarantee ... not if something else
> won't win any prizes but won't be that bad for phase noise either.
is going on (parasitic oscillations that they didn't bother to eliminate).
> I've been thinking of building a simple S9 calibrator using a DIPIf it's HF AM/CW/SSB and all you care about is amplitude, phase noise
> oscilator.
matters little, I think.
> A little math gets the correction for the ratio of the fundamental19th harmonic might be a bit much, depending on rise/fall times. Some
> peak/peak voltage to the square wave peak to peak to be 0.827 (using
> up to the 19th harmonic).
output buffers don't have the fastest edges, and aren't symmetrical.669 2007-04-21 17:34:24 larry allen Re: DIP Oscillators - A Cautionary Tale
or crystal accurate.. They are devices meant to just bring you in....
close enough, to start with...
Larry ve3fxq
----- Original Message -----
671 2007-04-25 01:27:48 Michael Smith DIP Oscillators - A Cautionary Tale
(how, I don't know) and fed 12volts into one of them. It absolutely LOVED
it!. I saw on the CRO there was 12v pp output, decided this was pretty
amazing for a thing with a 5v supply, and put my finger on its little metal
case....the old adage "if its too hot to touch, its too hot" immediately
came to mind. Anyway, fixed the regulator circuit, the DIP oscillator
survived ok, much to my surprise.
Andrew VK3BFA.
_________________________________________________________________
Advertisement: Its simple! Sell your car for just $30 at carsales.com.au
http://a.ninemsn.com.au/b.aspx?URL=http%3A%2F%2Fsecure%2Dau%2Eimrworldwide%2Ecom%2Fcgi%2Dbin%2Fa%2Fci%5F450304%2Fet%5F2%2Fcg%5F801577%2Fpi%5F1005244%2Fai%5F838588&_t=754951090&_r=tig&_m=EXT672 2007-04-25 02:53:28 Greg W:-) DIP Oscillators - A Cautionary Tale
Well thats one theory....
OR---
Maybe you tapped the ether , the free energy of the cosmos.
The voltage went up and the device got hot ,,sure thats proof of the
powerful ether at work. :-P
Hehe ,,,those new age weeirdoze eat talk like that like kids and candy.:-D
I just couldn't resist.
gregW:-) OH2FFY
http://www.swdxer.co.nr/
=====================================================================